Control Valve Selection and Sizing

Transcription

1 Control Valve Selection and Sizing APPLICATION AND SELECTION OF CONTROL VALVES 1050 Introduction 1050 Orientation Table 1051 Control Valve Trends 1051 Globe vs. Rotary Valves 1051 Valves vs. Other Final Control Elements 1051 Control Valve Sizing 1051 Collecting the Process Data 1053 Determining the Valve Pressure Drop 1053 Characteristics, Gain, and Rangeability 1054 Characteristics and Gain 1054 Valve Rangeability 1057 Actuator Selection 1058 Piston Actuators 1058 Actuator Speeds of Response 1059 Actuator Power 1059 Valve Failure Position 1059 Positioners 1060 When to Use Positioners 1060 When Not to Use Positioners 1060 Positioners to Eliminate Dead Band 1061 Split-Range Operation 1061 Accessories 1061 Process Application Considerations 1061 Pressure Considerations 1061 High-Temperature Service 1063 Low-Temperature Service 1066 Cavitation and Erosion 1068 Methods to Eliminate Cavitation 1068 Control Valve Noise 1072 Flashing and Erosion 1073 Corrosion 1075 Viscous and Slurry Service 1075 Valves That Can Be Sterilized 1076 Valve Leakage 1076 Installation 1080 Climate and Atmospheric Corrosion 1080 Control Valve Specification Form 1080 References 1080 Bibliography ACCESSORIES AND POSITIONERS 1087 Introduction 1089 Smart Valves 1089 Positioners 1090 When to Use Positioners 1091 When Not to Use a Positioner 1093 Positioner Performance 1093 Positioner Designs 1094 Positioner Accessories 1095 Position Indicators 1095 Transducers 1096 I/P (Electropneumatic) Transducers 1096 Digital Electropneumatic Transducers 1096 Relays 1096 Booster Relays 1096 Reversing and Other Relays 1097 Quick-Exhaust Relays 1098 Relays to Lock-up Valve Position 1099 Failure Position Guaranteed by Stored Air

2 1046 Control Valve Selection and Sizing Energy Supplies 1100 Air Sets 1100 Hydraulic (High-Pressure) Operation 1101 Hydraulic (Water) Operation 1101 Limit Switches 1101 Solenoid Valves 1101 Three-Way Solenoids 1101 Four-Way Solenoids 1102 Solenoid Capacity 1102 Handwheels 1102 Limit Stops 1103 Bypass Valve 1103 References 1104 Bibliography ACTUATORS: DIGITAL, ELECTRIC, HYDRAULIC, SOLENOID 1105 Introduction 1107 Selection and Application 1107 Actuator Types 1107 Actuator Features 1107 Digital Valve Actuators 1109 Electromechanical Actuators 1110 Reversible Motor Gear Actuators 1111 Rotary Output Actuators 1111 Linear Output Actuators 1112 Electrohydraulic Actuators 1114 External Hydraulic Source 1114 Hermetically Sealed Power Pack 1114 Motor and Pump Combinations 1115 Solenoid Valves 1118 Modulating Solenoid Valves 1120 Smart Actuators 1121 Applications 1121 References 1122 Bibliography ACTUATORS: PNEUMATIC 1124 Introduction 1126 Definitions 1126 Actuator Features and Selection 1126 Spring/Diaphragm Actuators 1126 Steady-State Force Balance 1127 Actuator Sizing Example 1128 Actuator Nonlinearities 1129 Dynamic Performance of Actuators 1129 Safe Failure Position 1131 Pneumatic Response Times 1132 Piston Actuators 1133 High-Speed Actuators 1133 Relative Merits of Diaphragm and Piston Actuators 1135 Rotary Valve Actuators 1136 Cylinder Type 1137 Rotation by Spline or Helix 1138 Vane Type 1138 Rotary Pneumatic Actuators 1138 Other Pneumatic Actuators 1139 Pneumohydraulic Actuators 1139 Electropneumatic Actuators 1140 Reliability 1141 Conclusions 1142 References 1142 Bibliography ADVANCED STEM PACKING DESIGNS 1144 Introduction 1144 History 1144 Bibliography CAPACITY TESTING 1150 Introduction Scope Purpose Nomenclature Test System General Description Test Specimen Test Section Throttling Valves Flow Measurement Pressure Taps Pressure Measurement Temperature Measurement Installation of Test Specimen Accuracy of Test Test Fluids Incompressible Fluids Compressible Fluids 1153 Summary CHARACTERISTICS AND RANGEABILITY 1154 Introduction 1154 Valve Gain and Loop Gain 1154 Nonlinear Processes 1154 Installed Valve Gain 1155 Theoretical Valve Characteristics 1155 Valve Testing 1155 Valve Characteristics 1155 Valve and Process Characteristics 1155 Selection Recommendations 1156 Installation Causes Distortion 1157 Distortion Coefficient 1157 Correcting the Valve Characteristic 1158

3 Contents of Chapter Rangeability 1158 Improved Definition of Rangeability 1159 Why Traditional Rangeability Is Wrong 1159 Conclusions 1160 Bibliography DIAGNOSTICS AND PREDICTIVE VALVE MAINTENANCE 1161 Introduction 1161 Diagnostics 1161 Instrumentation Used 1161 Diagnostic Methods 1162 Characteristics Tests 1162 Valve Signatures 1162 Analyzing Valve Signatures 1164 Conclusions 1164 Bibliography DYNAMIC PERFORMANCE OF CONTROL VALVES 1165 Valve Response 1165 Definitions 1165 Discussion 1166 Valve System 1166 Install Positioner 1167 Increase Force 1168 Reduce Friction 1168 Defining Response 1168 Measuring Response 1168 Relationships 1169 Determine Required Response Specifications 1169 Application Examples 1169 Flow Control 1169 Reactor Mixing 1170 Neutralizing Waste Water 1170 Antisurge Valve 1170 Delay or Slowdown Valve Action 1170 Safety Solenoid Valves 1170 Troubleshoot Valve Response 1171 Bibliography 1171 References EMERGENCY PARTIAL-STROKE TESTING OF BLOCK VALVES 1172 Introduction 1172 The Partial-Stroke Test 1172 Mechanical Limiting 1173 Position Control 1173 Solenoid Valve 1173 Impact of PST on SIL 1173 Block Valve Analysis 1175 Overall SIS Performance 1178 Single Block Valve Case 1178 Dual Block Valve Case 1179 Conclusions 1181 Bibliography FIELDBUS AND SMART VALVES 1182 Introduction 1182 Benefits and Savings 1183 Hart, Foundation Fieldbus, and Profibus-PA 1183 Valve Calibration and Configuration 1185 Safety and Pollution 1188 On-line Plant Asset Management 1189 Digital Valve Instrumentation 1189 Second Generation 1189 Conclusions 1192 References INTELLIGENT VALVES, POSITIONERS, ACCESSORIES 1193 Introduction 1193 Advantages of Intelligent Positioners 1193 Typical Performance Specifications 1194 Generating the Pneumatic Output 1194 Valve Performance Monitoring 1195 Controlling the Process 1195 Changing the Valve s Characteristics 1195 Operation of Smart Positioners 1196 Maintenance and Calibration 1197 Accessories 1197 Flow Control by Smart Valve 1197 Limitations 1198 References 1198 Bibliography MISCELLANEOUS VALVE AND TRIM DESIGNS 1199 Introduction 1200 Miscellaneous Valve Designs 1200 Dynamically Balanced Plug Valves 1200 Positioned Plug In-Line Valves 1201 Expansible Valve Designs 1202 Fluid Interaction Valves 1205 Special Valve Application 1206 Cavitation and Flashing 1206 Dirty Process Services 1208 High Noise 1209 High-Capacity Valves 1210

4 1048 Control Valve Selection and Sizing Cryogenic Valves 1210 High-Temperature Valves 1210 Steam Conditioning Valves 1211 Tank-Mounted Valves 1211 Bibliography VALVES: NOISE CALCULATION, PREDICTION, AND REDUCTION 1213 Introduction 1213 Sound and Noise 1214 Speed of Sound 1214 The Human Ear 1214 Loudness Perception 1215 Limiting Valve Noise 1215 Valve Noise 1216 Control Element Instability 1217 Resonant Vibration 1217 Hydrodynamic Noise 1218 Aerodynamic Noise 1218 Controlling Noise 1218 Path Treatment 1219 Source Treatment 1220 Aerodynamic Noise Prediction 1223 Standards 1224 Calculations 1224 Noise Calculation Example 1230 Applying Distance Corrections 1232 Hydrodynamic Noise Prediction 1232 Bibliography SIZING 1234 Introduction 1234 About This Section 1234 Standards 1234 General Principles 1235 The Flow Coefficient 1235 Liquid Sizing 1237 Relative Valve Capacity Coefficient (C d ) 1237 Factors F L, F F, F P, and F LP 1237 Example 1241 Units Used in Valve Sizing 1241 Sizing Example for Liquids 1243 The Cavitation Phenomenon 1244 Flashing 1248 Laminar or Viscous Flow 1250 Gas and Vapor Sizing 1252 Equations for Turbulent Flow 1252 Constants for Engineering Units 1253 Expansion Factor (Y ) 1253 Choked Flow 1253 Velocity of Compressible Fluids 1254 Sizing for Compressible Fluids (Example 12) 1255 Two-Phase Flow 1256 Liquid-Gas Mixtures 1257 Liquid-Vapor Mixtures 1258 Conclusions 1259 Nomenclature 1259 References 1260 Bibliography VALVE TYPES: BALL VALVES 1262 Introduction 1264 Throttling Ball Valves 1264 Conventional Ball Valves 1265 The Valve Trim 1266 Flow Characteristics 1267 Characterized Ball Valves 1268 Construction 1268 Characteristics 1269 Ball and Cage Valves 1269 Sizes and Other Features 1270 Ball Unseated by Stem 1271 Ball Gripped by Cage 1271 References 1271 Bibliography VALVE TYPES: BUTTERFLY VALVES 1273 Introduction 1274 Conventional Butterfly Valves 1275 Operation 1276 Construction 1276 High-Performance Butterfly Valves 1276 Tight Shut-off Designs 1278 Leakage Ratings 1279 Fire-Safe Designs 1280 Torque Characteristics 1280 Noise Suppression 1282 Bibliography VALVE TYPES: DIGITAL VALVES 1284 Introduction 1284 History 1285 Balanced Piston Digital Control 1285 Top-Entry Design 1285 Flow Metering 1288 Gas Flow 1288 Liquid Flow 1288 Conclusions 1288 Reference 1289 Bibliography 1289

5 Contents of Chapter VALVE TYPES: GLOBE VALVES 1290 Valve Trends 1291 Trim Designs 1292 Trim Flow Characteristics 1294 Rangeability 1295 Standard Trim Configurations 1296 Special Trim Configurations 1296 Trim Materials 1298 Leakage 1298 Plug Stems 1299 Bonnet Designs 1300 Bolted Bonnets 1300 Pressure Seal Bonnets 1301 Bonnet Classification 1302 Bonnet Packing 1303 Body Forms 1308 Double-Ported Valves 1309 Single-Seated Valves 1309 Three-Way Valves 1315 Lined and Thermoplastic Valves 1315 Valve Connections 1316 Flanged Ends 1316 Welded Ends 1317 Threaded Ends 1317 Special End Fittings 1318 Materials of Construction 1318 Trademarks 1321 Reference 1321 Bibliography VALVE TYPES: PINCH VALVES 1323 Introduction 1323 The Sleeve 1326 Pinch Valve Types 1328 Pressure Limitations 1328 Shell and Tube Design 1328 Throttling Characteristics 1334 Applications 1336 Wastewater 1336 Flue-Gas Desulfurization 1337 Mine Slurries 1337 Paper and Tile Manufacturing 1337 Toxic Gas Applications 1337 Pigments, Paint, and Ink 1337 Glue 1337 Food 1337 Powders and Grinding Compounds 1337 Chemicals 1337 Cavitation 1337 The Phenomenon 1338 The Pinch Valves 1338 Limiting or Eliminating the Damage 1338 Conclusions 1340 Bibliography VALVE TYPES: PLUG VALVES 1341 General Characterisics 1342 Plug Valve Features 1343 Throttling and Actuator Considerations 1343 Design Variations 1343 Characterized Plug Valves 1344 V-Ported Design 1344 Adjustable Cylinder Type 1345 Semispherical Plugs for Tight Closure 1345 Expanding Seat Plate Design 1345 Retractable Seat Type 1346 Overtravel Seating Design 1346 Multiport Design 1347 Bibliography VALVE TYPES: SAUNDERS DIAPHRAGM VALVES 1348 Introduction 1348 Saunders Valve Construction 1348 Materials of Construction 1350 Straight-Through Design 1351 Full Bore Valve 1352 Dual-Range Design 1352 Bibliography VALVE TYPES: SLIDING GATE VALVES 1353 Introduction 1354 Sliding Gate Valve Designs 1354 Knife Gate Valves 1354 Positioned-Disc Valves 1355 Plate and Disc Valves 1356 Bibliography 1357

6 6.1 Application and Selection of Control Valves B. G. LIPTÁK (1970, 1985, 1995, 2005) A. BÁLINT (2005) Subjects Covered in this Section: Actuators; see also Sections 6.3 and 6.4 Cavitation; see also Section 6.15 Characteristics; see also Section 6.7 Corrosion; see also Section 6.19 Erosion; see also Section 6.19 Fire safety; see also Sections 6.16 and 6.17 Flashing and erosion; see also Section 6.15 Gain; see also Section 6.7 High pressure drop applications; see also Section 6.15 High-pressure services High-temperature services Installation considerations Intelligent valve features; see also Section 6.12 Jacketed valves Leakage; see also Sections Low-temperature services (cryogenics); see also Section 6.23 Noise abatement; see also Section 6.14 and 6.17 Packing designs; see also Section 6.19 Positioners; see also Section 6.2 Process data Rangeability; see also Section 6.7 Selection chart for control valves Sequencing, split-ranging Sizing; see also Section 6.15 Small flow applications; see also Section 6.23 Specification forms for control valves Toxic applications; see also Section 6.23 Vacuum services Viscous and slurry services INTRODUCTION In the field of control valve design, the most important developments of the last decade occurred in the areas of electric and digital actuators (Section 6.3), in valve diagnostics (Section 6.8), dynamic performance evaluation (Section 6.9), safety shutdown systems (Section 6.10), fieldbus interaction (Section 6.11), intelligent positioners (Section 6.12), valve status detection and use for control (Section 6.13), and in the increased availability of special valve designs (Section 6.12). Because each of these topics are covered in the noted separate sections, they are not treated in detail in this section. While this section attempts to discuss all basic aspects of control valve selection and application, in this area too, there exists some overlap with other sections. For example, as noted in the alphabetic listing above, the topics of valve characteristics and rangeability are discussed in more detail in Section 6.7; noise and its reduction in Section 6.14; sizing in Section 6.15; valve actuators and accessories including positioners in Sections 6.2 and 6.4; and the features of the particular valve designs in Sections 6.16 to Therefore, the reader is advised to treat this section only as an overview of the subject of control valve applications and refer to the individual sections of this chapter for the detailed discussion of its many specific aspects. It should also be noted that in most of the sections in this chapter English units are used with their SI equivalents given in parenthesis. An exception is Section 6.14 on valve noise calculation, which follows the general practice of acoustics and uses SI units. Appendices 1 and 2 (at the end of this handbook) give all the conversion factors that are required to go from one to the other system of units. So, for example, 1 lb/in. 2 equals 6.89 kpa (kilo Pascals) or bars. Hence, a 3 15 PSIG range is the approximate 1050

7 6.1 Application and Selection of Control Valves 1051 equivalent of kpa. The valve capacity coefficient used in this chapter is the C v, which is unity, if the valve passes 1.0 gpm of cold water at a specific gravity of 1.0 at a pressure drop of 1.0 psid. The metric equivalent is the K v, which corresponds to a valve passing 1 m 3 of cold water per hour at a pressure drop of 1 bar. Therefore, if the reader wishes to convert any Cv value into K v, the multiplier is 1.17, and therefore C v = 1.17 K v. Orientation Table In order to provide some overall orientation about the relative merits of the different valve designs, an orientation table, Table 6.1a, was prepared. In this table some of the more common applications are described, together with some indications of the suitability of the various valve designs. At the end of this section, a standard control valve specification form, prepared by Instrumentation, Systems, and Automation Society, is provided together with an explanation of each of the entries in that form. The discussion of the various topics related to valve selection follows the approximate order of the entries in that form. Therefore, after some introductory remarks, the discussion begins with topics related to the service conditions (process data) and then continues with topics related to the features and accessories of the valve and its installation. CONTROL VALVE TRENDS When this handbook was first published some 35 years ago, the overwhelming majority of throttling control valves were the globe types, characterized by linear plug movements and actuated by spring-and-diaphragm operators. At that time, the rotary valves were considered to be on/off shut-off devices. Today, globe valves are still widely used, but their dominance is being challenged by the less expensive rotary (ball, butterfly, plug) valves, which are usually actuated by cylinder operators. This trend represents a mixed blessing and, therefore, is worth further discussion. Globe vs. Rotary Valves The main advantages of the traditional globe design include the simplicity of the spring-and-diaphragm actuator; the availability of a wide range of valve characteristics; the relatively low likelihood of cavitation and noise; the availability of a wide variety of specialized designs for corrosive, abrasive, and high/low temperature and pressure applications; the linear relationship between control signal and valve stem movement; and the relatively small amounts of dead band and hysteresis in its operation. These features make the globe valve usable without positioners, which on fast processes is an advantage. The main reason why rotary valves have been increasing their market share is their lower manufacturing cost and higher relative flow capacity (C d = instead of C d = 10 15, as for globe). They also weigh less, can act as both control and shut-off valves, and are easier to seal at the stem to meet OSHA and EPA requirements. The limitations of globe valves, in addition to their higher cost per unit C v (in Europe, the equivalent term K v is used), include their relatively slow speed and low stiffness (plug position is affected by dynamic forces in the process fluid), both of which can be improved by using hydraulic cylinder actuators operating at higher pressures. Major disadvantages of rotary valves are their higher tendency to cavitate and to produce excessive amounts of noise. They are also more likely, due to their smaller size per unit C v (K v ), to have larger pipe reducers with the associated waste of pressure drop and distortion of characteristics. Their control quality can suffer from the nonlinear relationship between actuator linear movement and valve rotation, plus from the linkages, which can introduce substantial hysteresis and dead play. These characteristics result, in most cases, in a definite requirement for using a positioner, which on fast processes can cause the deterioration of control quality. Valves vs. Other Final Control Elements Before proceeding through the steps of selecting a control valve, one should evaluate if a control valve is truly needed in the first place, or if a simpler and more elegant system will result through some other means. For example, an overflow weir can suffice to keep levels below maximum limits, and choke or restriction fittings can serve the function of pressure letdown at constant loads. In other locations, it might be possible to reduce the investment by using regulators instead of control valves. The advantages of regulators include their high speed (high gain) and their self-contained nature, which eliminates the need for power supplies or utilities. If remote set point adjustment is needed, regulators can be provided with airloaded pilots to accommodate that requirement. While all regulators (being proportional-only controllers) will display some offset as the load changes, the amount of offset can be minimized by maximizing the regulator gain. In still other applications, it is prudent to replace whole flow control loops with positive-displacement metering pumps or to replace the control valve with variable-speed centrifugal pumps. The cost-effectiveness of the approach is usually found to be in lowered pumping costs, because the pumping energy that was burned up in the form of pressure drop through the control valve is not being introduced, and therefore it is saved. CONTROL VALVE SIZING Control valve sizing is discussed in depth in Section 6.15, and therefore only a few general recommendations are made here. One should first determine both the minimum and maximum C v (K v in Europe) requirements for the valve, considering not only normal but also start-up and emergency conditions. The selected valve should perform adequately over

8 TABLE 6.1a Orientation Table for Selecting the Right Control Valves for Various Applications Features & Applications Features: ANSI class pressure rating (max.) Ball: Conventional Ball: Characterized Butterfly: Conventional Butterfly: Highperformance Digital Globe: Singleported Globe: Doubleported Control Valve Types Globe: Angle Globe: Eccentric disc Pinch Plug: Conventional Plug: Characterized Saunders Sliding gate: V-Insert Sliding gate: Positioned disc Special: Dynamically balanced Max. capacity (C d ) Characteristics F G P F, G E E E E G P P F, G P, F F F F, G Corrosive Service E E G G F, G G, E G, E G, E F, G G G, E G G F, G G G, E Cost (relative to single-port globe) Cryogenic service A S A A A A A A A NA A S NA A NA NA High pressure drop (over 200 PSI) High temperature (over 500 F) Leakage (ANSI class) A A NA A E G G E A NA A A NA NA E E Y S E G Y Y Y Y Y NA S S NA NA S NA V IV I IV V IV II IV IV IV IV IV V I IV II Liquids: Abrasive service C C NA NA P G G E G G, E F, G F, G F, G NA E G Cavitation resistance L L L L M H H H M NA L L NA L H M Dirty service G G F G NA F, G F G F, G E G G G, E G F F Flashing applications P P P F F G G E G F P P F P G P Slurry including fibrous service G G F F NA F, G F, G G, E F, G E G G E G P F Viscous service G G G G F G F, G G, E F, G G, E G G G, E F F F Gas/Vapor: Abrasive, erosive C C F F P G G E F, G G, E F, G F, G G NA E E Dirty G G G G NA G F, G G F, G G G G G G F G 1052 Control Valve Selection and Sizing Abbreviations: A = Available C = All-ceramic design available F = Fair G = Good E = Excellent H = High L = Low M = Medium NA = Not available P = Poor S = Special designs only Y = Yes

9 6.1 Application and Selection of Control Valves 1053 a range of 0.8 C vmin to 1.2 C vmax. If this results in a rangeability requirement that exceeds the capabilities on one valve, use two or more valves. Control valves should not be operated outside their rangeability. Driskell (see Bibliography) properly points to the fact that all fat settles in the control valve. In constant speed pumping systems, each design engineer will add their own safety margin in calculating pressure drops through pipes and exchangers, and finally in selecting the pump. Therefore, the control valve will end up with all these safety margins as added pressure drops, resulting in a muchoversized valve. A highly oversized valve will operate in a nearly closed state, which is an unstable and undesirable operating condition. In variable-speed pumping systems, this problem does not exist, because there the pump speed is adjusted to meet the load, and therefore the effect of accumulated safety margins is eliminated. Control valve pressure drop Pump pressure rise At low flow rates Pipe pressure drop Static head Collecting the Process Data In order to select the right control valve, one must fully understand the process that the valve controls. Fully understanding the process means not only understanding normal operating conditions, but also the requirements that the valve must live up to during start-up, shutdown, and emergency conditions. Therefore, all anticipated values of flow rates, pressures, vapor pressures, densities, temperatures, and viscosities must be identified in the process of collecting the data for sizing. In addition, it is desirable to identify the sources and natures of potential disturbances and process upsets. One should also determine the control quality requirements, so as to identify the tolerances that are acceptable in controlling the particular variable. The process data should also state if the valve needs to give tight shut-off, if the valve noise needs to be limited, or any other factors that might not be known to the instrument engineer. These can include subjective factors, such as user preferences, or objective ones, such as spare parts availability, delivery, life expectancy, or maintenance history. Lines 1 12 in the Specification Form for Control Valves (at the end of this section) describe the service conditions (process data) that must be provided for the control valve. It is important to carefully determine not only the normal values for this data but also the minimum and maximum values, because the valve must operate properly throughout its range not just under normal conditions. Determining the Valve Pressure Drop Assigning the sizing pressure drop for the valve is more complex than picking a number like 10% or 25% of the total system drop or a number like 10 or 25 psi (0.69 or 1.72 bar). It requires an understanding of the interrelationships that exist in pumping, fan, or compressor systems. If a system consists of nothing else but a pump, a control valve, piping, vessels, and an elevated destination, the energy profile through the system will be as illustrated in Figure 6.1b. Control valve pressure drop Pump pressure rise At high flow rates Static head FIG. 6.1b Pressure profiles of a pumping system at high and low flow rates. Note that the pressure drop available for the control valve drops as the flow rate rises. This is because at higher flows, the pump discharge pressure will be lower, while the pressure drop through the piping will rise. In other words, the control valve does not work with a fixed pressure drop, nor with a fixed percentage of the total system drop, but it simply takes whatever is left over from what is available and what is required by the rest of the system. Therefore, as shown in Figure 6.1c, the valve energy loss at any particular flow rate (load) is the difference between the corresponding points on the pump and system curves. As can be seen in Figure 6.1c, the available pressure drop increases as the load (flow rate) drops, and it becomes the minimum when the process flow rate (load) is the maximum. This means that the actual control valve rangeability (in terms of C v ) must be much larger than the ratio of maximum and minimum flows. There is a similar effect on the valve characteristics, which will be discussed (together with rangeability) in both the paragraphs that follow and also in more detail in Section 6.7. Because of the complexity of the problem, no simple rule of thumb can be used in assigning the valve pressure drop; thus, it is important for the engineer to have a good understanding of the pump and system curves together with the rangeability and characteristics (gain) requirements of the valve. In the process of deciding what pressure drop to assign, you must acknowledge the conflict between control quality and energy conservation. The higher the pressure drop through the valve (relative to the rest of the system), the more

10 1054 Control Valve Selection and Sizing Head FT Pmax 0 0 F min System curve 20 P Pipe friction pressure loss Pump curve Static head (pressure) 40 F norm F max Flow GPM FIG. 6.1c The difference between the pump discharge pressure curve and the system curve (which is the sum of the static head and the pipe friction loss) is the available valve differential. impact it will have on the process while, at the same time, the more pumping energy it will waste. Some might argue that, in fact, no pressure drop needs to be assigned to the control valve during the design phase because as safety margins accumulate, the pump will be oversized anyway. There is some practical wisdom in this attitude, because it is true that by the time the pump service conditions pass from the process engineer to the mechanical engineer, then are sent out for bidding to the manufacturers, and finally a pump is selected, its flow and pressure capabilities will always much exceed the originally specified requirements, and there will be plenty of pressure drop for the valve. While this argument sounds convincing and convenient, it is wrong. It is wrong because it results in unpredictable performance (possibly high noise) and usually also results in oversized control valves, which tend not only to be unstable, but also to have low rangeability. Therefore, the proper approach to the selection of valve pressure drop is to first determine the total friction energy loss (excluding static energy) of the system at normal load (flow) and assign 50% of that to valve pressure energy drop. Based on that assignment, one should next determine the resulting valve drop at minimum and maximum loads (flows) and select a valve that can handle the required C v rangeability. As will be discussed later, one should also select a valve characteristic that, after being distorted by the change in valve drop as the load varies, will give acceptable (stable) loop performance. CHARACTERISTICS, GAIN, AND RANGEABILITY Sum, pipe and static pressure Good control valve performance usually means that the valve is stable across its full operating range, it is not operating near Pmin to one of its extreme positions, it is fast enough to correct for process upsets or disturbances, and it will not be necessary to retune the controller every time the process load changes. In order to meet the above goals, one must consider such factors as valve characteristics, rangeability, installed gain, and actuator response. These topics will be separately addressed here and in Section 6.7. Characteristics and Gain Characteristics and gain are discussed in more detail in Section 6.7, but they are also briefly covered here. The reason process control engineers must be concerned with selecting the right valve characteristics (Figure 6.1d) is because the valve is part of the control loop, and the loop will be stable only if the products of all its gain components (the gains of the process, sensor, controller, and valve gains) is constant. Usually, the controller is tuned so that this gain product is 0.5, in order to give quarter-amplitude damping. This was discussed in some detail in connection with Figure 2.1x in Chapter 2. If the gains of the loop components do not vary with load, but are constant, the desirable choice is to use a constant gain control valve. A constant gain valve is a linear valve whose theoretical gain (change of flow per unit change of lift) is 1. If the gain of any of the loop components (such as the process) decreases with load (flow through the valve), the proper choice of control valve gain is 1, which increases with load (equal percentage), because this combination will keep the gain product of the loop relatively constant. Installation Causes Distortion As was pointed out in connection with Figure 6.1c, in mostly friction systems, such as pumping through long pipes (where only a small portion of the total pump energy is used to overcome constant static pressure), % C v or flow at constant P Linear (Gain = 1.0) 20 Equal percentage* % Lift or rotation FIG. 6.1d Inherent characteristics of control valves. *Gain increases at a constant slope. The rate of rise is a fixed % of the actual flow.

11 6.1 Application and Selection of Control Valves Available P Available valve differential pressure Required C v Flow GPM Flow GPM FIG. 6.1e In mostly friction systems, an increase in load (flow rate) results in a drop in the pressure drop, which is available for the control valve. Therefore, the same amount of increase in flow rate requires a larger increase in the valve capacity coefficient C v (K v ). For such application, an equal-percentage valve is needed. the pressure drop available for the control valve is dropping as the load (flow rate) is increasing, and therefore more C v is needed for unit flow increases as the flow rises (Figure 6.1e). This is a different condition from the condition at which the valve characteristics were established in the testing facility of the manufacturer, where the flow rate through the valve was measured under constant pressure drop conditions. Therefore, when such valves are installed, their gain (characteristics) shift, as will be discussed in Section 6.7. One way to correct for such distortion is to obtain near-linear installed characteristics by installing a valve with ideal inherent equal-percentage characteristics. As will be discussed in connection with positioners (Section 6.2), it is also possible to use a cam in the positioner to modify the installed characteristic of the valve, but this can dramatically both change the loop gain of the positioner and limit its dynamic response. Therefore, it is preferred to change the inherent characteristic of the valve trim than to install cams in the valve positioner. The installed characteristic of the valve can also be modified by characterizing the control signal going to the positioner. This characterization occurs outside the positioner feedback loop, and therefore it has the advantage of not changing the loop gain of the positioner. This method also has its dynamic limitations. For example, a 1% change in the controller output signal may be electronically narrowed to a change in the valve signal of only 0.1% (in the flat regions of the valve characteristic), but such change is too small for some valves to respond to at all. Therefore, the best solution to obtaining a constant loop gain is to select the inherent characteristics of the valve trim to compensate for the nonlinearity of the process and, thereby, arrive at an installed flow characteristics, which is nearly linear over the operating range of the valve. Selecting the Valve Characteristics Different engineers have approached the problems caused by process nonlinearity (drop in process gain) in different ways. One approach, that of the old school, was to oversize the pump so that the ratio between the minimum and maximum energy loss in Figure 6.1c will not be large, and therefore the gain of the process will not change much with load. This approach works, but it wastes pumping energy. Different engineers began to develop different rules of thumb to be used in selecting valve characteristics for the various types of control loops. These recommendations vary in complexity. Shinskey, for example, recommends equal percentage for temperature control and the use of linear valves for all flow, level, and pressure control applications (except vapor pressure, for which he recommends equal percentage). According to Driskell, one can avoid a detailed dynamic analysis by just considering the ratio of the maximum and minimum valve pressure drops ( p max / p min ) and follow the rule of thumbs listed for the most common applications in Table 6.1f. Lytle s recommendations are summarized in Table 6.1g; they are more involved, as they take more variables into account. TABLE 6.1f Valve Characteristics Selection Guide Service Valve ( p max / p min ) Under 2:1 Valve ( p max / p min ) Over 2:1 but Under 5:1 Orifice-type flow Quick-opening Linear Flow Linear Equal % Level Linear Equal % Gas pressure Linear Equal % Liquid pressure Equal % Equal %

12 1056 Control Valve Selection and Sizing TABLE 6.1g Recommendations on Selecting Control Valve Characteristics for Flow, Level, and Pressure Control Loops LIQUID LEVEL SYSTEMS Control Valve Pressure Drop Constant P Decreasing P with increasing load, P at maximum load > 20% of minimum load P Decreasing P with increasing load, P at maximum load < 20% of minimum load P Increasing P with increasing load, P at maximum load < 200% of minimum load P Increasing P with increasing load, P at maximum load > 200% of minimum load P Best Inherent Characteristic Linear Linear Equal-percentage Linear Quick-opening PRESSURE CONTROL SYSTEMS Application Liquid process Gas process, small volume, less than 10 ft of pipe between control valve and load valve Gas process, large volume (process has a receiver, distribution system, or transmission line exceeding 100 ft of nominal pipe volume), decreasing P with increasing load, P at maximum load > 20% of minimum load P Gas process, large volume, decreasing P with increasing load P at maximum load < 20% of minimum load P Best Inherent Characteristic Equal-percentage Equal-percentage Linear Equal-percentage FLOW CONTROL PROCESSES Best Inherent Characteristic Flow Measurement Signal to Controller Location of Control Valve Relation to Measuring Element Wide Range of Flow Setpoint Small Range of Flow but Large P Change at Valve with Increasing Load Proportional to Q In series Linear Equal-percentage Proportional to Q In bypass* Linear Equal-percentage Proportional to Q 2 (orifice) In series Linear Equal-percentage Proportional to Q 2 (orifice) In bypass* Equal-percentage Equal-percentage * When control valve closes, flow rate increases in measuring element. Most common. From Reference 1. Process Nonlinearity Yet another approach in overcoming the process nonlinearity caused by the variation in valve pressure drop is to modify the controller output signal to eliminate that nonlinearity (discussed in Section 6.7) or to replace the control valve with a complete flow control loop (Figure 6.1h). By selecting a linear flow transmitter, the characteristics of the slave loop will also be linear, and therefore its gain will be 1.0. This approach works fine, but it also increases the system cost. In addition, while it eliminates a nonlinearity, it introduces a slave loop, which on fast processes can degrade the quality of control quality. By using intelligent control valves (Section 6.12), one can obtain flexibility in implementing one or the other approach. Process nonlinearity also exists for reasons other than variations in the available valve pressure drop. The gain of heattransfer processes, for example, always drops as the load increases, because the heat-transfer surface is constant, and therefore the heat transfer is more efficient when small amounts of heat need to be transferred. Consequently, in order to keep the gain product of the control loop constant, it is necessary to compensate the dropping process gain with an increasing valve SP FC FY FT FIG. 6.1h Valves can be linearized by replacing them with a complete slave control loop.

13 6.1 Application and Selection of Control Valves 1057 gain. For this reason, all temperature control valves are always equal percentage. Composition processes can also be nonlinear, but their nonlinearity is usually more complex, such as the titration curve of a ph control system. In these situations, the likely solution is to use a linear valve (constant gain) and a nonlinear controller (Figure 2.19d), one whose gain varies as the mirror image of the process gain. Valve Rangeability Control valve rangeability is also discussed in more detail in Section 6.7. Here, it should suffice to state that the required rangeability should be calculated as the ratio of the C v (K v ) required at maximum flow (and minimum pressure drop) and the C v (K v ) required at minimum flow (and maximum pressure drop). The decision on whether a particular control valve is capable of providing the required rangeability should be evaluated on the basis of a plot of valve gain vs. valve C v. If the actual valve gain is within 25% of the theoretical valve gain between the minimum and maximum C v, the rangeability is acceptable. As will be discussed in Section 6.7, the rangeability definitions used by manufacturers are usually not based on valve gain. One way to increase the rangeability is to have the controller operate more than one control valve. As will be discussed below, such multiple valves can be split-ranged, sequenced, or operated in a floating mode. about 65%), providing a much more favorable span for the calibration of the positioner. In order to have the two valves act as one without disturbing the smooth equal-percentage characteristics at the points of switching, only one valve must be open at any one time. Therefore, the large valve must be prevented from operating at low flows, because in its nearly closed position its characteristics are not equal-percentage. For these reasons, only one valve must be open at any one time. In the scheme shown in Figure 6.1i, the small valve alone is manipulated until the controller output reaches the value corresponding to its full opening. At this point the pressure switch energizes both three-way solenoid valves, venting the small valve and opening the large to the same flow that the small had been delivering. Switching takes place in 1 sec or less, adequate for all but the fastest control loops. When the controller output falls to the point of minimum flow from the larger valve (35%), the solenoids return to their original position. Thus, the switch has a differential gap adjusted to equal the overlap between valve positioners (30%). The range of the positioner for the large valve is found by locating its minimum C v in Figure 6.1i. A rangeability of 50 would give a minimum C v of Control Valve Sequencing When the rangeability requirements of the process exceed the capabilities of a single valve, control valve sequencing loops must be designed that will keep the loop gain constant while switching valves. This requires careful thought. Assuming that the task is to sequence two linear valves, with sizes of 1 and 3 in. (25 and 75 mm) having C v s (K v s) of 10 (8.62) and 100 (86.2), respectively, Shinskey s recommendation is the following: If the large valve were to operate from 9 to 15 PSIG (0.6 to 1.0 bar) and the small one from 3 to 9 PSIG (0.2 to 0.6 bar), the loop gain would change by 10 when passing through 9 PSIG (0.6 bar). The only way to keep the loop gain constant in this example would be to operate the small valve from 0 to 10% and the large valve from 10 to 100% of controller output. This would result in a PSIG ( bar) range for the 1 in. (25 mm) and a PSIG ( bar) range for the 3 in. (75 mm) valve. Therefore, if more than one valve is required to increase rangeability, in most cases equalpercentage valve characteristics are needed. C v Small valve Large valve % 65% Controller output % Controller output PS 65% Differential-30% Sequencing Equal-Percentage Valves The sequencing of equal-percentage valves is done as follows: If the small valve had a C v of 10 and a rangeability of 50:1, its minimum C v would be 10/50 = 0.2. A line drawn on semilogarithmic coordinates connecting C v (K v ) 100 (86.2) and 0.2 (0.172) appears in Figure 6.1i. Observe that the C v of 10 of the small valve falls slightly above the mid-scale of the controller output (to Small Large FIG. 6.1i Sequencing two equal-percentage valves, while minimizing the upset caused by switching valves.

14 1058 Control Valve Selection and Sizing 3 9 # 9 15 # 1:2 1:2 AMP AMP ± ( 3 # ) ± BIAS 3 15 # P FC 3 15 # P ( 15 # ) BIAS FC Split ranging (Gain plus BIAS) (2) 1:2 AMP + BIAS P Small Set at 50% P Large Floating control of large valve in that it uses only one four-way solenoid pilot instead of two three-way units. The purpose of the 1:1 amplifier is to eliminate the bounce when the solenoid switches. As the purpose of valve sequencing is to increase the rangeability of the loop without upsetting its stability, the existence of two or more valves should not be noticeable by the controller. In other words, in a well-designed sequencing system, the controller would operate as if its final control element were a single valve, having the desired gain characteristics and a very wide rangeability. In order to keep the gain characteristics of the valve pair correct, there should be no bumps when the larger valve is opened. This requires that only one valve be throttled at a time and the other be closed. From this perspective, the performance of the small/large valve selection scheme in Figure 6.1j is superior to that of the split-ranging or the floating methods. 65% PSH S Vent Diff.: 30% B 1:1 AMP A P Small Large Small/Large valve select FIG. 6.1j Alternate methods of obtaining high turndown through the use of multiple control valves. This same approach can be used to sequence three or more valves. If linear characteristics are required, one should insert a 10:1 multiplier relay in the controller signal to the small valve, so that a 0 10% controller output will result in a 0 100% signal to the small valve. Split-Ranging or Floating Some process control engineers feel that the switching scheme in Figure 6.1i is too abrupt or could cause a maintenance problem. For these reasons, they prefer the methods of valve sequencing illustrated in Figure 6.1j. The split-ranging loop shown in the upper left of the figure contains both gain and bias relays to provide the added rangeability. Even better response can be provided by the use of a floating valve position controller, shown on the top right of Figure 6.1j. This controller slowly moves the larger valve so as to keep the smaller one near its 50% opening. This way, the small valve provides sensitivity and fast response to the loop within its capacity. The large valve is a sort of an automatic bypass, which sets the capacity and has a limited frequency response. The large/small valve selection scheme shown at the bottom of Figure 6.1j differs from the one in Figure 6.1i only P ACTUATOR SELECTION Sections 6.3 and 6.4 discuss the applications and relative advantages of the different pneumatic, electric, digital, and hydraulic actuator designs. The popularity of the spring-and-diaphragm actuator is due to its low cost, its relatively high thrust at low air supply pressure, and its availability with fail-safe springs. By trapping the pressure in the diaphragm case, it can also be locked in its last position. It is available in various designs: springless, double diaphragm (for higher pressures), rolling diaphragm (for longer strokes), and tandem, which provides more thrust. One of the limitations of this design is the lack of actuator stiffness (resistance to rapidly varying hydraulic forces caused, for example, by flashing). For such applications, hydraulic or electromechanical (motor gear) actuators are preferred, although a stiffer spring (6 30 PSIG, which corresponds to 0.41 to 2.59 bars) in a spring-and-diaphragm unit is sometimes sufficient to correct the problem. Piston Actuators Linear piston actuators provide longer strokes and can operate at higher air pressures than the spring-and-diaphragm actuators. When used to operate rotary valves, the linear piston or spring-and-diaphragm actuator does not provide a constant ratio of rotation per unit change in air signal pressure. Therefore, the use of positioners is always advisable. Rotary piston actuators operate at higher air pressures and can provide higher torque, suitable for throttling large ball or butterfly valves. The double-acting version of this actuator does not have a positive failure position, but such a position can be added by extending the piston case and inserting a helical spring. For higher torque (over 1000 ft lb f, which corresponds to 1356 Nm), heavy-duty transfer linkages are required (Scotch yoke or rack and pinion); such units cannot be easily disassembled and maintained in the field. These actuators

15 6.1 Application and Selection of Control Valves 1059 Stem position Velocity limit Large-signal response Small-signal response Amplitude Input Output Time Time FIG. 6.1k Response of velocity-limited actuators to a step change (left) and to a high-amplitude sine wave (right) in the control signal, according to Shinskey. also require positioners, because the relationship between air signal change and resulting rotation is not linear. The use of positioners on fast loops can deteriorate the loop s performance, as it will need to be detuned. Actuator Speeds of Response In the family of pneumatic actuators, the spring-and-diaphragm actuators are the slowest (1 30 sec/stroke), spring-returned pistons with supply/exhaust at both ends of the piston are the fastest. Dual pistons can stroke valves in 0.5 sec. In the family of electric actuators, the slowest are the electromechanical motor-driven valves (5 300 sec/stroke); hydroelectric actuators can move at 0.25 in./sec (6 mm/sec) or faster if hydraulic accumulators are used. The fastest are the small, on/off solenoids, which can close in 8 12 ms, while throttling solenoids require about 1 sec to stroke. Valve actuators can be velocity limited because they cannot move faster than their maximum design speed. This is true of both electric and pneumatic motors or actuators. In case of the latter, maximum speed is set by the maximum rate at which air can be supplied or vented. If the full stroking (100%) of a valve takes 4 sec, then its velocity limit is 25% per second. Valve signal changes usually occur in small steps, and therefore the velocity limit does not represent a serious limitation because, for example, the time required to respond to a 5% change is only 0.2 sec. This is fast enough for most loops. Figure 6.1k illustrates the response of velocity-limited actuators to various types of control signals. Actuator speeds can be increased by enlarging the air flow ports and by installing booster relays. On on/off valves, the addition of a quick-dump valve (Figure 6.2o) will dramatically increase the venting rate. The dynamic performance of the actuator can also be affected by modifying the tare volume, pressure range, or dead band. In order to reduce the dead band, one usually needs to modify the piston seals, linkages, or rack and pinion connections. Most valve actuators display some dead band or hysteresis band due to packing friction (Figure 6.1l). This can cause instability if the change in the control signal is small enough to fall within the hysteresis band width. Actuator Power The actuator is sized on the basis of the power or thrust required to overcome the unbalanced forces in the valve body and the seating force, and on the basis of the stiffness necessary for stability. In pneumatic actuators, the thrust is a function of piston or diaphragm area times air pressure. While the control signal is usually 3 15 PSIG (0.2 to 1.0 bar), the actuating pressure can be as high as the air supply pressure, if positioners or amplifier relays are installed. If their costs can be justified, electrohydraulic actuators will give the highest power and speed of response. Valve Failure Position In pneumatic actuators, the fail-safe action can be provided by a spring or from an air reservoir, but the latter represents added expense, complexity, and space requirements. Electric actuators are usually more expensive, except for such designs as the spring-loaded or modulating solenoid valves discussed in Section 6.3. Electric actuators are therefore most often used where air is not available, where the thrust required is less than 1000 lb f (4448 N), where it is acceptable to have the valve fail in its last position, and where slow response is not a drawback. Like most generalizations, however, these are not completely true: Electromechanical motor and hydraulic actuators are available with high thrusts and can be provided with positive failure. Hydraulic actuators are also available with Stem position Dead band Motor pressure FIG. 6.1l Dead band in the valve (left) can result in limit cycling (right) when the loop is closed. Flow Time

16 1060 Control Valve Selection and Sizing fast response speeds. To gain an in-depth understanding of the capabilities of actuators, refer to Sections 6.3 and 6.4. It is the responsibility of the process control engineer to specify the valve failure position. It is the general practice to fail energy supply valves (steam, hot oil, and so on) closed and energy-removing valves (cold or chilled water) open. The flow sheet abbreviations can be FC (fail closed), FI (fail indetermined), FL (fail in last position), and FO (fail open). Spring-loaded actuators are the most convenient means of providing FC or FO action, while two-directional air or electric motors will naturally tend to fail in their last positions. In addition to anticipating the consequences of actuator power failure, one should also consider the results of other component failures, such as the spring, diaphragm, piston, and so on. When such failures occur, the ultimate valve position will not be a function of the actuator design, but of the process fluid forces acting upon the valve itself. The choices are FTO (flow to open), FTC (flow to close), or FB (friction bound tends to stay in last position). FTO action is available with globe valves. FTC action can be obtained from butterfly, globe, and conventional ball valves. Rotary plug, floating ball, and segmented ball valves tend to be friction bound, with the flow direction possibly affecting the torque required to open the valve. POSITIONERS The positioner is a high-gain plain proportional controller that measures the valve stem position (to within 0.1 mm), compares that measurement to its set point (the controller output signal), and, if there is a difference, corrects the error. The open-loop gain of positioners ranges from 10 to 200 (proportional band of %), and their periods of oscillation range between 0.3 and 10 sec (frequency response of Hz). In other words, the positioner is a very sensitively tuned, proportional-only controller. Positioners that are electronically and digitally controlled, or are intelligent and are capable of self-diagnostics, communication on fieldbuses, and other advanced features, are not discussed here, because they are described in detail in Section When to Use Positioners The main purpose of having a positioner is to guarantee that the valve does, in fact, move to the position where the controller wants it to be. The addition of a positioner can correct for many variations, including changes in packing friction due to dirt, corrosion, or lack of lubrication; variations in the dynamic forces of the process; sloppy linkages (dead band); or nonlinearities in the valve actuator. The dead band of a valve/actuator combination can be as much as 5%; when a positioner is added, it can be reduced to less than 0.5%. It is the job of the positioner to protect the controlled variable from being upset by any of the above variations. In addition, the positioner can also allow for split-ranging the controller signal between more than one valve, can increase the actuator speed or thrust by increasing the pressure or volume or the actuator air signal, and can modify the valve characteristics by cams or electronic function generators. While the above positioner capabilities can be convenient, they can also be obtained without the use of positioners. For example, split-ranging can also be done by the use of splitranged valve springs or by multiple/biasing relays in the air signal line to the valve. Similarly, increasing the speed/thrust of the valve can be achieved by booster relays, and changes to the control valve characteristics can be obtained, not only by replacing the plug, but also by pneumatic or electronic characterizing of the controller signal. Therefore, these reasons do not necessitate the use of positioners. Actuators without springs always require positioners. When a valve is in remote manual (open loop) operation, it will always benefit from the addition of a positioner, because a positioner will reduce the valve s hysteresis and dead band while increasing its response. When the valve is under automatic (closed loop) control, the positioner will be helpful in most slow loops, which control analytical properties, temperature, liquid level, blending, slow flow, and large volume gas flow. A controlled process can be considered slow if its period of oscillation is three times the period at which the positioned valve oscillates. In such installations as B in Figure 6.2b, the addition of a positioner increases the open-loop gain and, therefore, the loop response. As a consequence, the tuning of the controller could also be improved by increasing the gain (making the proportional band narrower) and adding more repeats per minute in the integral setting. When Not to Use Positioners In the case of fast loops, positioners are likely to degrade loop response, contribute to proportional offsets, and cause limit cycling (fast flow, liquid pressure, small volume gas pressure). The positioner in effect is the cascade slave of the loop controller. In order for a cascade slave to be effective, it must be faster than the speed at which its set point, the master output signal, can change. The rules of thumb used in this respect suggest that the time constant of the slave should be ten times shorter (open-loop gain ten times higher) than that of the master and the period of oscillation of the slave should be three times shorter (frequency response three times higher) than that of the primary. The criteria for positioners need not be this stringent, but still, it is recommended not to use positioners if the positioned valve is slower than the process variable it is assigned to control. A controlled process can be considered fast if its period of oscillation is less than three times that of the positioned valve. In such situations, the positioned valve is one of the slowest components in the loop and, therefore, slows down the load (by limiting the open-loop gain of the loop and lengthening the period of oscillation). Part A in Figure 6.2b illustrates such a situation, where the loop can be tuned more tightly (higher gain, more repeats/minute) and, therefore, responds better without a positioner. It might also be noted that after a

17 6.1 Application and Selection of Control Valves 1061 new steady state is reached, the positioned installation gives more noisy control because of the hunting and limit cycling of the positioner, which cannot keep up with the process. Some will argue that all loops can be controlled using positioned valves if they are sufficiently detuned. This is true, but detuning means that the controller is made less effective (the amount of proportional and integral correction is reduced), which is undesirable. A 0.2 gain (500% proportional) setting on a level controller means that the tank can be flooded or drained before the controller fully strokes the valve. Also, there are cases where the gain must be so low (the proportional band so wide) that it is outside the available PB setting range of the controller. Positioners to Eliminate Dead Band All valves and dampers will display some dead band because of friction in their packing, unless positioners are used. Whenever the direction of the control signal is reversed, the stem remains in its last position until the dead band is exceeded, as shown in Figure 6.1l. This figure on the right shows that if a sine wave control signal is driving the valve actuator (motor), it produces a stem motion that is distorted and shifted in phase. This phase shift, when combined with the integrating characteristic of certain processes and with the reset action of a controller, causes the development of a limit cycle. According to Shinskey, widening the proportional band will not dampen the oscillation, but only make it slower. The limit cycle will not appear if a proportional-only controller is used, and if the process has no integrating element. Processes that are prone to limit cycling in this way are liquid level, volume (as in digital blending), weight (not weight-rate), and gas pressure all of which are related to the integrals of flow. Whenever one intends to control such a process with a proportional and integral (PI) controller, the use of positioners should be considered. In case of level control, one can accomplish the same goal by using a plain proportional controller and a booster or amplifier instead of a positioner. Positioners in general will eliminate the limit cycle by closing a loop around the valve actuator. Positioners will also improve the performance of valves on slow processes, such as ph or temperature. On the other hand, dead band caused by stem friction should not be corrected by the use of positioners on fast loops, such as flow or fast pressure. The positioner s function as a cascade slave, as was explained earlier, can cause oscillation and cycling on fast loops if the controller cannot be sufficiently detuned (minimum gain is not low enough). Similarly, negative force reactions on the plug require an increase in actuator stiffness and not the addition of a positioner. Actuator stiffness can be improved by increasing the operating air pressure or by using hydraulic actuators. Split-Range Operation The use of positioners for split-range applications is usually accepted regardless of the speed of the process. This is not entirely logical, because on fast loops the control performance can be degraded by the use of positioners. In such cases, some instrument engineers do discourage the use of positioners to implement split-ranging. Instead, they recommend gain-plus-bias relays so that the positioner (the lessaccurate device) will operate over its full range (Figure 6.1j). This also eliminates the need for a special calibration. One can also consider accomplishing the split-range operation through the use of different spring ranges in the valve actuators. In addition to the standard 3 to 15 PSIG (0.2 to 1 bar) range spring, valves can also be obtained with other spring ranges. These include 3 7 PSIG ( bar), 4 8 PSIG ( bar), 5 10 PSIG ( bar), 7 11 PSIG ( bar), 8 13 PSIG ( bar), and 9 13 PSIG ( bar). Lastly, if split-range positioners are installed on fast processes, the resulting degradation of control quality can be limited by adding a restrictor or an inverse derivative relay in the control signal to it and, thereby, artificially making the controller appear to be slower than it really is. This technique is not highly recommended (except to reduce wear and tear on the valve in noisy loops), because restrictors are prone to plugging or maladjustment and because they both degrade the loop performance. Accessories If the need is to increase the speed or the thrust of the actuator, it is sufficient to install an air volume booster or a pressure amplifier relay, instead of using a positioner. Boosters will give better performance than positioners on fast processes, such as flow, liquid pressure, or small-volume gas pressure control, and they will not be detrimental (nor will offer advantages) if used on slow processes. If the reason for adding a positioner is to alter or modify the control valve characteristics, this is not a valid justification on fast processes, because this aim can be satisfied by the use of dividing or multiplying relays in the controller output, which will not degrade the quality of control (see Section 6.7). PROCESS APPLICATION CONSIDERATIONS In selecting control valves, the properties of the process fluid must be fully considered. The process data should be carefully and accurately determined because even small variations in temperature or pressure can cause flashing or cavitation. Considerations include such obvious variables as pressure, temperature, viscosity, slurry, or corrosive nature, or the less obvious factors of flashing, cavitation, erosion, leakage, sterilization, and low flow rates. These are discussed in the paragraphs that follow below. Pressure Considerations The available design pressures for each valve type are listed in the feature summaries in the front of Sections

18 1062 Control Valve Selection and Sizing Selfdrag trim FIG. 6.1m High-pressure valve designs. In selecting the control valve for a particular application, one should pay particular attention to high pressure, high differential pressure, and vacuum services. These will be discussed in the paragraphs below. High-Pressure Services When designing valves for highpressure services, the following features are of particular importance: 1. Increased physical strength 2. Selection of erosion-resistant material 3. Use of special seals Valve bodies can usually withstand higher pressures than can the piping. Valve bodies for high-pressure services are usually forged to provide homogeneous materials free of voids and with good mechanical properties. The loads and stresses on the valve stem are also high. For this reason, higher strength materials are used with increased stem diameters. As shown in Figure 6.1m, the stems are usually kept short and are well-guided. High-pressure services will also increase the probability of noise, vibration, and cavitation, which will be discussed in later paragraphs. High Differential Pressure With high flow rates and high pressure drops, a large amount of energy is dissipated in turbulence. A fraction of this energy is radiated as noise (see Section 6.14). For most (but not all) gases and conditions, one result of the high pressure drop can be a very low outlet temperature. It is not always proper to use the gas laws to predict these temperatures, and instead, actual thermodynamic properties should be used. The very low temperature will cause some valve materials to become brittle, and careful selection of alloys and other materials is always required. Depending on the gas involved and other materials in the flowing fluid, hydrates or other solids may form in the valve. Liquid droplets may develop and cause erosion. It is necessary to investigate for any peculiarity of the flowing fluid. A high-velocity jet leaving the valve can erode downstream piping. Very high forces are developed on the valve body and internal parts and can cause valve instability. With the change in magnitude and, often, direction of the fluid, substantial reaction forces are developed. Serious damage may result if the valve and piping are not properly restrained. High operating pressure frequently involves high pressure drops. This usually means erosion, abrasion, or cavitation at the trim. These will be discussed in detail in the coming paragraphs, but it should also be mentioned here that cavitation and erosion resistance are usually not properties of the same metal. Materials resistant to erosion and abrasion include 440C stainless steel, flame-sprayed aluminum oxide coatings (Al 2 O 3 ), and tungsten carbide. On the stem, where the unit pressure between it and the packing is high, it is usually sufficient to chrome-plate the stem surface to prevent galling. Special self-energizing seals are used with higher pressure valves (above 10,000 PSIG, or 69 MPa, service) so that the seal becomes tighter as pressure rises. Popular body seal designs for such service include the delta ring closure and the Bingham closure (Figure 6.1m). As discussed in more detail in Section 6.19, the selfenergizing seals are used in connecting the high-pressure valves into the pipeline. These designs depend on the elastic or plastic deformation of the seal ring at high pressures for self-energization. Special packing designs and materials are also required in high-pressure service, because conventional packing would be extruded through the clearances. To prevent this, the clearance between stem and packing box bore is minimized, and extrusion-resistant material, such as glass-impregnated Teflon, is used for packing. Some of the likely causes of valve failure in high pressure drop services include: 1. Elastomer elements in Saunders or pinch valves (particularly if they fail to open) can be ruptured. 2. The stem thrust can be excessive for globe valves. If globe valves are flow to close, high p can damage the seat or prevent the actuator from opening the valve. If they are flow to open, they might open against the actuator.

19 6.1 Application and Selection of Control Valves High p can exceed the capabilities of plug or floating ball valves, and it can bend the shafts of butterfly valves, damage the bearings of trunion-type ball valves, or damage the seat of floating ball valves. 4. Pressure cycling generated by positive-displacement pumps can cause bolt fatigue if the number of cycles is excessive. Vacuum Service Low pressures can prevent some pressure-energized seals from properly operating, or they can cause leakage. In some processes, the in-leakage from the atmosphere results in overloading the vacuum source; in others, it represents a contamination that cannot be tolerated. Potential leakage sources include all gasketed areas and, to an even greater extent, the locations where packing boxes are used to isolate the process from the surroundings. For vacuum service, valves that do not depend on stuffing boxes to seal the valve stem generally give superior performance. Such designs include Saunders valves and pinch valves. These designs, unfortunately, are limited in their application by their susceptibility to corrosion and their temperature and control characteristics. Their applicability to vacuum service is further limited by their design. The jacketed pinch valve versions, for example, require a vacuum source on the jacket side for proper operation, and the mechanically operated pinch and Saunders designs are limited in their capability to open the larger-size units against high vacuum on the process side. The vacuum process tends to keep the valve closed, and this can result in the diaphragm s breaking off the stem and rendering the valve inoperative. For services requiring high temperatures and corrosionresistant materials, in addition to good flow characteristics and vacuum compatibility, conventional globe valves can be considered, with special attention given to the type of packing and seal used. One approach to consider is the use of double packing, as shown in Figure 6.1n. The space between the two sets of packing is evacuated so that air leakage across the upper Connection to vacuum source FIG. 6.1n Double packing is used to seal the valve stem if the process is under vacuum. packing is eliminated. The vacuum pressures on the two sides of the lower packing are approximately equal, and therefore there is no pressure differential to cause leakage across it. Usually, the space between the two packings is exposed to a slightly higher vacuum than the process, so that no in-leakage is possible. Double packing provides reasonable protection against in-leakage under vacuum, but it does not relieve the problems associated with corrosion and high temperatures. When all three conditions exist (vacuum, corrosive flow, and high temperature), the use of bellows seals (discussed in Section 6.19) can be considered. The bellows are usually made of 316 stainless steel and are tested by mass spectrometers for leakage. They not only prevent air infiltration but also can protect some parts of the bonnet and top-works from high temperature and corrosion. Like all metallic bellows, these too have a finite life, and therefore it is recommended that a secondary stuffing box and a safety chamber be added after the bellows seal. A pressure gauge or switch can be connected to this chamber between the bellows and the packing to indicate or warn when the bellows seal begins to leak and replacement is necessary. Considerations similar to those noted for high-vacuum service would also apply when the process fluid is toxic, explosive, or flammable. High-Temperature Service The temperature limitations of each valve design are listed in their feature summaries in Sections All process conditions involving operating temperatures in excess of 450 F (232 C) are considered high temperature. The maximum temperatures at which control valves have been successfully installed are up to 2500 F (1371 C). High operating temperatures necessitate the review of at least three aspects of valve design: 1. Temperature limitations of metallic parts 2. Packing temperature limitations 3. Use of jacketed valves Metallic Parts High temperatures can cause galling, can affect clearances, and can soften hardened trims. Temperature cycling can cause thermal ratcheting and stress, resulting in body or bolting rupture if the rate or frequency of temperature cycles is high. The high operating temperatures are considered in selecting materials for both the valve body and trim. For the body, it is suggested that bronze and iron be limited to services under 400 F (204 C), steel to operation below 850 F (454 C), and the various grades of stainless steel, Monel, nickel, or Hastelloy alloys to temperatures up to 1200 F (649 C). For the valve trim, 316 stainless steel is the most popular material, and it can be used up to 750 F (399 C). For higher temperatures, the following trim materials can be considered: 17 4 ph stainless steel (up to 900 F, or 482 C), tungsten

20 1064 Control Valve Selection and Sizing carbide (up to 1200 F, or 649 C), and Stellite or aluminum oxide (up to 1800 F, or 982 C). At high temperatures, the guide bushings and guideposts tend to wear excessively, and this can be offset by the selection of proper materials. Up to 600 F (316 C), 316 stainless steel guideposts in combination with 17 4 ph stainless steel guide bushings give acceptable performance. If the guideposts are surfaced with Stellite, the above combination can be extended up to 750 F (399 C) service. At operation over 750 F (399 C), both the posts and the bushings require Stellite. Packing Designs The ideal packing provides a tight seal while contributing little friction resistance to stem movement. With TFE packing, which is industry standard, the required stem finish is between 6 and 8 µ in. RMS (0.15 and 0.2 µm). A common packing design might consist of Teflon V-rings, which are discussed in more detail in Section Double packing with leak-off connection in between can be used on toxic or vacuum service (Figure 6.1n). On toxic services, an added seal can be provided by the injection of high-viscosity silicone or plastic packing, but this is rarely done. Solid rings or ribbons of pure graphite (Graphoil), while more expensive than Teflon, are also popular because they are suited for higher temperatures. On the other hand, they require more loading to energize the packing than does Teflon, and the resulting friction can cause stem lockup. On less demanding services, O-rings are also used, but not frequently because under pressure these elastomers will absorb gases, which can destroy the O-ring when depressurized rapidly. Metallic bellows-type seals are seldom used because of their pressure limitations and unpredictable lives. On toxic services, they should be provided with automatically monitored guard packing for security. The bonnets are usually flanged and are extended on hot or cold services so as to bring the operating temperature of the packing closer to the ambient. Screwed bonnets are not recommended for severe duty, and welded bonnets are not used at all, except as an extreme precaution on hazardous services. The sliding stems can sometimes drag atmospheric contaminants or process materials into the packing, but this can be overcome by close tolerance guide bushings or wiper rings. Packing contamination is less likely with rotary valves. In case of LPG, the packing should be isolated from outboard roller bearings and the intervening space vented to protect the lubricant. Packing Limitations Packing and bonnet designs in general were discussed in the previous paragraph and will also be covered in Section Here only their suitability for high-temperature service is reviewed. The packing temperature limitation for most nonmetallic materials is in the range of F ( C), the maximum temperature for metallic packing is around 900 F (482 C), and Teflon should not be exposed to temperatures above 450 F (232 C). Pure graphite (Graphoil) can be used from 400 to 750 F ( 240 to 399 C) in oxidizing service and up to 1200 F (649 C) in nonoxidizing service, with an ultimate potential of 3000 F (1649 C). Bonnets can be screwed, welded, or flanged. Screwed bonnets are not recommended for high-temperature service. Finned bonnet extensions were used in the past on hightemperature services, when packing material capabilities were more limited. These finned designs were not effective, and therefore with the introduction of Graphoil, their use was largely discontinued on rotary valves. For sliding stem valves, Teflon V-rings within extension bonnets are frequently selected and used up to 850 F (454 C). On high-temperature services, it can be effective to mount the bonnet below the valve. In liquid service, with the bonnet above the valve, the packing is exposed to the full process temperature due to the natural convection of heat in the bonnet cavity. If the bonnet is mounted below the valve, no convection occurs, and the heat from the process fluid is transferred by conduction in the bonnet wall only. Therefore, by this method of mounting, the allowable process temperature can be substantially increased in some processes. This is not the case for all applications, because some processes do not generate effective condensate seals, and in other services the liquidvapor interface line can cause metallurgy problems. Figure 6.1o provides a method for determining packing temperature, in gas or vapor service, with the bonnet above the valve for one particular valve design. With vapor service, it is likely that vapors will initially condense on the wall of the bonnet, lowering the temperature to the saturation temperature of the process fluid, but in some cases the heat conducted by the metallic bonnet wall will be sufficient to prevent this condensation from occurring. In short, the packing temperature will be at or above saturation temperature (T s ) in vapor service, but if the bonnet is mounted below the valve, the packing temperature is substantially reduced, due to the accumulated condensate. If the bonnet is below the valve, the relationship between process and packing temperature is not affected by the phase of the process fluid. In case of ball or plug valves with double-sealing, it is important to vent the space between the seals to the line, so that damage will not be caused by thermal expansion. Jacketed Valves A number of control valve designs are available with heat-transfer jackets. Others can be traced or jacketed by the user. Jacketed valves can be installed for either cooling or heating. When a cooling medium is circulated in the jackets, this is usually done to lower the operating temperature of the heat-sensitive working parts. Such jacketing is particularly concentrated on the bonnet, so that the packing temperature is reduced relative to the process. For certain operations at very high temperatures, intermittent valve operation is recommended, such that when the valve is closed it is cooled by the jacket, and when it is opened, it is kept open only long enough to prevent temperature equalization between the valve and the process.

21 6.1 Application and Selection of Control Valves 1065 Inlet pressure psia* A C Air steam B D.1 0 E F G H Process fluid temperature F * J Tp less than 450 F (232 C) (Use teflon packing) 450 F (232 C) Tp more than 450 F (232 C) (See instruction 3b) Tp -Packing temperature *See Section A.1 for SI units Bonnet Characteristics Instructions 1. Determine constants from table at right corresponding to bonnet Valve Size Bonnet Material * Bonnet Factors Standard Bonnet Extension Bonnet Radiation Fin Bonnet selection to locate points on scales D and F. 1" CS D Solve nomograph Key: F Line Up Straight Edge On Locate Intersection On SS D A TO B C F C TO D E 1 1 / 2 " CS D E TO F G F G TO H J SS D F a. If Tp 450 F (232 C), use Teflon packing. 3b. If Tp > 450 F (232 C), either use high-temperature packing or select 2" CS D another bonnet with smaller F value and recheck packing F temperature. SS D F " CS D F SS D F " CS D F SS D F " CS D F SS D F " CS D F SS D F * CS: carbon steel, SS: stainless steel. FIG. 6.1o Nomograph for packing temperature determination on hot gas or vapor services. (From Determination of Proper Bonnet and Packing for High-Temperature Processes, R.F. Lytle, Fisher Controls, Emerson Process Management.)

22 1066 Control Valve Selection and Sizing Heating medium inlet Drain FIG. 6.1r Special steam jacket for retrofit installation on valve. FIG. 6.1p Jacketed control valve for high-temperature service. Heating jackets with steam or hot oil circulation are used to prevent the formation of cold spots in the more stagnant areas of the valve or where the process fluid otherwise would be exposed to relatively large masses of cold metal. Figure 6.1p shows one of these valves, designed to prevent localized freezing or decomposition of the process fluid due to cold spots. Many standard globe pattern valves can be fitted with a jacket to allow heating or cooling as required (Figure 6.1q). Normally, these jackets are for services requiring steam. FIG. 6.1q Steam-jacketed valve. (Courtesy of Flowserve Corp.) Dowtherm or similar heating fluids prevent solidification or crystallization of certain fluids. Often, the manufacturer can provide these jackets, but where this is not available, there are firms that specialize in designing and installing such jackets on valves and other equipment. These special jackets can be designed either to weld to the valve as a permanent fixture (Figure 6.1r) or as separate devices bolted or clamped to the valve body. In the latter case, it may be necessary to use a heat-transfer paste between jacket and valve body to give efficient transfer by eliminating the air gap. Low-Temperature Service Cryogenic service is usually defined as temperatures below 150 F ( 101 C). Properties of some cryogenic fluids are listed in Table 6.1s. Valve materials for operation at temperatures down to 450 F ( 268 C) include copper, brass, bronze, aluminum, 300 series stainless steel alloys, nickel, Monel, Durimet, and Hastelloy. The limitation on the various steels falls between 0 and 150 F ( 17 and 101 C), with cast carbon steel representing 0 F ( 17 C) and 3 1 / 2 % nickel steel being applicable to 150 F ( 101 C). Iron should not be used below 0 F ( 17 C). Conventional valve designs can be used for cryogenic service with the proper selection of construction materials and with an extension bonnet (as described in detail in Section 6.19) to protect the packing from becoming too cold. The extension bonnet is usually installed vertically so that the boiled-off vapors are trapped in the upper part of the extension, which provides additional heat insulation between the process and the packing. If the valve is installed in a horizontal plane, a seal must be provided to prevent the cryogenic liquid from entering the extension cavity. When the valve and associated piping are installed in a large box filled with insulation ( cold box ), this requires an unusually long extension in order to keep the packing box in a warm area.

23 6.1 Application and Selection of Control Valves 1067 TABLE 6.1s Properties of Cryogenic Fluids Methane Oxygen Fluorine Nitrogen Hydrogen Helium Boiling point ( K) ( C) 259 ( 162) 297 ( 183) 307 ( 188) 320 ( 196) 423 ( 253) 452 ( 269) Critical temperature ( F) ( C) 117 ( 83) 181 ( 118) 200 ( 129) 233 ( 147) 400 ( 240) 450 ( 268) Critical pressure (psia) [bar(a)] Heat of vaporization at boiling point (BTU/lbm) (J/kg) 219 ( ) 92 ( ) 74 ( ) 85 ( ) 193 ( ) 9 ( ) Density (lbm/ft 3 ) gas at ambient conditions (kg/m 3 ) (0.673) (1.33) (1.57) (1.153) (0.080) (0.16) Vapor density at boiling point (1.778) (4.74) (4.614) (1.346) 1.06 (16.98) Liquid density at boiling point 26.5 (424.5) 71.3 (1142) 94.2 (1509) 50.4 (807.4) 4.4 (70.5) 7.8 (125) Cryogenic Valves A special design variation on the globe valve is the cryogenic valve. Section 6.19 shows a number of cryogenic service designs, including the Y-valve design. The most common design for this service is shown in Figure 6.1t. This design is specific to cryogenic (down to 454 F, or 270 C) service and no other. Body configurations are straight through, as shown, or angle body. Because of the need for Charpy impact for the extremely cold service, the materials are limited to bronze and austenitic stainless steels such as 304, 316, and 316L. Normally, the valves are welded into the piping or soldered in some cases with bronze. Small valves, 1 in. (25 mm) and 2 in. (50 mm), can be socket weld or butt weld, but butt weld in larger sizes up to the 10 in. (250 mm) maximum are available. Body ratings through ANSI 600 and flange ends, either integral or separable, are available depending upon manufacturer. Seat rings may be integral hard-faced with Stellite, screwed-in metal, or soft seat for tight shut-off. The metal shutoffs will be ANSI Class III or IV, depending upon manufacturer, and the soft seat using Teflon or Kel-F will provide Class VI. Upper packing Bonnet Bonnet gasket Cold box extension Upper guide Packing spacer Bonnet bolts Bonnet flange Plug Cold Box Valves Cold box valves are designed to have a low body mass for fast cool-down and reduced heat transfer. The long extended bonnet is provided with a plug stem seal to minimize liquid refluxing into the bonnet and packing area, thereby minimizing the heat loss due to conduction and convection. Actually, the small amount of liquified gas passing into the bonnet vaporizes and provides a vapor barrier between the liquified gas and the packing area. In addition, the pressure resulting from the vaporization of the liquid prevents additional liquid from passing into the bonnet area. Excess pressure vents back into the body. It is possible to fit these valves with vacuum jackets where the application requires this additional insulation. Plug seal Body Integral seat FIG. 6.1t Cold box valve with weld ends and welded bonnet. (Courtesy of Flowserve Corp.)

24 1068 Control Valve Selection and Sizing F L SYM disc Butterfly valve Seat Seat downstream upstream Globe valve Ball Gate valve Valve sizes 2" 36" BFV C d = C v /d 2 FIG. 6.1u Vacuum jacketing of cryogenic valve. Features that are desirable for cryogenic valves include small body mass, which ensures a small heat capacity and, therefore, a short cool-down period. In addition, the inner parts of the valve should be removable without removing the body from the pipeline, and if the valve is installed in a cold box, no leakage can occur inside this box because there are no gasketed parts. The most effective method of preventing heat transfer from the environment into the process is by vacuum jacketing the valve and piping (Figure 6.1u). The potential leakage problems are eliminated by the fact that there are no gasketed areas inside the jacket. For cryogenic services where tight shut-off is required, Kel-F has been found satisfactory as a soft seat material because cold can cause many other elastomer materials to harden, set, or shrink. If used as seals, this can cause leaking. Cavitation and Erosion The cavitation and erosion phenomena in connection with globe valves are discussed in detail in Section The choking effect of cavitation and its influence on valve sizing is covered in Section Sections 6.14 and 6.15 describe how the liquid pressure recovery factor (F L ) is related to the ratio between the valve pressure drop and the difference between the inlet and the vena contracta pressure. As was shown in Section 6.15, the cavitation coefficient K c is the ratio between the valve pressure drop at which cavitation starts and the difference between the inlet and the vapor pressure of the application. Section 6.15 also shows how the F LP factor can be calculated if the valve is placed within reducers, and it also shows the valve pressure differential at which choking starts can be calculated. The allowable maximum p before cavitation begins is p = K c (p 1 p v ). As the F L and K c values of the different valve FIG. 6.1v In this figure, the pressure recovery factor (F L ) of different valve designs is shown as a function of their discharge coefficients (C d values, which in the metric system are defined as ε = K v / DN 2 ) as these valves are throttled from their full open positions. 2 designs drop, the probability of cavitation increases. F L and K c values for fully open valves are also given in Section 6.15, and F L values for throttled valves are given in Figure 6.1v. The high flow velocity at the vena contracta of the valve is reached by obtaining its energy to accelerate from the pressure energy of the stream. This causes a localized pressure reduction that, if it drops below the fluid s vapor pressure, results in temporary vaporization. (Fluids form cavities when exposed to tensions equal to their vapor pressure.) Cavitation only occurs when the pressure in the vena contracta region drops below the vapor pressure of the flowing fluid. The vapor pressure is a function of fluid temperature and chemical structure. Cavitation damage always occurs downstream of the vena contracta when pressure recovery in the valve causes the temporary voids to collapse. Destruction is due to the implosions that generate the extremely high-pressure shock waves in the substantially noncompressible stream. When these waves strike the solid metal surface of the valve or downstream piping, the damage gives a cinder-like appearance. Cavitation is usually coupled with vibration and a sound like rock fragments or gravel flowing through the valve. Cavitation damage always occurs downstream of the vena contracta at the point where the temporarily formed voids implode. In case of flow-to-open valves, the destruction is almost always to the plug and seldom to the seat. Methods to Eliminate Cavitation Because no known material can remain indefinitely undamaged by severe cavitation, the only sure solution is to eliminate cavitation completely. Even mild cavitation over an extended time will attack the metal parts upon which the bubbles impinge. Hard materials survive longer, but they are not an economical solution except for services with mild intermittent cavitation. Cavitation damage also varies greatly with the type of liquid flowing.

25 6.1 Application and Selection of Control Valves 1069 P 1 P 2 Flow Pressure P 1 = Inlet pressure P 2 = Outlet pressure P v = Vapor pressure P vc = Venacontracta pressure Distance Vaporization starts as pressure drops below the vapor pressure of the flowing fluid Cavitation starts as bubbles collapse when pressure rises above P v Microjet FIG. 6.1w Cavitation occurs when downstream of the vena contracta the pressure rises. When it reaches the vapor pressure of the process fluid, the vapor bubbles implode and release powerful microjets that will damage any metallic surface in the area. The greatest damage is caused by a dense pure liquid with high surface tension (e.g., water or mercury). Density governs the mass of the microjet stream, illustrated in Figure 6.1w, and surface tension governs the more important jet velocity. Mixtures are least damaging, because the bubble cannot collapse as suddenly. As the pressure increases, partial condensation in the bubble changes the vapor composition, leaving some vapor to slow the collapse. Some applications of cavitating mixed hydrocarbons show no mechanical damage or high noise level. Cavitation can be reduced or eliminated by several methods, listed in the following paragraphs. Revising the Process Conditions A reduction of operating temperature can lower the vapor pressure sufficiently to eliminate cavitation. Similarly, increased upstream and downstream pressures, with p unaffected, or a reduction in the p can both relieve cavitation. Therefore, control valves that are likely to cavitate should be installed at the lowest possible elevation in the piping system and operated at minimum p. Moving the valve closer to the pump will also serve to elevate both the up- and downstream pressures. If cavitating conditions are unavoidable, then it is preferred to have not only cavitation but also some permanent vaporization (flashing) through the valve. This can usually be accomplished by a slight increase in operating temperature or by decreasing the outlet pressure. Flashing eliminates cavitation by converting the incompressible liquid into a compressible mixture. Revising the Valve Design Where the operating conditions cannot be changed, it is logical to review the type of the valve in terms of its pressure recovery characteristics. The more treacherous the flow path through a particular valve, the less likelihood exists for cavitation. Inversely, the valves most likely to cavitate are the high recovery valves (ball, butterfly, gate) having low F L and F c coefficients (Section 6.15). Figure 6.1x illustrates some of the ways available to eliminate cavitation. Figure 6.1y shows a number of anticavitation valve designs that combine multiple-port and multiple-flow-path features. If cavitation is anticipated, the engineer should select valves with low recovery and, therefore, high F c and F L coefficients. Different valve designs react differently to the effects of cavitation, depending upon where the bubbles collapse. If the focus is in midstream, materials may be unaffected. For example, in the Swiss cheese -type design, small holes in the skirt or cage are arranged in pairs on opposite sides of the centerline of the valve. Streams from opposing holes impinge on each other, causing the cavities to collapse in the liquid pool (theoretically). This method, illustrated in Figure 6.1z has been used successfully for mild cavitation.

26 1070 Control Valve Selection and Sizing P P 1 P 1 P 1 P P P P 2 Valve with less recovery higher F L & K C P 2 Two valves in series P 2 Reduce valve P P V P V P V P VC P VC P VC P VC P 1 P 1 P P P 2 Move valve closer to pump or to lower elevation P 2 Lower the temperature P V P V P VC P VC P V FIG. 6.1x The pressure profiles shown in dotted lines illustrate some of the options available to the process control engineer to eliminate cavitation. Labyrinth-type valves avoid cavitation by a very large series of right-angle turns with negligible pressure recovery at each turn, but the narrow channels are subject to plugging if particulate matter is in the stream (Figure 6.1aa). The multistep valves at the bottom of Figure 6.1aa can avoid cavitation by replacing a single and deep vena contracta, as would occur in a single-port valve, with several small vena contracta points as the pressure drop is distributed between several ports working in series. If the vapor pressure of the process fluid is below the outlet pressure of the valve (Condition A), this valve is likely to work. On the other hand, if P v is greater than P 2 (Condition B), this valve is likely to cavitate in its noted intermediate port. One might note that if Condition B occurred in the conventional valve (noted by the dotted line), no cavitation would occur, because some of the vapor formed at the vena contracta would never recondense but would stay in the vapor state (flashing). Therefore, these multistep valves are not recommended for flashing applications. Gas Injection Another valve design variation that can alleviate cavitation is based on the introduction of noncondensible gases or air into the region where cavitation is anticipated. The presence of this noncompressible gas prevents the sudden collapse of the vapor bubbles as the pressure recovers to values exceeding the vapor pressure, and instead of implosions, a more gradual condensation process occurs. As shown in Figure 6.1bb, the gas may be admitted through the valve shaft or through downstream taps on either side of the pipe, in line with the shaft and as close to the valve as possible. Because the fluid vapor pressure is usually less than atmospheric, the air or gas need not be under pressure. Revising the Installation In order to eliminate cavitation, it is possible to install two or more control valves in series. Cavitation problems can also be alleviated by absorbing some of the pressure drop in restriction orifices, chokes, or in partially open block valves upstream or downstream to the valve. The amount of cavitation damage is related to the sixth power of flow velocity or to the third power of pressure drop. This is the reason why reducing p by a factor of two, for example, will result in an eightfold reduction in cavitation destruction. In some high-pressure let-down stations, it might not be possible to completely eliminate cavitation accompanied by erosion or corrosion. In such installations, one might consider the use of inexpensive choke fittings (shown in Figure 6.1cc) instead of (or downstream to) control valves. A single, fixed-opening choke fitting is applicable only when the process flow rate is relatively constant. For variableflow applications, one can provide several choke fittings of different capacities isolated by several full bore on/off valves, providing a means of matching the process flow with the opening of the required number of chokes. If the chokes discharge into the vapor space of a tank, this will minimize cavitation damage because the bubbles will not be collapsing near to any metallic surfaces. Material Selection for Cavitation While no material known today will stand up to cavitation, some will last longer than others. Table 6.1dd shows that the best overall selection for cavitation resistance is Stellite 6B (28% chromium, 4% tungsten, 1% carbon, 67% cobalt). This is a wrought material and can be welded to form valve trims in sizes up to 3 in. (75 mm). Stellite 6 is used for hard-facing of trims and has the same

27 6.1 Application and Selection of Control Valves 1071 Cage retainer Balance holes in plug Plug Cavitrol cage Cage retainer Valve plug Valve stem Bonnet Lower stem guide Bonnet flange Bolting Seal ring Plug seals Cage Cartridge Seat ring Packing spacer Packing Sleeve gasket Sleeve Bonnet gasket Plug Seat gasket Body Frictional losses Seat ring Seat ring Cavitrol trim element and valve installation. (Courtesy of Fisher Controls, Emerson Process Management.) Sudden expansion, turbulent mixing, mutual impingement Inner stage Middle stage Outer stage Channel stream trim detail and valve installation. (Courtesy of Flowserve Corp.) VRT (Variable Resistance Trim) element, plates, and assembly in a valve. (Courtesy of Masoneilan, Division of Dresser Flow Control.) P 2 = 800 psi (5.5 Mpa) 800 (5.5 Mpa) 1500 (10.4 Mpa) 2200 (15.2 Mpa) 2900 (20.0 Mpa) 3600 (24.8 Mpa) 4300 (29.7 Mpa) 5000 (34.5 Mpa) P 1 = 5000 psi Flow (34.5 Mpa) Typical pressure drop chart Cascade-Trim step type plug for high pressure breakdown. (Courtesy of Copes Vulcan, Inc., SPS Process Equipment.) Flash-Flow trim element and valve installation. (Courtesy of Hammel Dahl ConoFlow.) FIG. 6.1y Control valve designs that are less likely to cavitate due to their multipath and multiturn flow paths. Step plug and orifice trim for liquid service. (Courtesy of Masoneilan, Division of Dresser Flow Control.)

28 1072 Control Valve Selection and Sizing Gland flange Packing follower Stem packing Yoke clamp Bonnet Stem Bonnet flange Bonnet seal Plug seal Plug Disk stack assembly Flex gasket Seat ring Body Hush trim element and plug flow is into plug bore and out. (Courtesy of Copes Vulcan, Inc., SPC Process Equipment.) Self-drag valve with example of disk element. (Courtesy of CCI-Control Components, Inc.) Turbo-cascade trim element and valve installation. (Courtesy of Yarway, Tyco Valves & Controls.) FIG. 6.1y (Continued). chemical composition but less impact resistance. Correspondingly, its cost is lower. In summary, the applications engineer should first review the potential methods of eliminating cavitation. These would include adjustment of process conditions, revision of valve type, or change of installation layout. If none of these techniques can guarantee the complete elimination of cavitating conditions, the design engineer should install chokes or special anticavitation valves that can last for some reasonable period, even if some cavitation is occurring. Control Valve Noise The calculation of noise levels generated by control valves and the methods of lowering these noise levels are both covered in Section 6.14 and, therefore, will not be repeated

29 6.1 Application and Selection of Control Valves 1073 here. As can be seen in that section, many of the features of low-noise control valves are similar to the features of the anticavitation valves shown in Figure 6.1y. Flashing and Erosion FIG. 6.1z The Swiss cheese design can withstand mild cavitation. Cavitation occurs when (in Figure 6.1w) p 2 > p v, while flashing takes place when p 2 < p v. When a liquid flashes into vapor, there is a large increase in volume. In this circumstance, the piping downstream of a valve needs to be much larger than the inlet piping in order to keep the velocity of the two-phase stream low enough to prevent erosion. The ideal valve to use for such applications is an angle valve with an oversized outlet connection. In Section 6.15, the method for calculating the exit velocity in such two-phase flashing applications is illustrated Section A-A A Lift A Labyrinth turn CCI-Courtesy Control Components, Inc. Conventional valve P 1 P v Multi-step valve P 2 Condition A P 1 Multi-step valve P v Conventional valve Condition B P 2 Courtesy Masoneilan, Division of Dresser Flow Control. FIG. 6.1aa The labyrinth (top) and multistep (bottom) valve designs help to reduce the probability of cavitation.

30 1074 Control Valve Selection and Sizing Air Gas inspiration Air Air Standard ball modified Special ball valve FIG. 6.1bb Cavitation can also be alleviated by the admission of air into the flowing stream. FIG. 6.1cc The probability of cavitation can also be reduced by installing a choke fitting downstream of the valve. by an example. In addition, the piping must be designed so that it is not damaged by slug flow. The impingement of liquid droplets can be erosive if the velocity is great enough (> 200 ft/s, or 60 m/s, across the orifice), such as in applications involving high-pressure letdown of gas or vapor with suspended droplets. On highpressure let-down applications, the ideal valve to use is the dynamically balanced plug valve provided with a hard-faced plug. Section 6.15 gives an example on how the exit velocity from a steam let-down valve can be calculated. Erosion caused by high exit velocities can also cause corrosion problems. TABLE 6.1dd Relative Resistance of Various Materials to Cavitation Trim or Valve Body Material Relative Cavitation Resistance Index Approximate Rockwell C Hardness Values Corrosion Resistance Aluminum 1 0 Fair Low Synthetic sapphire 5 Very high Excellent High Brass 12 2 Poor Low Carbon steel, AISI C Fair Low Carbon steel, WCB Fair Low Nodular iron 70 3 Fair Low Cast iron Poor Low Tungsten carbide Good High Stellite # Good Medium Stainless steel, type Excellent Medium Stainless steel, type Good Medium Aluminum oxide Fair High K-Monel Excellent High Stainless steel, type 17-4 ph Excellent Medium Stellite # Excellent Medium Stainless steel, type 440C Fair High Stainless steel, type 329, annealed Excellent Medium Stellite # Excellent Medium Stellite #6B Excellent High Cost

31 6.1 Application and Selection of Control Valves 1075 Some metals do not corrode due to a self-regenerating protective surface film; however, if this film is removed by erosion faster than it is formed, the metal corrodes rapidly. For lining, either nobler metals or ceramics should be considered in such situations. The preferred arrangement for flashing service is to use a reduced port angle valve discharging directly into a vessel or flash tank. Corrosion Some information data on corrosion is contained in Table 6.1dd. A detailed tabulation of the chemical resistance of materials is given in Appendix A.3 of this volume. In evaluating a particular application, one should also consider the facts that most process fluids are not pure and that the corrosion rate is much influenced by flow velocity and the presence of dissolved oxygen. On corrosive services, one can also consider the use of lined valves (tantalum, glass, plastics, and elastomers), but one should consider the consequences of lining failure. Damage can be caused by accidentally exposing the lining to high concentrations of inhibitors or line cleaning fluids. Gas absorption in elastomer linings can also cause blistering. Viscous and Slurry Service When the process stream is highly viscous or when it contains solids in suspension, the control valve is selected to provide an unobstructed streamline flow path. The chief difficulty encountered with heavy slurry streams is plugging. Conditions that can contribute to this include a difficult flow path through the valve, shoulders, pockets, or dead-ended cavities in contact with the process stream. Valves with these characteristics must be avoided because they represent potential areas in which the slurry can accumulate, settle out, and gel, freeze, solidify, decompose, or as most frequently occurs, plug the valve completely. The ideal slurry valve is one that 1. Provides full pipeline opening in its open position 2. Provides for unobstructed and streamlined flow in its throttling position 3. Has high pressure and temperature ratings 4. Is available in corrosion-resistant materials 5. Is self-draining and has a smooth contoured flow path 6. Will fail safe 7. Has acceptable characteristics and rangeability 8. Has top works that are positively sealed from the process Unfortunately, no one valve meets all of these requirements, and the instrument engineer has to judge which features are essential and which can be compromised. If, for example, it is essential to provide a full pipe opening when the valve is open, there are several valves that can satisfy this requirement. They include the various pinch valves (A in Figure 6.1ee), the full opening angle valves (B control signal Connection for flushing Flexible sleeve Metallic jacket A Jacketed pinch valve B Full opening angle valve C Saunders valve Inlet D Characterized ball valve E Self-draining valve F Eccentric rotating plug valve G Sweep angle valve FIG. 6.1ee Valves for viscous and slurry services.

32 1076 Control Valve Selection and Sizing in Figure 6.1ee), some of the Saunders valve designs (C in Figure 6.1ee), and the full-ported ball valves (Section 6.16). While these units all satisfy the requirement for a fully open pipeline when open, they differ in their limitations. The pinch valves, for example, are limited in their materials of construction, pressure, temperature ratings, flow characteristic, speed of response, and rangeability, but they do provide self-cleaning streamlined flow, which in some designs resembles the characteristics of variable venturi. The Saunders valves and pinch valves have similar features, including the important consideration that the sealing of the process fluid does not depend on stuffing boxes. They are superior in the availability of corrosion-resistant materials, but they are inferior if completely unobstructed streamline flow is desired. Pinch valves are suitable for very low pressure drop services only, while Saunders and wedge plug valves can operate at slightly higher pressures. Lined butterfly valves are a good choice if the process pressure is high while the valve drop is low. The angle valve with a scooped-out plug satisfies most requirements except that its flow characteristics are not the best, and it is necessary to purge it above the plug in order to prevent solids from migrating into that area. Full-ported ball valves in their open position are as good as an open pipe section, but in their throttling positions both their flow paths and their pressure recovery characteristic are less desirable. Valves that do not open to the full pipe diameter but still merit consideration in slurry service include the following designs: characterized ball valves (D in Figure 6.1ee), various self-draining valve types (E in Figure 6.1ee), the eccentric disc rotating globe designs (F in Figure 6.1ee), and the sweep angle valves (G in Figure 6.1ee). Each of these has some features that represent an improvement over some other design. The characterized ball valve, for example, exhibits an improved flow characteristic in comparison with the full-ported ball type. It is well-suited for process fluids containing fibers or larger particles. The selfdraining valve allows slurries to be flushed out of the system periodically. Complete drainage is guaranteed by the fact that all surfaces are sloping downstream. The sweep angle valve, with its wide-radius inlet bend and its venturi outlet, is in many ways like the angle slurry valve. Its streamlined nonclogging inner contour minimizes erosion and reduces turbulence. In order to prevent the process fluid from entering the stuffing box, a scraper can be furnished, which if necessary can also be flushed with some purge fluid. The orifice located at the very outlet is built like a choke fitting. Both orifice and plug may be made of abrasion-resistant ceramic or hard metals. For slurries with large solid particles, the ideal orifice shape is a circle, such as that of an iris valve or a jacketed pinch valve (A in Figure 6.1ee). Orifice size, particle size, and rangeability are interrelated. For any particle size and orifice shape, there is a minimum opening below which plugging can be expected. To get good rangeability (control at low flow rates) the valve p should be made small. One way to accomplish this is by use of a head box (Figure 6.1ff). The selection of the valve style and the piping configuration around the valve inlet must be guided by the intractability of the particular slurry. Valves That Can Be Sterilized In food processing applications, in addition to the above, valves must not contain pockets where process material can be retained, and they should be constructed so that they can be easily sterilized and disassembled for cleaning. Materials of construction should not contain compounds that are prohibited by the FDA, such as some of the elastomer compounding materials. Valve Leakage Head box FIG. 6.1ff In slurry service, the use of a head box can provide a small and constant pressure drop across the control valve. Any flow through a fully closed control valve when exposed to the operating pressure differentials and temperatures is referred to as leakage. It is expressed as a cumulative quantity over a specified time period for tight shut-off designs and as a percentage of full capacity for conventional control valves. According to ANSI B16.104, valves are categorized according to their allowable leakage into six classes. These leakage limits are applicable to unused valves only: Class I valves are neither tested nor guaranteed for leakage. Class II valves are rated to have less than 0.5% leakage. Class III valves are allowed up to 0.1% leakage. Class IV valves must not leak more than 0.01% of their capacity.

33 6.1 Application and Selection of Control Valves 1077 TABLE 6.1gg Valve Seat Leakage Classifications per ANSI B (FCI 70 2) Class I II III IV V VI in. Nominal Port Diameter Maximum Leakage No test required 0.5% of rated valve capacity 0.1% of rated valve capacity 0.01% of rated valve capacity ml/minute of water per inch of orifice diameter per psi differential ml per minute of air or nitrogen vs. port diameter per the following tabulation mm Maximum Seat Leakage, ml/minute / Class V valves are specified to have a leakage of ml/min water flow per inch (25.4 mm) of seat diameter, per 1 psi (0,0685 bar) differential pressure. Class VI is for soft-seated valves, and leakage is expressed as volumetric air flow at rated p up to 50 psi (3.45 bar). Generally the functions of tight shut-off and those of control should not be assigned to the same valve. The best shut-off valves are rotary on/off valves, which are not necessarily the best choices for control. Table 6.1gg gives a summary of the ANSI leakage classes, and Table 6.1a lists the leakage capabilities of the various control valve designs. For added details on softseated designs and other special designs, refer to the feature summaries at the front of each section. Some valve manufacturers list in their catalogs the valve coefficients applicable to the fully closed valve. For example, a butterfly valve supplier might list a C v of 13.2 (K v = 11.4) for a fully closed, metal-to-metal seated 24 in. (600 mm) valve. It should be realized that such figures apply only to new, clean valves operating at ambient conditions. After a few years of service, valve leakage can vary drastically from installation to installation as affected by some of the factors to be discussed. It should also be noted that some fluids are more difficult to hold than others. Low-viscosity fluids such as Dowtherm, refrigerants, or hydrogen are examples of such fluids. Soft Seats One of the most widely applied techniques for providing tight shut-off over reasonable periods of time is the use of soft seats. Standard materials used for such services include Teflon and Buna-N. Teflon is superior in its corrosion resistance and in its compatibility to high-temperature services up to 450 F (232 C). Buna-N is softer than Teflon but is limited to services at 200 F (93 C) or below. Neither should be considered for operating conditions such as static pressures of 500 PSIG (34.5 bar) or greater, for use with fluid containing abrasive particles, or if critical flow is expected at the valve seat. The leakage of double-ported valves is much greater than that of single-ported ones, and it can be as high as 2 3% of full capacity in metal-to-metal seated designs. Temperature and Pipe Strain It is frequently the case that either the valve body is at a different temperature than the trim or the thermal expansion factor for the valve plug is different from the coefficient for the body material. It is usual practice in some valve designs (such as the butterfly) to provide additional clearance to accommodate the expansion of the trim when designing for hot fluid service. The leakage will, therefore, be substantially greater if such a valve is used at temperatures below those for which it was designed. Temperature gradients across the valve can also generate strains that promote leakage. Such gradients are particularly likely to exist in three-way valves when they are in combining service and when the two fluids involved are at different temperatures. This is not to imply that three-way valves are inferior from a leakage point of view. Actually, their shut-off tightness is comparable to that of single-seated globe valves. Pipe strains on a control valve will also promote leakage. For this reason, it is important not to expose the valve to excessive bolting strains when placing it in the pipeline and to isolate it from external pipe forces by providing sufficient supports for the piping. Seating Forces and Materials The higher the seating force in a globe valve, the less leakage is likely to occur. An average valve has a seating force of 50 lb f per linear inch (8750 N/m) of seat circumference. Where necessary, a much increased seating force will create better surface contact by actually yielding the seat material. Seating forces of this magnitude (about ten times the normal) are practical only when the port is small. Seating materials are selected for compatibility with service conditions, and Stellite or hardened stainless steel is an appropriate choice for nonlubricating, abrasive, high-temperature, and high pressure drop services. These hard surface materials also reduce the probability of nicks or cuts occurring in the seating surface, which might necessitate maintenance or replacement. Small-Flow Valves Valves with small flow rates are found in laboratory and pilot plant applications. Even in industrial

34 1078 Control Valve Selection and Sizing installations, the injection of small quantities of neutralizers, catalysts, inhibitors, or coloring agents can involve flows in the range of cubic centimeter per minute. Valves are usually considered miniature if their C v is less than 1 (K v = 0.862). This generally means a 1 / 4 in. (6.25 mm) body connection and a 1 / 4 in. (6.25 mm) or smaller trim. The top works are selected to protect against oversizing, which could damage the precise plug. The field of small-flow control valves is a highly specialized area, unlike any other application. The mechanical design and fluid flow constraints encountered essentially make these valves custom applied for the service. Small-flow valves are used in laboratories, process pilot plants, and some areas of full-scale plants. The design and building of these valves test the ingenuity of any manufacturer. There are several design types available, and great care must be taken to match the application with the valve design. While the manufacturers publish C v (K v ) ratings for their valves and trims, these should be treated with extreme caution and used for reference purposes only (below 0.01 C v [ K v ]). In many cases, the flow may shift from laminar to turbulent with flowing conditions and valve stroke changes. It is quite common for a laminar flow pattern to predominate, particularly with viscous fluids or in low-pressure applications. Laminar flow means the flowing quantity will vary directly with pressure drop instead of with the square root of pressure drop. It is wise to devise a test procedure simulating the service application to evaluate performance of a specific valve before using it in actual service. It is not uncommon for two identical small valves to exhibit somewhat different C v (K v ) capacities and flow curves under the same test conditions. Needle Valves There are at least three approaches to the design of miniature control valves: (1) the use of smoothsurfaced needle plugs, (2) the use of cylindrical plugs with a flute or flutes milled on it, and (3) positioning the plug by rotating the stem. One of the most common designs looks like a miniature version of the standard globe control valve (Figure 6.1hh). The trim consists of a precision honed and close-fitting plug fitted into an orifice made of a hard alloy. The control area consists of a fine taper slot milled into the outer surface of the piston-shaped plug or a long shallow taper plug. In smaller C v (K v ) trims, this slot may be a calibrated scratch in the surface. It is not uncommon to find up to 30 trims available in a given body to cover the C range of to 0.1 (K from to K v ). The rangeability, i.e., the ratio between maximum and minimum controllable flow, can be limited for this type of valve due to the inherent leakage flow between piston and orifice. The smaller the trim size, the lower the rangeability. Needle plugs give more dependable results than the ones with grooves, scratches, or notches because the flow is distributed around the entire periphery of the profile. This results FIG. 6.1hh Needle-type small-flow valve and plugs with flutes milled on cylindrical surface. in even wear of the seating surfaces and eliminates side thrusts against the seat. The trim is machined for very small clearances, and hard materials or facings are recommended to minimize wear and erosion. Needle plugs are available with equal-percentage (down to C v = 0.05 [K v = 0.043]), linear, and quick-opening characteristics (Figure 6.1hh). Some manufacturers claim the availability of valves with coefficients of C v = (K v = ) or less. At these extremely small sizes it is very difficult to characterize the plugs (equal-percentage is not available), and the valve rangeability also suffers. It is easier to manufacture the smaller cylindrical plugs with one or more grooves (Figure 6.1hh) and obtain the desired flow characteristic by varying the milling depth. Both the needle and the flute plugs are economical, but it is difficult to reproduce their characteristics and capacity accurately. Ball Valves In contrast to the relatively long stroke piston and orifice valve discussed, there is another valve design for small-flow applications that has a very short and variable adjusted stroke (Figure 6.1ii). Here, a synthetic sapphire ball is allowed to lift off a metal orifice and throttle the flow. The particular advantage of this valve is that the diaphragm stroke can be adjusted to produce various stem lifts with a standard 3 15 PSIG ( bar) signal. Two versions are available. One can be adjusted to cover 0.07 to C v (0.06 to K v ) and the other can be adjusted from 1 to C v (0.862 to K v ). These can be used for high-pressure, high-drop applications ranging from 3000 to 30,000 PSIG (207 to 2068 bar), depending upon the model.

35 6.1 Application and Selection of Control Valves 1079 H FIG. 6.1ii Low-flow ball valve with adjustable short-stroke actuator. (Courtesy of A.W. Cash Co.) L h Stem Rotation Type Another short-stroke valve especially suitable for high-pressure service (up to 50,000 PSIG, or 3447 bar) is shown in Figure 6.1jj. Variation in C v (K v ) rating of identical plugs and seats is achieved by mechanical adjustment of a toggle arrangement. The toggle can change the valve stroke between and in. (0.25 to 3.75 mm). With different trim inserts, a C v range of 1 to (K v range of to ) can be covered. In this design, the lateral motion of the plug is achieved by rotating the stem through a lead screw. The linear diaphragm motion is transferred into rotation by the use of a slip ball joint. Valve capacity is a function of orifice diameter (down to 0.02 in., or 0.5 mm), number of threads per inch in the lead screw (from 11 to 32), amount of stem rotation (from 15 to 60 ), and the resulting total lift, which generally varies from to 0.02 in. (0.125 to 0.5 mm). FIG. 6.1kk Low-flow valve using laminar flow element. (Courtesy of Emerson Process Management.) The extremely short distance of valve travel makes accurate positioning of the plug essential, and this necessitates a positioner. The combination of a long stem and short plug travel makes this valve sensitive to stem load and temperature effects. Because this differential thermal expansion can cause substantial errors in plug position, this valve is limited to operating temperatures below 300 F (149 C). Laminar Valve Finally, the latest addition to the low-flow valve family is shown in Figure 6.1kk. This valve is designed Slip ball joint FIG. 6.1jj Small-flow plug positioned by stem rotation.

36 1080 Control Valve Selection and Sizing to operate on the laminar flow principle. The flow is controlled by forcing the fluid through a long and narrow path, formed between two parallel surfaces. The actuator varies the laminar gap through a diaphragm and O-ring-sealed hydraulic ram. Because, in the laminar regime, flow varies linearly with pressure drop across the valve and with the third power of the gap width between the two surfaces (valve travel), this design has extremely high control rangeability along with a wide C v range of 0.02 to (K v from to ). The laminar principle also eliminates cavitation effects with liquids and sonic choking velocities with gases. This design can be used with inlet pressures and pressure drops up to 2675 PSIG (185 bar). While it is not commonly thought of for low-flow valve application, the labyrinth disc valve design (see Figure 6.1aa) can be manufactured for this purpose. Its most common application is for reducing high-pressure fluid samples for analyzers, but obviously it can be used in other services. One common factor for this design, as well as for all of the others discussed, is the need for the fluid to be extremely clean. These valves are not tolerant of dirt or sediment due to the small passage and close clearances. Unless the fluid is known to be clean, it is necessary to provide for a high level of filtration upstream. INSTALLATION When a valve is larger than 4 in. (100 mm) or sometimes when it is more than one size smaller than the pipe, it is advisable to use pipe anchors to minimize force concentrations at the reducers and more frequently to relieve flange stress loading due to valve weight. The end connections on the valve should match the pipe specifications. If welded valves are specified, the nipples should be factory-welded and the welds should be stress-relieved. If lined valves are specified, their inside diameters should match that of the pipe to avoid extrusion. On flangeless valves, the bolting and the tightness of the gaskets can be a problem if the valve body is long. If valves are fast closing (or fail) in long liquid lines, water hammer can result in the upstream pipe or vacuum can develop in the downstream line. Fast-opening steam valves can thermally shock the downstream piping. Steam traps should be provided at all low points in a steam piping network. Anchors should be provided in all locations where sudden valve repositioning can cause reaction forces to develop. Flow-to-close single-seated valves should not be used because if operated close to the seat, hydraulic hammer can occur. If the damping effect of the actuator alone will not overcome the vertical plug oscillation, then either the actuator should be made stiffer (higher air pressure operation) or hydraulic snubbers should be installed between the yoke and the diaphragm casing. Climate and Atmospheric Corrosion In humid environments such as the tropics, moisture will collect in all enclosures, and therefore drains should be provided. Electrical parts should all be encapsulated where possible or be provided with suitable moistureproof coating. Vent openings should be provided with storage plugs and insect screens. Even with such precautions, the vents and seals will require preventive maintenance and antifungus treatment in some extreme cases. Cold climates can produce high breakaway torque of elastomers, and in general, metals and plastics will become more brittle. Electrohydraulic actuators will require heating because oils and greases can become very viscous. In high-temperature environments, the weak link is usually the actuator, but liners, plastic parts, and electric components are also vulnerable. The damage is not only a function of the temperatures but also of the lengths of time periods of exposure. Diaphragm temperature limits are a function of their materials: Neoprene 200 F (93.3 C), Nordel 300 F (148.7 C), Viton 450 F (232.2 C), silicone glass 500 F (260 C). For higher temperatures, one can replace the diaphragms with pistons (or with metallic bellows in some extreme cases) or add heat shields. In power plant applications, valves frequently must be designed to withstand anticipated seismic forces. If the atmosphere contains corrosive gases or dusts, it is desirable to enclose, purge, or otherwise protect the more sensitive parts. The stem, for example, can be protected by a boot. In hazardous areas, all electrical devices should either be replaced by pneumatic ones or be made intrinsically safe or explosionproof. CONTROL VALVE SPECIFICATION FORM Compiling the information necessary to specify a control valve is best done with the aid of a tabulation sheet. Many large companies have their own customized forms. Figure 6.1ll shows a general-purpose form (ISA Form S20.50 Rev. 1) standardized by the Instrumentation, Systems, and Automation Society. After the form, general instructions are provided to assist in completing the form. A similar data sheet standard has been published by IEC as publication IEC , Industrial Process Control Valves, Part 7: Control Valve Data Sheet. References 1. Lytle, R. F., Equipment Selection for Control System Performance, paper presented at PUPID-SECON Conference in Birmingham, U.K., Rahmeyer, W., The Critical Flow Limit and Pressure Recovery Factor for Flow Control, InTech, November 1986.

37 6.1 Application and Selection of Control Valves 1081 FIG. 6.1ll ISA S20 Specification Form For Control Valve * Copyright ISA 1981, reprinted with permission of the Instrumentation, Systems, and Automation Society (ISA).

38 1082 Control Valve Selection and Sizing

39 6.1 Application and Selection of Control Valves 1083

40 1084 Control Valve Selection and Sizing

41 6.1 Application and Selection of Control Valves 1085 Bibliography Adams, M., Control Valve Dynamics, InTech, July ANSI/ISA-TR , Control Valve Response Measurement from Step Inputs, Research Triangle Park, NC: Instrumentation, Systems, and Automation Society, Aubrun, C., Robert, M., and Cecchin, T., Fault Detection in a Control Loop, Control Engineering Practice, Vol. 3, No. 10, pp , October Arant, J. B., Fire-Safe Valves, InTech, December Arant, J. B., Positioner Use Is Myth-Directed, InTech, November Ball, K. E., Final Elements: Final Frontier, InTech, November Baumann, H. D., A Case for Butterfly Valves in Throttling Applications, Instruments and Control Systems, May Baumann, H. D., How to Assign Pressure Drop across Control Valves for Liquid Pumping Services, Proceedings of the 29th Symposium on Instrumentation for the Process Industries, Texas A&M University, Baumann, H. D., Effect of Pipe Reducers on Valve Capacity, Instruments and Control Systems, December Baumann, H. D., The Introduction of a Critical Flow Factor for Valve Sizing, ISA Transactions, Vol. 2, No. 2, April Baumann, H. D., Control-Valve Sizing Improved, InTech, 46(6): pp , June Beard, C. S., Final Control Elements, Philadelphia, PA: Chilton, Biegacki, S. and VanGompel, D., The Application of DeviceNet in Process Control, ISA Transactions, Vol. 35, No. 2, pp , Bishop, T., Chapeaux, M., Jaffer, L., et al., Ease Control Valve Selection, Chem. Eng. Prog. 98, (11): November Boger, H. W., Recent Trends in Sizing Control Valves, 23rd Annual Symposium on Instrumentation, Texas A&M University, Boger, H. W., Flow Characteristics for Control Valve Installations, ISA Journal, November Bolek, W., Sasiadek, J., and Wisniewski, T., Two-Valve Control of a Large Steam Turbine, Control Engineering Practice, Vol. 10, No. 4, pp , April Borden, G. and Zinck, L., Control Valve Seat Leakage, InTech, November Brunone, B. and Morelli, L., Automatic Control Valve-Induced Transients in Operative Pipe System, Journal of Hydraulic Engineering, Vol. 125, No. 5, p. 534, May Buckley, P. S., Design of Pneumatic Flow Controls, Proceedings of the 31st Annual Symposium on Instrumentation for the Process Industries, Texas A&M University, Buckley, P. S., A Control Engineer Looks at Control Valves, Proceedings of 1st ISA Final Control Elements Symposium, Wilmington, DE, Buresh, J. F. and Schuder, C. B., The Development of a Universal Gas Sizing Equation for Control Valves, ISA Transactions, Vol. 3, No. 4, October Cain, F. M., Solving the Problem of Cavitation in Control Valves, Advances in Instrumentation and Control Proceedings of the ISA/91 International Conference, Vol. 46, Part 2, pp , Paper # , Cain, F. M. and Barnes, R. W., Testing for Cavitation in Low Pressure Recovery Control Valves, ISA Transactions, Vol. 25, No. 2, pp , Control Valves Globe, Plug, Pinch, Needle, Gate, Measurements and Control, February Control Valves, Regulators, Measurements and Control, June Control Valve Market to Reach nearly $3.5 Billion by 2008, Control Eng., April Coughlin, J. L., Control Valves and Pumps: Partners in Control, Instruments and Control Systems, January Control Valves Deserve Your Respect, Control Engineering, Vol. 49, No. 4, p. 64, April Cunningham, E. R., Solutions to Valve Operating Problems, Plant Engineering, September 4, Digital Hydraulic Valves Gaining Momentum, Design News, Vol. 58, March 3, Dobrowolski, M., Guide to Selecting Rotary Control Valves, InTech, December Driskell, L. R., Sizing Control Valves, ISA Handbook of Control Valves, ISA, Fagerlund, A. C., Recommended Maximum Valve Noise Levels, InTech, November Gassmann, G. W., When to Use a Control Valve Positioner, Control, September George, J. A., Sizing and Selection of Low Flow Control Valves, InTech, November Gibson, H. I., Variable-Speed Drives as Flow Control Elements, ISA Transactions, Vol. 33, No. 2, pp , July Growth Forecasted for the Control Valve Market, Control Solutions 74 (12):8, December Hammitt, D., Key Points about Rotary Valves for Throttling Control, Instruments and Control System, July Hammitt, D., How to Select a Valve Actuator, Instruments and Control Systems, February Hanssen, A. J., Accurate Valve Sizing for Flashing Liquids, Control Engineering, February Hanson, C. L. and Clark, J. C., Fast-Closing Vacuum Valve for High-Current Particle Accelerators, Review of Scientific Instruments, January Hegberg, M. C., Control Valve Selection for Hydronic Systems, ASHRAE J., 42 (11): 33, November Hegberg, M. C., Control Valve Selection Response, ASHRAE J., 43(3): 24, March Hill, A. G. and Lau, K. H., Artificial Intelligence in Control Valve Selection, ISA Transactions, Vol. 28, No. 1, pp , Horch, A., A Simple Method for Detection of Stiction in Control Valves, Control Engineering Practice, October 1999, pp Hutchison, J. W. (ed.), ISA Handbook of Control Valves, 2nd ed., Research Triangle Park, NC: Instrumentation, Systems, and Automation Society, Installing Smart Positioners A Wise Move, Chemical Engineering Magazine, December 1, Intelligent Actuator Allows Programmable Valve Control, Water World, Vol. 15, No. 5, p. 101, May Jämsä-Jounela, S.-L., Dietrich, M., Halmevaare, K., and Tiili, O., Control of Pulp Levels in Flotation Cells, Control Engineering Practice, Vol. ll, No. 1, pp , January Jury, F. D., Positioners and Boosters, Instruments and Control Systems, October Kam, W. Ng., Control Valve Noise, ISA Transactions, Volume 33, No. 3, pp , September Karpenko, M., Sepehri, N., and Scuse, D., Diagnosis of Process Valve Actuator Faults Using a Multilayer Neural Network, Control Engineering Practice, Vol. 11, No. 11, pp , November Kayihan, A. and Doyle, F. J., III, Friction Compensation for a Process Control Valve, Control Engineering Practice, Vol. 8, No. 7, pp , July Keagle, J., It s 18 ma, Do You Know Where Your Positioner Is? InTech, May Keles, O. and Ercan, Y., Theoretical and Experimental Investigation of a Pulse-Width Modulated Digital Hydraulic Position Control System, Control Engineering Practice, Vol. 10, No. 6, pp , June Keskar, P. Y., Analysis of Lightning-Related Damages to Instrumentation and Control System for Water and Wastewater Plants, ISA Transactions, Vol. 35, No. 1, pp. 9 15, May Kimura, T., Hara, S., Fujita, T., and Kagawa, T., Feedback Linearization for Pneumatic Actuator Systems with Static Friction, Control Engineering Practice, Vol. 5, No. 10, pp , October Kirsner, W., Control Valve Selection, ASHRAE J., 43(3): 24, March Lee, J. G. and Kim, O. H., Development of a New Hydraulic Servo Cylinder with Mechanical Feedback, Control Engineering Practice, Vol. 7, No. 3, pp , March Lipták, B. G., Control Valves in Optimized Systems, Chemical Engineering, September 5, 1983.

42 1086 Control Valve Selection and Sizing Lipták, B. G., Control Valves for Slurry and Viscous Services, Chemical Engineering, April 13, Lipták, B. G., How to Size Control Valves for High Viscosities, Chemical Engineering, December 24, Lipták, B. G., Valve Sizing for Flashing Liquids, ISA Journal, January Louleh, Z., Cabassud, M., and Le Lann, M.-V., A New Strategy for Temperature Control of Batch Reactors: Experimental Application, Chemical Engineering Journal, Vol. 75, No. 1, pp , August Monsen, J. F., Spreadsheet Sizes Control Valves for Liquids/Gas Mixtures, InTech, December Moore, R. L., Flow Characteristics of Valves, in ISA Handbook of Control Valves, 2nd ed., Pittsburgh, PA: Instrumentation, Systems, and Automation Society, Morgenroth, J., Quarter-Turn Plug, Ball, and Butterfly Valves, Plant Engineering, July 24, Catalog of U.S. & Canadian Valves & Actuators, Value Manufacturers Association of America. O Keefe, W., Learn Fluid-Handling Lessons from Nuclear Isolation Valves and Actuator Systems, Power, January Page, G. W., Predict Control Valve Noise, Chem. Eng., New York, 107 (9): 23 26, August Pham, D. T., Jennings, N. R., and Ross, I., Intelligent Visual Inspection of Valve-Stem Seals, Control Engineering Practice, Vol. 3, No. 9, pp , September Price, V. E., Smart Valve Intelligence Takes Many Forms, InTech, August Pyotsia, J., A Mathematical Model of a Control Valve, 1992 ISA Conference, Houston, TX, October Rahmeyer, W., Cavitation Testing of Control Valves, Instrument Society of America, paper no. C.I. 83-R931, presented at the International Conference in Houston, TX, October Rahmeyer, W., The Critical Flow Limit and Pressure Recovery Factor for Flow Control, InTech, November Recommended Voluntary Standard Formulas for Sizing Control Valves, Fluid Controls Institute, Inc., FCI 62-1, May Revised Control Valve Standard, Hydrocarb. Process., 82(9): 114, September Restrepo, A., González, A., and Orduz, S., Cost-Effective Control Strategy for Small Applications and Pilot Plants: On/Off valves with Temporized PID Controller, Chemical Engineering Journal, Vol. 89, No. 1 3, pp , October Riveland, M. L., The Industrial Detection and Evaluation of Control Valve Cavitation, Instrument Society of America, paper no. C.I , presented at International Conference in Philadelphia, PA, October Roth, K. W. and Stares, J. A., Avoid Control Valve Application Problems with Physics-Based Models: Kinetic Energy Criteria have Many Limitations, Hydrocarb. Process., 80(8):37, August Ruel, M., Control Valve Health Certificate, Chem. Eng., New York, 108(12):62 65, November Sanderson, R. C., Elastomer Coatings: Hope for Cavitation Resistance, In Tech, April Scott, A.B., Control Valve Actuators: Types and Application, InTech, January Sharif, M.A. and Grosvenor, R.I., The Development of Novel Control Valve Diagnostic Software Based on the Visual Basic Programming Language, P. I. Mech. Eng. I.-J. Sys., 214(12):99 127, Shinskey, F. G., Control Valves and Motors, Foxboro Publication No Singleton, E. W., Control Valve Sizing for Liquid Viscous Flow, Engineering Report No 7, Introl Limited. Spreter, R.,, Valve Application Benefit from Technology Upgrades, Power Engineering, Vol. 103, No. 9, p. 46, September Standard Control Valve Sizing Equations, ANSI/ISA-S Thornhill, N. F., Cox, J. W., and Paulonis, M. A., Diagnosis of Plantwide Oscillation through Data-Driven Analysis and Process Understanding, Control Engineering Practice, Vol. 11, No. 12, pp , December Tullis, J. P., Hydraulics of Pipelines: Pumps, Valves, Cavitation, Transients, New York: John Wiley and Sons, Valve Actuator Roundup, InTech, January Wendy, E. J. and Stanton, C. C., Industry Corner: A World View of Industrial Valves, Business Economics, Vol. 32, No. 2, p. 56, April Weir, W., Control Valve Market Analyzed in New Study, Hydrocarbon Processing, Vol. 75, No. 9, p. 27, September Weirauch, W., Electric Actuators Leading the Way in Device Network Use, Hydrocarbon Processing, Vol. 75, No. 10, p. 31, October Weirauch, W., Valve Automatization Gains Renewed Interest, Hydrocarbon Processing, Vol. 80, No. 7, p. 25, July What s New in Valves and Valve Operators, Pipe Line & Gas Journal, Vol. 227, No. 11, p. 43, November What s New in Valves and Valve Operators, Pipelines & Gas Journal, Vol. 229, No. 11, p. 61, November Wilton, S. R., Control Valves and Process Variability, ISA Transactions, Vol. 39, No. 2, pp , April Wolter, D. G., Control Valve Selection, InTech, October Yang, J. C. and Clarke, D. W., The Self-Validating Actuator, Control Engineering Practice, Vol. 7, No. 2, pp , February Yu, W. X., Lin, J., Wang, C.W., et al., A Thermo-Control Valve Fabricated from Shape-Memory Alloy for Use in the Oil Field, Rare Metal Mat. Eng., 31(5): , October Yu, F. C., Easy way to estimate realistic control valve pressure drops, Hydrocarb. Process., 79(8): 45 48, August 2000.

43 6.2 Accessories and Positioners H. D. BAUMANN (1970) C. G. LANGFORD (1985, 2005) B. G. LIPTÁK (1995) P S From safety interlocks 1:1 Control valve with pneumatic positioner XLS Side-mounted handwheel Valve with solenoid pilot Control valve with booster relay Valve with limit switch Top-mounted handwheel or limit stop P I P Control valve with electropneumatic positioner Control valve with current to pneumatic transducer Flow sheet symbols Types: Materials of Construction: Supply Pressure (Gauge): Inaccuracy: Air accumulators, air sets, handwheels, I/P transducers, limit stops, limit switches, positioners, relays (biasing, booster, lock-up, quick-exhaust, reversing), smart (microprocessor-based) valve electronics, solenoid pilots, stem position transmitters, servo solenoid valve. Bodies and cases: die-cast zinc, white metal, aluminum. Bellows: copper alloys. Diaphragms: elastomer-coated fabric, thin metal. Misc. parts: steel, brass, aluminum. Varies from 20 to 25 PSIG (140 to 170 kpa) for signal relays to 60 to 100 PSIG (400 to 700 kpa) for positioners or boosters, some to 150 PSIG (1000 kpa). Positioners: repeatable to ± 0.1 to 1% and accurate to ± 0.5 to 2% of span. Transducers: ±0.5 to 1% of span. Boosters: ±0.1 to 1% of span. Signal Ranges Pneumatic: 3 9, 3 15 (preferred), 6 30, and 9 15 PSIG (20 60, , , (Gauge Pressure): and kpa). Electronic: Digital Fieldbus and other, 1 5, 4 20 (preferred), and ma DC. Others are used such as 0 10 and 1 10 VDC. Costs: Partial List of Suppliers: Air sets, $40; handwheels, $500 for larger size; I/P transducers, $300; limit stops, $100; limit switches, $50 to $150; positioners, $400 (pneumatic), $600 (electronic); $1500 (special) solenoid pilots, $80; stem position transmitter, $250 to $500. Air Sets: Adams Valve Inc. ( ControlAir Inc. ( Cashco Inc. ( Fisher Controls ( ITT Conoflow ( Masoneilan Dresser ( subunits/masoneilan/index.cfm) Moore Products Co. ( 1087

44 1088 Control Valve Selection and Sizing Handwheels and Limit Stops: Anchor/Darling Valve Co. (see Flowserve) Auma Actuators Inc. ( Copes-Vulcan Inc. ( Daniel Valve Co. (www. danielvalve.com/) Duriron Co. ( Fisher Controls ( Grinnell Supply Sales Co. (www. grinnell.com/) Keystone Controls Inc. ( Leslie Controls Inc. ( Limitorque Corp. ( Pacific Valves ( Mastergear Div. ( Neles-Jamesbury Inc. ( PBM Inc. ( Rotork Controls Inc. ( Velan Valve Corp. ( Valtek ( Xomox ( I/P Transducers, Positioners, Stem Position Transmitters: ABB Kent Inc. ( Bailey Controls Co. (see ABB) Bray Valve and Controls ( DeZurik ( Fisher Controls International Inc. ( Foxboro Co. ( Honeywell Industrial Controls ( ITT Conoflow ( Jordan Controls Inc. ( Kammer Valves Inc. ( Leslie Controls Inc. ( Masoneilan Dresser, ( subunits/masoneilan/index.cfm) Moore Products Co. ( Neles (METSO) Controls ( Valtek Inc. (www. flowserve.com/valves/index.htm) Limit Switches: Allen Bradley ( Go Switch Inc. (Topworx) ( Masoneilan, Dresser Valve & Controls Div. ( businessunits/measurement/subunits/masoneilan/index.cfm) Micro Switch/Honeywell (content.honeywell.com/sensing/) Pepperl + Fuchs Inc. (www. pepperl-fuchs.com/pa/welcome_e.html;) Proximity Controls Inc. ( Relays (Biasing, Booster, Reversing): ABB Kent-Taylor, Bailey Controls Co. ( Fairchild Industrial Products Co. (flw.com/fairchild/) Fisher Controls International ( Foxboro Co. ( Moore Products Co. ( Robertshaw Controls Co. ( Smart (Microprocessor-Based) Valve Electronics: EIM Controls ( Kaye & MacDonald ( Limitorque ( com/valves/index.htm) Rotork Controls ( Valtek ( Solenoid Pilot: Automatic Switch Co. (

45 6.2 Accessories and Positioners 1089 Crosby Valve and Gauge Co. ( Gilmore Valve Co. ( Keystone International Inc. ( Leslie Controls Inc. ( Parker Hannifin Corp. (Skinner) ( Richards Industries Valve Group Inc. ( Spirax Sarco Inc. ( Tom Wheatley Valve Co. ( Servo Solenoid Valve: Target Rock Valve ( Hydraulic Relays(Water): GA: Industries ( INTRODUCTION This section describes some of the traditional valve accessories, including positioners. For the valve accessories that provide self-diagnostic capability or can communicate on the fieldbuses, the reader is referred to Section 6.12, which is fully devoted to an in-depth discussion of intelligent valves, positioners, and accessories. The word accessories applies to the many devices added to control valves. Many of these are more necessities than accessories, depending on the application. This section lists and discusses some of the many accessory devices that can be attached to both throttling and on/off valves. The purpose is to improve their performance or to obtain remote feedback on their status. The positioner is the single most important valve accessory, and this section will emphasize the design and the application of positioners. Before starting that discussion, a brief summary will consider the contribution of microprocessors to the development of smart valve accessories. See also Section 6.9, Dynamic Performance of Control Valves, for a more detailed look at valve dynamics. SMART VALVES Section 6.3 discusses some of the more common applications of intelligent actuator systems, and Section 6.12 is fully devoted to an in-depth discussion of intelligent valves, positioners, and accessories. The microprocessor can provide logic to improve the positioning and the performance of the valve. Additional services may include operational and maintenance communications with the valve. Microprocessor-based systems are available in watertight or explosionproof (NEMA 6 or 7) construction and can tolerate ambient temperatures from 40 to 185 F ( 40 to 85 C). They can also be used where any or all of high humidity, fungus, or dust are present. As shown in Figure 6.2a, these devices have become quite complex, and any proposed application for smart positioners should review the operating conditions and the application requirements with the manufacturer. Some of the positioners emphasize maintenance services; some make improved control their main goal. Microprocessor-based systems can incorporate self-tuning and self-calibrating electronic positioners or electronic proportional, integral, derivative (PID) controllers, which respond to a digital or analog external set point and commands. Locating the controller logic at the valve and dedicating it to full-time control is an advantage on critical loops, providing fast response for applications such as compressor surge protection, where speed of response is a major consideration. There may also be an improvement in control robustness and reliability if the system can continue to control without the central control system. This does not mean that positive control is lost, because the system is still under orders from the central system, only the continuous cycle of measurecompute-act-feedback is local. Commands for set point, mode of operation, and data collection reside in the central system. Local control eliminates the time lost during communications and response of the central control system like a programmable logic controller (PLC) or distributed control system (DCS). Smart positioners can modify the characterization of the valve to reduce control nonlinearities in the valve and actuator and to provide special action. It is possible to match the valve to the process requirements to reduce the change in process gain over the range of operation or to provide special response or action as required. Smart actuator circuitry can also protect electric motor operators from electrical overload when the valve is jammed or from reverse phase operation. Smart systems can also take advantage of the lower cost of digital communications over a single loop of two-conductor or fiber-optic cable. This method of communications can substantially reduce the wiring cost of installations that can control a number of actuators, pumps, and other devices. The smart valves can have calibration changed and reconfigured and can detect such performance changes as pressure drop changes resulting from a fouled pump or pipeline. They can also limit the valve travel to stay within the range where the required characteristics (gain) are guaranteed or to limit maximum or minimum flows.

46 1090 Control Valve Selection and Sizing PC for calibration, configuration, diagnostics or data acquisition Serial digital communication (RS-485) Process controller, distributed control system, host computer for controller setpoint or actuator position 4 20 ma Air supply 24 VDC Power supply Select loop type P P 1 P 2 temp. flow position Starpac processor/ software Local display I/P Positioner P4 P5 Valve and actuator Position sensor Non-volatile memory for calibration & configuration info storage Temperature T Flow Discrete digital signals Secondary 4 20 ma signals Input (2) Output (2) Input Output Upstream pressure Downstream pressure Stem position P1 P2 FIG. 6.2a Smart valve packages may have local display and sensors for temperature, flow, pressures, pressure differential, and stem position. (Courtesy of Valtek, Flowserve Corp.) Some of the smart valve packages incorporate additional sensors installed on or in the valve. These measurements may include upstream and downstream process pressure, pressure difference, flow, process fluid temperature, stem position, and actuator operating pressures (Figure 6.2a). This allows both for total distributed control and for the flexibility of manipulating the valve to control any one of the variables. These valves may also include local indicators and the capability of calibration or reconfiguration of the valve, transmitters, or controller. Diagnostic capabilities include selftuning and the ability to evaluate the valve and process responses to step or to ramp changes in the valve position. Upon failure of the electrical power supply the valve functions can be supported by battery backup. The user is cautioned that some of these more sophisticated accessories may require more than the usual two-wire 4 20 ma DC power/signal wiring. Some are four-wire with separate power required. Some are two-wire, but may present a higher impedance load to the system than the simpler field instruments. This higher source voltage requirement may prevent wiring two positioners in series for split ranging. Primarily used with on/off valves, on-line partial stroke testing may be provided for safety applications to provide verification of the valve actuation system. The user is advised to check the current versions of applicable standards because there is significant interest and activity in this area. POSITIONERS The positioner along with the associated actuator constitutes a simple position control loop. The valve begins to act only after the error is detected and the controller output has changed. Inherently, this is a lagging action. With the typical air supply pressure used, PSIG ( kpa), any large change in valve signal will result in a critical pressure drop within the positioner. When the internal flow restriction has a pressure drop ( P) greater than 50% of the absolute inlet pressure, then sonic flow velocities will exist. This sets a constant and limiting flow rate through the pilot valve. This maximum flow rate capacity is probably the one listed in the product specifications. As the valve stem approaches the desired position, the P decreases and the flow decreases to be more nearly proportional to the signal error. This results in some slowing of the response at the end of the stroke change and may help to stabilize response.

47 6.2 Accessories and Positioners 1091 When to Use Positioners The purpose of a positioner is to improve the accuracy of control valve response. This means that the valve position will more closely approach the position commanded by the control system. A positioner can reduce the effects of many dynamic variations. These include changes in packing friction due to dirt, corrosion, lubrication, or lack of lubrication; variations in the dynamic forces of the process; sloppy linkages (causing dead band); and nonlinearities in the valve actuator. The dead band of a good valve/actuator is 2% (Section 6.4), but it has been measured at up to 5%. Large plug valves and ball valves with less than perfect linkages and inadequate actuators may be far worse. A better positioner with the proper actuator can often have a dead band of less than 0.5% of stroke. The positioner increases the actuator speed or thrust by increasing the actuator pressure or airflow volume, and can modify the valve characteristics through the use of mechanical links and cams or electronic function generators. While these positioner capabilities are very important, some of these capabilities can also be obtained or approximated with other accessories. For example, split-ranging is possible using pneumatic relays. Multiplier/biasing relays in the air signal line to the valve modify the relationship between controller output and actuator air pressure. The response speed/thrust of the valve can be increased with the use of airflow booster relays, and changes to the control valve characteristics can be obtained, not only with a different plug characteristic shape, but also by pneumatic or electronic characterizing of the controller signal. Consider that the computing relays may have a lower flow capacity than a positioner and the speed of response may be affected. When the valve is in remote manual (open-loop) operation, the positioner will reduce the effects of the valve hysteresis and dead band and improve the accuracy of the response. When the valve is under automatic (closed-loop) control, the positioner will normally improve control, but will do that only if the loop response is slow when it is compared to the control valve response (analysis, temperature, liquid level and blending, slow flow, large volume gas flow). An imperfect positioner may degrade loop response, contribute to proportional offsets, and cause limit cycling in fast loops where the valve cannot keep up (fast flow, liquid pressure, small volume gas pressure). Older and simpler positioners did limit response and create problems. The issue is more of the speed of the valve relative to the controlled process rather than absolute speed (Figure 6.2b). It is not uncommon for the valve to be the slowest part of the loop and limit process response. This may be a problem, or it may be perfectly acceptable. The needs of the process should be the basis for determining the performance requirements of the control system. The very worst situation occurs when the valve system and the process have equal or similar time constants. Time response of a fast process (0.03 seconds process time constant) Controller settings: PB = 45% (gain = 2.2) I = 20 rep./min. Response with a positioner Response with a positioner Setpoint change (5%) (A) Time response of a slow process (100 seconds process time constant) Response without a positioner Response with a positioner Setpoint change (5%) (B) PB = 500% (gain = 0.2) I = 20 rep./min. Controller settings: PB = 5% (gain = 20) I = 0.5 rep./min. PB = 1% (gain = 100) I = 1.0 rep./min. FIG. 6.2b The response of a fast process (A, on top) is better without a positioner, while the response of slow processes (B, on bottom) is better with a positioner. 2 Actuators without springs or equivalent padding pressure to provide the spring return function will usually require positioners with dual outputs to drive both sides of the diaphragm or piston. The spool valve or relay valve will have a second output, with its action the reverse of the first one. Because the opposing air pressure below the diaphragm decreases when the air pressure above the diaphragm increases, the actuator has a greater net force available without the opposing spring force. It is necessary to have a packing gland or O-ring seal around the valve stem for the lower air pressure. A positioner with only a single output can be used with a reversing relay to provide the second output. A number of different control design requirements can be accomplished with positioners. A reverse-acting positioner (increase in input causes a decrease in output) makes an airto-open, spring-to-close valve function as an air-to-close, spring-to-close valve. This combination may be specified for systems with interlocks or batch operations. The use of reverse acting positioners is discouraged by some users for maintenance reasons. Beware of terminology here: Fail Open has been

48 1092 Control Valve Selection and Sizing Controller signal P in = 9 15 # (3 15 PSIG) P 1:2 Amp 1:2 in = 3 9 # with bias reverse R1 P out =15 3 # R2 (P out = (P in 9) 2 + 3) with bias P out = 3 15 # P out = (18 [(P in 3)2 + 3)] P P Valve 1 FO Value 2 FC FIG. 6.2c Split-range operation of two valves can be obtained by the use of amplifying and reversing relays. They can be either be pneumatic (shown here) or electronic. confused with Flow to Open, and Fail Closed has been confused with Flow to Close. These are different and must be defined and understood especially in safety applications. Heating, ventilation, and air-conditioning (HVAC) systems with low-pressure pneumatic thermostats may use some sort of simple positioner to increase actuator pressures. The alternative is electric signals and electric motor-activated valves. These services rarely need fast response or tight control. Split-Range Operation Split-ranging is the use of two valves controlled from one controller output signal. With positioners, each positioner moves its valve over only a part of the controller output range. One typical application is for temperature control. Here, the cooling valve is full open at 0% controller signal and closed at 50% signal and beyond, while the heating valve is closed over the 0 50% range and begins to open at 50% controller signal and fully opens at 100%. See Figure 6.2c and Table 6.2d. The example is a simple one and has many possible difficulties. The specific application may require a different sort of calibration. A dead zone with both valves closed at a controller output of 50% will have no total flow over that dead band. This may be used to reduce costs of heating and cooling. Other applications may require that both valves be partly opened at the 50% controller output point. For a system where the valves add the temperature control fluid into a circulating system, it might work well. But, as an example, if the system is once through where the heat transfer rate is proportional to the heating/cooling fluid velocity, the valves are usually set up with some overlap in valve position. That is, the cooling flow might not shut off until the Controller Output (CO) is 75%, and the heating flow might start at 25% CO. The total flow and thus heat transfer rate will remain more nearly constant and stabilize the operation. With this approach, maintenance is simplified because standard positioner calibrations are used and the full positioner accuracy is retained. Note that the specifications for the positioner are defined for the full range of the positioner and that if only half the output range is used, the effective dead band for each valve is doubled. For a maximum in flexibility, control each valve from a separate control system output. TABLE 6.2d As it is Shown in Figure 6.2c, it is Possible to Use Relays instead of Split-Range Positioners Controller Output Output Controller 3 (0.2) 9 (0.6) 15 (1.0) Controller Output Output Relay Output Relay 1 15 (1.0) 3 (0.2) 3 (0.2) Signals in % Output Relay Signals in PSIG (bar) Output Relay 2 3 (0.2) 3 (0.2) 15 (1.0) Control Valve Control Valve 1 100% 0% 0% Signals in % (with overlapping flows) Output Relay Output Relay Control Valve Control Valve Control Valve 2 0% 0% 100% Control Valve Tight Shut-Off In some applications, there is a need for special tight shut-off. A positioner calibration can provide the full actuator pressure applied to the valve stem when the valve is commanded to be closed. High Rangeability A special application might require accurate flow control for a small percentage valve opening. The positioner calibration can provide that the valve moves very little for a large fraction of the low end of the signal range but still opens fully at 100% of signal. Examples include a drastic difference between operating conditions and start-up or emergency shutdown, or a large butterfly valve. An application as simple as a level control may be very poor because the valve is oversized for special operating problems. A very simple flow control loop will add considerable accuracy at a modest cost. The primary requirement for the flow sensor is to be reliable; even poor accuracy is adequate. A simple field controller, even pneumatic, will constitute an inexpensive secondary control loop requiring little attention. Gap Action If there is a great problem with packing wear, it is possible to set up the controller or the positioner for gap action. Here, the valve signal changes only when the control error exceeds a certain limit.

49 6.2 Accessories and Positioners 1093 Condensate pots that accumulate steam condensate are an example of this. A partly opened valve will experience damaging cavitation. A valve that quickly switches between fully open at a high condensate level and fully closed at a low level will avoid cavitation. Boiler blowdown valves are a similar application. When Not to Use a Positioner The positioner/actuator can be defined as the secondary control loop in the cascade loop, working with the process controller output providing the control signal. In order for a cascade secondary to improve control, it must respond more quickly than the primary loop. The ideal situation would be if the time constant of the secondary was one tenth (openloop speed of response ten times as fast) of that of the primary and if the period of oscillation of the secondary loop was three times that of the primary. Even if the valve response is not as fast as these ideals, there still may be very helpful isolation of the process from process fluid changes. No process control response time can be faster than the slowest element in the control loop. The advantages of a cascade system are achieved with good secondary response. A general principle is to use a cascade system when the advantages of isolating the primary control loop from the nonlinearity, and variations in the secondary system, justify the added cost and complexity. In such installations (B in Figure 6.2b shows a gas pressure control process), the addition of a positioner increases the open-loop (steady state) gain and improves the loop response. For this situation, the process response may be further improved because the primary controller can be tuned with increased gain (narrower proportional band) and increased integral action (reset) to speed up the integral action. It is clear that poor valve response reduces the quality of control much more than the poor valve response would imply. In this situation the controller tuning must be modified (gain reduced, integral lengthened) to avoid oscillation due to dead band and response delay. In typical real applications controller tuning is conservative and avoids any hint of oscillation, and the response is even worse than it might be. Operators usually object to cycling processes. Part A in Figure 6.2b illustrates a fast liquid flow process, where the loop without a positioner can be tuned more tightly (for higher gain and more repeats/minute); such a loop responds better without a poor positioner. It might also be noted that after a new state is reached, the positioned installation gives better and noisier control because of increased speed of response. Some argue that all loops can be (and should be) controlled with positioned valves because they provide improved response. This is true, but there are situations where the positioner will add unjustified cost. The very fast flow loop of Figure 6.2b is one example. Another is a small valve with a relatively large actuator and good available force; it has a small valve stem and little friction and a small plug with low fluid forces. Situations where cost is primary and quality of control is not highly important may also omit positioners. Examples are HVAC where forces are low and signal changes are slow, but there is a desire to control the cost. Positioners are not normally required for on/off service. Positioner Performance The valve positioner is a servo-amplifier acting with the valve actuator to control the position of the valve stem. Without a positioner, the stem position may be changed or motion restrained by varying fluid pressures on the plug and by unpredictable friction forces. See Section 6.9 for more details on control valve response. The pneumatic input signal is typically 3 15 PSIG ( kpa), and the typical stroke is between 1 / 2 and 3 in. (12 75 mm). Errors of 0.05 in. (0.1 mm) are typical. The usual input air signal is the output of a pneumatic controller in the field or from a control panel or control system. The controller signal is the result of comparing a set point pressure to a measured variable as presented by a 3 15 PSIG value and the output is intended to reduce the error. Most positioners use air both as the operating fluid and as the source of power. Hydraulic actuator/positioners use high-pressure oil and are used on applications requiring large valves for high differential pressure services. These are units requiring actuator pressures higher than the normally available PSIG ( MPa) instrument air supply pressure. Hydraulic actuator/positioners are considerably more expensive; and they may be acoustically noisy and require specialized maintenance. Electric motor operators are also available (Section 6.3). They are used where air is not available and where their typically slow operating speed is acceptable. There are some specialized electric valve actuators that can provide very high position precision; others can provide very fast response. The analog electropneumatic positioner input section responds to a standard milliampere signal. Such units usually consist of a pneumatic positioner plus an integral I/P. Some users prefer to use a pneumatic positioner with a separate I/P (current to pressure converter), usually located near the valve. A few systems provide for packaging all of these accessories into one unit. An all-electric server solenoid motor-type positioner/actuator valve is also available. At least one control valve design uses a servo solenoid valve (SV) to achieve very fast and accurate response using the process fluid pressure for actuation. This approach requires a relatively clean fluid. The positioner normally provides a substantial improvement in valve and control loop performance, with the greatest improvement realized on slow control loops with low controller gains typically level, temperature, or analytical control (Figure 6.26).

50 1094 Control Valve Selection and Sizing Positioners have typical open-loop gain (change in output pressure per change in input signal, with the valve stem locked in position) of 10:1 up to 200:1. Dead band (the minimum input change for a detectable output change) is claimed as % of span. Vendors claim positioning accuracy of in. ( mm) under bench test conditions. Positioner Designs The control valve positioner as used with a globe-style valve is typically mounted on the valve yoke and has mechanical linkage, which is connected to the valve stem to sense position. The pneumatic positioner is powered by compressed air at pressures between 25 and 150 PSIG (170 and 1000 KPa). Tubing conveys the positioner output air to and from the actuator. There are some standards defining standardized mounting details for positioners, solenoid valves, and other accessories. Other designs mount the positioner directly in line with the valve stem, and installed on the end of the actuator. In this design the linkage to the valve stem is direct in a straight line. In this scheme, the positioner provides air directly to the top of the actuator. Figure 6.2e illustrates such a forcebalance positioner design. For rotary valves, the positioner is mounted above the valve body. In every case, the mountings must be solid and not allow any relative motion, which adds directly to the dead band. The installation should be so arranged that there is little temptation for workers to damage the connecting tubing and linkages by climbing on the valve. The exterior of the positioner should have a finish compatible with the chemical environment. The interior is (mostly) protected from corrosion by the steady bleed of dry clean air from the pilot valve. If the valve will be exposed to a very high level of vibration or acoustic noise, the purchase specifications should note this. Some positioner designs will withstand vibration better than others. Pipeline applications may use the pipeline gas to operate the pneumatic instruments. For these applications, the instrument vents are purchased with tapped connections to carry the vented gas to a safe location. Because the gas used may not be totally dry, there may be a need for filters and separators to protect the instruments. Consult the valve and instrument manufacturers for requirements. Also consider the area electrical classification for electropneumatic positioners. Because all positioners use small restrictions to reduce the air consumption at steady state, it is vital that the air be clean. Some positioner designs include a small integral filter. Satisfactory long-term operation of the instruments is considerably improved if all the dirt and moisture are removed in a good external supply filter before it reaches the positioner. Liquids are efficient transporters of dirt and solids in supply lines to places where they will cause trouble. Liquids also can lead to ice, which can block flow in restrictors resulting in positioner malfunction, and then it may vanish without a trace after causing an unexplained upset. A simple mechanical separator will normally not remove enough moisture to properly protect the instruments. Force-Balance Positioners The force-balance positioner shown in Figure 6.2e has an element that compares the force generated by the input signal with the force generated by the feedback spring connected to the valve stem. Figure 6.2f shows the electropneumatic force-balance positioner. Motion-Balance Positioners The motion-balance positioner in Figure 6.2g compares the motion of an input bellows or diaphragm with linkage attached to the valve stem. Either can be very accurate. Bellows-type input elements are generally thought to be more accurate than diaphragms, and although slightly more likely to fail in fatigue, both types are used successfully. New and mostly electronic positioners differ widely in design and performance. Valve actuator Spool valve Signal diaphragm Feedback spring Air supply Exhaust 3 15 PSIG input signal Ring magnet Input current Restriction Coil Nozzle Relay valve Exhaust Pivot Balanced beam Feedback spring Control valve stem FIG. 6.2e Force-balance positioner. Positioner cam 20 PSIG supply Control valve stem FIG. 6.2f Electropneumatic force-balance positioner.

51 6.2 Accessories and Positioners PSIG ( bar) input signal Spring Bellows deflection Bellows Restriction 20 PSIG (138 kpa) supply Nozzle Exhaust Relay valve Valve stroke Stepping motor Traverse nut assembly Range spring Nozzle (1st pneumatic stage) Flapper beam Feedback spring Air supply Restriction Feedback cam Pilot (2nd pneumatic stage) Control valve FIG. 6.2g Motion-balance positioner. Control valve stem FIG. 6.2h Digital-to-pneumatic valve positioner. Positioner Accessories Electrohydraulic Positioners The electrohydraulic positioner addresses the need to control actuators on very large valves and for high-pressure differential applications. These systems typically require an external motor-driven pump, reservoir, pressure control regulators, high-pressure accumulators, and filters. The positioners may use high pressure oil jet tubes that are directed by linkages to regulate the oil pressure in the actuator. Hydraulic pressures are typically 750 1,500 PSIG (5,000 10,000 kpa). A hydraulic pressure accumulator is often provided to store the energy for peak requirements. The valve vendors can suggest or supply the hydraulic system packages. Specification issues to be determined include the use of flammable or nonflammable oils, electrical classification, peak needs, noise, space required, costs, and maintenance. If the area has several large valves, a larger central hydraulic system may be used. Digital to Pneumatic Positioners The digital valve actuators are described in Section 6.3. Some of the digital to pneumatic positioner designs have used rotary motors to control the pilot system. These tend to feature lock-in-place on loss of input signal. Others use a small fast solenoid valve that switches rapidly between open and closed to create an average air pressure for the actuator. Still others may use a piezoelectric valve (electrical signals cause a deflection in a special crystal structure), either proportional or pulsating. A pulse stepping motor may rotate a shaft to set a follower, which develops an input signal force or position (Figure 6.2h). The remainder of the positioner is pneumatic. Stroke speed may be limited by stepping motor response. The only way to make decisions during the selection process with the many different designs with their many subtle differences may be to rely on proven performance. A number of options are available for most positioner designs, such as gauges to display the supply, signal, and output pressures. These add little cost but are a great help during checkout and maintenance. Bypass air switches will connect the controller output directly to the actuator. These were once popular when 15 PSIG (100 kpa) pressure was adequate to move the springdiaphragm actuator if the positioner was direct acting (increase input increase output). The bypass is rarely specified now because this feature permits only very limited maintenance on the positioner and most modern high-performance positioners do not offer this option. There are also safety implications in providing the possibility of changing the valve response. At one time, the usual justification for the bypass valve was to provide a way to get a poor performing positioner out of the loop for field troubleshooting. In some corporations the only control valves now purchased without positioners are small (1 in. [25 mm] and less), for low-pressure differential (0 20 PSIG [0 140 kpa]), and less critical service (comfort heating and ventilation), and where absolutely minimum cost is required. Mechanical switches are used to provide a signal to the operator or control system that the valve stem is at or beyond a specified position. The specifications for these switches should consider the voltage and current surges caused by lightning strikes on long wire runs. These are available to suit the local electrical classification requirements. For some manufacturers of piston-type actuators, springs to provide for fail-safe action on loss of air are considered an accessory and are not supplied except as specified. Position Indicators Globe-style valves normally have a simple stem position pointer mounted on the stem with a simple scale on the yoke. A local valve position indicator may be mounted on rotary valves to

52 1096 Control Valve Selection and Sizing allow the operator to see the state of the valve. Stem position transmitters, usually derived from positioner designs, are used to provide independent remote indication of valve stem position. Coil ( voice coil ) Nozzle/baffle Zero set TRANSDUCERS I/P (Electropneumatic) Transducers A large variety of converters (transducers) are discussed in detail in Section 3.3. The electropneumatic transducer (I/P) converts electrical signals (usually 4 20 ma DC) into a pneumatic signal, usually 3 15 PSIG ( barg). The most common application is the interface between an electronic controller output and the pneumatic control valves. It is also used between digital control systems and control valves. A few designs use signal feedback to improve accuracy (Figure 6.2i). Most I/P transducer designs are of the motion-balance type (Figure 6.2j), where the small force developed by the milliampere current through a coil in a magnetic field causes motion in a nozzle-baffle assembly, resulting in a changing pneumatic pressure. The nozzle-baffle system is an amazingly accurate mechanism for measuring small distances. When an I/P is used within a control loop between the controller and the control valve, its error is combined with the valve error, which is detected by the loop controller and then driven towards zero error. Repeatability and reasonable linearity are required, and most I/Ps have advertised accuracies of % of full scale. Most I/Ps have relatively low air capacity, and a booster relay may be needed to drive a pneumatic actuator unless a positioner is used. Digital Electropneumatic Transducers A variety of devices are used to convert digital signals in addition to the electronic digital-to-analog converter. One device uses a stepper motor, as in the digital positioner mentioned above. Permanent magnet FIG. 6.2j In a motion-balance electropneumatic transducer, the coil is of similar construction to the voice coil of a loudspeaker. Leads are arranged to go through a pivot point to reduce their effect on output. A second device has a number of wires connected to it and receives data on all these wires simultaneously. This parallel data signal is converted to an analog current, and this controls the output pressure. Another device responds to a string of pulses ( serial data) to set the output pressure. RELAYS Booster Relays A booster relay, Figure 6.2k, is a device that amplifies pneumatic signals in volume (capacity) or in pressure, or both. Most booster designs were derived from a pressure regulator, with the input signal providing the loading force in place of the regulator spring. Internal pressure feedback Input signal Pivot Restriction Air supply Output Feedback bellows + DC input signal Zero adjustment N S N S Nozzle To valve actuator Output Relay valve Exhaust Restriction 20 PSIG (138 kpa) supply FIG. 6.2i Electropneumatic force-balance transducer design and components. Air supply FIG. 6.3k Volume booster relay. Booster output

53 6.2 Accessories and Positioners 1097 may improve accuracy. The downstream-facing pitot tube compensates for flow-related pressure drop. The feedback diaphragm has a smaller area than the input diaphragm so that greater pressure is required to achieve force balance. Volume boosters with a pressure ratio of 1:1 (gain of 1) are sometimes used to speed up a valve actuator response if a positioner is not used. The booster cannot overcome inaccuracies due to friction or forces on the valve plug, but it will reduce their effects because of faster response. The stroking times of pneumatic piston actuators (Section 6.4) are a function of their size and of the size of the connecting tubing. A size 300 actuator with 3 / 8 in. (9.5 mm) tubing has a stroking time of 27.7 sec. If the output of the positioner operating that actuator is piped to one booster on each side of the piston, the stroking time is reduced to 5.4 sec, and if two boosters in parallel are used on each side, the striking time is reduced to 2.7 sec. Large control valves and large actuators may require flow or pressure boosters installed between the positioner and the actuator to achieve the required speed. One special dead band booster design, Figure 6.2l, does not respond until an (approximately) 1 PSIG (6.9 kpa) difference between input and output is exceeded. A built-in needle valve allows a limited airflow to bypass the booster gain portion and provides adjustable damping. Springs provide the 1 PSIG dead band with this relay. Very large volume amplification occurs for fast input signal change, but it does not amplify a slow change. One common application for this relay is the centrifugal compressor antisurge control, where the control valve must open very quickly (1 3 sec) but then is required to throttle smoothly. The needle valve is adjusted with the complete control system assembled and operating. Careful tuning is vital to proper operation. With the booster needle valve fully closed, the positioner-valve system may be unstable. Best operation occurs when the needle valve is open only enough to smoothly dampen the oscillations. If more than one of these relays is used to add capacity for one side of the piston actuator, it is absolutely necessary to set the needle valves as identically as possible to avoid a complex and probably unstable interaction between the relays. It is worth repeating here that no air device will operate properly without an adequate air supply. With their large airhandling capacities, valve booster relays require in. (12 20 mm) air supply and output tubing to the valve operator. Large filters are used to minimize any air supply restriction. In one special case, a large air tank 10 ft 3 (0.3 m 3 ) was installed near the valve in order to provide adequate local air surge capacity. Reversing and Other Relays Besides the booster function, pneumatic relays can meet other control valve requirements. The fixed-gain-plus-adjustablebias relay for split-ranging was mentioned in connection with Figure 6.2c. Another 1:2 reversing relay is shown in Figure 6.2m. This design has a reverse gain of two (0% input results in 100% output, 50% input results in 0% output). It will reverse the operation of a valve and provide the gain and bias needed for split-ranging, while retaining the native failure action on loss of the supply air. Signal in Bypass needle valve Set bias Exhaust Signal input Exhaust Output Output Air supply Supply FIG. 6.2l The design of a dead band booster in which the springs provide the 1 PSIG (6.9 kpa) dead band. FIG. 6.2m 2:1 reversing relay.

54 1098 Control Valve Selection and Sizing The relay is a force-balance device. The output pressure generates a force on a diaphragm with one half the area of the input diaphragm so that the output pressure must be twice the input pressure. A wide variety of relays are available, amplifying, reducing (gains of less than one), computing (averaging, square root, and so on.), selecting (highest, lowest), some with provision for biasing the output. Complex control concepts have been implemented entirely with pneumatic logic. This functionality is more commonly accomplished today with electronic circuits or in the control system. Very successful override control systems have been designed with one valve position controlled by several process variables such as pressure, level, and temperature. Table 6.2n lists the air capacity for a few selected pneumatic components to show the range of typical devices. The data was found in various places and some is old. If the exact capacity of a relay or the stroking speed of a valve is important, it would be wise to check with the manufacturer. Quick-Exhaust Relays The quick-release valve or quick-dump valve, Figure 6.2o, is a pilot valve that opens a high-capacity vent when the input pressure drops below a set pressure. This is used where it is needed to quickly open a valve that vents a vessel or system during a shut-down. It is not unusual for a positioner to bleed excess air from the actuator for some seconds before a valve begins to move, because the actuator is completely filled (saturated) with supply pressure air. The quick exhaust helps reduce the lost time. The controller continues to call for valve action after it reaches its mechanical stop. TABLE 6.2n Air Capacity of Selected Pneumatic Components Model Air Capacity PSI Supply Function Data Source Moore 750E 7.5 SCFM 60 PSIG Positioner Catalog Moore 71 4 sec, 3 15, 350 in 3 18 PSIG Positioner (1) Catalog Moore I/P 0.1 SCFM 20 PSIG I/P converter Catalog Moore 50M 3.7/2.1 SCFM 20 PSIG Controller Catalog Moore 61H 10.5 SCFM 9 PSIG Booster relay Catalog Moore 61L 4.5 SCFM 9 PSIG Booster relay Catalog Moore 61F 2.4 SCFM 9 PSIG Booster relay Catalog Moore 610F 2.4 SCFM 9 PSIG Booster relay Catalog Moore SCFM 9 PSIG Multi-input low-signal select Catalog Moore 66BA2 2.2 SCFM 9 PSIG 1:2 A relay (2) Catalog Moore GC SCFM 9 PSIG Amp + bias relay Catalog Moore SCFM 9 PSIG Computing relays (1) Catalog Moore GC 77 NA (low) E/P transverter Catalog Moore 750 I/P 7.5 SCFM 60 PSIG I/P positioner Catalog Moore SCFM 60 PSIG Positioner Catalog Moore SCFM (3) 25 PSIG Positioner Moore paper ASCO SCFM(4) 30 PSIG Solenoid valve Catalog Valtek Booster 2300/1350(5) 60 PSIG High capacity Catalog Fisher SCFM 60 PSIG Standard capacity Catalog Fisher SCFM 60 PSIG High capacity Catalog Valtek Beta 11 SCFM 60 PSIG Positioner Catalog Notes: 1. Typical 2. Typical for family of amplifying and reducing relays 3. Higher available as option 4. Typical three-way, universal, no minimum pressure difference requirement, used for valve interlock 5. Supply/exhaust

55 6.2 Accessories and Positioners 1099 Inlet To actuator Exhaust FIG. 6.2o Quick-exhaust valve in which the elastomer flapper normally allows air to flow to the actuator while sealing the exhaust; if the inlet pressure is below that in the actuator, it allows the air from the actuator to be exhausted. Control pressure Quick-exhaust three-way solenoid valves having a large exhaust port for a similar effect can also be used. To, from valve Relays to Lock-up Valve Position To, from controller FIG. 6.2p Lock-up relay: If control pressure drops below the spring setting of this relay, then the relay valve closes and seals in the last air pressure in the actuator. A lock-up relay or fail in last position relay, Figure 6.2p, is available to seal in the existing actuator pressures if the air supply is lost. It is intended to hold the valve in the last controlled position. If the valve is partly open at the time of air failure, this frozen position is not absolutely predictable because the servo-action is lost, variable plug forces may move the stem, and air may leak out in an unpredictable manner. When a piston actuator is to be locked in its last position when the air supply fails, it is necessary to install two lockup valves at the two ends of the piston and use a three-way switching valve to vent their actuators when the air supply pressure drops to some preset limit (Figure 6.2q). (Plugged) Actuator F In EXH CYL N Lock-up valve (normally closed) Lock-up valve (normally closed) F N 3-way switching valve Vent Air supply N - Normal F - Fail In F N CYL EXH (Plugged) 3 15 psi signal ( bar) Output 2 Output 1 Instrument Supply FIG. 6.2q A valve with piston actuator, which is provided with fail-in-place lock-up controls. (Courtesy of Valtek, Flowserve Corp.)

56 1100 Control Valve Selection and Sizing Check valve Volume tank Actuator F In CYL EXH N Lock-up valve (normally closed) Lock-up valve (normally closed) Bleed valve F A N B 3-way switching valve D Vent C N - Normal F - Fail CYL F Vent 3 15 psi signal ( bar) N In EXH Output 2 Output 1 Instrument Supply Positioner Air supply FIG. 6.2r External air tank can provide the energy source to guarantee a positive failure position for piston-actuated valves. (Courtesy of Valtek, Flowserve Corp.) Failure Position Guaranteed by Stored Air To air header Check valve Signal Tank Air set Supply FIG. 6.3s A temporary air supply can be provided by an air accumulator tank, which thereby will guarantee orderly shutdown. In piston-operated valves, the air stored within the piston can also be used as the energy source to move the valve to a safe position when the air supply is lost. This can be achieved by locking the air in on one side of the piston while venting the other, when the air supply pressure drops to a predetermined value. If there is concern that the piston volume is insufficient to fully stroke the valve under all conditions when the air supply might fail, an external tank can be used as the source of driving energy. Figure 6.2r illustrates such an installation for a fail-closed (air-to-open) valve. Some piston-type actuators have no spring, and others have a spring too weak to guarantee valve action on loss of air supply. For these, an air tank (typically 0.5 ft 3, or m 3 ) with a check valve on the inlet may be used to allow limited operation or orderly shutdown (Figure 6.2s). Because everything leaks, this will only be reliable for a reasonable time. Longer term requirements may be satisfied if the process pressure tends to hold the valve plug in the desired direction. ENERGY SUPPLIES Air Sets The air set is the air regulator with filter and drip pot used to supply air to the positioner or other instrument. It is often purchased with the valve, mounted, and piped. The regulator must have the pressure range to cover the spring range of the actuator, and it should have a built-in overpressure vent to protect the actuator; a gauge is valuable to aid in setting the output pressure and for discovering a failed or plugged supply. Flow capacity is a problem only with very large valves. Some piston actuators and positioners will operate with up to 150 PSIG (1 MPa) supply, and there is a trend toward eliminating the regulator, but the filter is retained for reliable valve operation. Most manufacturers recommend that each control valve have its own individual air set. The risk in providing air sets for high-pressure actuators such as air cylinders is that field operators can thereby limit the supply pressure, which in turn reduces actuator stiffness (resistance to the dynamic forces of the process). An air set must be used when the pressure rating of the actuator or positioner is lower than the air supply pressure. The use of air filters is always recommended. They should be installed in the air supply serving the positioner

57 6.2 Accessories and Positioners 1101 and should be designed for the maximum air supply pressure. Their purpose is to remove moisture, oil, and all particles that are 5 µ or larger. In industry it is axiomatic that dirt is everywhere, and it will always find small restrictions to plug. Hydraulic (High-Pressure) Operation High-pressure, high-performance, hydraulic oil-filled systems are used for special needs, such as very large valves or those that require forces beyond 150 PSIG air supply. These function much like larger versions of the hydraulic power steering used on automobiles. Very high speed performance and power is available, all at added cost. Most of the accessories and functions commonly found in pneumatic systems are available. These include: positioners, flow boosters, and pressure regulators. Hydraulic (Water) Operation There are two sorts of hydraulic operations. The first was addressed above. The second hydraulic system is the ubiquitous but rarely considered municipal water supply and wastewater systems, which have their own special set of requirements and solutions. Actuators, pilot valves, regulators, and control relays are available that operate using water pressure just like the pneumatic systems used elsewhere. These control the water to fill elevated storage tanks and stop flow when the tanks are filled, and then implement the desired operating logic when water pressure is lost. Even fluid flow rate flow control is possible. Self-contained water operation eliminates the need for electrical or air power and signals. Relays and valves to control water pressure surges can be provided. Users should contact the appropriate specialized suppliers for further information. LIMIT SWITCHES Switches are installed on electric motor-driven valves to open the circuit and stop driving the motor when the valve is at its limit (fully open or closed) or on motor over-torque. The name limit switch is also used to describe switches installed to signal when a valve is at or beyond a predetermined position. These switches are used for operator information, interlock inputs, or computer feedback. It is necessary to consider the mounting problems, electrical classification of the area, the electrical characteristics of circuit, overtravel of actuating arm, and corrosive nature of the area. Usually, it is easiest to purchase the valve complete with the required switches already installed. Because of environmental problems, some users have been using sealed magnetically actuated or proximity switches. Note that it is difficult to adjust limit switches closer than ±5 10 % and that dead bands of 2 5% can occur. Problems with failure of small contacts have been traced to voltage surges picked up on long field wiring. FIG. 6.2t Installation of proximity switches to detect the open and closed position of a valve or a linear valve actuator. (Courtesy of Valtek, Flowserve Corp.) Plantwide standardization on limit switch specifications for makes and model number will reduce spare part storage and simplify maintenance. When specifying the limit switches, one should specify the required contact ratings, the contact configurations (SPDT, DPDT, and so on), and the type of housing required. Typical choices include weatherproof, explosionproof, or hermetically sealed explosionproof. Figure 6.2t illustrates a proximity switch installed on a linear actuator. The maximum spacing allowable between the switch and the sensed surface is 0.11 in. (3 mm). The switch is available in UL- and CSA-approved explosionproof designs. SOLENOID VALVES The solenoid valve (see Section 6.3 for more details) as a control valve accessory is used (1) to operate on/off pneumatic actuators or (2) to interrupt the action of modulating valves by switching air or hydraulic pressures. It is common practice to use a solenoid valve as the pilot for a pneumatically operated on/off valve because of the wide choice of features and capabilities available in the solenoid valve. Solenoid valves are primarily used as parts of start-up or shutdown, interlock, or batch systems to cause the control valve to take some predetermined action under certain conditions. Three philosophies are in common use. Three-Way Solenoids φ φ S In the first, the 3 15 PSIG ( bar) signal to the positioner is blocked and the downstream tubing either vented or connected to some other preset pressure (Figure 6.2u). This approach is reliable because the solenoid valve is lightly stressed, the positioner and valve have been in continuous use, and any failure or poor operation should have been detected during normal operation. With this scheme, the substituted signal can be any value over the operating range, the

58 1102 Control Valve Selection and Sizing From controller N - Normal I - Interlock Preset pressure Ext. feedback to controller FIG. 6.2u Under normal conditions the solenoid passes the controller signal to the positioner. Under abnormal conditions the interlock solenoid valve blocks the controller signal and opens the path for a manually preset air signal to reach the positioner. valve will go to the desired opening, and the advantages of the positioner are retained. If the actual positioner input is fed back to the controller external feedback connection, then a smooth return to normal control may be expected when the control loop is returned to normal operation. In the second philosophy, the solenoid valve is installed in the tubing between the positioner and the actuator (Figure 6.2v). Solenoid valves with adequate pressure rating and flow capacity are required. Only three control valve actions are possible: fully closed, fully open, or lock-up existing pressures. A third philosophy is to shut off the air supply to the positioner and let the valve act as designed on loss of air supply. This is one scheme used in batch operations where the valves are switched into service when a process unit or pump is turned on. In order for the system to shut down (valve to close) in case of loss of power or emergence of an unsafe condition, it is desirable for the solenoid valve to be continuously energized during normal operation. This will guarantee that any failure, loss of power, or a broken wire will cause a fail-safe action. As a control valve accessory, usually a three-way (three ports) solenoid valve is required. Some designs require that pressure be always applied to one certain port and that another certain port always be used as the vent. This does not always suit the required logic, but valves can be found designed for universal operation where there is more freedom in assigning port function. S Vent N I P Air supply Also, note that clean, dry, oil-free instrument air provides no lubrication, and some types of solenoid valves (spooltype) will have a short life or become unreliable without lubrication. See also the discussion of the impact of solenoid valve operation on valve response in Section 6.9. Four-Way Solenoids For on/off cylinder-operated valve actuators, four-way solenoids are often used (Figure 6.2w). They are fast, provide positive operation, and are available for a variety of AC or DC voltage services and with Class F coils for up to 310 F (154 C) temperature services. Solenoid Capacity Solenoid S D Airset E E: Energized D: Deenergized FIG. 6.2w On/off cylinder actuator operated by four-way solenoid. E D Vent Actuator Each approach must consider the flow capacity of these solenoid valves. The desired solenoid valve C v must be greater than the C v of the positioner to avoid a reduction in stroke speed. If they are equal, then the valve speed will be roughly half of that without the solenoid. It is typical of solenoid valves that small valves are directly operated and the larger ones are pilot operated. In pilot-operated valves, a small direct-operated valve uses air pressure to switch the larger main valve. Pilot-operated valves require a certain minimum air pressure differential in order to operate the main valve. If the solenoid valve is tripped while the air pressure is less than this pressure, the main valve will not change state and tripping the pilot will have no effect. The valve will trip later when the pressure differential becomes high enough to operate it. Where full pressure is always present this is not an issue. From controller N - Normal I - Interlock S I FIG. 6.2v Interlock solenoid valve controlling actuator directly. N P HANDWHEELS Handwheels are used to provide for partial or complete manual control of the valve and to override the pneumatic actuator. Some, mounted on top of the actuator, Figure 6.2x, can only push on the valve stem to close (or open with inverted trim). Others can be configured for continuous, bidirectional operation with force amplification ratios from 40:1 up to over 100:1. These top-mounted designs are illustrated in Figure 6.2y.

59 6.2 Accessories and Positioners 1103 FIG. 6.2x Top-mounted handwheel can be used to manually operate a springopposed pneumatic actuator, but only in one direction. Side-mounted handwheels have engagement clutches to allow the handwheel to fully stroke the valve open and closed, Figure 6.2z. In the continuously connected handwheel design, the handwheel is provided with a neutral position. When placed in that position, the handwheel does not interfere with automatic operation of the valve. FIG. 6.2z The side-mounted handwheel, if engaged, can fully stroke a springopposed pneumatic actuator in both directions. Turning the handwheel one way forces the stem to extend, and turning it the other way forces the stem to retract. Adjusting the handwheel screw away from the normal position introduces a limit stop on the valve travel in one direction or the other, but not both. Consider that some handwheels may interfere with interlock shutdown operation of the valve. For the manual operation of rotary valves, either hand levers in. (38 56 cm) in length or clutch-equipped gearbox-type handwheels are used. If manual throttling control is intended, the plant design must consider how the human operator will know how to set the valve and if the process can be safely controlled manually. This is not always the case, and handwheels have limited application in modern continuous process plants. LIMIT STOPS It is possible to install fixed limit stops to limit valve stem motion to either ensure a minimum opening or limit a maximum opening (Figure 6.2aa). These are usually purchased with the valve. Consider how to document the purpose and settings for these for maintenance purposes. BYPASS VALVE FIG. 6.2y Top-mounted, continuously operated bidirectional handwheel. (Courtesy of Valtek, Flowserve Corp.) Perhaps not thought of as an accessory, the manual bypass for steam shut-off valves are critical to start-ups. The manual valves are cracked open to pressurize and heat up the steam

60 1104 Control Valve Selection and Sizing 6. ANSI/ISA-TR , Control Valve Response Measurement from Step Inputs. 7. ISA , Method of Evaluating the Performance of Positioners with Analog Input Signals and Preumatic Output. Bibliography FIG. 6.2aa Externally adjustable limit stop, shown in a valve body subassembly. header. Opening the large valve could result in serious damage downstream as slugs of water are propelled into equipment. Other applications may have similar issues. References 1. Coughran, M. T., Measuring the Installed Dead Band of Control Valves, ISA, Gassmann, G. W., When to Use a Control Valve Positioner, Control, September Langford, C. G., A User s View of Process Control and Control Valve Positioners, ISA, Langford, C. G., A Method to Determine Control Valve Dynamic Requirements, ISA, ANSI/TSA , Test Procedure for Control Valve Response Measurement from Step Inputs. Arant, J. B., Positioner Use Is Myth-Directed, InTech, November Carey, J. A., Control Valve Update, Instruments and Control Systems, January Fernbaugh, A., Control Valves: A Decade of Change, Instruments and Control Systems, January Gassman, G. W., When To Use a Control Valve Positioner, Control, September Brayton, P., Installing Smart Positioners A Wise Move, Chemical Engineering Magazine, December 1, Jury, F. D., Positioners and Boosters, Instruments and Control Systems, October Keagle, J., It s 128 ma, Do You Know Where Your Positioner Is? InTech, May Lipták, B. G., Control Valves in Optimized Systems, Chemical Engineering, September 5, Lloyd, S. G., Guidelines for the Use of Positioners and Boosters, Instrumentation Technology, December Miller, T. J., Pneumatic Valve Actuators Continue to Dominate as Electronics Move into Process Control, Control Engineering, Vol. 28, No. 10, p. 88, Price, V. E., Smart Valve Intelligence Takes Many Forms, InTech, August Pyotsia, J., A Mathematical Model of a Control Valve, 1992 ISA Conference, Houston, TX, October Schuder, C. B., Fluid Forces in Control Valves, Instrumentation Technology, May Scott, A. B., Control Valve Actuators: Types and Application, InTech, January Shinskey, F. G., Dynamic Response of Valve Motors, Instruments and Control Systems, July Parkinson, G., Smart Actuator Incorporates All the Controls in One Package, Chemical Engineering Magazine, January 1, 2003.

61 6.3 Actuators: Digital, Electric, Hydraulic, Solenoid R S C. S. BEARD (1970) B. G. LIPTÁK (1985, 1995) Spring-loaded pneumatic cylinder actuator Dual acting pneumatic cylinder actuator Electric solenoid actuator shown with manual reset P. M. B. SILVA GIRÃO (2005) P H E H Pneumatic-hydraulic and electro-hydraulic actuators M M Flow sheet symbols Pneumatic and electric rotary motor actuators Types: 1. Digital 2. Electromechanical (linear and rotary) 2a. Stepping motors in smaller sizes 2b. Reversible motor gears for larger sizes 3. Hydraulic and electrohydraulic (the pump can be driven by stepping or servomotors) 4. Solenoid Energy Sources: Speed Reduction Techniques: Electric, hydraulic or both Worm gear, spur gear, or gearless Torque Ranges: ft lb f ( N m) for type 2a, and 1 250,000 ft lb f ( ,954 N m) for type 2b actuators with gearboxes Linear Thrust Ranges: Speeds of Full Stroke: Costs: The maximum of about 500 lb f (2224 N) output force can be obtained from type 2a actuators; ,000 lb f (445 44,500 N) can be obtained from type 2b ones. The thrust capability of type 3 actuators can exceed 100,000 lb f (445,000 N). Small solenoids can close in 8 12 msec. Throttling solenoids can stroke in about 1 sec. Electromechanical motor-driven valves stroke in sec. Electrohydraulic actuators generally move at 0.25 in./sec (6.35 mm/sec) but can be speeded up by the use of hydraulic accumulators in. (6.35 mm) solenoid pilot in stainless steel costs about $50; with a twoway design in explosionproof plastic construction it costs about $250, and when built for high-temperature service it costs about $300. An electromechanical, type 2a actuator for in. ( mm) small ball valve can be obtained in explosionproof construction with limit switches for on/off service for $750; with positioner and feedback potentiometer such an actuator costs about $

62 1106 Control Valve Selection and Sizing A typical rotary actuator with 300 ft lb f (407 N m) torque rating costs about $2500. The cost of larger type 2b or type 3 actuators can exceed $10,000. Partial List of Suppliers: ABB Group (1, 2) ( Aeroflex Incorp. (2) ( Allenair Corp. (2, 3, 4) ( Auma Actuators Inc. (1, 2) (www. auma-usa.com) ASCO-Automatic Switch Co. (Division of Emerson) (4) (www. asco.com) Bafco Inc. (3) ( Barksdale, Inc. (3) ( com) Bettis-Emerson Process Management (2, 3) ( Bodine Electric Co. (2) ( Bosch-Rexroth Corp. (2, 3, 4) ( Bray Controls-Bray International, Inc. (2, 4) ( Bürkert Fluid Control Systems (1, 4) ( Butler Automatic, Inc. (3) ( Center Line-Crane Co. (2) ( Circle Seal Controls Inc. (4) ( Clark-Cooper Corp. (4) ( Curtiss-Wright Flow Control Corp. (3, 4) ( Danaher Corp. (2, 3) ( Danfoss (4) ( Detroit Coil Co. (4) ( DeZurik/Copes-Vulcan (1, 2, 3) ( Eaton Corp. (2) ( El-O-Matic-Emerson Process Management (2) ( products/index.html) Engineering Measurements Comp. (Emco) (1) ( Emerson Process Management (4) (www. emersonprocess.com) ETI Systems, Inc. (1, 2) ( Exlar Corporation (1, 2) ( Flo-Tork Inc. (3) ( GE Water Technologies (1, 2) ( Hoke, Inc. (2) ( Honeywell Automation and Control (1, 2) ( Humphrey Products Co.(4) ( Invensys-Eurotherm (2, 3) ( Jordan Valve (2) ( Kammer Valves-Flowserve Corp. (1, 2) ( stm) Keane Controls Corp. (4) ( Keystone Valve USA Inc. (2, 3) ( Leslie Controls, Inc. (1, 2) ( Limitorque Corp.-Flowserve Corp. (1, 2) ( Metso Automation Inc. (1, 2, 3) ( McCanna Inc.-Flowserve Corp. (2) ( Micro Mo Electronics Inc. (2) ( Nihon Koso Co. Ltd; (1, 2, 3) ( Norgren-Herion (4) ( OCV Control Valves (1, 2, 3, 4) ( Oil City Valve Automation (2, 3, 4) ( Oilgear Co. (3) ( Oriental Motor USA Corp. (2) ( Parker Hannifin Corp. (1, 2, 3, 4) ( Plast-O-Matic Valves Inc. (2, 4) ( Regin Hvac Products Inc. (2) ( Rotork Controls Inc. (2, 3) (

63 6.3 Actuators: Digital, Electric, Hydraulic, Solenoid 1107 Saint-Gobain Performance Plastics-Furon Fluid Handling Div. (4) ( Samson Controls, Inc. (1, 2, 3) ( Servo Systems Co. (2) ( Servotronics Inc. (4) ( Shafer-Emerson Process Management (3) ( Shore Western Mfg. Inc. (3) ( com) Siemens AG (1, 2, 3) ( SMAR International Corp. (1, 2) ( Snap-Tite Inc. (4) ( Sonceboz Corp. (2) ( Spirax Sarco (2) ( Superior Electric Co. (2) ( Thunderco Inc. (3) ( Tyco Valves (2, 3) ( Valcor Scientific (4) ( Valtek-Flowserve Corp. (3) ( Vetec Ventiltechnik GmbH (2, 3) ( Worchester Controls-Flowserve Corp.(2, 3) ( Note: More information may be obtained for instance at com, particularly at categories.html, at or at INTRODUCTION This section covers a variety of valve actuators, except the pneumatic ones, which are discussed in Section 6.4. The various accessory items such as positioners, handwheels, limit switches, potentiometers, and the like are also discussed in Section 6.2. The discussion of digital actuators in this section is intended to complement Section 6.12 on intelligent valves and Section 6.18 covering digital control valves. The discussion in this section begins with some selection and application guidelines, which is followed by the description of the five actuator categories: 1) digital, 2) electromechanical, 3) electrohydraulic, 4) motors and pumps, and 5) solenoids. The section ends with a brief discussion of the trends in valve actuation, including the features of microprocessor-based smart and intelligent actuators. SELECTION AND APPLICATION The following are some of the characteristics to consider in the application and selection of all types of actuators. Table 6.3a gives a summary of advantages, disadvantages, and applications for some of the designs. Actuator Types Valve actuator types discussed in this section belong to one of the following categories: (a) electric actuators, which use a motor to drive a combination of gears to generate the desired torque or thrust level. This category includes (i) rod linear actuators, whose output rod provides linear motion via a motor-driven ball or ACME screw assembly. In this design, the actuator s load is attached to the end of a screw, or rod, and is often unsupported; (ii) rodless linear actuators, whose load is attached to a fully supported carriage. Rodless linear actuators provide linear motion via a motor-driven ball screw, ACME screw, or belt drive assembly; and (iii) electric rotary actuators that use a motor to drive components rotationally. The next category (b) are the hydraulic and electrohydraulic valve actuators, which convert fluid pressure into motion. They include (i) linear actuators or hydraulic cylinders, which use a cylinder and hydraulic fluid to produce linear motion and force, and (ii) hydraulic rotary actuators that use pressurized hydraulic oil to rotate mechanical components. The third category (c) are the linear solenoids that convert electrical energy into mechanical work via a plunger with an axial stroke in either a push or pull action. They can be rated for continuous duty (100% duty cycle operation, continuous duty solenoids) or for off-on applications, less than 100% duty cycle (intermittent duty solenoids). The last category (d) are digital actuators, which now include all types of valve actuation solutions where the valve is digitally controlled. Pneumatic actuators are still the technology favored by valve actuators buyers. Nevertheless, the market share of electric, hydraulic, and electrohydraulic actuators is increasing, while the overall use of solenoid values has dropped. 1 Actuator Features Speed and Torque Ranges Speed requirements vary from less than 1 rpm to about 160 rpm. The upper limit of available torque is about 250,000 ft lb f (339,000 N m) with gearboxes.

64 1108 Control Valve Selection and Sizing TABLE 6.3a Applications, Advantages, and Disadvantages of Various Actuator Designs 2 Actuator Types Advantages Disadvantages Applications Electromechanical High thrust High stiffness coefficient Powered by electricity or pneumatics High thrust High stiffness Fast speeds Powered by electricity; no pneumatic source required Direct interface with computer system Complex design No mechanical fail safe Large, heavy structure Complex design Large, heavy structure Hydraulic temperature sensitive Linear or rotary valves 2 36 in. body size Electrohydraulic Linear or rotary valves 2 in. to unlimited Electric (servomotor or stepping motor) Large structure Low thrust No mechanical fail safe Slow speed Linear valves 1 / 2 2 in. body size Stem thrusts and rotational drive torques are limited only by the size of the motor used and the ability of the gear, bearings, shafts, and so on to carry the load. Speed of operation depends on the gear ratios, adequate prime move power, and means of overcoming the inertia of the moving system for rapid stopping. This is most important for proportional control uses. Some actuators have a limited selection of drive speeds, while others are furnished in gear ratios in discrete steps of 8 20% between speeds. Manual Operation Manual operation is sometimes necessary for normal operational procedures, such as start-up, or under emergency conditions. Only units that are rotating very slowly or those with low output should have continuously connected handwheels. Most units have the handwheel on the actuator with a clutch for demobilizing the handwheel during powered operation. Clutches are manual engage and manual disengage, or manual engage and automatic disengage upon release of the handwheel. Others are manual engage when the handwheel is rotated, with the motor reengaging when the handwheel is not being rotated; or power re-engage, which takes the drive away from the operator upon energization, leaving the handwheel freewheeling. Electrical Equipment Most of these actuators include much of the electric gear within the housings of the unit. Components such as limit, auxiliary, and torque switches and position or feedback potentiometers are run by gearing to stem rotation, so they must be housed on the unit. Installation is optional concerning pushbuttons, reversing starters, lights, control circuit transformers, or line-disconnect devices in an integral housing on the unit. Any or all of these components may be located externally, such as in a transformer, switch, or control house. The enclosures must be designed to satisfy NEMA requirements for the area. High Breakaway Force Resistance to opening requires a method of allowing the motor and gear system to develop speed to impart a hammer blow, which starts motion of the valve gate or plug. Selection of motors with high starting torque is not always sufficient. The dogs of a dog-clutch rotate before picking up the load, or a pin on the drive may move within a slot before picking up the load at the end of the slot. Systems are used that delay contact for a preselected time or until the tachometer indicates the desired speed of rotation. Torque Control for Shutdown Torque control for shutdown at closure or due to an obstruction in the valve body is accomplished in numerous ways, but each one uses a reaction spring to set the torque. When rotation of the drive sleeve is impeded, the spring will collapse, moving sufficiently to operate a shut-off switch. Position Indication An indicator can be geared to the stem rotating gear, but it becomes a problem when actuator rotation varies from 90 to as much as 240 revolutions. Gearing of a cam shaft operating the position indicator or auxiliary switches must be calculated to obtain a fairly uniform angle. Upon correct gearing, an indicating arrow or transmitting potentiometer can be rotated. Maintenance of Last Position This is no problem when the actuator includes a worm gear or stem thread. Use of spur gears can cause instability when positioning a butterfly or ball valve. Status quo is obtained by use of a motor brake or insertion of a worm gear into the system. Protection against Stem Expansion The status quo ability of a thread or worm gear is detrimental when the valve itself is subjected to temperatures high enough to expand the stem. This expansion, when restrained, can damage the seat or plug, bend the stem, or damage the actuator thrust bearings. One of the original patents for this type of actuator included Belleville springs to allow the drive sleeve to move with the thermal expansion and relieve the linear force.

65 6.3 Actuators: Digital, Electric, Hydraulic, Solenoid 1109 Mounting Methods Industry has dictated a set of dimensions for the mounting flanges and bolt holes for newly manufactured valves. Retrofit mounting requires adaptation to existing valves. Mounting requires a plate to match the existing valve, which is screwed into the yoke upon removal of the manual drive sleeve, welded or brazed to the yoke, or, for a split yoke, bolted to the yoke. Adaptability to Control This feature includes adaptability to many voltages and to single- or polyphase supplies. Polyphase motors of V and 60 Hz predominate. Single-phase motors up to about 2 horsepower (hp) (1,492 W) are used. Reversing starters with mechanical interlocks are used for both proportional and on-off service. The coils that open and close the contacts are energized by an open-center double-pole switch, which can be incorporated in the automatic control circuit. For manual control, the starter may be of the type that maintains contact, requiring the open or close button to be held in position. Some units are also wired for momentary depression so that the actuator runs until it reaches the limit switch, or until a stop button is depressed. s Inputs x Analog computing amplifier s x kr SCR A Meter Differential amplifier s x kr=0 Link Phase-shift network Driving motor SCR B 0 Zero reference Sampling program circuitry Isolation transformer 115 V 50/60 cycles Proportional Control Proportional control of these large units can be accomplished by including the coils of the reversing starter in a proportional control circuit. This requires a position feedback, which may be a potentiometer. For this type of control, a Wheatstone bridge circuit would be used. The reversing starter controls any voltage or phase required by the motor. A transducer to transform any of the accepted electronic controller outputs (e.g., 1 5 ma) to a resistance relative to controller output permits use of the actuator in these systems. A smaller unit, with a force output of 1460 lb f (6494 N) at stall and 500 lb f (2224 N) at a rate 0.23 in./sec (5.84 mm/s), uses solid-state control of the motor. At the same time it eliminates stem position feedback into the controller. A DC signal x (Figure 6.3b) from a process transmitter provides the loop with its measurement, such as temperature, as well as such high-responds systems as flow. A differential amplifier responds to the magnitude and the polarity of an internally modified error signal, which triggers silicon-controlled rectifiers (SCRs) to obtain bidirectional drive. The synchronous motor can be driven in either direction, depending upon the relation to the set point. DIGITAL VALVE ACTUATORS Digital actuators can accept the output of digital computers directly without digital-to-analog converters. Only simple onoff elements are needed for their operation. The number of output positions that can be achieved is equal to 2 n, where n is the number of inputs. Accuracy of any position is a function of the manufacturing tolerances. Resolution is established by the number of inputs and by the operating code selected for Alternate loads Load A Load B s = instantaneous set point x = instantaneous transmitter signal k = adjustable transmitter compensation r = instantaneous rate of change of transmitter signal FIG. 6.3b Proportional motor control circuit with position feedback. a given requirement. The smallest move achievable is called a 1-bit move. The code may be binary, complementary binary, pulse, or special purpose. A three-input piston adder assembly produces eight discrete bit positions. The adders in Figure 6.3c are shown in the 6-bit extended position. The interlocking pistons and sleeves will move when vented or filled through their selector valves. This same adder can be used to position a four-way spool valve, with a mechanical bias to sense position. The spool valve controls the position of a largediameter piston actuator or force amplifier. Use of a DC motor featuring a disc-armature with low moment of inertia has created another valve actuator particularly adaptable to a digital input. Brushes contacting the flat armature conduct current to the armature segments. Incremental movement is caused by half-waves at line frequency for rotation in either direction. The rotation of the armature is converted to linear stem motion by use of a hollow shaft internally threaded to match the valve stem. Actuator output is 5,000 lb f (22,240 N) maximum for noncontinuous service at a rate of 0.4 in./sec (10 mm/s) through a valve stroke of 3 in. (76.4 mm). The actuator is de-energized at stroke limits or at power overloads by thermal overload relays.

66 1110 Control Valve Selection and Sizing Normally open solenoid Adders (hydraulic D/A converters) Solenoids Normally closed solenoid Mechanical feedback beam Four-way valve Return Bias piston Pressure FIG. 6.3c Digital valve actuator. Actuator (force amplifier) Operation of the actuator requires application of a thyristor (SCR) unit designed for this purpose. This unit accepts pulses from a computer or pulse generator. The thyristor unit consists of two SCRs with transformer, triggers with pulse shift circuit, facility for manual actuator operation, and the previously mentioned thermal overload relays. The output consists of half-waves to pulse the armature of the actuator. Modules are also available for process control with the necessary stem position feedback, slow pulsing for accurate manual positioning, and full-speed emergency operation. The digital-to-pneumatic transducer shown in Figure 6.3d is used to convert a controller output to a pressure signal for operating a pneumatic valve. The concept of digital valve actuator has been extended to other valve actuating systems such as stepping motors, which are discussed in the following paragraph dedicated to Zero adjust 25 PSIG (172.5 kpa) filtered supply pressure Zero spring Range spring Beam FIG. 6.3d Digital-to-pneumatic transducer. Pivot Fixed orifice Motor Stops Pivot Non-rotating nut Nozzle 3 15 PSIG (0.2 1 bar) Output pressure electromechanical actuators, and to the so-called smart or intelligent actuators briefly introduced below in Smart and Intelligent Actuators. The discussion of digital actuators in this section is intended to complement Section 6.12 on intelligent valves and Section 6.18 covering digital control valves. It is in this context that some of the suppliers in the partial list of suppliers at the beginning of this section identified as providing digital valve actuators indeed provide modern smart valve actuating solutions. ELECTROMECHANICAL ACTUATORS Electric actuators can utilize a reversible electric motor provided with an internal worm gear to prevent drive direction reversal (back-drives) by unbalanced loads. These units can operate both linear and rotary valves. Servomotor drives can position valves in response to feedback signals from linear or rotary encoders. Two-phase AC servomotors are available with up to 1 HP rating, while direct current servomotors can meet higher loads. Stepping motors, which rotate the shaft by a discrete step angle when energized electrically, can also be used in valve actuators. The electromechanical device that rotates the shaft can be a solenoid used to operate a star-wheel or ratchet device; it can be the stepping movement of a permanent magnet, the flux of which causes poles or teeth to align and thereby affect rotation; or it can be a variable reluctance unit, where the rotor-stator poles or teeth are aligned by electric fields. When used as valve actuators, the stepping motors are well suited for direct digital control, and pulse feedback can be provided for accurate closed-loop positioning.

67 6.3 Actuators: Digital, Electric, Hydraulic, Solenoid 1111 Of the above designs the actuators using reversible motors are the most often used variety and therefore they will be discussed in more detail in the following paragraphs. Reversible Motor Gear Actuators Lever-operated valves and dampers are available with torques of ft lb f ( N m) and with full stroke speeds of seconds. For very large valves or dampers, this type of actuator can deliver torques of ,000 ft lb f (203 13,558 N m) and a full stroke speed of seconds. The power consumption of these larger units ranges from kva operating on 460 V three-phase power supply. Some of these designs can be obtained with springs to return the valve to a safe condition upon electric failure. The motor gear actuator consists of an electric motor connected to a gear train; the gear train rotates a stationary drive nut, which in turn drives the threaded or keyed valve stem up or down. As the valve plug contacts the seat, the resistance is transmitted to a Belleville spring, which at a preset limit interrupts the motor power circuit. For throttling applications, the actuator can be provided with a positioner that compares the external control signal (analog or digital) with an internal position feedback signal and keeps turning the motor until the error between the two signals is eliminated. Rotary Output Actuators Worm Gear Reduction The actuator shown in Figure 6.3e is an example of the use of a double worm gear reduction to obtain output speeds of around 1 rpm with an input motor speed of 1800 rpm. The worm gear is self-locking, so it prevents the load from moving downward by back-driving the motor. However, the worm gear is less than 50% efficient, so more power is used compared to spur gears. Also shown is a handwheel for manual operation during power loss. Spur Gear Reduction The spur gear actuator in Figure 6.3f has very low power loss through the relatively efficient spur Hammer blow Drive sleeve Geared limit switches Torque control Prime mover Gear drive FIG. 6.3e Electric actuator with worm gear reduction. De-clutching handwheel Travel limit switches FIG. 6.3f Spur gear reduction. Output shaft Adjustable cams (Optional) potentiometer Auxiliary switches gears. Ordinarily the load could back-drive this system, but a friction member at the motor end of the gear train minimizes this undesirable action. Of course, if the load has a frictional characteristic, it will not impose a back-driving torque on the actuator. Actuators for larger outputs are externally mounted motors. Two motors operating a single gear train have been used to obtain 3,750 ft lb f (5084 N m) of torque through 90 rotation in 75 sec. An actuator has been designed for accurate rotary positioning that develops 5000 ft lb f (6800 N m) at stall and will rotate 90 in 10 sec. SCRs energize the motor as commanded by a servotrigger assembly housed separately. Electrical gear may include adjustable cams to operate limit and auxiliary switches. A potentiometric feedback calibrated to the rotation of the actuator is required for use with a control circuit. The unit shown in Figure 6.3f was developed to slide over and be keyed to the shaft of a boiler damper. Actuators of this type must be adaptable to mounting on and operating a variety of quarter-turn valves. Because of the difficulty in setting limit switches to accurately stop the valve in the shut position, it is advisable to incorporate a torquelimiting device to sense closure against a stop. Opening can be controlled by a limit switch. One compact unit contains the features noted with an output of 750 ft lb f (1017 N m) of torque. The unit is powered by a motor with a high-torque capacitor and includes a mechanical brake, feedback potentiometer, limit switches, and a de-clutchable hand wheel. Flex-Spline Reduction Figure 6.3g is an example of a unique single-stage, high-reduction system. Instant breakaway and efficient transfer of prime mover power is obtained

68 1112 Control Valve Selection and Sizing Reversible low-inertia AC motor Continously connected handwheel Motor drive pinion gears Load connection linear motion FIG. 6.3g Flex-spline gear reduction. with a modified concentric planetary system consisting of a semiflexible gear within a rigid gear. A three-lobed bearing assembly transmits power to the gears by creating a deflection wave transmission that causes a three-point mesh of the gearing teeth on 30% or more of the external gearing surface. The semiflexible geared spline has fewer teeth than the nonrotating internal gear it meshes with. The spline slowly rotates as it is pressed into the larger gear by the bearings on the motor shaft. Linear Output Actuators FIG. 6.3h Rack and pinion assembly converts the rotation of a horizontal, motor-driven shaft into vertical, linear motion. or a lever arm fixture that fits the shaft configuration of a plug cock or ball valve. Maximum output is 1600 lb f (7117 N) at 5 in./min (127 mm/min). Rotating Armature An internally threaded drive sleeve in the armature of the motor is used to obtain a linear thrust up to 6,600 lb f (29,360 N) at a rate of 10 in./min (254 mm/min). Bearings in the end cap support the drive assembly (Figure 6.3j). The drive stem is threaded to match the drive sleeve and is kept from rotating by a guide key. Motors and Rack Figure 6.3h uses a worm and a rack and pinion to translate horizontal shaft motor output to vertical linear motion. Maximum force output is approximately 1500 lb f (6672 N) at about 0.1 in./min (2.5 mm/min). A continuously connected handwheel, which must rotate the rotor of the motor, can be used when there is short stem travel and relatively low force output. The actuator is designed with a conventional globe valve bonnet for ease of mounting. Units operate on 110 V and have been adapted to proportional use with a 135 ohm Wheatstone bridge or any of the standard electronic controller outputs. Terminal strip Motor Electric brake Shaft Screw drive Eye Motor and Travelling Nut Linear unit consists of the motor, gears, and a lead screw that moves the drive shaft (Figure 6.3i). A secondary gear system rotates cams to operate limit and auxiliary switches. The unit may have a brake motor for accurate positioning and a manual handwheel. The bracket on the rear end allows the actuator to rotate on the pin of a saddle mount, so that the drive shaft can be pinned directly to the lever arm of a valve Swivel flange Traveling nut Auxiliary switcheslimit switches located on opposite side FIG. 6.3i Electric quarter-turn actuator with linear output and with limit switches.

69 6.3 Actuators: Digital, Electric, Hydraulic, Solenoid 1113 FIG. 6.3j Electric actuator with rotating armature has been adapted for proportional control. Thrust-limit switch assemblies are mounted in each end of the housing to locate the hollow shaft in mid-position. When the linear movement of the drive stem is restricted in either direction, the limit switch involved will operate to shut down the unit. Thermal cut-outs in the motor windings offer additional overload protection. Strokes are available from 2 48 in. ( mm). The unit has been adapted for proportional control by use of an external sensing position for feedback. For use as a valve actuator, it must be mounted so that the drive stem can be attached to the valve stem, or a suitably threaded valve stem must be supplied. Rotary to Linear Motion An electric proportional actuator (Figure 6.3k) is designed for continuous rotation of a drivesleeve on a ball-screw thread. 3,000 lb f of thrust (13,335 N) is obtained at a stem speed of 1 in./min (25 mm/min). One or two DC signals are used separately or numerically added or subtracted. Triacs operate on position error to control a DC permanent magnet motor that positions a stem within an adjustable dead band. Degree of error and rate of return are sensed by a lead network to determine the direction and time that the motor must run. Stem reversals are almost instantaneous. The back emf of the motor is used as a velocity sensor and is fed into a circuit that allows adjustment of the speed of drive sleeve rotation. Gain can be adjusted to control oscillation of the stem. Stem position feedback is by a linear variable differential transformer (LVDT). Use of the DC motor allows for torque control through sensing of motor current. A manual handwheel is furnished that can only be used when the unit is deenergized by a manual/automatic switch. Application of the linear requirement of a valve necessitates a linkage for translation from rotary to linear motion. Use of a linkage (Figure 6.3l) provides a thrust for operating the valve. Motor Lift scale Motor shaft Linkage arm Linkage arm clip Lift adjustment locknut Locknut Shaft adjustment screw Washer A Relief spring mechanism Washer B Bolt Cylinder Screw Stem button clamp Stem button Stem Bracket Allen set screw FIG. 6.3k Proportional electrical drive converts rotary to linear motion. FIG. 6.3l Electric actuator with linkage to convert rotary motion to linear.

70 1114 Control Valve Selection and Sizing ELECTROHYDRAULIC ACTUATORS The main features of electrohydraulic actuators have been summarized in Table 6.3a. They are often used where instrument air is unavailable or where the required actuator stiffness or thrust cannot be obtained from pneumatic actuators. They are heavier than pneumatic actuators; therefore, in order to avoid straining the bonnet, vertical upright installations are recommended. Electrohydraulic actuators are superior to electromechanical ones in the areas of speed, positioning without overshoot, actuator stiffness, and complexity. Electrohydraulic actuators can be stepping motor or servovalve driven. In the servovalve designs the pumps are running continuously, while in the stepping-motor configuration they run only when the valve needs repositioning. Electrohydraulic actuators can be mounted directly onto the stems of valves that are 6 in. (152 mm) or smaller and can be provided with springs to guarantee a fail-safe position in case the hydraulic fluid is lost. The actuator usually consists of a hydraulic cylinder, a pump with motor, some feedback linkage on valve position, and a balancing arrangement. The balancing or positioning mechanism compares the external control signal with the valve position and actuates the hydraulic system if repositioning is required. In spring-loaded designs the hydraulic fluid might drive the valve in one direction (opening it, for example), while the spring drives it in the other direction (closing it, for example). Open, closed, or last-position failure positions are available. When higher thrusts, longer strokes, or rotary valves are involved, the fail-safe position can be provided by storing energy in a hydraulic accumulator instead of using a spring. Hydraulic actuators are sensitive to viscosity changes caused by ambient temperature variations and, therefore, in subfreezing temperatures must be provided with heaters, which can require substantial heat-up periods before the valve can be operated. Because the hydraulic fluid is incompressible, the actuator is stiff and provides stable and accurate positioning with hysteresis and dead band within 5% of span. Their speed of operation is similar to pneumatic actuators in. ( mm) per second of stem movement but can be speeded up substantially by the use of high-pressure accumulators. Intelligent, programmable electrohydraulic actuators can be provided with bidirectional hydraulic gear pumps, which are driven by microprocessorcontrolled stepping motors. In the broad sense, the use of two three-way solenoids or one four-way solenoid externally mounted to the actuator constitutes an electrohydraulic system. More extensively, electrohydraulic applies to a proportionally positioned cylinder actuator. This requires a servosystem, which is a closed loop within itself. A servo-system requires one of the standard command signals, which is usually electrical but can be pneumatic. This small signal, which often requires amplification, controls a torque motor or voice coil to position a flapper or other form of variable nozzle. This positions a spool valve or comparable device to control the hydraulic positioning of a high-pressure second-stage valve. The second-stage valve directs operating pressure to the cylinder for very accurate positioning. Closing the loop requires mechanical (Figure 6.3m) or electrical feedback to compare the piston position with the controller output signal. The electrical feedback can be a servoamplifier, illustrated in connection with a linear hydraulic actuator shown in Figure 6.3n. Hermetically Sealed Power Pack A much more compact electrohydraulic actuator combines the electrohydraulic power pack with the cylinder in one package. Many of these actuators are designed as a truly integral unit. An electric motor pump supplies high-pressure oil through internal ports to move the piston connected to the stem (Figure 6.3o). The small magnetic relief valve is held closed during the power stroke until de-energized by an external control or Command signal Feedback springs Feedback cam & follower T Hydraulic supply Torque motor Hydraulic supply Servoloop Spool valve External Hydraulic Source The term electrohydraulic has been applied to actuator systems in which the hydraulic pressure to one or more actuators is supplied by a hydraulic mule. The hydraulic power is supplied to the actuator by electrical control means. FIG. 6.3m Two-stage servovalve with mechanical feedback. Actuator position

71 6.3 Actuators: Digital, Electric, Hydraulic, Solenoid 1115 Integral manifold end cap Protective cap Analog absolute LDT (position sensor) (B) Permanent magnet Pressure compensated variable volume pump Pressure filter Servovalve (A) Servovalve command Position feedback Position transducer (B) Actuator (C) Servoamplifier 4 20 ma command signal Servovalve (A) Piston with low friction seals Highly polished cylinder bore (C) Complete plumbing Sonic wand Industry standard mounting flange Split stem clamp for easy disassembly Valtek standard yoke-to-body mounting Sectioned view Exterior view FIG. 6.3n Servovalve-operated electrohydraulic linear valve actuator. (Courtesy of Valtek, Flowserve Corp.) emergency circuit to allow the spring to cause a down stroke. The same unit can be used to cause the spring to return to the up position using a Bourdon switch to produce force limit. Motor and Pump Combinations Reversible Motor and Pump A reversible motor can be used to drive a gear pump in a system to remove oil from one side of the piston and deliver it to the other side (Figure 6.3p). M Piston Travel limit pull type FIG. 6.3o Hermetically sealed electrohydraulic power pack. M Piston Force limit push type The check valves allow the pump to withdraw oil from the reservoir and position the directional control valve in order to pressurize the cylinder. Reversing the motor (and pump) reverses the direction. When the motor is de-energized, the system is locked up. For proportional control, feedback is necessary from stem position to obtain a balance with the control signal. Jet Pipe System A very old control system for a cylinder, the jet pipe, is employed in an electrohydraulic actuator. An electromechanical moving coil in the field of a permanent magnet is used to position a jet that can direct oil to one end or the other of the cylinder actuator (Figure 6.3q). Forcebalance feedback from stem position creates the balance with the controller signal. Hydraulic Control of Pinch Valves Controlled hydraulic positioning of a sleeve valve is obtained with a moving coil and magnet to position the pilot (Figure 6.3r), which controls pressure to the annular space of the valve. Feedback is in the form of a Bourdon tube, which senses the pressure supplied to the valve and moves the pilot valve to lock in that pressure. Multiple Pump The multiple pump system consists of three pumps running on the shaft of one prime mover (Figure 6.3s). There is one pump for each side of the piston and one for

72 1116 Control Valve Selection and Sizing Piston Directional control valve Centering springs Bourdon element Hydraulic pressure to valve Internal relief valve Check valve #1 Check valve #2 Motor running forward Check valve #4 Check valve #3 Tank line Internal relief valve DC + Input AC Input Capacitor Electric motor Feedback spring Relief valve Filter Gear pump Start-up filter Rubber sleeve type slurry valve Piston Cylinder Piston hydraulically locked Oil reservoir FIG. 6.3r Electrohydraulic control of a jacketed pinch valve. Check valve #2 Check valve #1 Reservoir Motor shut off Check valve #4 FIG. 6.3p Pump with reversible motor combination, where the actuator is locked when the pump is de-energized. Input signal Force motor Jet pipe Shut-off valve Feedback spring Feedback linkage to jet pipe FIG. 6.3q Electrohydraulic actuator with jet pipe control. Check valve #3 (Oil) Motor Internal relief valve (as required) Pump Suction the control circuit. The force motor tilts a flapper to expose or cover one of two control nozzles. The flow through a restricting nozzle allows pressure to be transmitted to one side of the piston or the other. Force-balance feedback is created by a ramp attached to the piston shaft, which positions a cam attached to the feedback spring. Upon loss of electric power, the cylinder shut-off valves close to lock up the pressure in the cylinder and assume a status quo. A bypass valve between the cylinder chambers allows pressure equalization to make use of the manual handwheel. Vibrating Pump A vibratory pump is used in a power pack mounted on a cylinder actuator (Figure 6.3t). A 60 Hz alternating source causes a plunger to move toward the core, and, upon de-energizing, a spring returns it. The pump operates on this cycle and continues until the piston reaches the end of its stroke, when a pressure switch shuts it off. The solenoid that retains the pressure is de-energized by an external control circuit. Maximum stem force is 2,500 lb f (11,120 N) at a rate of about 1 in./min (25 mm/min), or 5,000 lb f (22,240 N) at 0.3 in./min (7.6 mm/min). Two-Cylinder Pump A two-cylinder pump, driven by a unidirectional motor, injects pressure into one end of a cylinder or the other, depending upon the positions of two solenoid relief valves (Figure 6.3u). The solenoid on the left is closed to move the piston to the right, with hydraulic pressure relieved through the other relief valve. Motion continues until the valve it is operating is seated. The build-up of cylinder pressure operates the pressure switch at a predetermined setting to de-energize the motor and both solenoid relief valves. This locks the hydraulic

73 6.3 Actuators: Digital, Electric, Hydraulic, Solenoid 1117 Input signal R -restriction Zero spring R 1 R 2 Force motor Cylinder shut-off valves Feed back spring Air bleed Feed back cam By-pass valve Drain off connection 3 section pump with biult-in relief valves Pressure to hydraulic amplifier Pressure to upper cylinder Pressure to lower cylinder Suction to pump FIG. 6.3s Hydroelectric valve actuator with three pumps running on the shaft of one mover. pressure in the cylinder. Switching of the three-way switch will start the motor and reverse the sequence of the relief valves to move the piston to the left. At full travel (which is the up position of a valve stem), a limit switch shuts down the unit. Open-centering the three-way switch at any piston position will lock the piston (and valve stem) at that point. Remote control is accomplished by manipulating the open-center switch. Automatic control is acknowledged by including this open-center function, which may be solid state, in the control circuit. A potentiometer or LVDT that senses the stem position is required for feedback. Stem output is 6,000 lb f (26,690 N) at a rate of 3 sec/in. (0.12 sec/mm). Stem travels up to 7 in. (178 mm) are available, Solenoid relief valve Intake R 2 R 1 Pump Motor (or solenoid air valve) Intake Relay Solenoid relief valve Actuator spring Actuate up Neutral Indicator stem Relief valve Signal Presssure switch Solenoid Solenoid valve Limit switch Limit switch To reservoir Actuator cylinder Pressure switch Actuator piston Check valve Valve stem Core Piston spring Plunger Coil Oil sump Power supply To reservoir To reservoir Three-way switch FIG. 6.3t Electrohydraulic actuator with vibratory pump. FIG. 6.3u Two-cylinder pump-type electrohydraulic actuator.

74 1118 Control Valve Selection and Sizing although longer travels are feasible. The entire system is designed in a very compact explosionproof package that can be mounted on a variety of valve bonnets. The speed of response to energizing and de-energizing the control circuit makes it feasible to adapt the unit to digital impulses. SOLENOID VALVES Solenoids move in a straight line and therefore require a cam or other mechanical converter to operate rotary valves. These actuators are best suited for small, short-stroke on-off valves, requiring high speeds of response. Solenoid-actuated valves can open or close in 8 12 ms. Their fast closure is not always an advantage: In water systems, it can cause water hammer. They are limited to pressure drops below 300 PSIG (20.7 bars), although when provided with pilots, levers, or double seats, they can handle higher pressure drops. They are available in two- or three-way designs, with power requirements ranging from W with VAC or VDC power supplies. Solenoids are reliable devices, and they can provide multimillion cycles on liquid service. Solenoids (consisting of a soft iron core that can move within the field set up by a surrounding coil) are used extensively for moving valve stems. Although the force output of solenoids may not have many electrical or mechanical limitations, their use as valve actuators has economic and core (or stem) travel limitations, and they are expensive. A solenoid valve consists of the valve body, a magnetic core attached to the stem and disc, and a solenoid coil (Figure 6.3v). The magnetic core moves in a tube that is closed at the top and is sealed at the bottom; this design eliminates the need for packing. A small spring assists the release and initial closing of the valve. The valve is electrically energized to open. FIG. 6.3w When valve packing is required, the increased friction can be overcome by a solenoid valve design with two or more strong return springs. Stronger springs are used to overcome the friction of packing when it is required (Figure 6.3w). Reversing the valve plug results in reverse action (open when de-energized). Even stronger stroking force can be obtained by using the force amplification effect of a mechanical lever, in combination with a strong solenoid (Figure 6.3x). Using a solenoid to open a small pilot valve (Figure 6.3y) increases the port size and allowable pressure drop of solenoid-operated valves. Solenoid coil Housing Shading coil Magnetic flux paths Stationary core (plug nut) Coil connections Core tube Movable core (plunger) Bonnet Spring Stem Spring retainer Body Spring compressed Inlet Outlet Intlet Outlet Disc Direction of flow through valve De-energized Orifice Energized FIG. 6.3v The operation of a direct-acting solenoid valve involves the lifting of the plunger when energized.

75 6.3 Actuators: Digital, Electric, Hydraulic, Solenoid 1119 Armature for push solenoid Pilot poppet Pilot flow Pilot exhaust Main poppet Pilot flow entrance Pilot metering orifice FIG. 6.3x The force generated by the solenoid is amplified in the lever-type actuator. FIG. 6.3z Pilot-operated in-line solenoid valve, where the pilot serves to apply pressure to the main poppet piston. Small solenoid pilot valves are widely employed to supply pressure to diaphragms or pistons for a wide range of output forces. Pilot operation applies pressure to a diaphragm or piston or may release pressure, allowing the higher upstream pressure to open the valve. A good example is the in-line valve (Figure 6.3z). Most solenoid valves are designed to be continually energized, particularly for emergency shutdown service. Thus the power output is limited to the current whose I 2 R-developed heat can be readily dissipated. Using a high source voltage and a latch-in plunge overcomes the need for continuous current. The single-pulse valve-closing solenoid is disconnected from the voltage source by a single-pulse delatch solenoid and hence does not heat up after it is closed. A pulse to the delatch solenoid permits the valve to be opened by a spring. Three-way solenoid valves with three pipe connections and two ports are used to load or unload cylinders or diaphragm actuators (Figure 6.3aa). Four-way solenoid pilot Cushioned closing Manual opening Return spring Outlet Inlet Vent De-energized position Open Closed FIG. 6.3y The solenoid valve can close against higher pressure drop, if the solenoid operates a small pilot valve. Energized position FIG. 6.3aa The operation of a normally closed three-way solenoid valve.

76 1120 Control Valve Selection and Sizing valves are used principally for controlling double-acting cylinders. Modulating Solenoid Valves Modulating magnetic valves (Figure 6.3bb) utilize springloaded low-power solenoids to provide throttling action. The only moving part in this design is the valve stem, which has the valve plug attached to one end and an iron core attached to the other. The valve opening is thereby a function of the voltage applied across the solenoid. (Figure 6.3cc illustrates the change in plunger force as a function of applied voltage and plunger air gap.) Such throttling solenoids are frequently used in the HVAC industry and in other applications where the valve actuators do not need to be very powerful. The actuator thrust requirements are lowered by balancing the inner valve through the use of pressure equalization bellows or floating pistons. The positive failure position of these valves is provided by spring action. Throttling solenoids are typically available in 1 / 2 8 in. ( mm) sizes and are limited by the pressure difference against which they can close. They are available in two- or three-way designs and can handle water flows up to 5000 gpm (19 m 3 /m). The use of throttling solenoids can completely eliminate the need for instrument air in the control loop. These valves ON/OFF valves position reed switches Modulating valves LVDT Magnet Core rod Indicator tube Return spring Solenoid assembly Plunger Fixed core Return spring Bonnet tube Valve inlet Inlet orifice Pilot disc Main disc Pilot valve discharge orifice Pilot disc Main disc Main disc Pilot disc Vent port Flow Flow FIG. 6.3bb Throttling solenoid valve design with LVDT used to measure the stem position for feedback.

77 6.3 Actuators: Digital, Electric, Hydraulic, Solenoid 1121 Plunger force pounds (Kg) 120 (54) 110 (50) 100 (45) 90 (41) 80 (36) 70 (32) 60 (27) 50 (23) 40 (18) 30 (14) 20 (9) 10 (4.5) (2.5).2 (5) B ON/OFF valve spring characteristics.3 (7.5).4 (10).5 (13) Modulating spring valve characteristics.6 (15).7 (18).8 (20) Plunger air gap in inches (mm).9 (23) 1.0 (25) FIG. 6.3cc The force available to the pilot plunger of a throttling solenoid depends on the voltage and the air gap. A 100 V 90 V 80 V 70 V 60 V Microprocessor control board Motor O-ring seals cover & shaft Anodized aluminum bracket Stainless steel shaft Brass nickel plating Anodized aluminum housing Position indicator Heavy duty stainless steel geartrain New universal coupler Nickel plated swivel adapter Valve VA valve actuator VK valve kit FIG. 6.3dd Microprocessor-based electric valve actuator. (Courtesy of ETI Systems, Inc.) offer a higher speed and better range than their pneumatic counterparts. Some manufacturers claim a range of 500:1 and a stroking time of 1 sec. The design illustrated in Figure 6.3bb can be driven directly from microprocessor-based building automation systems. The design shown in Figure 6.3bb also includes a separate positioner, which accepts a 4 20 ma DC input from the controller and delivers a DC output signal to the throttling solenoid. Valve position feedback is obtained through the use of a linear variable differential transformer mounted directly on the valve. SMART ACTUATORS In addition to specifying the thrust, travel, failure position, control signal, power supply, speed, electrical area classification, and ambient conditions that an actuator must meet, one can also consider the use of microprocessor-based smart systems and of some other types of intelligent systems. One of the trends in process automation is the use of distributed control and, in some cases, to locate the controller as close as possible to the final control elements (control valves). Often the control signal is digital (digital valve control, or DVC 3 ). It is in this respect that an increasing number of features have been added to valves, valve actuators, and positioners through the integration of microprocessors with them. Applications Microprocessors have been included in valve actuators and valve positioners for some time (Figure 6.3dd), but the degree of exploitation of the possibilities of the smart and intelligent actuators for the purposes of process control is highly dependent on data access and thus on digital networking capabilities. The tendency is to provide new actuators with some sort of network connection. 4 As discussed in more detail in Sections 4.16 and 6.11, Foundation fieldbus, HART, and Profibus are some of the leading suppliers. Open architectures, such as Foundation fieldbus, have made positive contributions, because by broadening the market they allowed more users to take advantage of better performing actuating systems at lower prices. Moreover, the compatibility problems between different fieldbus systems are in the process of being resolved, and it is hoped that through international standardization plug and play capabilities 3 will soon be available for such devices. The intelligence provided by microprocessors can and is used for valve tuning, semiautomatic calibration, and data collection for maintenance and diagnostic purposes. It is also used for standalone control when that might improve the positioning, the protection, or the communication of the valve. Table 6.3ee lists some of the more common applications of smart and intelligent actuator systems.

78 1122 Control Valve Selection and Sizing TABLE 6.3ee Applications for Integrated Intelligent Systems 5 Under-instrumented loops where additional information is required for optimization or improved control. Loops requiring very tight control with a fast update for the PID algorithm. Applications requiring a large turndown in flow measurement and control capabilities. Processes that need local supervisory and control capabilities to provide continuous operation or controlled shutdown upon loss of control signal. New or retrofit installations where there is insufficient space for conventional instrumentation systems. Critical systems requiring continuous monitoring and predictive diagnostics on the valve or process. Applications where changing conditions require the loop to be reconfigured to control diverse variables for best control. Standalone remote applications requiring programmable operation and remote interface for monitoring. New applications where an intelligent system can be used as a fully integrated system, reducing engineering, cost, and maintenance requirements. Specialized control functions such as gap control for very large turndown, or self-contained minimum pump recirculation systems for pump or compressor protection. Master station Microprocessor-based systems are available in watertight or explosionproof (NEMA 6 to 7) construction and can tolerate ambient temperatures from 40 to 185 F ( 40 to 85 C), as well as the presence of moisture, fungus, or dust. They can incorporate sensors, an electronic positioner, or an electronic proportional, integral, derivative (PID) controller, which can operate off a digital or analog external set point. Locating the controller card at the valve and dedicating it to full-time control is an advantage on very fast critical loops, such as compressor surge protection, where otherwise surge could evolve while the central DCS control system is scanning/updating other loops. Smart systems can also take advantage of the lower cost of digital communications over a single loop of two-conductor or fiber-optic cable (Figure 6.3ff). This method of communication substantially reduces the wiring cost of installations that can control up to 250 actuators, pumps, or solenoids. The smart valves can be remotely calibrated and reconfigured and can be used to detect such performance changes as pressure changes resulting from a fouled pump or pipeline. They can also limit the valve travel to stay within the range where the required characteristics (gain) are available. It is perhaps at the maintenance level that the most benefits can be obtained through the use of smart and intelligent systems. 6 Smart actuator circuitry can protect electric motor operators from burning out when the valve is jammed or from reverse phasing. More importantly, by close monitoring of FIG. 6.3ff Multiple intelligent actuators can be monitored and controlled by a two-wire multiplexer loop. (Courtesy of Rotork Controls Ltd.) the valve, the required information can be obtained to evaluate the valve and actuator performance and thereby replace preventive and corrective maintenance by predictive maintenance with the benefits of improved process performance and products quality at a lower cost. References 1. Harrold, D., A Changing Landscape, Control Engineering, December 2003, text = valve + actuators. 2. Scott, A. B., Control Valve Actuators: Types and Application, InTech, January Harrold, D., Making Valve Controllers/Positioners Smarter is Smart Business, Control Engineering, January 2003, net/ctl/index.asp?layout=articleprint&articleid=ca Hoske, M. T., Implementing Industrial Networks, Control Engineering, July 1998, valve+actuators. 5. Price, V. E., Smart Valve Intelligence Takes Many Forms, InTech, August Merritt, R., A Real (Valve) Turn-On Control Design, August 2002,

79 6.3 Actuators: Digital, Electric, Hydraulic, Solenoid 1123 Bibliography Barnes, P. L., Protect Valves with Fire-Tested Actuators, Instrument and Control Systems, October Baumann, H. D., Trends in Control Valves and Actuators, Instrument and Control Systems, November Berris, R., Hazony, D., and Resch, R., Discrete Pulses Put Induction Motors into the Stepping Mode, Control Engineering, Vol. 29, No. 1, 1982, p. 85. Colaneri, M. R., Solenoid Valve Basics, Instruments and Control Systems, August Cottell, N., Electrohydraulic Actuation-Still in Control?, IEE Colloquium on Actuator Technology: Current Practice and New Developments, May 1996, London, U.K. Fitzgerald, W. V., Loop Tuning and Control Valve Diagnostics, Paper # , 1991 ISA Conference, Anaheim, CA, October Hammitt, D., How to Select Valve Actuators, Instruments and Control Systems, February Hoerr, D., Low-Power DC-Solenoid Valves, M&C News, September IEC Ed. 1.0 b:1985, Industrial Process Control Valves. Part 6: Mounting Details for Attachment of Positioners to Control Valve Actuators, IEC Ed. 1.0 b:1997, Industrial Process Control Valves. Part 6: Mounting Details for Attachment of Positioners to Control Valves Section 1: Positioner Mounting on Linear Actuators, IEC Ed. 1.0 b:2000, Industrial Process Control Valves. Part 6-2: Mounting Details for Attachment of Positioners to Control Valves. Positioner Mounting on Rotary Actuators, Kervin, D., Zero-Crossing Triac Drivers Simplify Circuit Design, Control Engineering, Vol. 29, No. 3, p. 76, Koechner, Q. V., Characterized Valve Actuators, Instrumentation Technology, March Liantonio, V., High Pressure Modulating Solenoid Valve for Steam/Gas Service, InTech, January Lipták, B. G., Control Valves in Optimized Systems, Chemical Engineering, September 5, Lockert, C. A., Low Load Technique Controls Motor Energy Losses, Control Engineering, Vol. 29, No. 2, 1982, p MSS SP , Guidelines for Metric Data in Standards for Valves, Flanges, Fittings and Actuators, Price, V. E., Smart Valve Intelligence Takes Many Forms, InTech, August Pyotsia, J., A Mathematical Model of a Control Valve, 1992 ISA Conference, Houston, TX, October Valve & Actuator User s Manual, British Valve & Actuator Association, Scott, A. B., Control Valve Actuators: Types and Application, InTech, January Shinskey, F. G., Dynamic Response of Valve Motors, Instruments and Control Systems, July Usry, J. D., Stepping Motors for Valve Actuators, Instrumentation Technology, March VDI/VDE 3845, Industrial Process Control Valves. Interfaces between Valves, Actuators, and Auxiliary Equipment, 1998.

80 6.4 Actuators: Pneumatic C. S. BEARD (1970) O. P. LOVETT (1985) B. G. LIPTÁK (1995) H. L. MILLER (2005) Diaphragm actuator Hand actuator Single-acting, cushion loaded piston actuator M Pneumatic rotary motor actuator Pressure balanced diaphragm actuator Spring-loaded pneumatic cylinder actuator Double-acting piston actuator Flow sheet symbols Types: Applicable to Valve Sizes: Standard Spring Ranges: Max. Actuator Pressure Ratings: Actuator Temperature Ratings: A. Linear A1. Spring-and-diaphragm A2. Piston B. Rotary B1. Cylinder with Scotch yoke B2. Cylinder with rack and pinion B3. Dual cylinders B4. Spline or helix B5. Vane B6. Pneumohydraulic B7. Air motor B8. Electropneumatic A to 8 in. (12 to 200 mm) A to 30 in. (12 to 750 mm) B. 2 to 30 in. (50 to 750 mm) A1. 3 9, 3 15, 9 15, 6 30 PSIG (20 60, , , kpa) A1. 60 PSIG (414 kpa); some higher A PSIG (1035 kpa); accessories may lower ratings B. 250 PSIG (1725 kpa) A1 and A2. 20 to 150 F ( 30 to 66 C); some higher B. 40 to 200 F ( 40 to 95 C); special up to 350 F (177 C) Actuator Areas: A1. 25 to 500 in. 2 (0.016 to m 2 ) A2 and B. 10 to 600 in. 2 (0.006 to 0.38 m 2 ); bore diameters from 2 to 44 in. (50 mm to 1.1 m) and strokes up to 24 in. (0.61 m) 1124

81 6.4 Actuators: Pneumatic 1125 Linear Thrust A to 45,000 lb f (22 to 10,100 N) (Stem Force) Ranges: A to 32,000 lb f (22 to 7,200 N); specials up to 186,000 lb f (41,800 N) Speeds of Full Stroke: Torque Ranges: Cost: Hysteresis, Dead Band, and Linearity: Partial List of Suppliers: A1. Small (25 to 100 in. 2 ) actuators, 1 to 5 sec. A to 24.0 in./sec (8 to 600 mm/sec) or less than 1 to 30 sec/stroke B to 6.0 in./sec (8 to 150 mm/sec) or 1 to 30 sec/stroke. Dual pistons (Figure 6.4aa) can stroke large rotary valves in 0.5 sec. For the same size units, spring-and-diaphragm units are the slowest, spring-returned pistons are faster, and double-acting pistons with supply/exhaust at both ends are the fastest (Values are given for double-acting actuators; for spring-return designs torque is about half.) B. 10 to 100,000 ft lb f (0.69 to 6850 N m); special units for 5 million ft lb f (343,000 N m) have been built Included in control valve cost A and B. Generally within 2%, but when spring-and-diaphragm or rotary piston actuators are operating rotary valves, the linearity is worse ABB Kent-Introl ( Actuation Valve & Control Ltd. ( Allenair Corp. ( Aluma Actuators Inc. ( Arca Regler Gmbh ( Bardiani Valvole SpA ( Bray Valve & Controls ( Cashco Control Valves & Regulators ( China Zhejiang Chaoda Valve Co. Ltd ( Circor International, Inc. (Leslie Controls) ( Combraco Industries Inc. ( Control Components Inc. (Bailey, CCI, BTG, STI, Sulzer, Shin Woo) ( com) Controlmatics Industrial Products ( Dresser (Leeden, Masoneilan) ( Emerson Process Management (Contex, Fisher, El-O-Matic, Shafer Valve) (www. Emersonprocess.com) Flo-Tork, Inc. ( Flowserve Corp. (Anchor/Darling, Kammer, Valtek, Worcester, McCanna, Limitorque) ( Jordan Valve ( Koso America Inc. (Hammeldahl) ( K-Tork International Inc. ( Larox Flowsys Inc. ( Nihon Koso ( Metso Automation USA, Inc. (Neles-Jamesbury, Valmet Automation) ( Mumatics Inc. ( Norriseal Controls ( Parcol SpA ( Red Valve Co. Inc. ( Rotork Controls Inc. ( Samson Regeltechnick.bv ( Severn Glocon Ltd. ( Spirax Sarco Inc. SPX Valves & Controls (Copes Vulcan, Dezurik) ( Tyco Flow Control Div. (Keystone, Morin, Biffi, Descoti, Sempell, Yarway, Grinnell, MCF) ( Welland & Tuxhorn Gmbh ( Wier Group plc (Atwood & Morrell, Batley Valve, Blakeborough Controls, Hopkinsons, Sebim, Flowguard) ( Xomax Corp. ( Yamatake Corporation (

82 1126 Control Valve Selection and Sizing INTRODUCTION Pneumatic valve actuators respond to an air signal by moving the valve trim into a corresponding throttling position. This section covers the two basic designs most frequently utilized: the diaphragm and the piston actuator. The discussion of diaphragm- and piston-type actuators is followed by the treatment of pneumatic-rotary and pneumatic-hydraulic actuators. In connection with the performance of these actuators, an analysis is presented of the various forces positioning the plug, including diaphragm, spring, and dynamic forces generated by the process fluid. An understanding of the interrelationships among these forces will allow the reader to properly size these actuators and make the correct spring selection. The failure safety of valve actuators and the relative merits of diaphragm vs. piston actuators and the topic of high-speed actuation using pneumatics is also discussed. This section is concluded with a summary of the results of a longterm evaluation of the performance of pneumatically actuated control valves in the field. DEFINITIONS An actuator is that portion of a valve that responds to the applied signal and causes the motion resulting in modification of fluid flow. Thus, an actuator is any device that causes the valve stem to move. It may be a manually positioned device, such as a handwheel or lever. The manual actuator may be open-closed, or it may be manually positioned at any position between fully open and fully closed. Other actuators are operated by compressed air, hydraulics, and electricity. The actuators discussed here are those capable of moving the valve to any position from fully closed to fully open and those using compressed air for power. Of such there are two general types: the spring-and-diaphragm actuator and the piston actuator. In a spring-and-diaphragm actuator, variable air pressure is applied to a flexible diaphragm to oppose a spring. The combination of diaphragm and spring forces acts to balance the fluid forces on the valve. In a piston actuator, a combination of fixed and variable air pressures is applied to a piston in a cylinder to balance the fluid forces on the valve. Sometimes springs are used, usually to assist valve closure. Excluding springs, there are two variations of piston actuators: cushion loaded and double acting. In the cushion-loaded type, a fixed air pressure, known as the cushion pressure, is opposed by a variable air pressure and is used to balance the fluid forces on the valve. In the doubleacting type, two opposing variable air pressures are used to balance the fluid forces on the valve. An actuator can be said to have two basic functions: (1) to respond to the external signal of a controller and cause an inner valve to move accordingly (with the proper selection and assembly of components, other functions can also be obtained, such as a desired fail-safe action) and (2) to provide a convenient support for valve accessory items, such as positioners, limit switches, solenoid valves, and local controllers. ACTUATOR FEATURES AND SELECTION Table 6.4a describes the applications and relative advantages of a variety of actuator designs. The table lists both the advantages and the limitations of the various designs. The popularity of the spring/diaphragm actuator is due to its low cost, its relatively high thrust at low air supply pressures, and its availability with fail-safe springs. By trapping the pressure in the diaphragm case, it can also be locked in its last position. It is available in springless designs, double diaphragm designs (for higher pressures), rolling diaphragm designs (for longer strokes), and tandem designs (for more thrust). One of the limitations of this design is the lack of actuator stiffness (resistance to rapidly varying hydraulic forces - for example, those caused by flashing). For such applications double-acting piston actuators are used and for extraordinary requirements hydraulic or electromechanical (motor gear) actuators may be preferred. A stiffer spring, 6 30 PSIG ( kpa), in a spring/diaphragm unit is sometimes sufficient to correct the problem. Linear piston actuators provide longer strokes and can operate at higher air pressures than can the spring/diaphragm actuators. When used to operate rotary valves, the linear piston or spring/diaphragm actuator does not provide a constant ratio of rotation per unit change in air signal pressure; therefore, the use of positioners is always a requirement. Rotary piston actuators operate at higher air pressure and can provide higher torques, suitable for throttling large ball or butterfly valves. The double-acting version of this actuator does not have a positive failure position, but this can be corrected by extending the piston case and inserting a helical spring. For higher torques (over 1000 ft lb f [68 Nm]), heavyduty transfer linkages are required (Scotch yoke or rack and pinion); such units cannot be disassembled and maintained in the field. These actuators also require positioners because the relationship between air signal change and resulting rotation is not linear. SPRING/DIAPHRAGM ACTUATORS This discussion is restricted to pneumatic actuators. The external signal, therefore, is an air signal of varying pressure. The air signal range from a pneumatic controller is commonly 0 18 PSIG (0 124 kpa). Signal or actuator input pressure starts at 0 PSIG, not 3 PSIG (21 kpa). A common mistake is to confuse the 3 15 PSIG ( kpa) range of transmitter output pressure with the signal to a valve. The higher value of 18 PSIG (124 kpa) is fixed only by the air supply to the controller (or positioner),

83 6.4 Actuators: Pneumatic 1127 TABLE 6.4a Features of Pneumatic Actuators Type of Actuator Advantage Disadvantage Application Linear spring-and-diaphragm Low cost Mechanical fail-safe Moderate thrust Small package Simple design Excellent control with or without control devices Moderate cost Moderate thrust Small package Simple design Excellent control with control device Long stroke High-speed options Moderate stiffness Moderate cost Mechanical fail-safe Small package Simple design Easily reversible Excellent control with control device Low cost Moderate thrust Small or large package Good control with control device Mechanical fail-safe option Slow speed Poor stiffness Instability Linear valves 1/2 8 in. ( mm) body size Linear piston Large spring compression when used for failure Linear valves 1/2 30 in. ( mm) body size Rotary spring-and-diaphragm Low thrust in spring cycle Instability Rotary valves 1 6 inch ( mm) body size Rotary pistons Slow speed Large spring Compression Rotary valves 1 24 inch ( mm) body size See Section 4.3 for a detailed treatment of the features of hydraulic and electric actuators. and it can easily be set to 20 PSIG (138 kpa) or higher. A variety of other input pressures are sometimes used, such as 0 30 or 0 60 PSIG (0 207 or kpa). Both the spring-and-diaphragm and the piston actuator produce linear motion to move the valve. These actuators are ideal for use on valves requiring linear travel, such as globe valves. A linkage or other form of linear-to-rotary motion conversion is required to adapt these actuators to rotary valves, such as the butterfly type. Steady-State Force Balance P A K X In spring-and-diaphragm actuators the stem positioning is achieved by a balance of forces acting on the stem. These forces are caused by the pressure on the diaphragm, spring travel, rubbing friction, and fluid forces on the valve plug (Figure 6.4b). Equation 6.4(1) can be derived from a summation of forces on the valve plug adopting the positive direction downward. PA KX P A =0 6.4(1) where A is the effective diaphragm area, A v is the effective inner valve area, K is the spring rate, P is the diaphragm v v Flow AT P v FIG. 6.4b Forces acting on a spring-and-diaphragm actuated, fail open (FO) control valve. pressure, P v is the valve pressure drop, and X is stem travel. Equation 6.4(1) applies to a push-down-to-close actuator and valve combination with flow under the plug. This type of actuator is commonly referred to as direct acting. A v

84 1128 Control Valve Selection and Sizing 100% (Closed) Stem travel Flow FIG. 6.4c Reverse-acting spring-and-diaphragm actuator, attached to a flow to open, spring to close (fail closed = FC) control valve. Another popular actuator configuration is one causing the stem to rise on an increase of air pressure. It is commonly called a reverse acting actuator (Figure 6.4c). By using the same sign convention, the force balance equation for this valve configuration is given in Equation 6.4(2). PA + KX P v A v = 0 6.4(2) If the flow direction is reversed in Figure 6.4b, the equation becomes: PA KX + P A =0 6.4 (3) Likewise, reversing flow direction in Figure 6.4c results in Equation 6.4(4): PA + KX + P v A v = 0 6.4(4) These equations are simplified because they do not consider friction and inertia. Friction occurs in the valve stem packing, in the actuator stem guide, and in the valve plug guide or guides. Usually, for static valve actuator sizing problems, negligible error is introduced by ignoring the friction terms. If Equation 6.4(1) is plotted as signal pressure vs. stem travel and if the case of no fluid forces on the plug (bench test) is assumed, then the curve shown in Figure 6.4d is obtained. Next, consider the case of plug forces due to fluid flow, assuming that the term P v is constant for all travel positions. This has the effect of shifting the straight line to the right to some position depending on the magnitude of P v. Curves similar to those in Figure 6.4d can readily be drawn for the other valve configurations represented by Equations 6.4(2), 6.4(3), and 6.4(4). The distance between the lines is the force resulting from A v is the effective inner valve area, K is the spring rate, P is the diaphragm pressure, P v is the valve pressure drop, v v (Open) (21) No plug forces intersecting the abscissa is the pressure needed to move the valve plug. Actuator Sizing Example With plug forces 6 (42) 9 (63) FIG. 6.4d In a fail open (FO) valve, as the air pressure on the diaphragm rises, the valve closes. The distance between the two sloping lines corresponds to the force acting on the plug (A v ), which force is generated by the differential between the up and downstream pressures (P v ), when these forces tend to keep the valve open, as in Figure 6.4b. Let us assume that the forces acting on a 1 in. (25 mm) singleported globe valve are to be evaluated. In that case: A = 46 in. 2 (0.03 m 2 ) X = 5/8 in. (15.9 mm) full travel K = 885 ft lb f /in. (7.83 N/mm) 12 (84) 15 (104) If no plug forces exist, Equation 6.4(1) reduces to PA = KX. Solving for the pressure change required to obtain full travel from open to closed: P = KX/A = (885)(5/8)/(46) = PSIG (83 kpa) Diaphragm pressure P (PSIG) (kpa) This is reasonably close to the 12 PSIG (0.83 kpa) desired operating span. Practical considerations of variations in spring constants and in actuator-effective areas usually prevent such a close approach to the desired span, and frequently a ± 10 % leeway is permitted. When there are plug forces, it is seen from Equation 6.4(1) that an additional actuator force is required to maintain balance. The actuator pressure required to begin stem motion can be calculated for the case of a 1 in. diameter (25 mm) plug (A v = π/4) and 100 PSIG (690 kpa) pressure drop. Equation 6.4(1) can be used to solve for P as follows (stem travel is zero, thus there are no spring forces): P = [(K = 885)(X = 0) + (P v = 100)(A v = π/4)]/(a = 46) = 1.7 PSIG (11.8 kpa) 6.4(5)

85 6.4 Actuators: Pneumatic 1129 Stem travel in. (mm) 2.4 (60) 2.0 (50) 1.6 (40) 1.2 (30) 0.8 (20) 3 PSIG * 1 PSIG * 15 PSIG * 0.4 (10) 1PSIG 0 75 (0.049) *1 PSIG = 6.9 kpa 3 PSIG = 21 kpa 15 PSIG = 104 kpa 100 (0.065) FIG. 6.4e The effective diaphragm areas vary with both the stem travel and with the pressure acting on the diaphragm. This means that the diaphragm pressure must increase to 1.7 PSIG (11.8 kpa) before stem travel begins. This is the distance between the two sloping straight lines in Figure 6.4d. Actuator Nonlinearities 15 PSIG 3 PSIG 125 (0.081) 1 PSIG Effective diaphragm area 150 (0.098) 15 PSIG 3 PSIG 175 in m 2 In practice we encounter many nonlinearities, and the ideal curves in Figure 6.4d are not obtained. These nonlinearities are due to several factors, such as the variable effective diaphragm areas. The effective diaphragm area varies with travel and with the pressure level on the diaphragm. Figure 6.4e illustrates this for three different sizes of diaphragms. Another source of nonlinearity is in the variation of the valve plug forces (P v A v ). Figure 6.4f illustrates the variations in these plug forces for two 4 in. (100 mm), single- and double-ported valves. The figure also shows the effects of flow over and under the plugs of a single-ported valve. Springs are also nonlinear in that the spring rates vary with travel. By judicious selection of springs, considering their spring rate and travel, the effects of their nonlinearity on the valve assembly can be minimized. When all of these nonlinearities are considered, a plot of actuator travel vs. diaphragm pressure would not be a straight line as shown in Figure 6.4d, but might be a curve such as the one shown in Figure 6.4g. A nonlinear curve, such as the one labeled actual in Figure 6.4g, is not necessarily objectionable. When used in an automatic control loop, the static nonlinearities are compensated for by the controller. This curve is actually a part of the gain term in the valve s transfer function, and the other part is the flow characteristic. When a valve positioner is used, the positioner overcomes these nonlinearities, and the result is similar to the ideal curve shown in Figure 6.4g. Stem force (tension or compression) in pounds* 1,300 1,200 1,100 1, FIG. 6.4f The forces acting on the valve plug are also nonlinear, because the pressure difference across the valve varies as a nonlinear function of stem travel. Dynamic Performance of Actuators (1) 4" size (100 mm), single - ported flow to close valve. Forces generated by the fluid flow create stem tension. (2) Same as (1), but flow to open design and therefore forces act to compress the stem (3) % Lift 0 (Stem travel) *lbs = kg 4" size (100 mm), doubleported valve with flow between the seats. At low lifts tension, at high lifts compression is generated. Several control valve subsystems must be analyzed in order to thoroughly evaluate their dynamic performance. The separate systems include: 1. The spring-mass system of the valve s moving parts. 2. The pneumatic system from controller output to valve diaphragm chamber. If a valve positioner is used, there are two separate pneumatic systems: one from the controller output to the positioner and another from the 100 Closed Stem travel (% lift) 50 Open 0 3 (21) Actual 5 (35) (2) (1) Ideal 10 (69) FIG. 6.4g Ideal and actual relationship between diaphragm pressure and stem travel. (3) 15 (104) 20 (138) Diaphragm pressure (PSIG) (kpa)

86 1130 Control Valve Selection and Sizing positioner output to the diaphragm chamber. The interconnecting tubing is consideration in all of the pneumatic systems. Spring-Mass System Dynamics Analysis of the spring-andmass system is only valid for linear systems. It is necessary either to neglect consideration of the nonlinear elements or have a system wherein the nonlinear effects are minor. In the case of control valves with sufficient power in the actuator, the latter case is approached. With such an understanding of the nonlinear effects, we proceed as though valve actuators were linear devices. The spring-mass system is represented by the following differential equation: M dx 2 b dx KX PA PV dt + 2 dt + = v 6.4(6) where b is the net friction force and M is mass. The net friction force would include friction due to seals, mechanical rubbing, and viscous friction on the plug. The static, time-independent terms of Equation 6.4(6) are identical with Equation 6.4(1). The transfer function of the valve actuator is the LaPlace transform of differential Equation 6.4(6): Xs () AK / = 2 Ps () ( wgk / ) s + ( bk / ) s (7) where g is the gravitation constant, s is the LaPlace operator, and w is the weight of moving parts. This can be written in terminology more useful to instrument engineers using the time constant τ (tau) and damping factor ζ (zeta). xs () Ps () = τ s + 2τςs (8) The coefficient of the s 2 term in Equation 6.4(7) is the square of the reciprocal of the undamped natural frequency of the spring-mass system. It is a useful number in understanding the relative importance of a control valve s dynamic components. Table 6.4h is a list of the natural frequencies of different size valves of the average design. It should be noted that even the largest valve with its un-damped natural frequency of Hz = 9 is ten times faster than the typical pneumatic performance of a control valve. For a more detailed discussion of the dynamics of diaphragm actuators and of the effect of standing pressure waves in the piping on that dynamics, the reader is referred to the discussion by Lynch. 1 Resolution and Valve Oscillation The minimum change in the stem position of a control valve is called its resolution, which limits the ability of the control signal to position the valve exactly at a specific point of its travel. Because of this limitation, valves can continuously oscillate, as the increment v TABLE 6.4h Natural Frequency of Control Valves (in.) Valve Size (mm) Undamped Natural Frequency (Hz) / of the stem position that can be delivered is larger than what is required. The result is a continuous sequence of overshooting and undershooting the stem position target. The valve s resolution, this minimum change in position, cannot be very accurately calculated, because of the variations between the valve designs. The definition of resolution for all pneumatic actuators is the ratio of the change in friction force to the spring rate. The relative resolution (R) is calculated by taking the difference between the static and dynamic friction forces (F s F d ) and dividing that with the difference between the mechanical spring rate and the spring rate of the trapped air (K a K s ). The resolution is calculated or percent of full stroke is then obtained by dividing by the stroke. In equation form the resolution, R, in dimensions of length is: R = (F s F d )/(K a K s ) 6.4(9) where: R is the resolution in units of either length or percentage of full stroke F s is the static spring friction F d is the dynamic (running) spring friction K a is the spring rate of the trapped air K s is the spring rate of the spring The spring rate of the trapped air (K a ) is shown by Equation 6.4(10): PA K = a V 6.4(10) where: A is the effective area of the actuator V is the volume of trapped air in each actuator cavity P is the absolute pressure in the actuator The main contributors to the friction forces are the packing friction on the actuator stem, the valve stem and the effect of any valve internal balancing seals. The motion of the actuator is faster than the time it would take for the air to be exchanged on both sides of the diaphragm or piston. Therefore, the air pressure conditions at each point of travel can

87 6.4 Actuators: Pneumatic 1131 be evaluated in a static manner, and the air exchange dynamics can be ignored. A number of observations can be made in connection with Equation 6.4(9). One such observation is that the dynamic or running friction (F d ) is always less than the static or breakaway friction (F s ). The difference usually is 25 35%. Another observation is that the friction forces for a PTFE (Teflon) seal are less than the friction forces generated by higher temperature seals made of fibrous graphite. The size of these friction forces is much affected by the amount of extra torque applied to tightening packing box seals during installation. Too much compression of the seals will result in high friction forces and in stem travel oscillations. The mechanical spring rate (K s ) is essentially constant. The air spring rate (K a ) can be increased by selecting a large effective area (A) in the actuator in combination with a small air volume (V). Because the air spring rate (K a ) is a function of the air volume and because the mechanical spring force changes with the stem movement, the resolution (R) will also vary with valve travel. The relationship between resolution and stem travel for a variety of actuator designs is shown Figure 6.4i. In most designs the use of a spring tends to reduce the resolution. In springdiaphragm combinations, the resolution is improved (reduced) when the spring opposes the direction of valve stem travel. Good maintenance is essential to minimize the frictional forces in valve actuators. The positioner, which delivers or exhausts air to/from the actuator, is slower than the speed at which small changes in stem position occur. Once the valve is installed the only means available to the user to change the resolution is to modify the supply air pressure, up to the limit of the design pressures. Because the air spring rate (K a ), which can be calculated by Equation 6.4(10), is small in comparison to the mechanical spring rate (K s ), an increase in the air supply pressure is not likely to have much impact on the resolution. For reasons of competition, the valve manufacturers usually provide the smallest actuator they can for the particular Normalized resolution Piston-no spring Piston w/spring Diaph-w/spring Diaph against spring Piston against spring Travel % FIG. 6.4i Actuator resolution curve vs. travel for a number of actuator designs. Resolution % FIG. 6.4j Resolution vs. diaphragm effective area. application. Yet, for a small additional expense the user can usually obtain an actuator with a larger effective area and obtain a noticeable improvement in resolution and, therefore, in the controllability of the loop. By so doing, the continuous oscillation of the valve can often be stopped. As it is shown in Figure 6.4j, the valve resolution can be much reduced (improved) by using a larger actuator. As can be noted from the figure, a doubling of the effective area of a diaphragm actuator cut the actuator resolution nearly in half. Safe Failure Position With spring Against spring Relative area The valve application engineer must choose between the two readily available fail-safe schemes for control valves, either fail open or fail closed. The choice will be based upon process safety considerations in the event of control valve air failure. Complete plant air failure, controller signal failure, and local air supply failure must all be considered. Local failure is significant when a valve positioner is being used and when piston actuators with cushion loading are used. The choice must be based on detailed knowledge of the valve application in the overall process or system. Two generalizations are that in a heating application, the valve should fail closed, and in a cooling application it should fail open. There are certainly applications where either failure mode is equally safe; then, considerations of standardization may be used. Fail-safe involves the selection of actions of actuator and inner valve. Both actuator and inner valve usually offer a choice of increasing air pressure to push the stem down or up, and pushing the stem down may open or close the inner valve. The proper choice of combinations may be made by fail-safe considerations. The process application of the valve must be investigated to determine whether, on instrument air failure, it would be better to have the valve go fully open, fully closed, or remain in its last position. There may not be much flexibility in the inner valve action. For example, a single-seated top-guided valve must

88 1132 Control Valve Selection and Sizing push down to close the plug. There is freedom of choice, however, in either single- or double-seated top- and bottomguided valves. Other valve bodies, such as the Saunders and pinch valve styles, must be of the push-down-to-close type. Rotary types, such as butterfly and ball valves, may be arranged either way. The inner-valve flexibility leads to two cases: one in which either inner-valve action is permissible and one in which the inner-valve must be push-down-to-close. When there is a choice of inner-valve action, overall valve action may be obtained by selecting the suitable inner-valve action and always using increasing air to push down the actuator. This is known as a direct actuator. A direct actuator is preferred because of economy reasons in spring-and-diaphragm actuators. The savings may be in purchase cost. It is also realized in maintenance costs, because there is no actuator stem seal to cause possible leakage and maintenance costs. When the inner valve must be push-down-to-close, it is necessary to use both direct and reverse actuators to accomplish the desired fail-safe actions. Figure 6.4k summarizes the available diaphragm failure options. The piston-type actuator is equally suitable for direct or reverse action. If it is the actuator to be used, the application engineer has complete freedom of the choice of selecting the valve action. The Role of the Positioner The above description provides a baseline for safe valve failure if the actuators are not provided with positioners. By the addition of a positioner, the topic of safe valve failure becomes quite complex. This is because in this case not only the pneumatic signal to the actuator can fail, but also the air supply to the positioner. In order to satisfy these requirements and also to make available the fail in place configurations, it is necessary to provide various accessories to either exhaust or trap the actuator air pressure. These accessories include pneumatic pilot valves (see Section 6.2) that, if air pressure is lost, will trip to provide a safe valve action. As far as digital systems are concerned, as of this writing only one digital positioner manufacturer provides a tight shut-off positioner. Valve failure (Overall) Fail Open Fail Closed Actuator Inner Valve Direct Direct Reverse Reverse Reverse Direct Direct Reverse FIG. 6.4l Forces acting on control valve with a spring-and-diaphragm type actuator. In the cases where only the positioner can send an air signal to the valve actuator, the valve failure position is usually unaffected by the addition of the positioner. An exception to this statement is if a nonbleeding digital positioner is used. Actually, it is questionable if a bleed can be considered as a means of providing positive failure position. This is because the time to bleed the air out through the positioner can be quite long. This long time that is required to reach a failure position is often unacceptable, and in such cases, one should provide pneumatic pilot valves to quickly trap or exhaust the air. Pneumatic Response Times Control P A P v An earlier discussion considered the transfer functions of the spring-and-diaphragm actuator between the actuator pressure to the resulting stem travel. Next we will consider the pneumatic transfer function of the air signal from the controller to the diaphragm actuator (Figure 6.4l). A short tube behaves linearly as a pure resistance, and the air volume in the actuator above the diaphragm behaves as a capacitance. So the combination is a resistance-capacitance time constant. Some time constant values obtained from tests with very short tubing are given in Table 6.4m. X A v K TABLE 6.4m Time Constants for Short Tube Sections FIG. 6.4k Overall valve failure positions, which can be achieved by various combinations of direct or reverse actuators and inner valves. Valve Size Time Constant (in.) (mm) (sec)

89 6.4 Actuators: Pneumatic 1133 Performance is usually limited by the controller s or positioner s ability to supply the required air fast enough. The time constant values in Table 6.4l were obtained from tests in which the air supply was not limiting. These figures show that the valves are capable of fast response. Section 3.1 contains a more detailed discussion of transmission lags and methods of boosting. PISTON ACTUATORS Piston actuators are either single or double acting. The singleacting actuator, shown in Figure 6.4n, utilizes a fixed air pressure, known as the cushion, to oppose the controller signal. This valve does not have spring or diaphragm area nonlinearities, but it is of course subject to the same plug force nonlinearities (Figure 6.4f) as the spring-and-diaphragm actuator. In order to use such an actuator for throttling purposes, it is necessary to have a positioner. The positioner senses the actuator motion and causes the valve to move accordingly. It cannot be used as a proportioning travel device without the positioner; consequently, its performance is that of the ideal curve in Figure 6.4g. A double-acting piston actuator is one that eliminates the cushion regulator and uses a positioner with a built-in reversing relay. Thus, the positioner has two air pressure outputs, one connected above the piston and the other below. The positioner receives its signal and senses travel in the same manner as a single-acting positioner. The difference is in the outputs; one pressure increases and the other decreases to cause piston travel. Table 6.4o provides typical actuator stroking times as a function of actuator sizes, stroking distances, and connecting tube sizes. For closed-loop control applications, the speed of response that is critical is not the full stroking time but rather the time it takes to move the valve about 5% of its full stroke, which is much faster than the values given in Table 6.4o. Therefore, velocity limiting of the loop usually occurs only when the valve is slow relative to the controlled variable Positioner Signal from controller As Feedback spring TABLE 6.4o Spring-Loaded Piston Actuator Stroking Times for Linear Control Valves * Actuator Size Time (seconds) for Maximum Stroke ** 1 / 4 in. Tubing 3 / 8 in. Tubing Stroke (inches) Actuation pressure: 60 psi * Courtesy of Valtek Flowserve Corp. ** Stroking time only (does not include time from receipt of signal and beginning of stem motion). and when the change in the controller signal to the valve is large. One advantage of the piston actuator is that higher pressures can be used for motive power. The higher pressure provides better stiffness and resolution. It also provides more force to keep the valve closed. The higher force between the valve plug and the seating surface ensures a tighter shutoff and helps to meet the leakage specifications of the original design. The piston actuators can also generate longer strokes, because their stroke is limited only by the length of the cylinder used. Such longer strokes are required for many special valve designs that are used in services where cavitation, noise, or pipe vibration is a problem (see Section 6.14). All of the failure modes are available with the piston actuator. A spring may be used inside or outside of the cylinder to cause the valve to fail in the desired position. The use of outside springs necessitates the use of additional seals and increases mechanical complexity. The use of an inside spring requires a larger cylinder volume that could impact the resolution of the valve. To obtain a positive failure position without a spring requires a standby air tank to provide the needed power for moving the valve into the failure mode. It also requires control accessories to ensure proper exhaust or lock-up of the air pressures. Seal FIG. 6.4n Single-acting piston actuator. Stem As cushion HIGH-SPEED ACTUATORS In the past, high-speed actuation was usually provided by hydraulic actuators. Applications that require fast control valve motion include compressor recycle, turbine bypass, and pressure relief applications. Control valves used in starting up or shutting down a process may also require fast actuation

90 1134 Control Valve Selection and Sizing Volume Booster Position feedback magneto potentiometer Direct drive pneumatic position module Piston Fail safe module Quick exhaust Lock-up Air inlet FIG. 6.4p High speed pneumatic circuitry. (Courtesy of Control Components Inc.) to protect equipment. Fast speed in these cases usually means full valve stroke in 1 to 2 sec. In case of pneumatic actuators, the factor that is limiting the speed is the speed at which air can be fed or exhausted from the cylinder. This is accomplished by accessories that direct the air to the cylinder while bypassing the positioner and by increasing the air feed line sizes and the air supply pressure. Because of the increased pressure, the piston actuator is usually preferred in comparison with the diaphragm design, because of its higher pressure rating. Diaphragm actuators are usually limited to 40 PSIG (275 kpa), with a few designs permitting 60 PSIG (410 kpa) operation. The high-pressure designs can operate with the usual plant air systems of 150 PSIG (1035 kpa) with normal operating pressures of 100 PSIG (690 kpa). This range of air supply pressures is readily available in most plants. The high-speed pneumatic actuation can be achieved by the use of boosters to feed and exhaust air from the cylinders. These boosters are actuated when the control signal calls for a 10% or so change in valve opening. An actuator that is being fed by two boosters located above and below the piston is shown in Figure 6.4p. In addition to the boosters, other accessories are also needed to guarantee the required failure positions. There are many disadvantages to these designs, including the difficulty of tuning and calibrating all of the component devices individually so as to maintain a stable operation. In addition, the tubing configuration requires a substantial amount of space around the actuator, and the accessories and tubing provide a lot of handhold and support points for operators, which can lead to damage. These systems can also be sensitive to vibration and to high temperatures radiating from a hot valve. A newer development is shown in Figure 6.4q. This design required the development of high-capacity servovalves that can be positioned to very close tolerance and of a programmable electronic controller. The controller can Valve yoke FIG. 6.4q High-speed piston actuator without boosters. (Courtesy of Control Components Inc.) receive the HART or fieldbus protocol of the control system, can maintain the travel of the actuator so that there is no visible overshoot even during the fastest transients, and can calibrate the actuation system within seconds. The higher pressure and stiffness of the actuator allows positioning within 0.25% of the total travel. Dead time and hysteresis are also considerably lower than for the more conventional pneumatic actuators. Figure 6.4r shows a response curve for an actuator with a 14 in. (355 mm) travel in which full stroke is achieved in 1.3 sec. The dead time is less than sec, and there is no visible overshoot when operating a relatively high friction valve. There are no mechanical linkages, tubing, or accessories Percent open 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Set point Travel 14 inch 355 mm Time-milliseconds FIG. 6.4r High speed actuator step change response. (Courtesy of Control Components Inc.)

91 6.4 Actuators: Pneumatic 1135 to fail from vibration or from other environmental conditions. According to the supplier, all failure positions can be provided and diagnostic software is also available. (For more details, see Sections 6.8 on valve diagnostics, 6.11 on fieldbus interaction, and 6.12 on intelligent valves.) The high-speed pneumatic actuators reduce the need to use hydraulic actuators, when the stroking time of a second or two is sufficient. One pneumatic actuator supplier feels that both electric and hydraulic high-speed actuators are more expensive than the pneumatic ones. (For a detailed discussion of hydraulic and electric actuators refer to Section 6.3.) RELATIVE MERITS OF DIAPHRAGM AND PISTON ACTUATORS Table 6.4a compared the features of diaphragm- and pistontype pneumatic actuators. When choosing between piston actuators and the spring-and-diaphragm type, the fail-safe consideration may be the reason for the final selection. If properly designed, the spring is the best way of achieving fail-closed action. Fail-open action is less critical. Piston actuators may depend upon air lock systems to force the valve closed on air failure. Such systems may work well initially, but there are possibilities for leaks to develop in the interconnecting tubes, fittings, and check valves. Therefore, such piston actuator systems are not considered reliable by many. However, according to one manufacturer, field data on reliability shows the opposite to be true. Air lock systems also add to the actuator s cost. Piston actuators may also be specified with closure springs to provide positive failure positions. Valve installation in the line is also a factor to consider. Flow over the plug assists in maintaining valve closure after air failure, but the considerations involving dynamic stability are more important. Therefore, the use of flow-to-open valves is recommended for most diaphragm actuators, as the actuator force is usually marginal. Piston actuators are larger and require more space than do diaphragm actuators. This is particularly the case when the piston is provided with a spring to provide a positive failure position. Both the diaphragm and piston actuators use manifolds and have become available in modular designs. This has somewhat reduced their costs and lowered their potential for leakage by reducing the number of connections that could leak. An example of the modular design of a linear actuator is shown in Figure 6.4s. The linear actuator includes a positioner, boosters, and adjustable quick exhaust. Figure 6.4t illustrates the modular design of a rotary piston actuator. The rotary design is the balanced pinion type that can be provided with a plug-in positioner, which can be either the conventional or the smart design. The smart positioners also are available with bus communication options. The modular designs provide flexibility, as they may be changed out in the field to accommodate process changes or FIG. 6.4s Piston actuator with modular manifolds. (Courtesy of Control Components Inc.) upgrades or to correct errors in the original design. These modular designs are usually provided with position indicators, which are enclosed within the actuator housing. Because of these housings, reliability is better than in designs where the linkages are exposed. Most control modules will eventually include bus control communication and self-diagnostics capability. Pneumatic vs. Hydraulic Actuators While the purpose of this section is to describe the features of pneumatic actuators and while a detailed discussion of hydraulic and electric actuator features is given in Section 4.3, a few comments FIG. 6.4t Rotary actuator provided with interchangeable control modules. (Courtesy of Emerson Process Management.)

92 1136 Control Valve Selection and Sizing Output 2 Output 1 Zero adjust. Zero adjust, lock knob Range adjust, lock screw Range adjust, screw Cylinder Piston Cam Signal 3-15 psi Feedback spring Instrument signal capsule Balance beam Follower arm Actuator lever arm Supply Plot valve spool Plot valve body Positioner FIG. 6.4u When linear actuators are used to operate rotary valves, a unit change in controller signal will not result in a unit change in rotation unless a positioner is used. (Courtesy of Flowserve Inc.) from the perspective of a pneumatic piston actuator manufacturer is included here. The main advantages of hydraulic actuators are speed and stiffness. This is the case because of the high density and the incompressibility of liquids in comparison to air. The speed difference between pneumatic and hydraulic actuators has been narrowed, as the stroking time of some pneumatic piston operators (particularly with dual pistons) is about 1 sec. With the use of 100 PSIG (690 kpa) or higher air pressures, the piston actuator stiffness and stability have also improved and approach that of the hydraulic actuators. The stem forces provided by a pneumatic piston can equal or exceed those of the hydraulic cylinders, because the area of an oversized air piston can provide higher stem forces than a standard hydraulic actuator. For example, if the hydraulic cylinder area is one 50th of the air piston, while the air pressure is one 20th of the oil pressure, the stem force produced by the hydraulic actuator is actually less than that of the pneumatic actuator. ROTARY VALVE ACTUATORS When linear spring-and-diaphragm, piston, or cylinder actuators are used on rotary valves, their performance will not be linear unless a positioner is used (Figure 6.4u). By the addition of a positioner, one can guarantee that the ratio between a unit change in controller signal and the resulting rotation will be uniform. On the other hand, loose-fitting actuator lever and follower arms can still create dead play in the actuator, which will lower the responsiveness of the loop when the direction of change in the control signal reverses. As can be seen in Figure 6.4v, the torques required to rotate ball, butterfly, or plug valves are not linear. These valves are usually used only for on/off applications. When considered for closed-loop throttling control, the nonlinear relationship between the air signal and the resulting rotation makes the use of a positioner essential. Figure 6.4w shows the nonlinearity in the torque characteristics of double-acting and spring-loaded cylinder actuators. Ball Butterfly Plug Stem torque + 0 Stem torque + 0 Stem torque + 0 Closed Open Closed Closed Open Closed Closed Open Closed FIG. 6.4v The torque characteristics of such rotary valves as ball, butterfly, and plug valves are not linear.

93 6.4 Actuators: Pneumatic 1137 Percentage of break torque 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% Double acting actuators Actuator rotation, degrees Break torque End torque Spring stroke Pneumatic stroke Spring return actuator FIG. 6.4w Torque characteristics of spring-return and double-acting cylinder actuators. (Courtesy of Rotork Controls Ltd.) Naturally, the piston actuator should be so selected that its break torque exceeds the peak torque requirement of the valve, which occurs at the beginning and the end of the stroke (break torque). The torque characteristics of the double-acting piston actuator of Figure 6.4x shows two maximums. One maximum occurs at the closed position and one near the position, which corresponds to the peak of the dynamic torque for butterfly valves. Naturally, this actuator too should be so selected that its break torque exceeds the peak torque requirement of the valve, which occurs at the beginning (break torque) and near the end of the stroke. Cylinder Type Increased use of ball and butterfly valves or plug cocks for control has bred a variety of actuators and applications of existing actuators for powering these designs. Positioning a quarter-turn valve with a linear output actuator using a lever arm on the valve resolves itself into a problem of mounting and linkages. The actuator can be stationary, with a bushing to restrain lateral movement of the stem. This requires a joint between the stem and a link pinned to the lever arm. The actuator can be mounted on a gimbal 1.4 FIG. 6.4y Pneumatic cylinder actuator with Scotch yoke. mechanism to allow required movement. The actuator can be hinged to allow free rotation to allow for the arc of the lever arm. Various Scotch yoke designs, such as the one shown in Figure 6.4y, can be used with one, two, or four cylinders. Use of rollers in the slot of the lever arm utilizes the length of the lever arm of the valve opening or closing points. A rack and pinion can be housed with the pinion on the valve shaft and the rack positioned by almost any linear valve actuator. The rack (Figure 6.4z) can be carried by a doubleended piston or by two separate pistons (Figure 6.4aa) in the same cylinder, where they move toward each other for counterclockwise rotation and away from each other for clockwise rotation. 1.3 Relative torque Minimum torque Rack Opening-Degrees FIG. 6.4x Double-acting piston actuators for rotary motion. (Courtesy of Metso Automation.) Pinion Valve shaft FIG. 6.4z Rack and pinion actuator with dual-acting cylinder.

94 1138 Control Valve Selection and Sizing FIG. 6.4cc Rack and pinion actuator with spring-loaded (fail-safe) cylinder. FIG. 6.4aa Rack and pinion actuator operated by two separate pistons. FIG. 6.4bb Rack and pinion actuator operated by two parallel pistons. Similar action is obtained (Figure 6.4bb) by two parallel pistons in separate cylinder bores. Dual cylinders are used in high-pressure actuators used to rotate ball valves as large as 16 in. (400 mm) in less than 0.5 sec. An actuator similar to the one shown in Figure 6.4cc can be spring-loaded for emergency or positioning operation. On/off operation of cylinders for quarter-turn valves requires solenoid or pneumatic pilots to inject pneumatic or hydraulic pressure into the cylinders. Open or closed position must be set by stops that limit shaft rotation or piston travel. Thereby the valve rotation is stopped and held in position until reverse action is initiated. Positioning action requires a calibrated spring in the piston or diaphragm actuator, a valve positioner, or a positioning valve system that loads and unloads each end of the cylinder. Rotation of the valve must be translated to the positioner by gears, direct connection, cam, or linkage. The valve positioner must be the type that includes the fourway valve. The positioning valve system can be a four-way valve with a positioner for use with a pneumatic controller. The piston is sometimes positioned by a servo-system consisting of a servovalve that accepts an electronic signal, a four-way valve to amplify and control the pressure to the cylinder, and a feedback signal from a potentiometer or LVDT. An electrohydraulic power pack or pneumatic pressure source may be used to furnish pressure to a pair of cylinders, one for each direction of rotation, as shown in Figure 6.4dd. Rotation by Spline or Helix A multiple helical spline rotates through 90 as pressure below the piston (Figure 6.4ee) moves the assembly upward. A straight spline on the inside of this piston extension sleeve rotates the valve closure member through a mating spline. The cylinder is rated at 1500 PSIG (10 MPa). The actuator will rotate 1 in. (25 mm) and 1 1 / 2 in. (37.5 mm) valve stems. The mounting configuration is designed to adapt to many quarter-turn valves. A nonrotating cylinder (Figure 6.4ff), with an internal helix to mate with a helix on a rotatable shaft, creates a form of rotating actuator. Hydraulic or pneumatic pressure in the drive end port (left) causes counterclockwise rotation; clockwise rotation is caused by pressure on the opposite side of the piston. There is a patented seal between the internal and the external bores of the cylinder and the external surface of the shaft. The unit is totally enclosed by seals to protect it from contaminated atmospheres. A hydraulic pump, reservoir, and necessary controls can be mounted integrally. Vane Type Injection of pressure on one side of a vane to obtain quarterturn actuation is straightforward and obtainable with a minimum number of parts. A single vane (Figure 6.4gg) can be used for 400 to 30,000 in. lb f (27 to 2060 N m). Units can be mounted together for double output, or a double-vane design (Figure 6.4hh) can also be used. The success of the vane actuator as a control device is dependent upon the control systems. By use of an auxiliary pneumatic pressure source, all types of fail-safe actions are possible, although not as positively as with spring loading. Use of line pressure to create hydraulic pressure on the vane is piloted by both manual and automatic methods. Use of a rotary potentiometer to sense position and complete a bridge circuit is necessary for proportional control. Rotary Pneumatic Actuators Pneumatic pressure is used to power a rotary motor to drive any of the large gear actuators. Control is by a four-way valve. The motor shown in Figure 6.4ii is running in one position. This will continue until the valve is repositioned or until a cam operates a shut-off valve at one end of the stroke. Reversal of the four-way valve causes reverse operation. An intermediate position causes the motor to stop. The four-way valve can be operated by pneumatic or electric actuators for remote automatic control. A position transmitter will allow adaptation to closed-loop proportional control.

95 6.4 Actuators: Pneumatic 1139 Pin Seal Cylinder Seal Roller Pressure opens Key Crank Piston head Pressure closes Vent or purge ports Guide bushing Top section Piston rod Piston seal Position-indicating switch Switch cover Actuating arm Bearing plate Seals Bearing retainer Switch wire outlet Set screw Tie rod Cylinder flange Bearing Actuator case Seal Stem Side section Mounting stud bolts FIG. 6.4dd Crank and roller actuator operated by a pair of cylinders. OTHER PNEUMATIC ACTUATORS Pneumohydraulic Actuators An actuator with two double-acting cylinders uses an integrally designed pneumatic or electric powerpack (Figure 6.4jj). This has the advantage of furnishing a constant hydraulic pressure to the cylinders regardless of the power source to the prime mover. Few cylinder sizes are needed to cover a wide range of torque outputs. The prime movers are sized and selected to obtain the actuator speeds desired with the pneumatic pressures or electric voltages available. Multiple auxiliary switches, position transmitter, and positioning devices are adapted to the unit. Gas pressure is used to create hydraulic pressure using two bottles (Figure 6.4kk). The stability of a hydraulically operated cylinder is utilized in this manner, using line gas pressure and the bottle size for amplification. The manual control valve can be replaced by a variety of electric or pneumatic pilot valves for automatic control. A hand pump is furnished that can take over the hydraulic operation in the absence of gas pressure or malfunction of the pilot controls. This selfsustaining approach to cylinder operation finds wide application for line break shut-off and for the various diverting and bypass operations of a compressor station. A hydrostatic system consisting of a pneumatic prime mover (Figure 6.4ll) on the shaft of a hydraulic pump to run a hydraulic motor has interesting features. This actuator incorporates many of the features of other high-force geared actuators that rotate a drive sleeve. Torque control consists of a relief valve in the hydraulic line to the motor. This eliminates the reactive force of spring-loaded torque controls. Starting torque occurs

96 1140 Control Valve Selection and Sizing FIG. 6.4gg Single-vane quarter-turn actuator. FIG. 6.4ee Helical spline actuator. because hydraulic slippage of the pump allows the motor to reach maximum speed. Direction and deactivating control is attained with a fourway valve. As many as 16 auxiliary switches, settable at any position, are housed in the unit. Limit switches can be pneumatic or electric. A wide range of torque outputs and speeds is obtained by selection of prime mover, pump, and motor combinations. Initial success of the unit was partially due to FIG. 6.4hh Double-vane quarter-turn actuator. Pneumatic pressure Vent Limit valve spool Spring Selector valve spool Motor Pinion FIG. 6.4ii Rotary air motor actuator. its adaptability for retrofit to existing valves. The unit can be manually operated if required. Electropneumatic Actuators FIG. 6.4ff Rotating helix actuator. An actuator that defies classification, except that it is pneumatically powered and electrically controlled for proportional

97 6.4 Actuators: Pneumatic 1141 Signal to close Vent Limit control (optional) Hydraulic cylinders Signal to open Vent Stem drive Valve stem Reservoir Pneumatic motor To open To close Torque control valves (2) Hydraulic Sump pump intake FIG. 6.4jj Actuator with two double-acting cylinders and power pack. Exaust Pneumatic inlet Exhaust 4-way pneumatic valve (optional) application, is described at this point. Operation of a threaded drive sleeve occurs when spring-loaded pawls create a jogging action on a drive gear. Pressure introduced through one of the external lines selects the pawl to become active when the rocker arm is repetitively rocked by the pneumatic motor. A lead screw positions a sliding block to operate control switches, potentiometer, and position indicators. The lead screw is driven from a small spur gear and bevel gears. Air is supplied from PSIG ( kpa) to give torque outputs up to 360 or 720 ft lb f (25 or 50 Nm) using two actuators. Air consumption is from 0.75 SCF 1.70 SCF ( m 3 ) per revolution at 140 PSIG (966 kpa). Maximum valve stem diameter that can be rotated is 2 in. (50 mm). Numerous motivating combinations are used for control, including electropneumatic, electric, and fully pneumatic. A wide variety of components can be used to build up these systems, as shown in Figure 6.4mm. RELIABILITY Some of the reliability studies on the performance and reliability of pneumatically operated valves were done in the offshore oil industry 2 and in the power industry. 3,5 These studies did not evaluate only the actuator but the complete valve and actuator system combined. The offshore study for pneumatically operated control valves was made in 1984 and reported the highest failure rate of 1 failure every 10,000 hours; or a little less than 1 failure per year. A 1990 study made for pneumatically operated control and on/off valves by the Institute of Nuclear Power Operation 4 (INPO) showed a failure rate of 116,000 hours. In 1997 a summary on the reliability of air-operated valves by the electric power industry was published. 3 The sample consisted of 525 failures from 4,726 component years of operation. This would equate to a failure every 9 years or 79,000 hours of operation. The data covered all types of valves with both piston and diaphragm actuators from eight manufacturers. The study looked at what was called high duty (e.g., control valves) vs. low duty (e.g., shut-off valves) as well as the impact of intrusive maintenance. A summary of the results Hydraulic snubber Hand pump Control valve Hydraulic lock Double-check valve Exhaust Pressure inlet Strainer Control junction box FIG. 6.4kk Pneumohydraulic actuator powered by line pressure. (Courtesy of Shafer Value Co.)

98 1142 Control Valve Selection and Sizing Optional limit control Hydraulic motor Gas motor Gas in to close Gas in to open Optional electrically controlled 4-way pneumatic valve Stem nut Valve stem or handwheel drive Pinion & output bull gear Exhaust Gas inlet Exhaust Hydraulic pump Sump intake Thrust control valves Reservoir FIG. 6.4ll Actuator system with pneumatically powered hydrostatic valve. (Courtesy of Ledeen Div. Textron.) from this study is given below: High duty cycle valves exhibit up to 13 times lower reliability than low duty cycle valves. Intrusive preventive maintenance causes reliability to decrease by a factor of 6 times. The reliability of piston-type actuators was 4 times better than that of diaphragm-type actuators. The reliability of diaphragm-actuated valves was 6 times more likely to be impacted by duty cycle than were piston actuators. Control valve operators Mechanical Cam, toggle, etc. Lever Push button Solenoid Differential pilot Solenoid pilot Pilot exhaust (bleed pilot) Thermal Diaphragm pilot Filter regulator lubricator 2 POS. 3-way N/C valves FIG. 6.4mm Electropneumatic valve actuator. (Courtesy of OIC Corp.) Spring Pilot Detent Pilot exhaust (bleed pilot) These results are a bit surprising because the common perception is that the diaphragm actuator is much more reliable that the piston actuator. It is also possible that by combining all types of valves (on/off and throttling) and by placing emphasis on duty cycle (opening and closing of the valves), the results of this evaluation would differ with one that evaluated only throttling control valves. CONCLUSIONS Before deciding on the type of control valve actuators to be used in a particular application, the reader is advised to also read Section 6.3, which discusses hydraulic, electric, and digital valve actuators, and to study Table 6.3a, which lists the advantages and disadvantages of electromechanical, electrohydraulic, and servo- or stepping-motor-operated electric actuators. Those who want to learn even more about the features and performance of available valve actuators are advised to study the test reports, books, and articles listed in the Bibliographies of both Sections 6.3 and 6.4. As far as actuator trends are concerned, pneumatic actuators are still the technology favored by the users of valve actuators. Nevertheless, from 1998 to 2003, the market share of electric actuators increased from 11% to 45% and the market share of hydraulic and electrohydraulic actuators increased from near 0% to 28%. During the same period, use of solenoid valves has dropped by 42%. 6 References 1. Lynch, J., Impedance-Coupled Valve and Fluid System Instability, American Nuclear Society, Pittsburg Meeting, May Offshore Reliability Data Handbook, OREDA participants, Hovik, Norway, 1984.

99 6.4 Actuators: Pneumatic Hinchcliffe, G. and Worledge, D., Preventive Maintenance Program Basics: Air Operated Valves, Report TR V1, Electric Power Research Institute, Palo Alto, CA, July Private communication between PLG, Inc. and Control Components Inc., IEEE Guide to the Collection and Presentation of Electrical, Electronic, Sensing Component, and Mechanical Equipment Reliability Data for Nuclear-Power Generating Stations, IEEE Standard , The Institute of Electrical and Electronics Engineers, Inc., Harrold, D., A Changing Landscape, Control Engineering, December 2003 ( actuators). Bibliography ANSI/ISA , Test Procedure for Control Valve Response Measurement from Step Inputs. ANSI/ISA-TR , Control Valve Response Measurement from Step Inputs. Barnes, P. L., Protect Valves with Fire-Tested Actuators, Instruments and Control Systems, October Baumann, H. D., The Need for Pneumatic Power Signals for Control Valves, InTech, January Baumann, H. D., Trends in Control Valves and Actuators, Instruments and Control Systems, November Carey, J. A., Control Valve Update, Instruments and Control Systems, January Fernbaugh, A., Control Valves: A Decade of Change, Instruments and Control Systems, January Ritz, G., Control Valve Actuator Options Part 1, Control, June Control Valve Dynamic Specification, Entech, Coughran, M. T., Measuring the Installed Dead Band of Control Valves ISA, Fitzgerald, W. V., Loop Tuning and Control Valve Diagnostics, Paper # , 1991 ISA Conference, Anaheim, CA, October Grabau, T., Superior Actuator Diaphragm Material for Control Valves in Low Temperature Applications, Valve Magazine, Fall Hammitt, D., How to Select Valve Actuators, Instruments and Control Systems, February Heney, P., Rotary Actuators: Turnabout Is Fair Play, Hydrolics and Pneumatics, April Holtgraver, E., Quarter-Turn Pneumatic Actuators: Torque 101, Valve Magazine, Summer Howleski, M. F., Adapting Electric Actuators to Digital Control, Instrumentation Technology, March Installing Smart Positioners A Wise Move, Chemical Engineering Magazine, December 1, Kalsi, M. S., Eldiwany, B., Sharma, V., and Somogyi, D., Dynamic Torque Models for Quarter-Turn Air, Operated Valves, Air Operated Valve Users Group, 20th Conference, Clearwater Beach, FL: Kalsi Engineering, Inc., Koechner, Q. V., Characterized Valve Actuators, Instrumentation Technology, March Langford, C. G., A Method to Determine Control Valve Dynamic Requirements, ISA, Lipták, B. G., Control Valves in Optimized Systems, Chemical Engineering, September 5, Ludwig, A. User-Friendly Control Valve, Valve World 98 Conference, KCI Publishing BV, the Netherlands, Merrick, R. C., Valve Selection and Specification Guide, Van Nostrand Reinhold, Miller, S.F., Electronic Valve Controller Replaces Conventional Pneumatic System, ISA Expo 2004, Paper ISA04-p, Instrumentation, Systems, and Automation Society, Research Triangle Park, North Carolina, Houston, Oct Mowrey, A., Advancements in Valve and Actuator Technologies, Valve Magazine, Fall Patel, A., Air Operated Actuators for High Cyclic Duty, Air Operated Valve users Group, Clearwater Beach, FL, Jan , Price, V. E., Smart Valve Intelligence Takes Many Forms, InTech, August Pyotsia, J., A Mathematical Model of a Control Valve, 1992 ISA Conference, Houston, October Scott, A. B., Control Valve Actuators: Types and Applications, InTech, January Samy, S. and Stemler, D., Linear Piston Actuators, Air Operated Valve Users Group, 21st Conference, Clearwater Beach, FL, Control Components Inc., Shinskey, F. G., Dynamic Response of Valve Motors, Instruments and Control Systems, July Valve Actuator Roundup, InTech, January Smart actuator incorporates all the controls in one package, Chemical Engineering Magazine, January 1, 2003.

100 6.5 Advanced Stem Packing Designs J. B. ARANT (2005) Designs and Sizes: Design Pressures: Design Temperatures: Chevron or V-rings, and rectangular rings. All valve and stem sizes available commercially. Special packing can be designed for any and all rotary or linear valve sizes. Up to 2500 PSIG ANSI 450 to 1000 F ( 268 to 538 C), depending upon the materials of construction and configuration of the packing system, and valve design. Teflon with a cooling bonnet can be used up to 850 F (454 C). Packing Materials: Teflon, Kalrez, Zymaxx, and Expanded Graphite Cost: Partial List of Suppliers: The costs of Teflon and Expanded Graphite packing sets and rings are moderate. The KVSP packing is more costly on first costs. However, they are more economical overall due to their greatly reduced maintenance needs, longevity, and overall performance, with regard to fugitive emissions and the resulting valve dynamics. Teflon, Kalrez, Zymaxx, and KVSP are all registered trademarks of E.I. Du Pont Co. (now, Du Pont- Dow Elastomers); Expanded or Flexible Graphite is a registered trademark of Union Carbide Co. INTRODUCTION Similarly to most industrial products, control valve packing has also evolved over time. The two main requirements of control valve stem packing are that 1) it should seal the valve stem to eliminate or minimize leakage, and 2) it should not influence the performance or dynamics of the control valve. The ability of a control valve to respond to small control signal changes without overshooting and cycling is inversely proportional to packing friction. Therefore, the ideal packing would be one that is as tight (as leakage free) as a bellows seal and yet completely friction free. Real packings do have some emissions (Figure 6.5g) and do have some friction. The lower the friction, the smaller the dead band (resolution) and the shorter the response time of the packing (Table 6.5h). Prior to 1950, stem leakage was a constant problem with these older packing designs. Today, both of the above requirements can be met. Modern seals can be extremely tight, providing performance approaching that of a bellows seal, and they can also be low friction. This guarantees not only the good dynamic response of the control valve, but also low maintenance and a long service life. Today s packing is no longer just a packing, but it is a well-engineered part of the control valve and its sealing systems. They can be installed to operate under wide temperature and pressure ranges. They can be used in applications involving temperatures from cryogenic to very high temperatures. They can also operate under process pressures from vacuum to several thousands of PSIG. History The early control valves were provided with only one type of packing, which consisted of braided asbestos rings. These rings were formulated with various braiding designs, utilizing various types of asbestos, binders, and lubricants. For control valves, the best versions were used, such as Blue African asbestos with premium binders and braid lubricants. These rings were installed in a very smooth, finished well inside a valve bonnet. The valve stem was well polished. The valve bonnet was fitted with a lubricator assembly that included a shut-off isolation valve and a ram screw. Depending on the process fluid, the appropriate choice of grease sticks was inserted and the lubricant was forced into the packing area by the ram. After this, the isolating valve was closed to prevent the pressurized process fluid from escaping. By today s standards, this was system, yet it was very workable. All instrument maintenance mechanics, and many of the instrument engineers and supervisors, routinely carried 1144

101 6.5 Advanced Stem Packing Designs 1145 Die-formed ribbon flexible graphite Braided filament graphite FIG. 6.5a Valve stem packing made out of PTFE V-rings with internal coil spring located below the packing. boxes of lubrication sticks in their pockets so as to be able to control stem leakage while guaranteeing a well-lubricated packing. It was a never-ending chore. Teflon V-Rings In the mid- to late 1940s, the Du Pont Co. developed a fluorocarbon polymer, Teflon, and it occurred to the control valve manufacturers that this material could be made into packing rings. These rings not only had the potential of providing a good seal, but were also inherently selflubricating and resistant to essentially all process fluids. So the Teflon V-rings, or chevron rings, as they came to be known, were born (Figure 6.5a). Because of its corrosion resistance to almost any process fluid, its self-lubrication, easy manufacturing via molding, and good temperature and pressure ratings, Teflon rapidly replaced asbestos as a packing material in many industries. Essentially the only problem that arose was that this packing material was so slick and had such low friction that the control valve had a tendency to overshoot and oscillate at times. This was partially because these early valves were designed for use with the higher friction asbestos packing. The solution to this problem was to add some friction back into the Teflon packing. So, the packing gland was tightened a little, and the instrument mechanics learned when to do it. Graphite Later, Union Carbide developed the Expanded Graphite, or flexible carbon, materials that could be formed into shapes such as packing rings (Figure 6.5b). This increased the choice of packing. Graphite can be used at higher operating pressures if the process fluids are the nonoxidizing-type fluids, although Teflon, by far, provided a superior control valve packing. Teflon could be used at temperatures up to 850 F, if the valve was provided with an extended cooling bonnet. FIG. 6.5b Graphite packing consisting of die-formed ribbons and braided filament graphite rings. Graphite packing works very well with manually operated valves, but it gives poor performance on control valves, because it affects their dynamics and resolution. This happens because it tends to plate out on the valve stem. This further increases stem friction and necessitates increased loading of the packing in order to seal the stem. In addition, there can also be a fugitive emission problem with graphite packing. The Enviro-Seal and the Dual Packing Another early development in the history of leak-resistant packing systems was one developed by Fisher Controls, using its Enviro-Seal stem packing system (see Figures 6.5c and 6.5d). While the development of the Enviro-Seal was a big step forward in valve stem sealing and in coping with leakage and fugitive emissions, it required about the same amount of packing seal maintenance as did the older packings. In addition its friction was much higher than that of Teflon, which negatively impacted its performance. This was reported in ISA (Instrumentation, Systems, and Automation Society) Appendix A, which evaluated the link between packing friction and valve performance. Nevertheless, the Enviro-Seal helped to hold the fort until better packing designs eventually came onto the marketplace. Another interim and early improvement in the field of packing designs was the use of dual Teflon packing sets, with a lantern ring between the sets. The lantern ring could be

102 1146 Control Valve Selection and Sizing Springs (N inconel 718) Antiextrusion washers Packing box ring (stainless steel) Packing follower (stainless steel) Lantern rings (stainless steel) Anti-extrusion ring (filled PTFE) Packing ring (PTFE) Anti-extrusion ring (filled PTFE) FIG. 6.5c The Enviro-Seal was developed by Fisher Controls and included a PTFE packing ring and external Belleville disc springs. used for sealing greases, or to allow piping to a leakage disposal system. While this design worked to some degree, it was also a somewhat clumsy and high maintenance system. Springs (N07718) Guide bushing (carbon) Packing washers (PTFE) Guide bushing (carbon) Guide bushing (carbon) Enviro-seal PTFE Enviro-seal graphite Packing follower (stainless steel) Packing ring (composite) Packing ring (flexible graphite) Packing ring (composite) Packing box ring (stainless steel) FIG. 6.5d Fisher s Enviro-Seal provided with flexible graphite packing rings and external Belleville disc springs. Industrial Practices In general, chemical and most other industries favored the Teflon packings, while the petroleum industry continued to favor asbestos and Expanded Graphite, because of their suitability for high-temperature and fire safe services, which are common in that industry. Their rationale was that fire safety should be their main concern, because they handled flammable fluids at high temperatures. This logic was questionable, as the chemical industries also handled highly flammable fluids along with high temperatures and were just as vulnerable to fires. In the chemical industry a woven graphite ring was added at the top of the Teflon stack, with which these packings passed the API 607 valve fire test. With such packing designs, the chemical industry had years of experience with no particular problems. This difference in philosophies between the two industries persisted until recent years when the Federal Clean Air Act changed everything. Asbestos was phased out by the federal government as a hazardous material, and the petroleum industries were left with few packing options other than Expanded Graphite, plus some unsuccessful attempts at developing synthetic fibers as asbestos substitutes. Development of the KVSP Packing In the late 1980s, the Du Pont Co. and industry in general, came under pressure from the Federal Clean Air Act. The response was to again develop a new valve stem packing based on the use of forming chevron rings out of Kalrez. This was a fluroelastomer material that had all of the corrosion-resistance properties of Teflon, plus a higher temperature rating along with the elastomeric characteristics. At that time, the primary use of Kalrez was in O-rings and other sealing shapes. During these years, because of air pollution concerns, the federal government was trying to eliminate or to minimize more of the fugitive emissions from valve and pipe joint seals. Therefore, extensive tests were conducted and it was found that Kalrez Valve Stem Packing (KVSP), this new valve stem packing, did an excellent job in reducing emissions far below the levels of the federal and state (California) clean air standards. According to the testing per the Instrumentation, Systems, and Automation Society and KVSP performed essentially as well as Teflon. Thus was born the KVSP (Figure 6.5e). After very extensive testing, using helium to simulate the most severe process plant leakage conditions, both the process industries and the control valve manufacturers started to use KVSP packing as a premium packing for solving their valve stem fugitive emission problems while still maintaining the dynamic response of Teflon packing. By replacing non- Teflon valve stem packing with KVSP, process plants and even California refineries in populated areas were able to meet EPA and Federal Clean Air Act fugitive emission standards. Figure 6.5f illustrates the improvement that KVSP represents in terms of packing friction relative to graphite. As can be seen from the figure, the variability has been more than cut in half by using KVSP. Further testing, development, and field experience with the KVSP packing led to the fire-safe design, per the API 607 Standard, in addition to the high-temperature (KVSP 500) and the creep- and extrusion-resistant (KVSP 400/500) variations. Specific combinations of Kalrez, Teflon, Zymaxx (a high strength Teflon composite reinforced with long carbon fiber structural composites), and expanded woven graphite rings provided the solutions to many specific applications. While Teflon and Expanded Graphite rings are still extensively used for control valve packing, KVSP packing is making

103 6.5 Advanced Stem Packing Designs 1147 Die-formed flexible graphite Kalrez V-ring seals Teflon or Zymaxx components KVSP 400 Series Kalrez/Teflon KVSP 500 Series Kalrez/Zymaxx KVSP Fire-Safe Kalrez/Zymaxx/graphite Note: consists of Kalrez and Teflon PTFE/Zymaxx Maximum continuous temperature use 204 C (400 F)/288 C (550 F) FIG. 6.5e The KVSP packing system consists of Teflon or Zymaxx components and Kalrez V-ring seals below a flexible graphite ring Graphite KVSP Avg. friction, Ib Max. friction, Ib Min. friction, Ib Actuator net pressure, psig Closed Spring rate, bfm Total travel, in Graphite packing process variability KVSP packing process variability 4 2 Graphite packing 1000 ppm KVSP packing 1 ppm Ideal valve travel Valve actuator travel, in. FIG. 6.5f In addition to better controlling fugitive emissions, KVSP also reduces the packing friction effects on process control (variability). The curves compare the variability of the valve travel in response to increasing actuator pressure, when graphite and KVSP packing is used.

104 1148 Control Valve Selection and Sizing 10,000 5,000 Kalrez packing system (live loaded) PTFE packing system (live loaded) Peaks to 10,000 ppm Emissions, ppm EPA requirement 1990 Clean Air Act amendment 2 State regulation (Calif.) 3 Company standards 1st 2nd 3rd Three temperature swings up to 177 C (350 F) 20 M 30 M full stroke cycles FIG. 6.5g Emissions caused by creep and cold flow during temperature swings are much reduced when Kalrez packing is used instead of PTFE. These emissions satisfy both EPA s Clean Air Act and the California state regulations. rapid inroads in many areas due to its superior capabilities and lower overall ownership costs. One could say that KVSP is the nearest thing to a universal valve stem packing material today. This packing works well on both linear and on rotary stem control valves, and is suitable to practically all process fluids. Some users have adopted it as their primary packing, which is used in all applications except for the more benign and innocuous ones such as water, where Teflon is more economical. Field retrofitting kits are available for valves to upgrade spring-loaded Teflon packings in existing valves with KVSP. KVSP Performance The major attributes of this packing, patented by Du Pont-Dow Elastomers, are as follows: When using it, the emissions are not measurable and exceed both the federal and California State Clean Air emission standards. Its resistance to HAPS and toxins is equivalent to that of Teflon. Its leakage containment is almost as good as that of a bellows. Its packing friction is very low. Its dynamic response is equivalent to that of Teflon. Its process variability of 0.1% is also equivalent to that of Teflon. When using it, the plant yield and throughput increases, because it has as little impact on process control dynamics and resolution as does Teflon. A major performance advantage of KVSP over Teflon is that it is far less sensitive to leakage caused by thermal cycling at high temperatures. Figure 6.5g shows the emissions resulting from temperature swings when using KVSP and Teflon. It can be noted that because of creep and cold flow, the emissions using KVSP (Kalrez) are much reduced. Table 6.5h provides a summary of the performance testing results that was published in Instrumentation, Systems, and Automation Society From the table one can see that the ability of a control system to respond to small control signal changes is inversely proportional to packing friction, and the lower it is, the better control dynamics can be expected. This is because with lower packing friction, the valve s step response is faster and its dead band (resolution) is smaller. Kalrez and PTFE V-rings are effective not only on control valves, but also for the stem packing of automatic on/off and manual valves, if fugitive emissions are to be minimized. As TABLE 6.5h Performance of Leading Packing Designs Packing Type Dead Band or Resolution Step Response Time Fisher s Graphite Enviro-Seal % sec Fisher s PTFE Enviro-Seal % sec PTFE V-Packing % sec KVSP Packing % sec Data Based on ISA Control Valve Performance Testing Summary

105 6.5 Advanced Stem Packing Designs 1149 to the number of packing rings, five-ring packing sets are the norm. Three-ring packing can be used in many rotary stem valve applications, because the rotary stem movement is easier to seal against leakage than are the linear stems. Yet, for reasons of convenience and standardization, the five-ring design is the most often used. Bibliography ANSI/ISA , Test Procedure for Control Valve Response Measurement from Step Inputs, ANSI/ISA-TR , Control Valve Response Measurement from Step Inputs, Brestal, R. et al., Control Valve Packing Systems, technical monograph 38, Marshalltown, IA: Fisher Controls, Control Valve Dynamic Specification, EnTech Control Inc., Toronto, Canada. Coughran, M. T., Measuring the Installed Dead Band of Control Valves, ISA TECH/ ISA S26, 1968, and ANSI MC 4.1, 1975, Dynamic Response Testing of Process Control Instrumentation, Langford, C. G., A User s View of Process Control and Control Valve Positioners, ISA 1996 Paper # Packing Selection Guidelines for Sliding Stem Valves, Bulletin 59.1:062, Marshalltown, IA: Fisher-Rosemount, March Senior, K. A., Technical Guidelines and Design Information, Using KVSP Packing Systems for Improving Process Control And Minimizing Fugitive Emissions, white paper, DuPont Dow LLC, Senior, K. A., Valve Packing Systems Improve Process Control, Chemical Processing, June 1997.

106 6.6 Capacity Testing (Copyright ISA 1996, reprinted with permission of ISA, The Instrumentation, Systems, and Automation Society. Standard formatted for publication in this handbook by A. Bálint, 2005) INTRODUCTION The Control Valve Capacity Test Procedure is covered by Instrumentation, Systems, and Automation Society s (ISA s) Standard S The international equivalent of these U.S. standards are IEC , Industrial Process Control Valves, Part 2-1: Flow Capacity Sizing Equations for Fluid Flow under Installed Conditions, and IEC , Industrial Process Control Valves, Part 2-3, Flow Capacity Test Procedure. The equations in the international editions are essentially equivalent to those in the U.S. standards, with the exceptions of using metric units. A revision of the ISA standard is anticipated in the near future, and for this reason only the front part of S is reprinted below. The reader is advised to refer the revised standard (when published) for the test procedures and calculations. Meanwhile, Section 6.15 can be used for all valve sizing applications, while Section 6.14 provides all the information needed for valve noise calculations. The reader should also note that in the ISA test procedure the pressure taps are not located at the in- and outlets of the control valve, but two pipe diameters upstream and six pipe diameters downstream. Therefore, the pressure differential measured during the test is the sum of the valve pressure drop ( P v ) and the drop through eight pipe diameters length of straight piping ( P p ). This can introduce errors, if the pressure drop in the pipe exceeds 5 10% ( P p /( P p + P v ) > 0.1). This section is reprinted from Control Valve Capacity Test Procedure, ANSI/ISA-S , with permission of the copyright holder, ISA, The Instrumentation, Systems, and Automation Society, 67 Alexander Drive, P.O. Box 12277, Research Triangle Park, NC SCOPE This test standard utilizes the mathematical equations outlined in ANSI/ISA-S75.01, Flow Equations for Sizing Control Valves, in providing a test procedure for obtaining the following: a. Valve flow coefficient, C v b. Liquid pressure recovery factors, F L and F LF c. Reynolds number factor, F R d. Liquid critical pressure ratio factor, F F e. Piping geometry factor, F P f. Pressure drop ratio factor, X T and X TP This standard is intended for control valves used in flow control of process fluids and is not intended to apply to fluid power components as defined in the National Fluid Power Association (NFPA) Standard T PURPOSE The purpose of this standard is to provide a procedure for testing control valve capacity and related flow coefficients for both compressible and incompressible fluids. This standard also provides a procedure to evaluate the major data. 3. NOMENCLATURE (For the actual control valve sizing equations for both compressible and incompressible fluids, the reader should refer to Section 6.15) 4. TEST SYSTEM 4.1 General Description A basic flow test system as shown in Figure 6.6a includes a. Test specimen b. Test section c. Throttling valves d. Flow-measuring device e. Pressure taps f. Temperature sensor 4.2 Test Specimen The test specimen is any valve or combination of valve, pipe reducer, and expander or other devices attached to the valve body for which test data are required (Figure 6.6a). Modeling of valves to a smaller scale is an acceptable practice in this standard, although testing of full-size valves or models is preferable. Good practice in modeling requires attention to significant relationships such as Reynolds number, the Mach number where compressibility is important, and geometric similarity. 1150

107 6.6 Capacity Testing 1151 Symbol C v d D F d F F F k F L F LP F P F R G f G g K m N 1, N 2, etc. p 1 p 2 Description Valve flow coefficient Value inlet diameter Internal diameter of the pipe Valve style modifier Liquid critical pressure ratio factor, dimensionless Ratio of specific heats factor, dimensionless Liquid pressure recovery factor of a valve without attached fittings, dimensionless Product of the liquid pressure recovery factor of a valve with attached fittings (no symbol has been identified) and the piping geometry factor, dimensionless Piping geometry factor, dimensionless Reynolds number factor, dimensionless Liquid specific gravity at upstream conditions (ratio of density of liquid at flowing temperature to density of water at 15.6 C [60 F]), dimensionless Gas specific gravity (ratio of flowing gas to density of air with both at standard conditions, which is equal to the ratio of the molecular weight of gas to the molecular weight of air), dimensionless Ratio of specific heats, dimensionless The number of similar flow paths (i.e., m = 1 for single-ported valves, m = 2 for double-ported, etc.) Numerical constants for units of measurement used Upstream absolute static pressure, measured two nominal pipe diameters upstream of valve-fitting assembly Downstream absolute static pressure, measured six nominal pipe diameters downstream of valve-fitting assembly p Pressure differential, p 1 p 2 p v Absolute vapor pressure of liquid at inlet temperature q Volumetric flow rate q max Maximum flow rate (choked flow conditions) at a given upstream condition Re v Valve Reynolds number, dimensionless T 1 Absolute upstream temperature (in K or degrees R) x Ratio of pressure drop to absolute inlet pressure ( p/p 1 ), dimensionless x T Pressure drop ratio factor of the valve without attached fittings, dimensionless x TP Value of x T for valve-fitting assembly, dimensionless Y Expansion factor, ratio of flow coefficient for a gas to that for a liquid at the same Reynolds number, dimensionless ν(nu) Kinematic viscosity, centistokes Subscripts: 1. Upstream conditions 2. Downstream conditions Temperature sensor Flow Test specimen Test section Downstream throttling valve FIG. 6.6a Basic flow test system. Source Upstream throttling valve Flow sensor Pressure taps

108 1152 Control Valve Selection and Sizing TABLE 6.6b Piping Requirements, Standard Test Section A* and ** B C D Standard Test Section Configuration At least 18 nominal pipe diameters of straight pipe 2 nominal pipe diameters of straight pipe 6 nominal pipe diameters of straight pipe At least 1 nominal pipe diameter of straight pipe Flow A B Pressure taps Test specimen*** C D *Dimension A may be reduced to 8 nominal diameters if straightening vanes are used. Information concerning the design of straightening vanes can be found in ASME Performance Test Code PTC , Applications, Part II of Fluid Meters, Interim Supplement on Instruments and Apparatus. **If an upstream flow disturbance consists of two ells in series and they are in different planes Dimension A must exceed 18 nominal pipe diameters unless straightening vanes are used. ***See Section 4.2 for definition of the test specimen. 4.3 Test Section The upstream and downstream piping adjacent to the test specimen shall conform to the nominal size of the test specimen connection and to the length requirements of Table 6.6b. The piping on both sides of the test specimen shall be Schedule 40 pipe for valves through 250 mm (10 in.) size having a pressure rating up to and including ANSI Class 600. Pipe having 10 mm (0.375 in.) wall may be used for 300 mm (12 in.) through 600 mm (24 in.) sizes. An effort should be made to match the inside diameter at the inlet and outlet of the test specimen with the inside diameter of the adjacent piping for valves outside the above limits. The inside surfaces shall be reasonably free of flaking rust or mill scale and without irregularities that could cause excessive fluid frictional losses. 4.4 Throttling Valves The upstream and downstream throttling valves are used to control the pressure differential across the test section pressure taps and to maintain a specific downstream pressure. There are no restrictions as to style of these valves. However, the downstream valve should be of sufficient capacity to ensure that choked flow can be achieved at the test specimen for both compressible and incompressible flow. Vaporization at the upstream valve must be avoided when testing with liquids. 4.5 Flow Measurement The flow-measuring instrument may be any device that meets specified accuracy. This instrument will be used to determine the true time average flow rate within an error not exceeding ±2% of the actual value. The resolution and repeatability of the instrument shall be within ± 0.5%. The measuring instrument shall be calibrated as frequently as necessary to maintain specified accuracy. 4.6 Pressure Taps Pressure taps shall be provided on the test section piping in accordance with the requirements listed in Table 6.6b. These pressure taps shall conform to the construction illustrated in Figure 6.6c. * A A Size of pipe Not exceeding Less than 50 mm (2 in.) 6 mm (1/4 in.) 50 mm to 75 mm (2 to 3 in) 9 mm (3/8 in.) 100 mm to 200 mm (4 to 8 in.) 13 mm (1/2 in.) 250 mm and greater 19 mm (3/4 in.) (10 in. and greater) FIG. 6.6c Recommended pressure connection. Minimum 2.5 A recommended 5 A A Not less than 3 mm (1/8 in.) 3 mm (1/8 in.) 3 mm (1/8 in.) 3 mm (1/8 in.) * Edge of hole must be clean and sharp or slightly rounded, free from burrs, wire edges or other irregularities. In no case shall any fitting protrude inside the pipe. Any suitable method of making the physical connection is acceptable if above recommendations are adhered to. Reference: ASME Performance Test Code PTC Applications, Part II of Fluid Meters, Interim Supplement on Instruments and Apparatus.

109 6.6 Capacity Testing 1153 Orientation: Incompressible fluids Tap center lines shall be located horizontally to reduce the possibility of air entrapment or dirt collection in the pressure taps. Compressible fluids Tap center lines shall be oriented horizontally or vertically above pipe to reduce the possibility of dirt or condensate entrapment. Multiple pressure taps can be used on each test section for averaging pressure measurements. Each tap must conform to the requirements in Figure 6.6c. TABLE 6.6d Allowable Misalignment between the Centerlines of the Test Specimen and of the Test Section Piping Pipe Size Allowable Misalignment mm 0.8 mm ( 1 / 2 1 in.) ( 1 / 32 in.) mm 1.6 mm (1 1 / 4 6 in.) ( 1 / 16 in.) 200 mm and larger 1% of the diameter 8 in. and larger) 4.7 Pressure Measurement All pressure and pressure differential measurements shall be made to an error not exceeding ±2% of actual value. Pressuremeasuring devices shall be calibrated as frequently as necessary to maintain specified accuracy. Pressure differential instruments are required in the measurement of the pressure differential across the test specimen to avoid additional inaccuracies resulting from taking the difference of two measurements. Exceptions to this are the procedures in Sections 6.2 and 8.2 for determining maximum flow rates for incompressible and compressible flow, respectively. (These sections are not included here, the reader is referred to ANSI/ISA for guidance in these procedures.) 4.8 Temperature Measurement The fluid temperature shall be measured to an error not exceeding ±1 C (±2 F) of actual value. The inlet fluid temperature shall remain constant within ±3ºC (±5 F) during the test run to record data for each specific test point. 4.9 Installation of Test Specimen The alignment between the centerline of the test section piping and the centerline of the inlet and outlet of the test specimen shall be as listed in Table 6.6d below: When rotary valves are being tested, the valve shafts shall be aligned with test section pressure taps. Each gasket shall be positioned so that it does not protrude in the flow stream Accuracy of Test Valves having an N 3 C v /d 2 ratio of less than 30 will have a calculated flow coefficient, C v, of the test specimen within a tolerance of ±5%. 5. TEST FLUIDS 5.1 Incompressible Fluids Fresh water shall be the basic fluid used in this procedure. Inhibitors may be used to prevent or retard corrosion and to prevent the growth of organic matter. The effect of additives on density or viscosity shall be evaluated by computation using the equations in this standard. The sizing coefficient shall not be affected by more than 0.1%. Other test fluids may be required for obtaining F R and F F. 5.2 Compressible Fluids Air or some other compressible fluid shall be used as the basic fluid in this test procedure. Vapors that may approach their condensation points at the vena contracta of the specimen are not acceptable as test fluids. Care shall be taken to avoid internal icing during the test. SUMMARY The international equivalent of this U.S. standard are IEC , Industrial Process Control Valves, Part 2-1: Flow Capacity Sizing Equations for Fluid Flow under Installed Conditions, and IEC , Industrial Process Control Valves, Part 2-3, Flow Capacity Test Procedure. A revision of the Instrumentation, Systems, and Automation Society standard is anticipated in the near future, and for this reason only the first five sections of S were presented here. The reader is advised to refer the revised ANSI/ISA standard (when published) for Sections 6 to 10 for the test procedures and calculations. Meanwhile, Section 6.15 can be used for all valve sizing applications, while Section 6.14 provides all the information needed for valve noise calculations.

110 6.7 Characteristics and Rangeability B. G. LIPTÁK (1995, 2005) A. BÁLINT (2005) INTRODUCTION The characteristics, gains, and rangeabilities of control valves are interrelated. The process control engineer should clearly understand these terms, because they describe the personality of the valve and as such play an important role in the closedloop performance of the loop. This section begins with a brief discussion of valve gains, followed by an explanation of the difference between theoretical and actual (installed) valve characteristics. The section ends with an explanation of valve rangeability. m 1 Controller gain G %m c = %e Dimensionless Valve gain G v = GPM m 3 /s % % Process fluid (u) (G v ) (G p ) CW TIC (G c ) Set pt. (r) TT (G s ) m + + Load (u) Process gain G F P = GPM C G P = m 3 /s VALVE GAIN AND LOOP GAIN The gain of any device is its output divided by its input. For a linear (constant gain) valve, the valve gain (G v ) is the maximum flow divided by the valve stroke in percentage (Fmax/100%). When a loop is tuned to provide quarter amplitude damping (Figure 6.7a), the controller gain (G c = 100/%PB) is adjusted until the overall loop gain (the product of the gains of all the loop components) reaches 0.5. If a linear controller and a linear transmitter are used, their gains are constant. Therefore, if the process gain (G p ) is also constant, a linear (G v = constant) valve is needed to keep the loop gain product constant at 0.5. If the transmitter is nonlinear, such as a d/p cell without square root extraction, the transmitter gain will rise in proportion to flow, and therefore the loop will be unstable at high flows and sluggish at low flows. The usual solution is to install a square root extractor, which makes the transmitter linear (G t = constant) with flow. One can also correct for the nonlinearity of the transmitter by using a nonlinear controller or a valve whose gain drops with flow (quick opening). Nonlinear Processes If the process is nonlinear (G p varies with load) while the other loop gains are constant, a change in load will result in the loop gain s shifting away from 0.5. Therefore, if the total gain rises, the loop will become unstable; if it drops, the loop will become sluggish. Therefore, if the process gain changes with load (a nonlinear process), the loop can remain stable Set point (r) + e b Sensor gain 100% % % G s = = span F C % GPM F % Loop gain =(G c )(G v )(G p )(G s ) = = % % GPM F = Dimensionless FIG. 6.7a The loop gain is the product of the gains of the loop component. In a properly tuned loop (decay ratio of 1 / 4 ), this gain product should be constant at 0.5. only if another gain component of the loop is also changing, and if that change is the inverse of the gain change of the process. Therefore, as the process gain (G p ) drops, this other gain must be rising, thereby keeping the loop gain constant at around 0.5. This other gain can either be the gain of the controller (G c ) or that of the control valve (G v ). If the controller gain varies with load, the controller is called a nonlinear controller. When the control valve gain varies with load (flow), it is named according to the type of nonlinearity that exists between the flow through the valve and the valve stroke. If, as the flow is increasing, the valve gain is also rising and that rate of rise is a constant rate with flow, the valve characteristic is called equal-percentage. If the valve gain is increasing at a variable rate with flow, it is named in accordance c 1154

111 6.7 Characteristics and Rangeability 1155 with the type of nonlinearity it provides (parabolic, hyperbolic, and so on). If the valve gain is dropping when the flow increases, it is called a quick-opening valve. Some processes, such as the ph process, are highly nonlinear. In such applications as ph control, the variation in the process gain (G p ) is compensated by an inverse variation in the controller gain (G c ), which is selected to drop when G p rises. Therefore, the controller of a ph process is a nonlinear controller. In heat transfer processes, because the heat transfer area is constant, its efficiency of heat transfer drops as the load (the amount of heat to be transferred across that fixed area) rises. Therefore the process gain (G p ) drops with the heat load. To compensate for this drop, the valve gain (G v ) must rise with the load. Such a valve is called an equal-percentage valve, which should be the valve characteristic selected for all temperature control valve applications. Installed Valve Gain As will be discussed in more detail later, the inherent valve gain changes after the installation of the valve, if the valve pressure differential varies with load. This is the case in all mostly friction pumping systems, because as the load (flow) rises, the pressure drop in the piping system also increases, which leaves less pressure drop for the valve. As the valve differential pressure drops with increasing flow rate, the valve gain (G v ) also drops. This tends to shift the installed gain of equal-percentage valves towards linear and the installed gain of linear valves towards quick opening. Therefore, on mostly friction pumping systems (Figure 6.1c), if it is desired to keep the valve gain relatively constant (linear characteristics), it is recommended to install an equalpercentage control valve. An even more effective method of keeping the valve gain (G v ) constant is to replace the valve with a linear control loop (Figure 6.1h). The disadvantage of this cascade configuration (in addition to the higher cost) is that this will degrade the control quality if the controlled process is fast, because in each cascade system, the gain of the outer loop must be smaller than that of the inner loop. This will necessitate the detuning (lowering the gain, increasing the proportional band) of the master controller (outer loop), which generates the set point for the flow controller (FC) in Figure 6.1h. THEORETICAL VALVE CHARACTERISTICS Valve Testing The standard methods of testing the capacity of valves are discussed in Section 6.6. It should be noted that the goal of this test is only to determine the valve C v (K v ) within an error of 5%. What is important to note is that the valve characteristics (G v characteristics) are neither tested nor defined by the standard. It should also be noted that during the testing, the pressure drop through the valve itself is not measured, because the P is detected across a pipe section, which includes the valve, plus a length of eight pipe diameters of straight pipe (Section 6.6). The end result of such a test is a valve characteristic curve, which describes the flow through the valve as it is stroked from 0 to 100% of its stroke. The C v (K v ) data provided by manufacturers is usually reliable within an error of about 10%, if the installation is identical to the test setup (usually it is not). Valve Characteristics The inherent characteristics of a control valve describes the relationship between the controller output signal received by the valve actuator and the flow through that valve, assuming that: 1. The actuator is linear (valve travel is proportional with controller output). 2. The pressure difference across the valve is constant. 3. The process fluid is not flashing, cavitating, or approaching sonic velocity (choked flow). Some of the widely used inherent lift to flow rate relationships are illustrated in Figure 6.7b. For example, in a linear valve, travel is linearly proportional to capacity and therefore the theoretical gain is constant at all loads. (The actual gain is shown in Figure 6.7e.) In equal-percentage valves, a unit change in lift will result in a change in flow rate, which is a fixed percentage of the flow rate at that lift. For example, in Figure 6.7b, each percentage increase in lift will increase the previous flow rate by about 3%. Therefore, the theoretical gain of equal-percentage valves is directly proportional to flow rate (actual gain shown in Figure 6.7e) and increases as the flow rate increases. On a logarithmic chart (left side of Figure 6.7b), the equal-percentage characteristics correspond to a straight line having a slope that corresponds to its fixed percentage. In quick-opening valves, the gain decreases with increasing flow rates. Figure 6.7b shows the quick-opening valve characteristics with the same total lift as for the other plug types. If the travel of the quick-opening plug is restricted so that the distance of 100% lift travel corresponds to only 1 / 4 of the seat diameter, then the valve characteristics will approach linear (if the hydraulic resistance is constant) with the gain being nearly constant. Valve and Process Characteristics Control loops are usually tuned at normal load levels (at normal flow rates through the control valve), and it is assumed that the total loop gain will not vary with process load. This assumption is seldom completely valid, because the process gain usually does change with load. Because one

112 1156 Control Valve Selection and Sizing % flow (Cv or Kv) Quick opening Linear 20 Equal percentage and butterfly % lift or stroke P is constant P % flow (Cv or Kv) Quick opening Linear Equal percentage and butterfly % lift or stroke FIG. 6.7b Inherent flow characteristics of quick-opening, linear, and equal-percentage control valves. cannot afford to retune the controller for each new load, it is desirable to select control valves that will compensate for these load change effects. For example, when a liquid-to-liquid heat exchanger is being controlled, the process gain and dead time (transportation lag) will both decrease as the load increases. Therefore, one should attempt to compensate for this inverse load-to-gain relationship by using a valve with a direct load-to-gain relationship, such as the equal-percentage valves. If one does that, as the heat-exchanger load increases and therefore the process gain drops, the valve gain will rise, thereby compensating for that effect and reducing the total change in the loop gain. Equal-percentage valves are not ideal, though, if high turndown is required or if there are solids in the throttled process fluid. An opposite example is a control loop whose sensor has an expanding scale, such as an orifice plate or a vapor-filled thermometer. With such detector, the loop gain is increasing with load, and therefore the gain of the selected valve should decrease with load. Therefore, a quick-opening control valve is often used. In a fairly large number of cases, the choice of valve characteristics is of no serious consequence. Just about any characteristic is acceptable for the following applications: 1. Processes with short time constant, such as flow control, most pressure control loops, and temperature controls through mixing a cold and a hot stream 2. Control loops operated by narrow proportional band (high gain) controllers, such as most regulators 3. Processes with load variations of less than 2:1 In general it can be said that the quick-opening characteristics are used in regulators and on orifice-type measurements, if no square root extractors are provided. The equal-percentage characteristics are most often used on heat-transfer type temperature control applications and on pumping systems where the valve differential pressure varies more than 2:1 as the load (flow rate) changes. Linear characteristics are used in most other cases. Selection Recommendations As was discussed in Section 6.1, different engineers have developed different rules of thumb in selecting valve characteristics for the various types of control loops. These recommendations vary in complexity. Shinskey, for example, recommends the use of equalpercentage valves for heat-transfer control and the use of linear valves for all flow, level, and pressure control applications, except vapor pressure, for which he recommends equalpercentage valves. Driskell suggests that for relatively constant valve differential pressures, quick-opening valves should be used for square root-type flow loops, equal-percentage valves for temperature and liquid pressure, and linear valves for all others. If the valve differential pressure varies with load, his quickopening recommendation shifts to linear, and his linear recommendation shifts to equal-percentage. Lytle s recommendations were summarized in Table 6.1g. They are even more involved, as they take into account more variables. As will be shown in Sections 6.16 to 6.23, the inherent characteristics of ball and butterfly valves are similar to the

113 6.7 Characteristics and Rangeability 1157 Destination Pump P System Ps D c = ( P t ) min ( P) ( P t ) max ( P s ) % Flow (Cv or Kv) % Flow or (C v or Kv) Pt Dc = 0.04 Dc = 0.10 Dc = 0.25 Dc = 0.50 Dc = Dc = 0.04 Dc = 0.10 Dc = 0.25 Dc = 0.50 Dc = % Lift * 0 or stroke Linear *Note that the minimum controllable flow increases as D c drops. * Equal percent % lift or stroke FIG. 6.7c These figures illustrates the effects of the distortion coefficient (D c ) on inherently linear (left) and on inherently equal-percentage valves (right), according to Boger. equal-percentage, while the characteristics of Saunders and pinch valves are closer to the quick-opening. Installation Causes Distortion As was pointed out in connection with Figure 6.1c, in mostly friction-type pumping systems, the pressure drop available for the control valve is dropping, as the load (flow rate) is increasing (Figure 6.1e). This is a different condition from the test conditions under which the valve C v (K v ) was measured by the manufacturer, because on the manufacturers test stand, the flow rate through a valve is measured under constant pressure drop conditions. Therefore, when such a valve is installed, its gain (characteristics) will shift as shown in Figure 6.7c. If that shift is substantial, one might obtain a near linear installed characteristic by installing a theoretically equal-percentage valve. Different engineers have approached the shift between inherent and installed valve characteristics in different ways. One approach, that of the old school, was to oversize the pump, so that the ratio between the minimum and maximum valve pressure drops (Figure 6.1c) will not be large, and therefore the gain of the process will not change much with load. This approach works, but it wastes pumping energy. Distortion Coefficient When the control valve is installed into the piping in a process plant, its flow characteristics are no longer independent of the rest of the system. This is because the flow through the valve will be subject to the frictional resistance, which is in series with the valve. The consequence is the type of distortion illustrated in Figure 6.7c. From the curves in Figure 6.7c, one can conclude if a particular installation will have a very substantial effect on both flow characteristics and rangeability of the valve, or not. Under conditions of excessive distortion, clearance flow alone can increase as much as tenfold, and equal-percentage characteristics can be distorted toward linear or even quick opening. It should be emphasized that Figure 6.7c assumes the use of a constant speed pump (Figure 6.1c). In variable-speed pumping systems, one might adjust the pump speed so as to keep the valve p constant, and therefore in such control systems the installed and theoretical valve characteristics are more similar, and less distortion is allowed to occur. Naturally, in variable-speed pumping systems one can completely eliminate the valve and just throttle the pump speed. The predictability of installed valve behavior is reduced, not only because the inherent valve characteristics deviate

114 1158 Control Valve Selection and Sizing Output % From controller A B C To valve Output % From controller B A C To valve 0 Input % Input % 100 FIG. 6.7d The valve characteristics can be modified by inserting a divider or multiplier relay into the controller output signal. from their theoretically prescribed character, but also because: 1. Actuators without positioners will introduce nonlinearity. 2. Pump curves will also introduce nonlinearity. It should also be recognized that in order to learn the true valve characteristics requirement, a full dynamic analysis of the process is required. Even if one took the trouble to perform such analysis, it would probably yield a valve characteristics requirement that is not commercially available in conventional air-operated control valves. For these reasons, one might consider any one of the following options: 1. Characteristics that are an intrinsic property of the valve construction, such as an equal-percentage ball or butterfly or a beveled (quick-opening) disc 2. Valves that are characterized by design, such as globe valves having linear or equal-percentage trims 3. Digital control valves that can be characterized by software 4. Characteristics that are superimposed through auxiliary hardware, such as function generators, characterized positioners, cams, and so on 5. Intelligent control valves (Section 6.12), which can electronically modify the control signal, which is received as a function of the inherent valve characteristics and of the desired valve gain Correcting the Valve Characteristic The linear valve has a constant gain at all flow rates, while the gain of the equal-percentage valve is directly proportional to flow. If the control loop tends to oscillate at low flow but is sluggish at high flow, one should switch the valve trim characteristics from linear to an equal-percentage. If, on the other hand, oscillation is encountered at high flows and sluggishness at low flows, the equal-percentage valve trim should be replaced with a linear one. Changing the valve characteristics can be done more easily by modifying the controller output signal or by inserting accessories into the control signal leading to the actuator than by replacing the valve. One approach proposed by Fehérvári/Shinskey is to insert a divider or a multiplier into the control signal line, as illustrated in Figure 6.7d. By adjusting the zero and span at port C, a complete family of curves can be obtained. The divider is used to convert an air-to-open equal-percentage valve into a linear, or an air-to-close linear valve into an equal-percentage one. The multiplier is used to convert an air-to-open linear valve into an equal-percentage, or an air-to-close equalpercentage valve into a linear. According to Shinskey, both devices are perfectly standard, sensitive, stable, easy to calibrate, and real lifesavers when one needs a linear butterfly valve. RANGEABILITY The conventional definition of rangeability used by most manufacturers is the ratio between maximum and minimum controllable flow through the valve. Minimum controllable flow (Fmin) is defined as the flow below which the valve tends to close completely. In other words, this widely held definition of Fmin refers not to the leakage flow (which occurs when the valve is closed), but to the minimum flow that is controllable in the sense that it can be changed up or down as the valve stroke is changed. Using this definition, manufacturers usually claim a 50:1 rangeability for equal-percentage valves, 33:1 for linear valves, and about 20:1 for quick-opening valves. These claims suggest that the flow through these valves can be controlled down to 2, 3, and 5% of the flow corresponding to their rated C v (K v ).

115 6.7 Characteristics and Rangeability % Flow Theoretical gain characteristics of equal % valve Gain FIG. 6.7e The theoretical vs. the actual characteristics of a 2 in. (50 mm) cage-guided globe valve, according to Driskell. The above definition of rangeability is based on the inherent C v (K v ) determined during testing. It can be seen in Figure 6.7c that the minimum controllable flow rises as the distortion coefficient (D c ) drops. At a D c value of 0.1, for example, the 50:1 rangeability of an equal-percentage valve drops to close to 10:1. This is because the valve pressure drop is much higher at low flows (Figure 6.1e), and therefore the minimum valve opening will pass much more flow. Thus, the required rangeability should be calculated as the ratio of the C v (K v ) required at maximum flow (and minimum pressure drop) and the C v (K v ) required at minimum flow (and maximum pressure drop). Improved Definition of Rangeability 1 The decision on whether a particular control valve is capable of providing the required rangeability should not be evaluated on the basis described above. This is because a loop is uncontrollable not only when the valve cycles between closed and some minimum flow but also when the loop gain product shifts away from 0.5 (the tuning target for quarter amplitude damping). Therefore, the acceptable flow range within which the valve can safely be used for closed-loop control must be based on a relationship between the theoretical and the actual valve gain. The valve rangeability should therefore be defined as the flow range over which the theoretical (inherent) valve 4 ±25% of = % Actual, inherent gain characteristics of equal % valve Theoretical gain characteristics of linear valve Actual, inherent gain characteristics of linear valve 3 2 ±25% of linear gain and the actual installed valve gain will stay within preset limits. Therefore, the rangeability of the valve can be defined as the ratio of the minimum and maximum C v s (K v s) bordering the region within which the actual valve gain is within ±25% of the theoretical valve gain. This advanced definition of valve rangeability establishes the point at which the flow-lift characteristic starts to deviate from the expected by more than 25%. (Figure 6.7e shows the points where the actual gain s deviation from the theoretical starts to exceed 25%.) If one defines intrinsic rangeability as that range of the ratios of C v (max) to C v (min) within which the values of the valve gain do not vary more than ±25% from the theoretical, then according to Figure 6.7e, the rangeability of a linear valve can be greater than that of an equal-percentage valve. Actually, if one uses this definition, the rangeability of equalpercentage valves is seldom more than 10:1. The rangeability of some rotary valves can be higher because their clearance flow tends to be less than that of other valves, and their body losses near the wide open position tend to be lower than those of other valve designs. Valve rangeability can also be limited by the sensitivity and accuracy of positioning. Why Traditional Rangeability Is Wrong The problem with traditional rangeability definitions is that they tend to overstate the range within which the control valve can be used. This results in poor control quality for the whole loop. The main reasons why the traditional definition for control valve rangeability is unacceptable are as follows: 1. The minimum controllable flow (F min ) is determined using a test during which the valve pressure differential is constant, while in most real-life installations the pressure differential is maximum when the flow is minimum. Therefore, the real value of F min should be higher than the claimed values. 2. The valve should not only be controllable within its rangeability but should have a gain (G v ) that is close to its theoretical gain. 3. The traditional definition of rangeability is based on dividing the maximum flow (F max ) by F min. This approach is also wrong, because while the flow through a nearly 100% open valve is controllable, its gain is nowhere close to the theoretical value. As was shown in Figure 6.7e, if the acceptable valve gain is defined as 1, which is within ±25% of the theoretical valve gain, F max of the linear valve should not exceed 60% and F max of the equal-percentage valve should not exceed 70% of the maximum flow through the valve. In terms of valve lift, these flow limits correspond to 85% lift for equal-percentage and 70% lift for linear valves.

116 1160 Control Valve Selection and Sizing CONCLUSIONS It is time to realize that smart transmitters and sophisticated control algorithms alone cannot result in properly functioning control loops. Stable and responsive closed-loop control also requires that the gain of the final control element (the valve) be much more predictable and better controlled. In many valve designs (digital valves excluded) this can only be achieved if the valve rangeabilities are redefined and thereby restricted (reduced). The manufacturers should contribute to the achievement of this improvement by publishing the characteristic curve for each valve showing its F min and gain values at different distortion coefficients (D c ). The user s contribution should be a better understanding of the role that the valve gain plays in process control and the realization that in most control valve designs the upper one third of the stroke is not usable for stable control. Bibliography Alcozer, S., Segovia, J. P., and Sbarbaro, D., Using Higher Order CMAC to Improve the Performance of Control Valves, ISA Transactions, Volume 36, Issue 1, pp , Arant, J. B., Positioner Use Is Myth-Directed, InTech, November Bajsic, I. and Bobic, M., Gain of control valve with polygonal flow characteristics, STROJ VESTN-J MACH E, 46(5), , Ball, K. E., Final Elements: Final Frontier, InTech, November Baumann, H. D., How to Assign Pressure Drop Across Control Valves for Liquid Pumping Services, Proceedings of the 29 th Symposium on Instrumentation for the Process Industries, Texas A&M University, Bialkowski, B., Coughran, M., and Beall, J., Control Valve Performance Update The Last 10 Years, PULP PAP-CANADA, 102 (): 21 22, November Boger, H. W., Flow Characteristics for Control Valve Installations, ISA Journal, November Borden, G. and Zinck, L., Control Valve Seat Leakage, InTech, November Byeon, H. W. and Ulrey, J. L., Designing a Standard Thermal Power Plant for Daily Startup/Shutdown: The HP Bypass Control and Safety Function, ISA Transactions, Volume 36, Issue 1, pp , Calibrating Control Valve Accuracy, MECH ENG, 122 (6): 33 33, June Campos, M. C. M., Satuf, E., and de Mesquita, M., Intelligent System for Start-up of a Petroleum Offshore Platform, ISA Transactions, Vol. 40, Issue 3, July Champagne, R. P. and Boyle, S. J., Optimizing Valve Actuator Parameters to Enhance Control Valve Performance, ISA Transactions, Vol. 35, Issue 3, 1996, Pages Control Valves, Regulators, Measurements and Control, June Coughlin, J. L., Control Valves and Pumps: Partners in Control, Instruments and Control Systems, January Coughran, M. T., Measuring the Installed Dead Band of Control Valves, ISA Transactions, Volume 37, Issue 3, pp , July Davis, J. A. and Stewart, M., Predicting Globe Control Valve Performance Part II: Experimental Verification, J FLUID ENG-T ASME, 124 (3): , September Exclusive Control Valve Performance Diagnostics, HYDROCARB PRO- CESS, 82 (12): 30 30, December Dobrowolski, M., Guide to Selecting Rotary Control Valves, InTech, December Driskell, L. R., Sizing Control Valves, ISA Handbook of Control Valves, Gassmann, G. W., When to Use a Control Valve Positioner, Control, September George, J. A., Sizing and Selection of Low Flow Control Valves, InTech, November Harrold, D., Calibrating Control Valves, Control Engineering Europe, Vol. 3, Issue 3, p. 43, June/July Hägglund, T., A Control-Loop Performance Monitor, Control Engineering Practice, Vol. 3, Issue 11, pp , October Kaya, A., Training in Process Control by Distributed Computers: A Hands- On Approach, ISA Transactions, Vol. 29, Issue 3, pp , Lipták, B. G., Control Valves in Optimized Systems, Chemical Engineering, September 5, Ogawa, K. and Kimura, T., Hydrodynamic Characteristics of a Butterfly Valve Prediction of Torque Characteristics, ISA Transactions, Vol. 34, Issue 4, December Price, V. E., Smart Valve Intelligence Takes Many Forms, InTech, November Simula, M., Improve Control with Smart Valve Technology; Benefits Include Reduced Maintenance and Better Stability, Hydrocarbon Processing, Vol. 78, Issue 8, p. 63, August Thanomsat, C., Taft, C. W., and Annaswamy, A. M., Level Control in Feedwater Heater Systems Using Nonlinear Strategies, ISA Transactions, Vol. 37, Issue 4, pp , September Tolliver, T. L., Continuous Control Experiences, ISA Transactions, Vol. 30, Issue 2, pp , Tullis, J. P., Hydraulics of Pipelines: Pumps, Valves, Cavitation, Transients, New York: John Wiley and Sons, Valenti, M., Improving Hydraulic Performance with Intelligent Valve, Mechanical Engineering-CIME, Vol. 118, Issue 4, p. 56, April 1996.

117 6.8 Diagnostics and Predictive Valve Maintenance K. BEATTY (2005) Types of Predictive Maintenance: Types of Diagnostic Systems: Special Features: Cost: Partial List of Valve Diagnostics Suppliers: a) Performance tests b) Valve characterization and signatures c) Valve response tests a) Field testing using specially mounted sensors b) Smart positioners with integral sensors c) Host system tools that measure response d) Partial stroke test for safety-related valves a) Actuator pressures, temperature, and position measurements b) Fully integrated with smart positioners, or field installed for individual tests c) Specially prepared valves for regular testing Premium for $300 $1000 for smart positioner, $ per test for field-mounted test systems ABB Inc. ( Dresser Flow Control, Masoneilan Operations ( Fisher Controls International ( Flowserve, Flow Control Div. ( Honeywell Industry Solutions ( Invensys Flow Control ( Metso Automation ( Samson Controls Inc. ( Siemens Energy and Automation ( Smar International Corp. ( Yamatke Corp ( Yokogawa Corporation of America ( INTRODUCTION There is currently a major change in the approach taken to equipment maintenance due to longer times between turnarounds and a reduction in the available personnel. Valves are typically one of the higher maintenance pieces of equipment in a process. This is because, unlike measurement instrumentation that only monitors the process, valves are dynamic devices, and in order for them to function properly, their main components, such as their actuators, packings, positioners, and I/P converters, all need to be operational. If any of these components fail, the result is that the process can no longer be remotely controlled. For this reason, control valve testing should evaluate the condition of all of these components. Valve diagnostics and predictive maintenance is relatively new. In the past, the approach to maintenance was to schedule the periodic rebuilding of critical valves based on the time that the valve has been in service and not on the basis of actual need. This type of preventative maintenance method has several drawbacks. One is that by using this approach, a lot of unnecessary maintenance is done. Secondly, if critical failures occur between maintenance intervals, they are overlooked. Finally, sometimes the act of maintenance itself causes problems in valves that had no problem before. DIAGNOSTICS Instrumentation Used Diagnostic information can be obtained from smart positioners, local portable sensors, and the evaluation of process control signals. As discussed in more detail in Section 6.12, smart positioners are relatively inexpensive and easy to use. For these reasons, they are used most commonly. The local sensors utilized in diagnosing the control valves are usually integrally mounted on the valve. In safety-related control valve systems (as discussed further in Section 6.10), 1161

118 1162 Control Valve Selection and Sizing specialized positioners are used to provide partial stroke tests, which can verify the operational status of the valve without interrupting the process. When diagnosing existing installations, the testing is often performed by analyzing the past performance and records of the process instrumentation. When time permits, portable sensors can be installed in the field to monitor the performance of the valves and to provide the required data for diagnostic tests. Intermediate signal control boxes can also be used, primarily in safety-related systems. These can partially stroke on/off safety and shut-off valves (Section 6.10) by briefly interrupting the signal to a solenoid. This way, the system can verify that the valve is operational, because it is beginning to operate. In order not to adversely affect the operating process, once the valve begins to respond, the solenoid is quickly re-energized. This way, the valve is tested without a need to shut down the process while testing. Diagnostic Methods Standards such as ISA , Method of Evaluating the Performance of Positioners with Analog Input Signals and Pneumatic Output, or ANSI/ISA , Test Procedure for Control Valve Response Measurement from Step Inputs, are useful in evaluating the performance capabilities of control valves. These standards describe the performance tests required to determine HRL hysteresis, repeatability, and linearity (HRL) of control valves. These tests provide very useful information on the performance of both the positioner and the valve actuator, but they do not provide much information that would be useful in predictive maintenance. Many smart positioners have built-in diagnostics that can provide usage information and current status that is very useful in troubleshooting existing problems. Most of these positioners do not provide on-board predictive diagnostic information. In order to perform predictive diagnostic tests, the positioner usually has to be accessed through a digital interface and its data downloaded into a PC for analysis. Many new host control systems can pass the information directly to a diagnostic PC over a network connection. Some positioners also have internal data historians that provide detailed information on the valve command, actual position, and actuator pressures that provide a level of detail not normally available in host historians for diagnosing transient problems. Characteristics Tests Another category of tests are the characteristic or signature tests, which are well suited for predictive diagnostics. They can be generated by a smart positioner or obtained by using a portable test rig. Valve signatures are obtained by ramping the valve through a range of travel and capturing information concerning the position command, the valve travel, and the actuator pressure(s). This information can be analyzed independently, or over time to observe changes in the characteristics of the control valve. For example, a signature can be the friction in a valve over a range of travel. Changes or abnormalities in the friction plot over the range of travel provide direct information related to seating problems or galling that may be occurring in the valve trim. A full travel signature can provide information on the travel calibration of pneumatic actuators or help to locate leaking or worn seals and fittings. New software tools are also available for the historical analysis of valve command signals and of the corresponding process response. This method of analysis is nonintrusive and can be applied with both smart and traditional positioners. Many of the analysis packages can access the required data directly from a historian in the DCS. This method of analysis involves the comparing of the controlled process measurement with the command signal received by the control valve. If the controlled process measurement is sensitive enough to respond to small changes in the valve opening, then this measurement can be correlated with the control signal received by the valve and, thereby, determine if there is sticking or lagging and quantify the magnitude of the problem. The main limitation with this type of analysis is that it provides an indication of the existence of a problem, but does not provide detailed information. Some equipment suppliers now offer diagnostic services on a contractual basis. In providing these services, the suppliers utilize large databases that they have compiled over time, using various types of valve designs in a wide variety of process services. These services often include a review of the existing installation to see if other valve designs might not be better suited for the applications at hand. Valve Signatures A valve signature is obtained by running a reproducible test on a control valve. During such a test, a repeatable signal is sent to the control valve, and the response to that signal is captured and then analyzed to see if the control valve operates correctly. The valve should be removed from the process and placed on a test stand while running the complete signature test. This is because during the test, the valve is forced to travel over a wide range, and if this was not done off-line, it could cause harm to equipment, personnel, or the process. For the signature of an emergency shutdown valve, refer to Figure 6.8a. However, a partial stroke signature can sometimes be obtained on-line. If a valve signature is obtained near the end of the valve stem travel, or even beyond the end of travel, it still can be of some value. For example, if the behavior of a normally closed valve is tested with a control signal range of 5 to 0%, one can obtain diagnostic data on the response

119 6.8 Diagnostics and Predictive Valve Maintenance Bar SOV initial operation (fault detectable) Volume booster/qev operation (fault detectable) Return stroke (fault detectable) Actuator and ESDV movement (fault detectable) 1 Bar Spring operation causes flattening off of pressure (fault detectable) Superimposed full depressurisation without ESDV movement FIG. 6.8a The signature of an emergency shutdown valve (ESDV). (Courtesy of Drallim Industries Ltd.) of the pilot valve in the positioner without opening the valve (moving the valve stem) at all. Similarly, limited signatures can be obtained, when control valves are in operation, by changing their control signals within a small range (1 2% of travel) around the required opening of the value. Valve Signature Types There are four types of valve signatures. These are obtained by detecting the valve s response to steps, ramps, partial strokes, or internal operation. The step and ramp signatures can be performed in either the opening or closing directions. Figure 6.8b shows the ramp test signature of a control valve with a double-acting actuator, and Figure 6.8c shows the same signature for a valve provided with a springloaded, single-acting actuator. When performing a ramp test, first the starting and ending valve positions are entered. The ramp rate and the sampling time is usually also entered. The data gathered at each sample point includes the value of the control command signal, the valve position, and actuator pressure. In order to obtain a step signature, first the step size is defined by setting the starting and ending valve positions, and then the recording time period is set, where time zero is the time at which the step command is issued. The data gathered at each sample point include elapsed time and the values of the control command, valve position, and actuator Percent Ramp test Time vs command Time vs position Time vs top pressure Time vs bottom pressure Delay in initial movement due to saturation of actuator Migration of pressure levels may indicate actuator leakage Pressure spikes indicate sticking or galling Offset due to fail-safe spring in actuator pressure above pistion pressure below pistion Valve command and response curves Seat contact and loading Pressure (psi) Time (seconds) Travel limits indicate calibration status FIG. 6.8b The ramp test signature of a control valve with a double-acting actuator. (Courtesy of Flowserve Corporation.)

120 1164 Control Valve Selection and Sizing Actuator pressure (psi) Closed Travel (in) X: 0.93 Y: FIG. 6.8c The ramp test signature of a control valve with a spring-loaded (single-acting) actuator. (Courtesy of Fisher Controls.) pressure. The result is a graph that shows the response of the valve to a step in its control signal including the approach and stabilization of the valve at the new position. A partial stroke test is similar to a step test, but it is based on a temporary and small change in the control signal. Instead of reaching and maintaining an end position, the test only lasts long enough for the valve to begin to move or sometimes not even that, but only begins to apply the necessary pressure to begin to move the valve stem before resuming normal operation. Internal operation signatures are primarily positioner diagnostic tools, because they test the operation of the pneumatic relay. Analyzing Valve Signatures Most valve manufacturers provide information and guidance for interpreting the above-described signature tests. Interpretation of valve signatures considers two different factors: One is the characteristics of type of valve used, the second is the past history of valve performance. In terms of valve characteristics, the control signal and valve travel relationship is a function of the type of valve used. Similarly, the minimum and maximum stem friction values depend on the type of packing used by the particular valve design (see Section 6.5 on packing designs). Other characteristics include the balancing pressure range of double-acting actuators and positioners, which, if it is outside the design range, is an indication of problems in the actuator. Also, there usually are inflection points in the valve signature, which indicate the particular seating and linkage characteristics of the valve. The second basis for evaluating the signature of a valve is by making a historical comparison with past performance. The reference signatures can best be obtained when the valve is in good working order. Once a set of baseline signatures for each valve have been obtained, subsequent signatures can be directly compared against that baseline reference signature, which was taken under the same conditions as the existing ones. The differences between the reference and actual signatures can point to possible problem areas. The Instrumentation, Systems, and Automation Society (ISA) committee SP , Valve Diagnostic Data Acquisition and Reporting is in the process of developing a standard for the methodology to be used in acquiring valve signatures. Their scope includes the defining of the test procedures to be used and the obtaining of reports that can be transferred from one diagnostic system to another. CONCLUSIONS The analyzing of valve diagnostic information is a new and evolving field. The diagnostic needs of a plant can depend on its age, on the type of control system being used, and also on the experience and attitude of the operating staff. Today, a variety of diagnostic services are offered by valve manufacturers, instrument suppliers, and service companies. With the passage of time, standardization is expected both in the methods of obtaining and in the interpretation of valve signature tests. Bibliography Essam, D., Final Element Testing, A Way Forward, Drallim Industries, October 26, 2001, Miller, L., Valve Diagnostics Past, Present, and Future, Fluid Handling Systems, November Merritt, R., Some Process Plants are Out of Control, Plant Services Magazine, December Fedenczuck, P., Fountan, P., and Miller, R., Profitable Performance, Hydrocarbon Engineering, November 2001.

121 6.9 Dynamic Performance of Control Valves C. G. LANGFORD (2005) VALVE RESPONSE This section first defines the terms associated with control valve mechanical response. It then explains how the speed and accuracy of the control valve response will determine the quality of control of the process. Then it provides suggestions for improving control valve response to achieve acceptable performance. Improved data collection and analysis with modern control systems make the quality of process control more visible and quantifiable. The business requirements for profit and improved quality control require better control performance. Modern control valve positioners, often incorporating computer technologies, can improve system performance (see Sections 6.2 and 6.12). See the Bibliography for test data from poorly behaving valves (see Figure 6.9a). The elimination of asbestos-based packings and the increased concerns over packing leakage had the result of improvements in packing design and materials. Developments in lowfriction packing materials provide a powerful tool for improved control (see Section 6.5). Some applications may have control response requirements that preclude the use of a conventional control valve. For these, a pressure or flow regulator may be the better choice. Regulators tend to be relatively fast and reasonably accurate. They lack the control options provided with a separate controller and the design and materials flexibility provided by a separate valve. Some regulators are available with a pressure receiver replacing the spring and screw top works. This provides for remote control of the loading pressure, and A B Signal 0 5 Seconds FIG. 6.9a The responses to a step change in control signal by valves A, B and C. C thus the set point. Many regulators use techniques that decrease the droop, which is the normal pressure drop at high flows (see Section 7.7). Definitions Valve: A valve is a device used for the control of fluid flow. It consists of a body that must satisfy the key requirement of containing the fluid with flow ports, and a moveable closure member, which opens, restricts, or closes the port(s) (see ANSI/ISA , Control Valve Terminology ). Actuator: An actuator is a device that supplies the force causing the movement of the valve closure member. Most commonly these are fluid or electrically powered (see ANSI/ISA ). Actuators most often use compressed air but may use electric, hydraulic, or electrohydraulic power. Motion conversion mechanism: A mechanism installed between the valve and the power unit of the actuator to convert between linear and rotary motion where required. The conversion may be from linear actuator action to rotary valve operation or from rotary actuator action to linear valve operation. Accessories: Additional devices used in the operation of the control valve. As described in ANSI/ISA , typical examples include a positioner, transducer, signal booster relay, air set, and snubber (see also Section 6.2). Dead band is the range through which an input signal may be varied, with reversal of direction, without initiating an observable change in output signal (ANSI/ISA-S [R1993]). In standard ANSI/ISA it is defined in percentage of input span. Note that in some other literature this definition is used for dead zone. See Figures 6.9b and 6.9c. The input signal is shown here as a linear ramp for clarity. Control valve response is the dynamic (considering time) relationship between the change in a control signal and the change in the valve stem position. The process operator is interested in the process response, not the valve stem position. It is important to remember the end purposes of the control valve functionality are business and safety. Process safety is easier if the process is under good control. Much of the discussion in this section uses the conventional globe style valve for the examples. However, the basic principles apply to all final control elements. The response of electric or air motors in a film windup, the performance of swirl vanes 1165

122 1166 Control Valve Selection and Sizing used on the inlet of a blower, and variable speed pumps all have dynamic characteristics similar to valves. DISCUSSION Valve position FIG. 6.9b Dead band is defined as the amount by which the controller output signal has to drop, before the valve position changes. This figure illustrates the response of a valve with high packing friction. Valve position Controller signal Controller signal Time Time Dead band Dead band FIG. 6.9c If the packing friction of the valve is reduced, the dead band required to initiate a change in valve position is also reduced. The function of a control valve is to change the closure member ( trim ) configuration in a way that the flow restriction coefficient, C v, changes in a controlled and predicted manner. Bernoulli s law states that except for losses and energy taken from or added to the fluid, the sum of the pressure energy, the kinetic energy, and the internal energy within a system will remain constant. The tortuous and varying fluid path in a control valve converts part of the pressure energy into kinetic energy, with a high-velocity jet and increased turbulence. Most of the turbulence dissipates into heat (internal) energy as the flow jet slows down into a more orderly velocity distribution downstream of the restriction. Some of the velocity energy will change back into pressure energy downstream of the flow jet at a lower pressure. The energy converted into heat will leave the valve with the flowing fluid or is lost through the piping wall. Except for inefficiencies, the quantity of energy lost is essentially equivalent to the energy required for a pump to increase fluid pressure by the same amount. The valve closure member connects to the actuator by way of a valve stem or shaft. The valve mechanical response is a function of the valve actuator and its various accessories and with the forces that act to cause or restrain the motion. The response of the process is a function of the response of the valve and the relationship is normally not a simple one. For difficult or sensitive applications, the valve response requirements should be determined from actual or calculated knowledge of the process response requirements rather than from a simple specification based only on common practices or habit. Valve System The installed control valve is a system, consisting of the body and trim, the actuator, and all the accessories. Air pressure supplied by a pneumatic controller or current to pressure (I/P) relay causes flow in to, or out of, the actuator and increases or decreases the pressure on the diaphragm (see Figure 6.9d). Information about valves and accessories is available in this handbook (Sections 6.2 and 6.12) and from catalogs or from various computer programs (see Table 6.2n). The manufacturers data can aid in developing practicable performance purchase specifications. There is no point in specifying an impossible valve or one that will require very special and expensive attention. Typically, a spring provides the opposing force in an actuator to move the stem in the opposite direction when the actuator air pressure decreases. The force is proportional to the change in stem position; this is Hook s law. The diaphragm actuator in the typical diaphragm changes effective area with pressure and stroke position. This results in an imperfect conversion of pressure into force. In the sliding stem valve, stem packing friction creates a force that opposes motion in either direction. This is often a large force, from hundreds to thousands of pounds. The result is to increase the pressure (mass of gas) required in the actuator to cause a change in stem position. Any source of friction will have a similar effect. With ball and plug valves, the friction of the liner and the seat is often much larger than stem packing

123 6.9 Dynamic Performance of Control Valves 1167 F Air friction. In this case, appreciable valve shaft windup may occur with shaft distortion from the high torque. This windup might not be apparent from outside the valve. Any freedom of motion between the closure member and the shaft will increase the dead band. Valve design details all play a part in this. Fluid forces on the closure member will vary with fluid conditions and the closure member design and position and may act in either direction. These forces are more significant in the larger valves. For valves smaller than 1 / 2 in. (12 mm), fluid forces are small. Balanced trim reduces these forces. Install Positioner Flow F Friction F Spring F Fluid FIG. 6.9d If both the process fluid and the valve spring are arranged to resist the closing of the valve, the air signal must generate a force, which will overcome these two, plus the force generated by the packing friction. The valve closure member will move when the sum of the forces acting to move the stem exceeds the sum of the forces that act to oppose motion. Because static (nonmoving) friction is higher than dynamic (moving) friction, the valve stem will not move in a smooth manner but will jump to a new position and then stop, because the actuator force decreases when it moves and the pressure drops. This new position is only rarely perfectly, and exactly, the position represented by the control signal. This valve response will result in a flow response and, then after that, in a process response. The process value will result in a change in the control system output, and the cycle starts again. The word cycle used here makes the point that cycling is almost inevitable. It is almost certain that the stem will move in steps. This discussion is about the size and significance of those steps. In the absence of other forces, the spring coefficient defines the stem position for a given actuator pressure. With other forces added in, the stem position will differ. Adding a positioner to the system reduces the effects of the other forces and raises the stiffness of the system as the positioner acts to oppose errors due to friction and fluid forces. The flow of air into or out of the actuator is an integrating (summing) process and will always lag or follow the input signal change. The process under control will also have characteristics of lag and dead time. The control system adds more lag. Despite all the efforts of the finest minds, it is not yet possible to see accurately into the future. Sophisticated modelbased control may help but nothing can be perfect. From this, it is clear that no real control system can control perfectly. The goal is to have something that is good enough to satisfy the system requirements. Valves, sensors, and the control system performance all together determine the quality of process control. Improvements from a base case will generally cost money, and the added costs must be justified and supported by the predicted desired improved performance. The relationship between these various times and magnitudes defines how well the process is controlled. Simply, a very fast process will require fast control response, and a process with high gain requires high accuracy. The catalog listings for most positioners and pneumatic relays list the available air flow capacity in SCFM (SCMM) or other mass units at the specified supply pressure (see Table 6.2n), the air capacity of selected pneumatic components. The positioner is a narrow proportional (high gain) pneumatic relay. This sensor and amplifier with the actuator comprises a servomechanism, with the valve stem as the controlled device. The positioner compares the difference between an input signal defined in terms of the stroke of the valve and the actual stem position as measured with a mechanical linkage, and it controls air to the actuator. The early simple designs had the disadvantage of cycling if the gain was too high for the response time of the valve assembly. That is, because the system is an integrating controller, the inherent phase lag in the stem response requires that the gain be limited or a low-frequency cycle will result. If the valve is the slowest element in the control loop, or behaves poorly, then it is the limiting control device. If the positioner slows the valve response, then it becomes the limiting device. If the volume of the actuator is large compared to the air flow rate, then the air flow rate is the limiting element. During the 1960s, typical positioners were very simple, with few options, and they had limited air flow capacity. In the early days of automatic control, special highperformance valves received special engineering attention. This led to papers that made the case that positioners could have a negative effect on control. The original positioners had a one-size-fits-all philosophy designed to work with any size of actuator. Price was a concern, and the possible improvement in control was not valuable enough to justify the extra costs, not only for purchase, but also for installation and operation. It is clear that in applications using larger valves, control improves with the use of positioners because of the servomechanism error correction. Concern with the speed of response led to higher air flow capacity. Later positioners were not linear in response, with low gain for a small position error and higher gain for larger errors. This also has the result that stem motion speed will differ with signal step size. Some modern positioners adjust to suit the specific application and may include

124 1168 Control Valve Selection and Sizing adjustable gain and integral action. Digital devices can improve accuracy, be more repeatable, and include the flexibility of configuration for the specific application. Positioners must have an air supply of in. (10 mm) tubing as a minimum, and each positioner requires a separate adequate supply filter-regulator. Control valves use substantial amounts of air, and severe control interaction can occur if an inadequate air supply affects the air supply pressure to other valves and transmitters. The interactions can be very complex, and the oscillations are impossible to stop. Increase Force The second available significant improvement is to increase the actuator force. A larger actuator will provide a greater force but adds cost and increases the air consumption. Increasing air supply pressure, up to the safe actuator operating pressure, improves response. An undersized or marginal actuator will exhibit slow response because it requires a greater change in pressure to start motion, and as actuator pressure approaches supply pressure, the flow rate will decrease. Many diaphragm actuators commonly used at 25 PSIG (170 kpa) can safely operate at pressures to 30 or 60 PSIG (200 or 400 kpa). It is necessary only to increase the air supply pressure and modify the calibration or use a pressure amplifying relay. If the change in actuator pressure required to move the stem is reduced, then the jump in position is reduced. The analogy might be using a wrench with a longer handle to loosen a frozen bolt. Control of the wrench is much easier if it is not necessary to exert your full strength on the wrench handle. There are practical limits to actuator sizes for each valve body size and type. Some actuator schemes use stacked actuators in line to achieve greater forces. Reduce Friction Perfect alignment of the valve actuator system and stem will minimize friction. The very close tolerances between the valve stem and the stem guide can only tolerate a small misalignment without binding up (locking) the assembly and prevent any movement. Packing friction will increase with misalignment. Careful attention during maintenance is important. Any looseness or play in the assembly will cause unpredictable friction. The proper tension on the packing follower bolts will prevent valve stem leakage without adding unnecessary friction. Careless overtightening may make the valve nearly inoperative and greatly reduce the life of the packing. Packing selection and design are important (see Section 6.5). Graphite packing or packing with graphite content, typically specified for high-temperature operation, will substantially increase friction. Stem packing friction will vary over a range of 10:1 or even more, depending on packing materials and designs. It is necessary to remember that the packing temperature may differ considerably from the process temperature towards the ambient and will depend on the design and installation. Valve manufacturers have data to help in these decisions. Optional extended packing designs provide even greater thermal isolation from the process. The low-friction TFE packings popular in the chemical industry are limited to moderate temperatures. Modern elastomeric low-friction packings can serve at temperatures far higher than the temperature limits of fluorocarbon packings and have passed standard fire tests in some valves (see Section 6.5). Packing selection is no longer a simple task. Defining Response The ANSI/ISA Standard, Test Procedure for Control Valve Response Measurement from Step Inputs, defines step sizes for signal input step change in terms of the resulting response. It defines four step sizes. First is a signal step so small that there is no response. The next larger step size results in some response, but it does not match the required specifications in amplitude or time. The next larger step size results in a response satisfying the specifications used. The largest step size is large enough that the amplitude is within specifications but the time is longer than specified. In each case, the specification used is a function of the application. As a useful definition for response time, the ISA (Instrumentation, Systems, and Automation Society) standard uses T 86, defined as the time for the valve to complete 86.5% of the final motion. Measuring Response Shop Test Valve testing on the bench at the factory or in a maintenance shop proves that the valve is mechanically functional, demonstrates valve response without fluid forces, and is much more convenient than field tests. A serious limitation is that there is no way to prove that the packing as installed will prevent stem leakage, and there is only a very limited ability to include fluid forces. (See Figure 6.9e.) In Situ Test In situ (in place) testing has the advantage of providing the desired information about the valve and process response under actual service conditions. The packing is fully functional; the test fluid pressures and temperatures are the service conditions. Control signal Actuator pressure stem Position: without friction with friction Time FIG. 6.9e Response curves: On the top step change in control air signal, second from top air pressure inside the actuator, third from top position of the (theoretical) valve stem w/o friction, fourth from top position of an actual valve stem with friction.

125 6.9 Dynamic Performance of Control Valves 1169 There are two kinds of in situ tests. Normal operation with normal flows, temperatures, and pressures provides the perfect model and provides the required data. Operating constraints may limit the allowable range of valve opening and testing. The alternative is to take data from a shutdown unit. This can determine only the valve mechanical response and may not be completely representative of the operating situation. In either case, it may be necessary to add temporary or permanent instrumentation to gather all of the desired information. Valve manufacturers and service companies can provide add-on instrumentation. Data collection may be as simple as personal field observation of stem movement and positioner gauges. It is easy to identify large stem position jumps and slow action by observing the signal gauges and the valve. If this approach does not provide adequate information, it may be necessary to introduce precise signal step changes of various sizes as shown in the above referenced standard and to take measured data from the valve stem along with the related flow meter and other process sensors. There is a general opinion that the usual standard process instrumentation is rarely adequate to collect all of the desired data. The user will want to take advantage of all the available resources and to develop and justify a plan to quantify the problem and determine the solution. Relationships A 1% change in stem position does not translate into a 1% change in flow coefficient and result in a 1% change in the process variable. The engineering requirement is to determine the effect of flow changes on the process and then derive the required control valve specifications from that. Determine Required Response Specifications In situ testing should provide enough information to define the required valve response specifications. The valve manufacturer can provide guidance for sound decisions for improving the performance of a specific valve design. The engineering analysis required to predict valve requirements is not difficult. For a number of reasons, the most common error in valve specifications is to oversize the valve. An oversized valve will increase the effective dead band by that same percentage. The definition of inherent rangeability of a control valve is: Rangeability: The ratio of the largest flow coefficient (C v ) to the smallest flow coefficient (C v ) within which the deviation from the specified inherent flow characteristic does not exceed the stated limits (Ref: ISA ). The ANSI/ISA Standard includes a definition for C vr that is a calculated valve coefficient for the installed valve. This provides a way to quantify the real valve performance and to determine the effective rangeability. The installed rangeability for an oversized valve may be much less than for a properly sized valve, because control is impossible at small valve openings due to the dead band and very high gain near the seat, and the excess capacity of the fully open valve has no effect on flow. Accurate service condition information is vital. It is far too common to add excessive contingency factors in sizing pumps and valves. A full process simulation will provide valuable data about process gain and valve gain and result in an understanding of how dead band will affect control. In the absence of the simulation, the user will need to estimate the valve gain at operating conditions and to estimate the response time needed. It is not economical to specify very tight control valve dynamics response specifications except where the process conditions justify this added cost. Application Examples Simple Tank Level Control Fluid enters the tank at a variable rate; an outlet valve controls level, and process considerations determine acceptable level limits. The analysis consists of considering the tank volume, the normal volume flow rates, and the allowable level deviation from set point. If the inlet flow can suddenly stop, calculate the required outlet valve response time from the rate of change in level. The required maximum allowable valve dead band is determined from the required precision for level. The allowable level dead band can be converted into flow dead band and then into the stem dead band at operating conditions in order to calculate the required valve dead band specification. A simple pressure control loop is a very similar problem. Heat Exchanger One fluid exchanges its heat energy with another fluid in a heat exchanger. The process engineer should provide the range of flow rates and temperatures and the impact of a change in the heating/cooling flow rate. In most cases, the response time is not an issue. The sensitivity, or gain, of the temperature from the flow rate changes will determine the allowable dead band. Because this is a slow process, the dead band effects become very visible. The common attempt to conceal the dead band by reducing controller gain and lengthening reset time will only make the cycle time so long that everyone forgets it. Any reduction in dead band will improve control. Knowledge of the expected valve dead band will permit calculating the required controller settings. Flow Control Most flow measurements are inherently noisy; the associated controller has low gain and fast reset function to filter the higher frequency valve control signals and still bring the flow towards the set point. The valve speed of response must be adequate for the anticipated pressure upsets. Dead band is not an issue because the controller signal changes rapidly and often. Usually, the control is adequate with a standard valve. In compressible flow, the downstream volume provides some smoothing. Heavy filtering at the transmitter will make everything look nice but the system will respond slowly to upsets.

126 1170 Control Valve Selection and Sizing Reactor Mixing A fast process reaction resulting when two fluids mix in a mixing volume is difficult to control. Downstream mixing will not make up for the errors. Past practice has been to make both flow measurements as similar as possible, and the control systems and valves as similar as possible, in order that the dynamics all are similar and, wishfully, all the dynamic errors will match and balance out. The process reaction is over and gone too quickly. The dead band must be very small, and the required precision of adjustment is very high. Special motor-operated small valves have satisfied these requirements. Process data is required to determine the allowable dead band. Neutralizing Waste Water This is another difficult control problem; flows and stream analysis may vary widely (see Section 8.32). The worst problem is that the neutralizing curve may have a very high gain over a narrow range, and this may vary with time. The region of very high gain is usually close to or at the set point value. The requirements are that the valve have enormously high precision and accuracy and then operate over a wide range. There are no real valves available to satisfy these specifications. The most popular solution is to use two valves in parallel. The small valve provides the precision and the larger bypass valve provides the bulk of the flow capacity. One reset-only controller adjusts the larger valve to cause the ph proportional-only controller to adjust the smaller valve to operate near 50% open. This solves both of the problems of precision and capacity. The natural dead band of the larger valve provides a gap action to allow the smaller valve to control over the small flow range. The price paid is a slow speed of response for large demand changes. Common practice is to use a downstream surge tank to mix and average the outflow ph. The design problem is to size the small valve large enough to be in control most of the time but to not oversize it and have too large a dead band. In real life, this means the study of the predicted neutralization curves and converting this information to the equivalent stem motion. For even greater rangeability, it is possible to extend this concept to a third valve. In each case, prediction of each dead band is valuable information. Antisurge Valve A centrifugal compressor requires a certain minimum mass flow through the machine in order to maintain a discharge pressure. If gas flow decreases below this value, then the compressor will lose flow and pressure and start to cycle internal flows and pressures. In larger machines, this surging puts excessive force on bearings and shafts and can quickly destroy the machine. Normal practice is to measure the gas mass flow into the machine and open an antisurge valve to discharge to atmosphere or to recycle enough gas around the machine to maintain the minimum flow. On a sudden decrease in gas demand, such as an interlock shutdown, the valve must open very quickly and then begin to throttle as needed in order to maintain stable operation. The compressor manufacturer can provide information on the required valve opening speed, or it is possible to estimate these values from the machine curve and knowledge of the pressurized volume. Adequate response may require the dead band high-volume booster relay described in Section 6.2. Other issues for these valves are associated with aerodynamic noise. Delay or Slowdown Valve Action There are a number of situations where it is desirable to delay or to slow valve action. For example, water hammer is the excessive pressure surge that is the result of a sudden change in fluid flow in a very long pipeline. Accurate control of the valve is still important; in some installations, the fluid flow will exert a high force and rapidly force the valve closed. When flow suddenly changes, a pressure surge wave will travel through the fluid at the speed of sound (c) in the fluid. This pressure wave will reflect off the first solid surface (elbow or tee) and travel back and forth. This wave may shake or damage the valve and the piping. The situation is strongly dependent on the exact flow and piping situation. Even a few percentages of entrained air or gas bubbles in a liquid will increase the compressibility enough to make a considerable difference. Reported problems include pipe hangers broken in a 2-in. pipe line at a high flow rate and very fast shut-off valves. A 4-ft.-diameter fiberglass waste water pipe system collapsed when the upstream valve closed too quickly. Some hydraulic power plants use surge tanks to absorb the energy. Possibly the best scheme is to use a reset-only controller to get a predictable, linear delay. With a digital control system, a ramp function configuration will provide an accurate and dependable rate of closure or opening. If the valve is on/off, an equal percentage-type trim will provide some smoothing of flow. Some authors have suggested a fast-acting pressure control loop can override the control signal to slow the valve to limit the overpressure. A very noisy process signal can lead to excessive valve or packing wear. The inverse derivative relay, which passes a fraction of a sudden input change immediately and the balance of the change later, will provide a low pass filtering effect (see Equation 1.2[14]). A needle valve, or restrictor, used with a volume chamber will provide a resistor/capacitortype filter. Safety Solenoid Valves Emergency shutdown systems (ESDs) may use a solenoid valve piped between the positioner and the valve actuator installed to reduce the chance that the positioner could interfere with positive valve action. Consider that the air flow capacity (C v ) of the solenoid might limit air flow and valve response. See Section 6.2 for more details.

127 6.9 Dynamic Performance of Control Valves 1171 Troubleshoot Valve Response Control valve response may be poor because of Oversize valve Incorrect trim Excessive packing friction Undersized actuator Loose linkage Inadequate mounting Inadequate air supply capacity or pressure, or plugged air filter Damaged or leaking tubing Leaking air lines or fittings Inadequate positioner flow capacity Broken actuator spring Incorrect calibration Leaking or damaged diaphragm or piston Positioner/actuator time constant equal to process time constant Manual valve handwheel or limit stops prevent free motion Bypass valve open Valve body plugged Broken valve stem Failed or plugged interlock solenoid valve Some of these possibilities are unlikely depending on the situation history; that is, did it once work well? Did it have recent maintenance? Are there changes to the flow system? In all the above, once the cause is identified, the solution is usually clear. Identification may be as simple as from observing the positioner gauges and the valve action. As in any control situation, instability occurs when phase (time) lag causes the control loop to react too late. A valve actuator has capacitance; that is, the valve does not change position when air flow starts to enter but moves only when enough air accumulates to build up pressure and cause motion. The classic control oscillation problem occurs when the valve time constant is the same as the rest of the process. A positioner has a small input capacitance, but because it causes large flow amplification, it reduces the effective actuator capacitance. A control loop tuned with very little gain and a long reset time may have a cycle time of hours and hide the cycle. It is often very difficult to discover whether a given control loop is causing a cycle or is just responding to another loop. Bibliography ANSI/ISA , Test Procedure for Control Valve Response Measurement from Step Inputs. ANSI/ISA-TR , Control Valve Response Measurement from Step Inputs. ANSI/ISA-S , Control Valve Terminology. ANSI/ISA-S (Reaffirmed 1993), Process Instrumentation Terminology. ANSI/ISA-S75.13, Method of Evaluating Performance of Positioners with Analog Input Signals and Pneumatic Output. Control Valve Dynamic Specification, EnTech Control Inc., Toronto, Canada, References Rich, G., Hydraulic Transients, Mineola, NY: Dover Publications, Inc., Shinskey, F. G., Process Control Systems, 4th ed., New York: McGraw-Hill, Coughran, M. T., Measuring the Installed Dead Band of Control Valves, ISA TECH/ Champagne, R. P. and Boyle, S. J., Optimizing Valve Actuator Parameters to Enhance Control Valve Performance, ISA Transactions 35 (1996) Instrumentation, Systems, and Automation Society-S26 (1968) and ANSI MC , Dynamic Response Testing of Process Control Instrumentation. Senior, K. A., Technical Guidelines and Design Information, Using KVSP Packing Systems for Improving Process Control and Minimizing Fugitive Emissions, white paper, DuPont Dow L.L.C., Senior, K. A., Valve Packing Systems Improve Process Control, Chemical Processing, June Brestal, R. et al., Control Valve Packing Systems, technical monograph 38, 1992, Marshalltown, IA: Fisher Controls. Packing Selection Guidelines for Sliding Stem Valves, Bulletin 59.0:062, Marshalltown, IA: Fisher-Rosemount. Langford, C. G., A User s View of Process Control and Control Valve Positioners, ISA, 1996, Paper # / Cho, C. H. Using Control Element Characteristics to Compensate for Process Nonlinearities, Control Engineering, August 1974.

128 6.10 Emergency Partial-Stroke Testing of Block Valves A. S. SUMMERS (2005) Features: Purpose: Discussion of partial-stroke testing of block valves Enhanced diagnostics can be used to achieve higher safety integrity level using partialstroke testing. Codes, Standards, and Instrumentation, Systems, and Automation Society (ISA), ANSI/ISA , Recommended Practices: Functional Safety: Electrical/Electronic/Programmable Electronic Safety-related Systems, Research Triangle Park, NC International Electrotechnical Commission (IEC), IEC 61508, Functional Safety: Safety-Instrumented Systems for the Process Industry Sector, Geneva, Switzerland International Electrotechnical Commission (IEC), IEC 61511, Functional Safety: Safety Instrumented Systems for the Process Industry Sector, Geneva, Switzerland INTRODUCTION ANSI/ISA (ISA 84) and IEC are new functional safety standards, covering the design, implementation, operation, maintenance, and testing of safety instrumented systems (SISs). Successful implementation of the safety life cycle model associated with these standards hinges on an essential design constraint: the safety integrity level (SIL). The SIL is a numerical benchmark, related to the probability of failure on demand (PFD). SIL is affected by the design quality, e.g., device integrity, voting, and common cause faults, and by the operation and maintenance strategy, e.g., diagnostics and testing interval. For many operating companies, one of the most difficult parts of complying with the standards is the testing interval often required for final elements, such as emergency isolation valves or emergency block valves. Traditionally, these valves have been tested at unit turnaround, using an off-line, fullstroke test to demonstrate performance. Thirty years ago, turnarounds were relatively frequent, occurring on average every 2 3 years. Due to successful mechanical reliability and preventive maintenance programs, many operating companies have been able to extend unit turnarounds. In some industries, it is now common to have turnaround intervals of 5 6 years. Extended turnaround intervals yield great economic returns through increased production. However, extended turnaround intervals also mean that block valves are expected to go longer between function tests, yet still achieve the same performance. This is simply not possible. When SIL 2 or SIL 3 performance is required, 5-year function tests are inadequate. Consequently, it is necessary to supplement the off-line full-stroke test. This involves implementation of valve diagnostics, such as partial-stroke testing, or alternate testing strategies, such as on-line full-stroke testing. Many users consider partial-stroke testing (PST) as a cost-effective alternative to on-line full-stroke testing (FST). The use of PST often eliminates the need for full flow bypasses, reducing engineering, capital, and installation costs, as well as potentially removing a bypass that could be inadvertently left open. Partial-stroke testing improves the block valve performance, as measured by the Average Probability of Failure on Demand (PFD AVG ). The amount of the reduction is dependent on the valve and its application environment. This section will discuss the method of determining the actual impact of the partial-stroke test on PFD AVG. It will also present a discussion of the three partial-stroke testing methodologies that are currently being evaluated and used by industry. THE PARTIAL-STROKE TEST There are three basic types of partial-stroke test equipment: mechanical limiting, position control, and solenoid valves. Each type involves different levels of sophistication and risk. 1172

129 6.10 Emergency Partial-Stroke Testing of Block Valves 1173 Mechanical Limiting Mechanical limiting methods involve the installation of a mechanical device to limit the degree of valve travel. When mechanical limiting methods are used, the valve is not available for process shutdown. The mechanical devices used for partial-stroke testing include collars, valve jacks, and jammers. Valve collars are slotted pipes that are placed around the valve stem of a rising stem valve. The collar prevents the valve from traveling any farther than the top of the collar. Any fabrication shop can build a valve collar, suitable for test use. A valve jack is a screw that is turned until it reaches a set position. The valve jack limits the actuator movement to the screw set position. The valve jack is ordered from the valve manufacturer when the valve is purchased. Valve jacks work with both rising stem valves and rotary valves. Mechanical jammers are integrated into the rotary valve design. They are essentially slotted rods that limit valve rotation when placed in position using an external key switch. Since the jammer is integrated into the rotary valve, the jammer must be purchased from a valve manufacturer. A contact can be provided for the key switch to allow annunciation in the control room whenever the key is used. Mechanical limiting methods are inexpensive in terms of capital and installation costs. These methods are manually initiated in the field and are personnel-intensive. A limit switch or visual inspection is used to confirm block valve movement. Successful test implementation and return of the block valve to normal operational status are completely procedure driven. For valve collars and jacks, bypass notification to the control room is entirely procedural. For the jammer, automatic notification using the key switch contact can be provided. Two of the biggest drawbacks to these methods are the loss of protection that occurs during the test and the lack of assurance that the valve is in or has been returned to normal status. There is no way to know for certain that the jack or jammer has been completely retracted without actuating the valve. Furthermore, unauthorized use of the valve jack or jammer cannot be determined by casual inspection. This means that the valve could potentially be out of service, with operations personnel unaware of the situation. These methods do not add to the normal operating spurious trip rate. However, there is the potential for a spurious trip during the partial-stroke test. For valve collars, the main culprit of spurious trips is improper installation, causing the collar to pop off the stem when the valve begins to move. Jacks and jammers must be placed in service by the technician; so procedural mistakes can result in the valve closing completely rather than just partially. Therefore, these methods are really only as good as the written procedures and technician training. Position Control Position control uses a positioner to move the valve to a predetermined point. This method can be used on rising stem and rotary valves. Because most emergency block valves are not installed with a positioner, this method does require installation of additional hardware. Consequently, cost is a major drawback for the position control method. A limit switch or position transmitter can be used to determine and document the successful completion of the tests. If a smart positioner is used for the position control, a HART maintenance station can collect the test information and generate test documentation. Of course, the use of a smart positioner and maintenance station further increases the capital cost. Some manufacturers have promoted the use of the positioner in lieu of a solenoid valve for valve actuation. However, most positioners do not have a large enough vent port (C v ) for rapid valve closure. Consequently, a solenoid valve should still be used for valve actuation. This solenoid valve must be installed between the positioner and the actuator. The positioner does contribute to the spurious trip rate during normal operation, because the positioner can fail and vent the air from the valve. When a solenoid valve is installed between the positioner and the actuator, the safety functionality is never lost during the partial-stroke test. De-energizing the solenoid valve will shut the valve, regardless of the positioner action. Solenoid Valve A partial-stroke test can be accomplished by pulsing a solenoid valve. The solenoid valve can be the same as the one used for valve actuation, resulting in lower capital and installation costs for this method than other methods. If the actuation solenoid valve is used, this method will also test the solenoid valve s capability to execute safe shutdown. Single Solenoid Valves The Minerals Management Service, which oversees safety for oil and gas operations for U.S. offshore waters, provides one method for partial-stroke testing. This method relies on the operator to pulse a single solenoid valve by turning a field-mounted switch, which de-energizes the solenoid coil for as long as the field operator holds the switch. The field operator monitors the valve position and releases the button when the operator confirms valve movement. When the valve moves, it can be inferred that the solenoid valve successfully vented. Of course, the main risk is that the operator may hold the switch too long, allowing the valve to close sufficiently to disrupt the process, resulting in unit shutdown. It is also possible to automate the single solenoid valve test using a pulse timer adjusted to achieve the desired valve travel. Valve travel confirmation is accomplished using a limit switch or position transmitter, allowing automatic documentation of test status. Because a failure of the solenoid valve or block

130 1174 Control Valve Selection and Sizing valve may result in excessive block valve travel, the pulse timer should be voted with the limit switch or position transmitter. If the valve reaches its desired travel point before the pulse timer is finished, the solenoid valve is reset. The test can be programmed in the SIS logic solver with the test implemented automatically on a programmed cycle time or initiated by the operator per a maintenance schedule. Another method is to measure the block valve position as related to air pressure in the actuator during the time of the solenoid pulse. This results in a fingerprint of the breakaway pressure for block valve closure, which can be compared with the original valve fingerprint. In order for this method to be effective, maintenance must have a specific procedure for examining the fingerprint to identify that the block valve is degraded and to respond with corrective action. When a simplex solenoid valve is being used to PST the block valve, the solenoid is de-energized and then re-energized. If the solenoid valve does not reset, the test becomes a trip. Using redundant solenoid valves can essentially eliminate this problem. TABLE 6.10a The Relationship among the Safety Integrity Level (SIL), the Average Probability of Failure on Demand (PFD AVG ), and Risk Reduction SIL PFD AVG Risk Reduction to ,000 to 100, to ,000 to 10, to to 1, to to 100 Redundant Solenoid Valves When arranged in a two-outof-two (2oo2) configuration, redundant solenoid valves provide improved reliability during normal operation and reduce the probability of a spurious trip during the PST. For processes that are sensitive to spurious trips, the reliability improvement is typically sufficient to justify the additional capital and installation costs. There are commercially available solenoid valve packages that provide on-line diagnostics of solenoid coil failure, facilitate on-line solenoid valve testing, and perform on-line, partial-stroke testing of the block valve. One particular package (patent pending) operates in a one-out-of-one hot standby (1oo1HS) configuration. During normal operation, the air signal passes through the package from the signal source to the valve actuator. When a trip occurs, the solenoid package vents the air from the valve actuator and allows the valve to move to its fail-safe position. With the 1oo1HS, one solenoid valve is used as the primary actuation solenoid and is confirmed on-line using a pressure switch. A secondary solenoid valve is off-line and confirmed in the vented state (off-line state) by a pressure switch. The safety logic solver is programmed so that if the primary solenoid valve goes to the vent state without being commanded (as detected by the pressure switch), the secondary solenoid valve is energized, preventing the spurious trip. Solenoid valve testing is performed by cycling the solenoid and by verifying that each solenoid valve successfully vents and resets using the pressure switches. The 1oo1HS can be used for PST by incorporating a PLC timer to pulse the power to the solenoids for just long enough to achieve the partial stroke. To verify the movement of the valve, a position transmitter or limit switch is used. The position indication is also used to prevent overstroking of the block valve; i.e., if the valve moves too far during the timed stroke, the solenoids are re-energized. For preventative maintenance activities, overstroke or understroke alarms can be configured to let maintenance know if the valve is moving too quickly or too slowly during the test. Impact of PST on SIL Safety integrity level (SIL) and MTTF are defined by the average probability of failure on demand for demand mode SIS. In Instrumentation, Systems, and Automation Society 84 and IEC 61511, there are four SIL classes. Each SIL class provides an additional order of magnitude risk reduction, as shown in Table 6.10a. The SIS standards require an examination of the PFD AVG to ensure that the risk reduction assigned to the SIS is met. It is important to note that SIL is not a property of a specific device. It is a SIS property, encompassing the field sensors through the logic solver to the final elements (e.g., solenoid valve, block valve, or pump motor control circuit). SIL is an important concept for safe operation, but plant management demands that the process plant operate at a high utilization rate. If a SIS is installed that has a low Mean Time to Failure Spurious (MTTF spurious ), the project will be considered a failure, regardless of the SIL that the SIS achieves. Consequently, any SIS assessment should include an analysis of the MTTF spurious. Let s examine a typical SIS, including transmitters, a redundant logic solver, solenoid valves, and block valves. In order to perform the PFD avg and MTTF spurious calculations, failure rate data is required. This data can be selected from various industry-published databases. The values used in this analysis are shown in Table 6.10b. For illustration purposes, the analysis will be presented in two parts: 1) the impact of PST on block valve performance, and 2) the impact of PST on the SIS performance. For the latter case, a dual solenoid valve package (1oo1HS) will be used for block valve actuation and PST. Other PST methods can be assessed using similar techniques. The results for other PST methods may provide very different results for PFD AVG and MTTF spurious. Consequently, the results presented in this section should not be generalized for all PST equipment. The reader is further cautioned to ensure that any data used during SIL verification is appropriate for their application. In other words, the results presented here may not be directly applicable to the reader s application.

131 6.10 Emergency Partial-Stroke Testing of Block Valves 1175 TABLE 6.10b Failure Rates Used in the Analysis Device MTTF D (Years) MTTF spurious (Years) Data Sources (1) Transmitters OREDA data for various transmitters Redundant Logic Solver 5000 Not included in analysis Vendor data for 1oo2D, 2oo3D, or 2oo4D logic solvers Solenoid Valve TR data Pressure Switch CCPS data TABLE 6.10d Average Probability of Failure on Demand of a Typical Block Valve Given as a Function of Testing Interval (Note: Does Not Include the Contribution of the Solenoid Valve) Testing Interval PFD avg 1 year 1.25E 02 2 years 2.50E 02 3 years 3.75E 02 4 years 5.00E 02 5 years 6.25E 02 Block Valve OREDA data (1) Due to the nature of the analysis presented in this chapter, values were selected to represent generic devices rather than specific devices. For the evaluation of an actual SIF, failure rate values representing the specific devices should be used. Block Valve Analysis Block valve components can be examined to determine which failures potentially result in the valve not operating when a process demand occurs. The causes of failures and associated modes of failure are shown in Table 6.10c. The PFD AVG can be calculated using the dangerous failure rate (λ D ) and the testing interval (TI). The mathematical relationship, assuming that the mean time to repair is small compared to the testing interval and that λ D TI is much small than 0.1, is as follows: PFD = λ D TI/2 6.10(1) Thus, the relationship between PFD and TI is linear. Longer test intervals yield a larger PFD AVG. The Offshore TABLE 6.10c Causes of Block Valve Failures and the Associated Modes of Failure Cause of Failures Actuator sizing insufficient for valve actuation Valve packing is seized Valve packing is tight Air line to actuator crimped Air line to actuator blocked Valve stem sticks Modes Valve fails to close (or open) Valve fails to close (or open) Valve is slow to move to closed or open position Valve is slow to move to closed or open position Valve fails to move to closed or open position Valve fails to close (or open) Reliability Data Handbook (OREDA) database has data for various valve types and sizes. For the purposes of illustration, a MTTF D of 40 years will be used. The failure rate λ D can be calculated from the MTTF D using the simplified equation: λ D 1 = MTTF 6.10(2) For a MTTF D of 40 years, this yields a dangerous failure rate of 2.5e 02 failures per year. The valve failure rate varies with type, size, and operating environment (e.g., process chemicals, deposition, polymerization, etc.). The reader should determine the appropriate failure rate for the specific application. The PFD, based on 2.5e 02 failures per year, is shown in Table 6.10d for various testing intervals. As expected, the valve performance at a 5-year testing interval is not the same as the valve performance at a 2-year testing interval. Due to the degraded performance at longer testing intervals, many companies have found that they must test the block valves on-line. Once facilities for on-line testing are installed, fullstroke testing can easily be performed. However, since a fullstroke test involves full contact of the valve seating members, frequent stroking can cause excessive wear to the block valve seat. This is a serious concern for soft-seated valves. Increased testing may achieve a higher integrity, but cause damage to the valve seat, leading to earlier valve failure. Another major concern is that the plant is unprotected while the block valve is in bypass for testing. The fraction of the time that the valve is in bypass must be considered in the PFD calculation. If the valve is bypassed every 6 months for testing and the test takes 1 hour, the PFD is increased by For longer bypass periods or more frequent testing, the impact on the PFD is even more significant. To maintain safety, operating procedures must include a list of the actions to be taken when the valve is in bypass, such as continuous monitoring of critical process variables and when manual shutdown should be initiated. D Valve seat is scarred Valve seat contains debris Valve seat plugged due to deposition or polymerization Valve fails to seal off Valve fails to seal off Valve fails to seal off Partial-Stroke Testing An alternative option to a full-stroke testing is a partial-stroke test. The test involves moving the valve from the fully open position. This tests a portion of the valve failure modes. The remaining failure modes are

132 1176 Control Valve Selection and Sizing TABLE 6.10e Dangerous Failures, Failure Modes, and Test Strategy Failures Failure Modes Test Strategy Actuator sizing is insufficient to actuate valve in Valve fails to close (or open) Not tested emergency conditions Valve packing is seized Valve fails to close (or open) Partial- or full-stroke Valve packing is tight Air line to actuator crimped Air line to actuator blocked Valve is slow to move to closed or open position Valve is slow to move to closed or open position Valve fails to move to closed or open position Partial- or full-stroke, if speed of closure or resistance to closure is monitored Partial- or full-stroke, if speed of closure or resistance to closure is monitored. Physical inspection Partial- or full-stroke Valve stem sticks Valve fails to close (or open) Partial- or full-stroke Valve seat is scarred Valve fails to seal off Full-stroke test with leak test Valve seat contains debris Valve fails to seal off Full-stroke test Valve seat plugged due to deposition or polymerization Valve fails to seal off Full-stroke test tested using a full-stroke test. The main purpose of the partialstroke test is to reduce the required full-stroke testing interval. How does partial-stroke testing affect the PFD? The valve failures are modeled in two parts: 1) those failures that can be tested using the partial-stroke (PS), and 2) those failures that can only be tested using a full-stroke (FS). For the calculation, the dangerous failure rate, λ D, must be divided into what can be tested at the partial-stroke (λ D PS) and what can only be tested with a full-stroke (λ D FS). To determine the percentage of failures that could be detected using PST, the failure mode distributions for various valve types and sizes contained in the Offshore Reliability Data Handbook were examined. This evaluation can be done for any valve type, based on the application environment and the shut-off requirements. Table 6.10e provides a listing of typical dangerous failures and failure modes for block valves. The test strategy indicates whether the failure mode can be detected by partialstroke testing or only by full-stroke testing. Based on the OREDA data, the typical percentage of the failures that can be detected by a PST is 70% for many valve types and services. Additional analysis can be performed to justify a higher percentage of detected failures. However, it is very difficult to substantiate a percentage greater than 85% for process industry applications. Those failures that are not detected during the PST are tested using an FST. An imperfect testing model is used for calculating the PFD AVG of the block valve when PST is utilized. The percentage of detected (PD) failures is used with the dangerous failure rate of the block valve (λ D ), testing interval (TI), and the mean time to repair (MTTR), as follows: PFD AVG = PD λ D TI PST /2 + (1 PD) λ D TI FST /2 + λ D MTTR 6.10(3) FST Supplemented by PST Table 6.10f provides a simple comparison of the PFD AVG for block valves tested using FST TABLE 6.10f Comparison of Average Probability of Failure on Demand (PFD avg ) for Block Valve(s) Undergoing Full-Stroke Testing (FST) Only or FST with Monthly Partial Stroke Testing (PST) (Note: Does Not Include Common Cause Failures or the Contribution of the Solenoid Valve) Single Block Valve Dual Block Valves FST Interval (Years) FST only PFD AVG FST and Monthly PST PFD AVG FST only PFD AVG FST and Monthly PST PFD AVG

133 6.10 Emergency Partial-Stroke Testing of Block Valves 1177 Probability to fail on demand (PFD) Partial stroke testing impact on PFD Valve is unavailable during partial stroke test Calculation assumes that 70% of the valve failures are tested at the partial stroke testing interval and that 30% of the valve failures are tested at the full stroke testing interval. The valve is unavailable for shutdown during the partial stroke test. Full stroke at 1 yr Full stroke at 2 yrs Full stroke at 3 yrs Full stroke at 4 yrs Full stroke at 5 yrs Partial stroke testing interval (HRS) FIG. 6.10g Relationship between the PST and PFD valves. Valve is unavailable during the test. only and using FST supplemented with PST. The PFD AVG is shown as a function of the full-stroke testing interval. For this illustration, the partial-stroke test is performed monthly. Similar calculations can be performed at other PST intervals. The results in Table 6.10f do not include the solenoid valve contribution to the PFD AVG. The PFD AVG for the single block valve is significantly reduced when PST is utilized. For double block valves, the PST lowers the PFD AVG by nearly an order of magnitude. The reader is cautioned that this breakdown is based on average valve performance and may not represent the breakdown for the reader s application. This evaluation should be done for each valve type, based on the application environment and the shut-off requirements. If the service is erosive, corrosive, or plugging, the failure rate and failure mode breakdown will be different from that shown in this chapter. If the valve is specified as tight-shut-off, the contribution of minor seat deformation or scarring may be more significant than shown in this chapter. Probability of Failure Using a dangerous failure rate of failures per year, Figure 6.10g shows the PFD when the test procedure requires bypassing the valve during the test. As expected, the PST does improve the PFD AVG performance of the valve. The star illustrates the point where the partialstroke test and full-stroke test are both conducted at a 1-year interval. This corresponds to the result shown for a 1-year full-stroke test in Table 6.10a. The downward trend of the curves for very frequent partialstroke testing is due to the valve being bypassed during the test. This removal results in the valve not being available for the fraction of time that the valve is being tested. The calculation assumes that the total test time is 30 minutes. If the actual test time is longer, the effect will be more pronounced. Figure 6.10h shows the PFD when the test procedure allows the valve to remain in service during the test. Very frequent partial-stroke tests improve the PFD substantially, because there is no loss of functionality during the test. Again, the star illustrates the point where the partial-stroke test and full-stroke test are both performed at a 1-year interval. For both test procedures, partial-stroke testing does improve the valve performance. For example, 5-year fullstroke testing achieved a PFD of (Table 6.10a). A 5-year full-stroke test supplemented with a 1-month (720 hours) partial stroke test achieved a PFD of , which is an approximately 30% reduction in PFD. In the cases of 1-year and 2-year full-stroke testing, a single block valve can potentially achieve SIL 2 performance when supplemented with frequent partial-stroke tests. For longer full-stroke testing intervals, the valve performance can increase from low SIL 1 to high SIL 1, depending on the partial-stroke testing interval. From the graphs, it is easy to see that no amount of partial-stroke testing is going to allow a single valve to achieve high SIL 2 performance, let alone SIL 3 performance, at full-stroke testing intervals of 1 year or more.

134 1178 Control Valve Selection and Sizing Probability to fail on demand (PFD) Partial stroke testing impact on PFD Valve is available during test Calculation assumes that 70% of the valve failures are tested at the partial stroke testing interval and that 30% of the valve failures are tested at the full stroke testing interval. The valve is available for shutdown during the partial stroke test. Full stroke at 1 yr Full stroke at 2 yrs Full stroke at 3 yrs Full stroke at 4 yrs Full stroke at 5 yrs Partial stroke testing interval (HRS) FIG. 6.10h Relationship between PST interval and PFD. Valve is available during the test. OVERALL SIS PERFORMANCE Let s examine a typical SIS, including transmitters, a redundant logic solver, solenoid valves, and block valves. For the PFD avg and MTTF spurious calculations, the failure rate data presented in Table 6.10b was used. In addition, the following assumptions were made: For redundant transmitters, it was assumed that analog signal comparison is performed and a fault alarm is initiated when the transmitters deviate unacceptably. D at the end of the voting architecture indicates the use of diagnostic coverage. The diagnostic coverage is assumed to be 80% for dual redundancy and 90% for triplicated redundancy. The PST can detect 70% of the valve dangerous failures. Common cause factor is assumed to be 2%. Mean time to repair is assumed to be 24 hours. The results of the analysis were plotted using bar charts to illustrate the relative contribution of each device on the PFD AVG. Single Block Valve Case For the simplex solenoid valve cases shown in Figure 6.10i, the SIS only achieved SIL 1 at annual function testing. When PST is used to supplement the full-stroke test, the SIS achieves very high SIL 1 to mid-range SIL 2. The individual bars can be examined to determine the major contributors to the PFD AVG. For the 1oo1 and 2oo2D cases, the transmitter is the major contributing cause to the PFD AVG. The impact of the transmitter can be decreased by changing the voting to 1oo2D or 2oo3D, which results in the single block valve being the major contributor. For the spurious trip rate calculation, the logic solver was not included in the calculation, because the spurious failure rate can vary so much dependent on the specific architecture. Some logic solvers have high spurious trip rates, and their contribution to the overall MTTF spurious can overwhelm the other SIS components. For this illustration, the logic solver was not included in this calculation to allow the examination of the field devices. For actual installations, the logic solver must be included in the PFD AVG calculation. The benefit of using the 1oo1HS is evidenced in Figure 6.10j. In each case, the spurious trip rate was reduced significantly when the 1oo1HS was used instead of a simplex solenoid valve. The simplex solenoid valve cases had a spurious trip rate of 10 years for the 1oo2D transmitter case and 15 years for the other transmitter cases. When the 1oo1HS is used, the MTTF spurious exceeds 45 years for the 1oo2D transmitter case and exceeds 420 years for the 2oo2D and 2oo3D transmitter cases. If the plant life is 20 years, the single solenoid valve cases yield an average of 1.5 trips during the life of the plant. Even if a spurious trip costs only $100,000, the simplex solenoid valve will result in the loss of $150,000 over the life of the plant. For the 1oo1HS cases, the MTTF spurious is greater than the plant life expectancy, yielding substantial savings.

135 6.10 Emergency Partial-Stroke Testing of Block Valves E 04 SIL E 03 Single safety shutoff valve PFDavg SIL E 02 SIL E 01 SIL E+00 1oo1 SOV FST only 1oo1 HS FST with PST 1oo1 transmitters including comparison of 1001 SOV with 1oo1 HS. 1oo1 SOV FST only 1oo1 HS FST with PST 1002D transmitters including comparison of 1001 SOV with 1oo1 HS. 1oo1 SOV FST only 1oo1 HS FST with PST 2002D transmitters including comparison of 1001 SOV with 1oo1 HS. 1oo1 SOV FST only Testing intervals: transmitters, logic solver, solenoid, and valves: annual. Valve partial stroke test (includes solenoid test): monthly 1oo1 HS FST with PST 2003D transmitters including comparison of 1001 SOV with 1oo1 HS. Redundant logic solver Transmitters (voting) FIG. 6.10i Plot of PFD AVG showing the benefit of using PST of a single block valve. Dual Block Valve Case Dual block valves are used in many installations to provide greater assurance that the process is isolated. When dual block valves are used, it is common for test intervals to be extended to unit turnaround. Consequently, for the analysis, a 5-year full-stroke test was assumed for the logic solver, solenoid valves, and block valves, while annual testing was used for the transmitters. For the simplex solenoid valve cases shown in Figure 6.10k, the SIS achieved high SIL 1 to the borderline between SIL 1 and SIL 2 at 1-year full-stroke testing. When PST is used to supplement the full-stroke test, the SIS achieves mid-range SIL 2 with 1oo1 and 2oo2D transmitter voting. For 1oo2D 1.000E E 03 Single safety shutoff valve STR(1/yr) 1.000E E 01 1oo1 SOV MTTFs: 15 yrs 1oo1 HS MTTFs: 80 yrs 1oo1 SOV MTTFs: 10 yrs 1oo1 HS MTTFs: 45 yrs 1oo1 SOV MTTFs: 15 yrs 1oo1 HS MTTFs: 420 yrs 1oo1 SOV MTTFs: 15 yrs 1oo1 HS MTTFs: 420 yrs Transmitters (voting) Expected unit life = 20 years 1.000E E+01 1oo1 transmitters including comparison of 1001 SOV with 1oo1 HS. 1002D transmitters including comparison of 1001 SOV with 1oo1 HS. FIG. 6.10j Plot of MTTF spurious showing the benefit of using PST of a single block valve. 2002D transmitters including comparison of 1001 SOV with 1oo1 HS. 2003D transmitters including comparison of 1001 SOV with 1oo1 HS. This comparison includes field devices only. The contribution of the redundant logic solver to the spurious trip rate is not shown.

136 1180 Control Valve Selection and Sizing 1.000E 04 SIL E 03 1oo1 HS FST with PST 1oo1 HS FST with PST Dual safety shutoff valve PFDavg SIL E 02 SIL E 01 SIL E+00 1oo1 SOV FST only 1oo1 HS FST with PST 1oo1 transmitters including comparison of 1oo1 HS. 1oo1 SOV FST only 1002D transmitters including comparison of 1001 SOV with 1oo1 HS. FIG. 6.10k Plot of PFD AVG showing the benefit of using PST of a dual block valve. 1oo1 SOV FST only 1oo1 HS FST with PST 2002D transmitters including comparison of 1001 SOV with 1oo1 HS. 1oo1 SOV FST only Testing intervals: Transmitters: annual. Logic solver, solenoid, and valves: five years. Valve partial stroke test (includes solenoid test): monthly 2003D transmitters including comparison of 1001 SOV with 1oo1 HS. Redundant logic solver Transmitters (voting) and 2oo3D voting, the PST makes it possible to achieve SIL 3. Again, the individual bars can be examined to determine the major contributors to the PFD AVG. As seen in the simplex block valve SIS, the transmitter is the major contributing cause to the PFD AVG for the 1oo1 and 2oo2D transmitter cases. The contribution of the transmitter is seriously reduced when the voting is changed to either 1oo2D or 2oo3D. In these latter cases, the major contributors to the PFD AVG are the dual block valves. The benefit of using the 1oo1HS is evidenced in Figure 6.10l. The MTTF spurious for the simplex solenoid valve cases is an average of 7 years. When the 1oo1HS is used for valve actuation, the MTTF spurious is more than 40 years for 1.000E E 03 Dual safety shutoff valve STR(1/yr) 1.000E E 01 1oo1 SOV MTTFs: 7 yrs 1oo1 HS MTTFs: 70 yrs 1oo1 SOV MTTFs: 6 yrs 1oo1 HS MTTFs: 40 yrs 1oo1 SOV MTTFs: 7 yrs 1oo1 HS MTTFs: 210 yrs 1oo1 SOV MTTFs: 7 yrs 1oo1 HS MTTFs: 210 yrs Transmitters (voting) Expected unit life = 20 years 1.000E E+01 1oo1 transmitters including comparison of 1001 SOV with 1oo1 HS. 1002D transmitters including comparison of 1001 SOV with 1oo1 HS. FIG. 6.10l Plot of MTTF spurious showing the benefit of using PST of a dual block valve. 2002D transmitters including comparison of 1001 SOV with 1oo1 HS. 2003D transmitters including comparison of 1001 SOV with 1oo1 HS. This comparison includes field devices only. The contribution of the redundant logic solver to the spurious trip rate is not shown.

137 6.10 Emergency Partial-Stroke Testing of Block Valves 1181 the 1oo2D transmitter case and more than 210 years for 2oo2D and 2oo3D transmitter cases. If the plant life is assumed to be 20 years, the simplex solenoid valve cases yield an average of 2.8 trips during the life of the plant. Again, if a spurious trip costs $100,000, the plant will lose more than $280,000 over the life of the plant due to spurious trips. CONCLUSIONS Partial-stroke testing does provide measurable improvement of the PFD over full-stroke testing alone. The amount of improvement is dependent on the specification, configuration, and application environment. The three partial-stroke testing methodologies offer choices between manual and automated testing. It is important to remember that partial-stroke testing is used to achieve diagnostics on the valve in lieu of on-line, full-stroke testing. Some of these methods presented in this chapter have a higher potential for spurious block valve closure (e.g., on-line spurious actuation or unintentional actuation during test) than others. For processes that are sensitive to spurious trips, the selection of specific method should take into account not only the diagnostic capability, but also the reliability. Whichever method is selected, procedures must be written to ensure that the block valve is not tripped during testing, the test is properly carried out, incorrect valve performance is documented, and maintenance is performed to return the valve to fully functional status. The primary issue is that partial stroke testing can reduce the full-stroke testing interval required to achieve the required SIL performance. Bibliography ANSI/ISA-ISA (IEC 61511), Functional Safety: Safety Instrumented Systems for the Process Industry Sector, Research Triangle Park, NC: Instrumentation, Systems, and Automation Society, IEC 61508, Functional Safety: Electrical/Electronic/Programmable Electronic Safety-Related Systems, Geneva, Switzerland: International Electrotechnical Commission, Guidelines for Process Equipment Reliability Data, Center for Chemical Process Safety, New York: American Institute of Chemical Engineers, IEC 61511, Functional Safety: Safety Instrumented Systems for the Process Industry Sector, Geneva, Switzerland: IEC, OREDA: Offshore Reliability Data Handbook, 4th edition, Norway: Det Norske Veritas Industri Norge as DNV Technica, ISA TR , Safety Instrumented Functions (SIF) Safety Integrity Level (SIL) Evaluation Techniques, Part 1: Introduction, Instrumentation, Systems, and Automation Society, Summers, A. E., Partial-Stroke Testing of Block Valves, Control Engineering, November Summers, A. E. and Zachary, B. A., Partial-Stroke Testing of Block Valves, ISA Expo 2002, Houston, TX, Instrumentation, Systems, and Automation Society, Summers, A. E. and Zachary, B. A., Improve Facility SIS Performance and Reliability, Hydrocarbon Processing, October 2002.

138 6.11 Fieldbus and Smart Valves J. BERGE (2005) Partial List of Suppliers: (Intelligent Positioners): Electric Actuator: Discrete Valve Coupler: ABB ( Biffi ( Dresser ( Emerson ( Flowserve ( Foxboro ( Metso ( Samson ( Siemens ( SMAR ( Tyco ( Westlock ( Yamatake ( Yokogawa ( EIM Controls ( Emerson ( Rotork ( Flowserve ( Ledeen ( Bürkert ( Emerson ( Flowserve ( Pepperl+Fuchs ( StoneL ( SMAR ( TopWorx ( INTRODUCTION The impact of fieldbus and smart technology on the design of control valves is profound. It changes how valves are installed, commissioned, set up, maintained, and operated. Ultimately, many benefits and savings can be achieved from using digital communication for valves. The brains of an intelligent control valve are in the digital valve positioner. An intelligent control valve is therefore primarily an issue of using digital valve positioners. For a basic introduction on positioners please refer to Section 6.2, for a discussion of intelligent valves and positioners see Section 6.12, and for a discussion of valve diagnostics and predictive maintenance, refer to Section 6.8. The brains of an on/off valve are in the discrete valve coupler (DVC). Therefore, using an intelligent on/off valve is an issue of using a discrete valve coupler. A Foundation TM Fieldbus or PROFIBUS-PA valve receives the command signal over the digital data highway or network. A smart control valve receives the signal via 4 20 ma but also provides digital communication for auxiliary functions. Actually, in the case of Foundation Fieldbus more often the control is done in the positioner itself, and the positioner receives the process variable from the transmitter over the fieldbus. Valve instrumentation based on HART, Foundation Fieldbus, and PROFIBUS-PA includes Control valve positioner Valve coupler Electrical actuator Fieldbus to current converter 1182

139 6.11 Fieldbus and Smart Valves 1183 Fieldbus to pneumatic signal converter Fieldbus remote I/O Position transmitter BENEFITS AND SAVINGS For valve users who spend too much effort and resources on maintenance, and see too much variability in product quality, network-enabled valves can significantly reduce the cost of operation. Unlike traditional hardwired valves, networked valve packages have microprocessors, which also allow for digital communication. Networking of valves and microprocessors has enabled several technological breakthroughs that in turn can result in savings. This is because all information can be accessed remotely and because all functions are performed in device firmware instead of hardware: flow characterization, soft limit switch, feedback, and automatic setup. HART, Foundation Fieldbus, and PROFIBUS-PA In Foundation Fieldbus-based systems all valves have a permanent network connection to the engineering tool and to the asset management tool (Figure 6.11a). The digital communication works across the same single pair of wires that is used for the desired valve position. Always on communications means that interrogation for monitoring, parameterization, and diagnostics can conveniently be done any time. Unlike in the past when smart valves, using HART communication, used temporarily connected handhelds, smart valves should have permanent communication with the control system in order to leverage the smart capabilities. A permanent communication link between valves and software enables on-line asset management. Operators can get on-line feedback of the control valve s true position and can monitor H1 fieldbus devices Operation Engineering Maintenance HSE fieldbus Conventional devices & HART Firewall I. S. barrier LIMS Execution Business FIG. 6.11a Illustration of a networked system architecture, which is always on. all critical parameters in real time without having to go into the field. A fair amount of data can be accessed also from a smart HART-based valve positioner. However, fieldbus with its higher speed and standardized data formats enables positioners to provide more diagnostics, and faster. Fieldbus-based positioners therefore can bring more diagnostics to the user than their smart counterparts. Digital valves can be configured remotely across the bus from a handheld or software tool. This makes it possible to adjust tuning, without having to leave the control room. A single type of digital valve positioner or discrete coupler can be configured for all control and on/off valves, respectively. Self-Diagnostics Digital valves have internal self-diagnostics that can be interrogated remotely across the bus from a handheld or software tool without having to go into the field. Diagnostics to this level of detail was never possible before the introduction of microprocessors and digital communication. More sophisticated valve diagnostics can be done by using on-line asset management or by utilizing proprietary software tools using information tapped from the positioner. In the case of Foundation Fieldbus the digital valve positioner or coupler can even issue alerts when failure is detected. The user can, from a single workstation, access all the positioners and other instruments using the engineering tools. The health of any device in the plant can thus be checked easily without having to venture into the field. In the past, valve diagnostics software was connected on a single unit basis. Diagnostics was performed after the valve had already failed, post-mortem. Additionally, digital valves internally collect operational statistics such as total travel and the number of reversals. The operational statistics are used for condition-based maintenance schemes. Interoperability Digital valves can be monitored remotely to see variables such as actuator pressure or torque, and to feed back actual valve position. In the case of Foundation Fieldbus and PROFIBUS-PA, the actual position feedback provided by the digital valve positioner or coupler is fast enough to be used as part of the control loop. The feedback signal travels digitally on the same wire as the desired valve position from the controller or logic, i.e., without additional wiring or external proximity switches. The HART, Foundation Fieldbus, and PROFIBUS-PA protocols are all developed and managed by independent organizations formed by several manufacturers. This means that many valve products, control systems, and other tools that communicate by using the same protocols are therefore compatible with them. It is possible to select valve positioners, couplers, and tools independent of the control system. All three protocols support electronics device description (DD, a.k.a. EDDL) in some form. The DD files are loaded in the configuration tool, permitting the tool to communicate with the digital valve and make all its special features accessible.

140 1184 Control Valve Selection and Sizing FIG. 6.11b Digital valve positioner. (Courtesy of SMAR.) DD for Foundation Fieldbus is very well supported in many control systems, while the DD for HART and PROFIBUS-PA is not so common. Thus better interoperability is achieved with Foundation Fieldbus, as the full set of information can be accessed from virtually any fieldbus system. PROFIBUS-PA is also taking two other approaches to interoperability; one is device profile, which defines a common basic set of functionality providing some degree of interoperability, and the other is FDT/DTM (field device tool/device type manager). Although FDT/DTM is also not widely supported in control systems, it does have some potential advantages when it comes to sophisticated valve diagnostics. Future enhanced versions of DD may also address this capability. Control Algorithm in Positioner A defining characteristic of Foundation Fieldbus is that it also is a standard function block diagram programming language to build the control strategy. In a modern network-based control system architecture control is often done in the field instrument (Figure 6.11b), and therefore it is possible to reduce or eliminate the number of control cards usually found in a DCS or PLC system. Fieldbus positioners perform control in the field by executing the PID control function block. Other blocks perform functions such as split-range, flow characterization, and the positioning servo itself. It is possible to adopt decentralized control using Foundation Fieldbus programming language instead of central controllers and proprietary languages. The failure of the electronics in the valve affects only a single loop. Diagnostics in the positioner and coupler detect loss of signal from the controller and subsequently bring the valve to its fail-safe position. Foundation Fieldbus and PROFIBUS-PA completely do away with 4 20 ma. Valves, transmitters, and other devices are multidropped on a network. Typically up to 16 devices are connected to the same pair of wires. Theoretically it is also possible to multidrop HART devices, but the communication speed is too slow for valve applications. Wiring Costs Using digital valves, the total installed cost can be reduced, achieving savings before the valve is even put into operation. Digital communications reduce installation cost in new construction and also in revamps. Networked control valve positioners require less wiring than their hardwired counterparts. For example, if actual position feedback is required from an analog positioner, it is necessary to run a second pair of wires back into the control room, where the DCS or PLC will also require an analog input point (on the left side of Figure 6.11c). For Foundation Fieldbus and PROFIBUS-PA no additional wiring is required because the feedback value is communicated on the bus together with the other values (at the center of Figure 6.11c). HART positioners also do not require an additional wire to run back into the control room. If the control system does not have native support for HART, it is possible to add a HART-to-current converter in the control room instead (Figure 6.11c right). The converter connects to the same two wires used for the positioner to receive the desired valve position signal from the DCS or PLC, and then it uses HART to poll the positioner to get the actual position digitally (Figure 6.11d). DCS/PLC Linking device DCS/PLC HART-tocurrent converter 4 20 ma 4 20 ma Bus Bi directional digital 4 20 ma with HART FIG. 6.11c Wiring requirements of providing position feedback in analog (left), Fieldbus (center), and HART (right) systems.

141 6.11 Fieldbus and Smart Valves 1185 together with other devices. Moreover, an interface module can handle as many as 64 devices, thus eliminating the need for a large number of conventional I/O modules. Significant savings in wiring and installation costs are therefore possible as compared to hardwired valves. In intrinsically safe installations the savings are even greater, because as many as four or eight devices share the same single safety barrier, depending on the scheme used. This too is a significant cost saver. To really benefit from Foundation Fieldbus and PROFIBUS- PA multidrop wiring, the valve electronics shall draw as little power from the bus as possible. Pneumatic valve positioners that use piezo technology to drive the flapper have significantly lower power consumption than the solenoid type. Lower current draw means lower voltage drop along the wires, which translates into more devices, longer wires, and fewer barriers. Therefore, by using lower power consumption positioners, such as 12 ma, further savings are possible. Positioners, couplers, and actuators based on HART, Foundation Fieldbus, and PROFIBUS-PA are available from multiple competing vendors. The competition among interoperable products brings prices down lower than for units using proprietary communication. Valve Calibration and Configuration FIG. 6.11d HART-to-current converter. (Courtesy of SMAR.) The converter generates an analog signal accepted by the analog input card in any control system or single loop controller. Additional safety barriers are also eliminated. If discrete open/close statuses are required, these are simply detected as alarms in the control system using the actual position signal. Similarly, networked discrete valve couplers require less wiring than their hardwired counterparts. For example, if open/ close statuses are required from a hardwired valve, it is necessary to run two additional pairs of wires back into the control room, where the DCS or PLC will also require two discrete input points. However, the discrete valve coupler provides it on the same bus (Figure 6.11e). An electric actuator may be even more extreme because the need for signals such as open/stop/close or desired valve position, shut-down, open/close statuses or actual position, local/remote, opening/closing, torque switch or torque value, and failures is not uncommon (Figure 6.11f). Using a bus, a lot of hardwire and I/O modules are eliminated. In this case, using bus technology, it is possible to reduce by anywhere from two to ten pairs of wires per valve even if one is only considering the auxiliary functions. Needless to say, the savings in cable and controller I/O cards and all the associated costs of engineering and installation is significant. Both Foundation Fieldbus and PROFIBUS-PA permit several devices to be connected on the same pair of wires Configuring and calibrating valves in the past required tedious mechanical adjustments. Using handheld terminal or computer software to access the valve, these tasks require less time and resources and thus result in savings (Figure 6.11g). For more detail on the subject of configuring intelligent devices, refer to Section 1.6 in Volume 1 of this handbook. Valve instrumentation without microprocessor and communication requires mechanical and electronics adjustments at the valve in order to change the way it operates. For example, the modification of flow characteristics previously had to be done by changing mechanical cams and springs. Using digital communication it can now be done remotely through software, thereby significantly reducing cost and time, particularly during commissioning. Other valve parameters, such as the opening and closing time of the valve, can also be set. To align mechanical feedback systems based on leavers, cams, and potentiometers in older positioners was tedious and time consuming, especially in the field. Using digital communication, a setup command can simply be sent from a handheld terminal or software in the control system to auto-calibrate the positioner. This significantly shortens the setup time, as stroking need not be done manually, achieving reduced installation cost. Once the setup is initiated the valve automatically finds its own fully opened and fully closed positions. A key advantage of Foundation Fieldbus is that control is also done in the field devices, typically in the valve. The decentralized architecture where control can be done in the field instruments is called Field Control System (FCS). Because control is done in the valve the number of loops in centralized controllers is reduced, and the number of

142 1186 Control Valve Selection and Sizing FIG. 6.11e Discrete valve coupler incorporates solenoid and position feedback. (Courtesy of TopWorx.) centralized controls can subsequently also be reduced. The savings achieved from the reduction in very expensive controllers are tremendous. To enjoy maximum flexibility, the positioner can have dynamically instantiable function blocks. This allows the user to freely select from more than a dozen blocks in a library containing scores of blocks including PID, splitter, and arithmetic. Valve Cycling and Stiction An inadequately functioning control valve will still degrade the overall loop performance and may cause the valve to wear out prematurely. Conversely, a healthy positioner ensures better control and less maintenance. The greatest savings potential for applying networked digital valves is in long-term operational costs such as maintenance and improved performance. The push for higher valve performance has been seen particularly in the pulp and paper industries, because of their need for increasing product uniformity. Digital valve solutions have higher performance than their analog counterparts, resulting in lower process variability, which in turn translates into higher quality, less rejects, higher yield and productivity, and lower raw material and fuel consumption. Information from the digital valve may be used to fine-tune the loop for optimum production output and tighter product uniformity. The higher valve performance ultimately results in savings because of the resulting boost in production and yield. Because aligning and calibrating analog positioners is tedious, this is often not done properly or not checked regularly. Automatic setup invoked through the fieldbus network makes sure it is done accurately, every time, so that positioning is precise. It is possible to remotely calibrate the positioner when conditions allow, e.g., while the process is down for any reason. As the valve seat and other parts wear and tear during operation, one can calibrate the valve to make sure that it operates and shuts off properly. It is commonly seen that control valve cycling can cause the control loops to cycle and that this can destabilize production. Studies show that as many as two out of three control loops are oscillating due to the dead band of control valves. The operational statistics such as total valve travel and number of cycles that now are provided by the positioners are extremely helpful in detecting oscillation problems and enable one to eliminate them (Figure 6.11h).

143 6.11 Fieldbus and Smart Valves 1187 gains and other parameters like the time required for opening and closing the valve can be set from a handheld or by a software tool to optimize positioner response to different valve sizes. For good control loop performance a control valve must be able to respond to small steps in the control signal by matching it accurately with its stem within 1% or less. If not, a problem called backlash or stiction will occur. It is typically caused by packing friction (Section 6.5), which in turn may be a result of wear and tear. In addition to being able to predict wear based on operational statistics, software is now available to analyze the valve s performance thanks to the improved communications capabilities to obtain data for such plots. The operational statistics may be used to more accurately predict the packing wear, as to strike the right balance of downtime due to replacement, and greater performance as a result of more frequent packing change. FIG. 6.11f Fieldbus electrical actuator. (Courtesy of Rotork.) One can eliminate the sources of variability one by one. A hunting valve is easy to catch because one can see a great number of reversals counted in the positioner or coupler and communicated over the fieldbus. If this occurs, one knows that either the positioner or loop is poorly tuned. The servo FIG. 6.11g The display of a handheld tool for configuring a digital valve positioner. (Courtesy of SMAR; the screenshot is of SMAR HPC301.) Position Feedback Fieldbus positioners provide an actual position feedback value over the communication network that is fast enough to be used as part of advanced control strategies and also for true bumpless transfer and true reset windup protection based on actual valve excursion end points. Because the valve position is available over the communication network, the actual position can be put into use at a later stage, without having to modify the valve positioner, running wires, or adding any AI points to the control system, as was the case in the days of analog feedback. Foundation Fieldbus thus makes actual position feedback possible for every control loop. This makes it feasible to provide true bumpless transfer back to the automatic control mode after the valve has been switched to manual control and was operated locally. Similarly, true integral windup protection can be provided when the valve has reached its endpoints or is unable to move. In the past this was done for some loops in conventional systems, but was not realistic for all. In fieldbus systems, this feature has no extra cost and therefore it is viable to optimize every loop and obtain the corresponding savings. Lastly, systems and devices using Foundation Fieldbus or PROFIBUS-PA eliminate the need for D/A and A/D conversions in the transmission path, as no analog signals are used. This results in higher accuracy. Because the actual valve position is transmitted back to the workstation using digital communication, inaccuracies caused by conversion to analog are eliminated. Additional points of calibration are also eliminated. Proactive or Predictive Maintenance For further details on valve diagnostics and predictive maintenance, refer to Section 6.8. Self-diagnostics done in the digital valves as well as using other information make it is possible to move to a proactive condition-based maintenance scheme that can lower maintenance expenses as compared to reactive or preventive maintenance schemes.

144 1188 Control Valve Selection and Sizing FIG. 6.11h Operational statistics, such as listed here, is easily reviewed and can trigger alarms notifying the operators that the valve is due for service. (Screenshot is courtesy of SMAR Syscon.) Asset management software can continuously poll digital valves over the network to find out their health immediately. This means that failed valves can be found faster and problems pinpointed with greater accuracy, making repair quicker and easier. Fewer resources spent means savings. For example, if the supply air to the positioner is lost, the operator is immediately notified. The operational statistics such as total valve travel and number of cycles that now are provided by the positioners are leading indicators of wear and tear. The operational statistics are used for proactive maintenance to determine the optimum time for valve overhaul. Based on valve manufacturer recommendation for stem packing change, alarm limits for these statistics can be set, notifying the technicians that maintenance is due. Spares can be ordered in advance, several valves can be fixed all in one shutdown, and maintenance can be done before the valve fails. Fewer unscheduled shutdowns mean higher productivity. Major maintenance savings are possible. In the past, many valves were removed, brought to the workshop, and torn down only to find nothing was wrong. Remote diagnostics reduce such unnecessary work by allowing the technicians to check the general health of the valve from the host before even going into the field. Valve instrumentation can be managed without having to leave the workstation. The cost that can be saved by not having to bring one valve into the workshop is substantial. Users can instead target their maintenance resources to the valves that really need attention. It primarily is the asset management software that allows technicians to find out what is really wrong, and based on the valve condition it is possible to schedule maintenance only when really necessary, prioritizing the most immediate needs. Rather than running until essential equipment fails, maintenance can be carried out when needed, and only when needed. Maintenance efforts can be focused on problem valves, not the healthy ones. Spare Parts and Flexibility The interoperability of HART, Foundation Fieldbus, and PROFIBUS-PA means that plants have many alternative third parties from which to obtain their replacement positioners and couplers. This ensures that spare prices will be market based, and not inflated, as was the case when proprietary protocols were used. To achieve maximum flexibility and enable the same positioner or coupler to be used with any pneumatic control valve or actuator in a plant, more and more functions are performed in software as opposed to hardware, making it easier to adjust and adapt to the application. Users can standardize on one positioner or discrete valve coupler for all different kinds of control valves, keeping spares and training to a minimum. Safety and Pollution Apart from the direct economical benefits, digital technology can also enhance safety and protect the environment. An

145 6.11 Fieldbus and Smart Valves 1189 important part about the actual position interlocks in Foundation Fieldbus is that they are handled by the PID and AO blocks, without the user having to configure separate logic for interlocks and reset windup protection on actual position, as was done in old DCSs. This is different from the old days, when some valves were connected to the DCS using 4 20 ma analog signals from the controller to the positioners. The DCS had no way of knowing the actual position of the valve. If the valve failed, nobody knew about it because there was no feedback or diagnostics. If somebody operated the valve by hand, nobody knew. This was very dangerous, because the operator was under the impression that automatic control was in operation, when it was not. Under these conditions, the controller can change its output, but the valve does not change, so the DCS still attempted to control even though there was no actuation. Also, if the output of the automatic controller was 50% (12 ma), while some technician in the field has been operating the valve by hand and opened it to 75%, when the loop was switched back to the automatic controller, the valve has to jump 25% (from 75 to 50%). To avoid this, some critical loops had a 4 20 ma feedback of actual valve position from the positioner to the DCS AI module. In the DCS, then, one had to program logic to detect the deviation and in this case put the PID in manual and make sure the PID output follows the actual position feedback from the valve. That is very complex logic, but for Foundation Fieldbus valves it is all automatically taken care of. Failure Detection Diagnostics from networked positioners and couplers detect failure in positioner, actuator, and valve including loss of supply air. Foundation Fieldbus communication will inform the operator of such events and will stop control. This allows for corrective action to be taken sooner, which might protect from dangerous situations and production loss. Thereby, Foundation Fieldbus makes systems safer. Another example is a thermocouple failure detected by a temperature transmitter propagated to the positioner, which in response shuts the loop down to a predetermined safe state without the need for any central control action. Another important form of diagnostics is partial stroke testing (Section 6.10). It verifies that the valve is not stuck, increasing the probability of a successful movement if called upon. This reduces the problem of valves stuck in one position. Digital communication in conjunction with asset management software can ensure that the installed base of digital valves is maintained well and is experiencing fewer surprise failures. Thus, asset management enables plants to run uninterrupted for longer periods of time, subsequently increasing productivity, keeping costs down. Diagnostics provide an early warning for abnormal conditions and may be used as an indication of problems to come, allowing technicians to solve problems before they adversely affect the process. Users can switch to a proactive maintenance program, thereby minimizing plant downtime. On-line Plant Asset Management An inadequately functioning control valve will upset the overall loop performance or may cause the valve to wear out prematurely. Conversely, a healthy positioner ensures better control and less maintenance. A fieldbus control valve positioner takes a total life cycle view of the valve and is designed to enable a longer life for the valve package. Positioners now come with built-in pressure sensors continuously monitoring the pressure at both the air supply input and the actuator chambers. This enhances simple text-based diagnostics such as loss of supply air with sophisticated chartand graph-based analysis, such as valve signature. Thus, the valve is provided with additional on-board sensors, which determine its condition, the ambient conditions, and such external factors as the loss of supply air, and gives this information to the microprocessor. The diagnostics is communicated to the on-line plant asset management tool using HART, Foundation Fieldbus, or PROFIBUS-PA communication. Asset management software takes the vast amount of diagnostics data in the positioner and turns it into information, which is useful to maintenance technicians. Together the asset management software and the positioners make it is easier to determine valve health and estimate their remaining life spans (Figure 6.11i). Valve Signature The Valve Signature plot traces the actuator pressure required to put the valve in a desired position (Figure 6.11j). As was shown in Section 6.8, comparing changes in this behavior over time helps identify problem areas. Other charts include Hysteresis (a.k.a. Positioner Signature or Dynamic Error Band ), step response (Section 6.9), and drive signal. It is a good idea to capture a base line signature for the valve package as it is new, and compare future signatures against this benchmark (Figures 6.8b and 6.10j). By using the above described techniques, maintenance can be done before the valves fail. DIGITAL VALVE INSTRUMENTATION The first generation of digital valve instrumentation was an electropneumatic device enhanced with a CPU and HART communication. Few other things changed. These devices still relied on the old lever and potentiometer for the actual valve position sensing for the feedback. They had the same old pneumatic gauges. High-current-consumption solenoids were used for flapper actuation. There was not much diagnostics to speak of. However, in the second generation the technology went beyond the mere incorporation of a CPU. Second Generation The microprocessor is at the center of every intelligent valve (Figure 6.11k). The firmware in digital valve instrumentation

146 1190 Control Valve Selection and Sizing FIG. 6.11i Fieldbus positioner diagnostics on asset management software. (Screenshot is courtesy of SMAR AssetView.) completely controls the valve. For example, instead of cam and spring mechanics on the feedback to apply flow characterization, these instruments do the characterization in software. Error due to mechanical imprecision is avoided, further improving performance. Likewise, split-range operation, digital display, and local operation are all provided using software. Position sensing is a key aspect of valve control and is an important means to acquire the data for sophisticated diagnostics and to tap the full potential of digital instrumentation.

147 6.11 Fieldbus and Smart Valves 1191 FIG. 6.11j A valve signature. (Screenshot is courtesy of SMAR AssetView.) See Section in Volume 1 of this handbook ( Linear and Angular Position Detection ) for details on position sensing. Noncontact position sensing technologies such as Hall-effect eliminates linkages that can stick and gall; they are also less sensitive to vibration, thus improving hysteresis and reliability. Accurate position sensing allows the valve to be able to respond to very small demands in position change as required for tight control. Power consumption is a key issue for digital valve instrumentation, and it is their only disadvantage as compared to analog instrumentation. For example, a traditional electropneumatic positioner has a voltage drop of only a few volts, Main circuit board Transducer board MAU & power supply Fieldbus modem PROM memory Local adjust CPU EEPROM RAM memory D/A converter Temp Piezo driver EEPROM memory A/D converter P_out1 P_out2 P_in Flapper Spool Valve Display Controller Display board Display FIG. 6.11k Block diagram of a Fieldbus pneumatic control valve positioner. (Courtesy of SMAR.) Hall Magnets

148 1192 Control Valve Selection and Sizing permitting the positioner to be driven from a regular controller output over very long wires. It is even possible to put two analog positioners in a split-range scheme in series without concern. However, smart positioners using HART typically have a voltage drop of 11 V or more and therefore Ohm s Law is putting a restriction on cable length, as controller open loop voltage may not be sufficient to drive the signal through the positioners plus cable resistance. Split-range applications generally require signal amplifiers. Current consumption is a concern in fieldbus installations where several devices are connected on the same network, causing voltage drops and restricting the number of devices on a safety barrier. Valve instrumentation that has I/P sections based on piezo technology typically have lower power consumption, making this less of a consideration. Additional auxiliary sensors can be included in digital valve instrumentation in order to provide inputs for more advanced diagnostics. For example, they might include internal ambient temperature sensing as well as built-in pressure sensors to continuously monitor the pressure at the air supply input and at the two pneumatic outputs. The on-board sensors are dedicated to providing continuous information to the microprocessor, so that it can determine the condition of the valve assembly and ambient conditions. CONCLUSIONS The market for digital valve positioners in general, and for fieldbus in particular, looks very promising and is perhaps the single most important event that has occurred in the control valve industry during the past several decades. Standards and flexible software have made it possible for users to choose third-party positioners when the positioners being offered by the control valve manufacturers cannot meet their performance requirement. With the right positioners it is very easy to upgrade an existing control valve to make it smart or fieldbus-compatible. Even when an old DCS system is converted into a fieldbus system, all the user needs to do is to remove the old 4 20 ma positioner and place a Fieldbuscompatible positioner in its place. References 1. Berge, J., Fieldbuses for Process Control Engineering, Operation and Maintenance, Research Triangle Park, NC: The Instrumentation, Systems and Automation Society (ISA), Merritt, R., Turned-On Valves: They re Using Embedded Intelligence and Digital Communication to Perform Data Acquisition, Diagnostics, and Control, Control Magazine, February 9, 2001.

149 6.12 Intelligent Valves, Positioners, Accessories H. D. BAUMANN (2005) I/P Control valve with electropneumatic positioner Flow sheet symbol Materials of Construction: Supply Pressure: Ambient Temperature: Inaccuracy: Positioner Input Signals: Costs: Partial List of Suppliers: Housing and cover: Low copper aluminum (stainless steel optional) 0.3 bar (5 PSIG) min., 10 bar (145 PSIG) max. 40 to 85 C ( 40 to 257 F). Min. dead band: 0.2% of span. Linearity: ±0.75% of span Analog: 4 20 ma, Digital: HART protocol or Fieldbus min. input current: 3.5 ma, load: 300 Ohms $ to $1, depending on functionality ABB Automation Products (Germany) ( ARCA-Regler (Germany) ( Dresser Valve Division (Masoneilan) ( Emerson Process Management (Fisher) ( Flowserve (Valtec) ( Foxboro/Invensys ( Metso Automation (Neles) ( Samson A.G. (Germany) ( Siemens A.G. (Germany) ( Yamataki (Japan) ( Young Tech Co. (Korea) ( Yokogawa (Japan) ( INTRODUCTION The reader is advised that various aspects of valve and positioner intelligence have also been discussed in the previous four sections, where control valve diagnostics (Section 6.8), dynamics (Section 6.9), safety (Section 6.10), and fieldbus interaction (Section 6.11) have been covered. Intelligent positioners are primarily used to control the position of the control valve stem in relation to the process control signal. They are also known as smart or digital positioners. In case of most designs, the digital designation is somewhat of a misnomer, because in most cases, it is only the signal processing portion of the positioner that is digital, while most of the input signals are still analog (4 20 ma) and most output signals are either electrical or pneumatic. Digital positioners arrived on the scene in the mid 1980s and thereafter rapidly replaced the former electronic, i.e., analog, positioners. Similarly to their analog predecessors, these intelligent devices also use air as their source of actuating power, and because of their enhanced electronics, they are about one third more expensive than their predecessors. Advantages of Intelligent Positioners Most intelligent positioners can be adapted to both sliding and rotary stem valves. On existing control valves, if built in conformance with IEC Standard , these intelligent positioners can replace pneumatic or analog electronic positioners of the same supplier. Conformance to a new 1193

150 1194 Control Valve Selection and Sizing Typical Performance Specifications FIG. 6.12a Typical digital positioner with explosionproof housing meeting FM and CSA approval requirements. (Courtesy of the Dresser Valve Group, Inc. Masoneilan.) German VDI/VDE Standard 3847 would even guarantee interchangeability between positioners made by different manufacturers. The control signals into the smart positioners can be analog (4 20 ma) or digital (via bus systems), among which Foundation Fieldbus predominates in the United States, while Prifibus is more commonly used in Europe. Similarly to electronic analog positioners, the digital units are also available in intrinsic safe designs or with explosionproof housings (see Figure 6.12a) and are protected from EMC interference. They typically meet U.S./CSA as well as CEN- ELEC (European) and CESI (Japanese) standards. In Figure 6.12a, one should note the presence of the explosionproof pushbuttons, which, in conjunction with the display window, enable the operator to perform on-site calibration and configuration. This is a feature that may not be available on all positioner models. In the illustrated design, calibration is done over the signal wires via HART protocol. Listed below are some of the advantages of digital positioners and other field devices relative to their analog counterparts: 1 Increased accuracy, 0.1 1% vs % for analog Improved stability, about 0.1% compared to 0.175% Wider rangeability (without sacrificing accuracy) up to 50:1 compared to 10:1 Capable of performing multifunctions Diagnostic and self-testing capability Ability for miniaturization Ability to be calibrated and adjusted without physical access to the instrument Utilize information that could not be measured or accessed by analog means Analog input signal: 4 20 ma DC (4 12 or ma split-range optional) Signal voltage for HART communication: 11 V Minimum current input: 3.5 ma Load: 300 Ohm Max. voltage: 30 V Power source: extra-low voltage (SELV) Reverse polarity protection Actions: Single-direct or single-reverse (double acting, optional) Hazardous area classifications: as required Temperature limits: C ( F) Typical pneumatic supply pressures: 0.3 bar (5 PSIG) to 8 bar (120 PSIG) Steady-state air consumption: 400 sl/hr (14 scfh) at 1.4 bar (20 PSIG) Max. output capacity: 10,600 sl/hr (375 scfh) at 1.4 bar (20 PSIG) Min. dead band: 0.2% of output span Independent linearity: ± 0.75% of output span. Generating the Pneumatic Output There are basically two ways to control the pneumatic output signal to the valve actuator. The first is to use a conventional I/P transducer coupled with an amplifying relay to increase the air flow to the valve actuator. The second is to generate the pneumatic output signal by pulsing either two piezoelectric valves, or two miniature solenoid valves (one for each stroke direction), and then amplify this output by a relay. The advantage of the pulsed system is that there is less air consumption, at least when the valve is not moving. Other advantages include that the valve travel time (time constant) can be changed quite easily by varying the pulse rate of the miniature valves. This could improve the stability of the valve and the control loop s performance. Another advantage is a more compact positioner size. A disadvantage is that these devices can have a larger dead band than their I/P-controlled cousins. This is necessary to keep the miniature valve from continuously pulsing if the signal is noisy. Unintended continuous pulsing can also happen, if there is a small air leak in the actuator or in the joints of the connecting air tubing. This could lead to an early failure of the on/off micro valves, which have a lifetime of about pulses. In a worst-case scenario, this lifetime limit could be reached in a little over a year, in case of sustained high frequency instability! Smart positioners offer a number of functional advantages, which are easily obtainable by digital processing and which could not be obtained from analog electronic positioners. Some of these features will be discussed in the following paragraphs.

151 6.12 Intelligent Valves, Positioners, Accessories 1195 Valve Performance Monitoring As was discussed in Section 6.8, the ability to monitor the performance of the control valve is a widely advertised feature of intelligent positioners. Valve performance is monitored by checking some of the basic valve calibration parameters, which include the zero position and the travel span of the valve. Additional tests can monitor the air pressure in the actuator against travel. This signature is compared against data obtained when the valve was newly installed. A major deviation from the desired characteristic could indicate potential problems such as the valve stuffing box being too tight, the valve stem being corroded, or the actuator spring being damaged. When such conditions evolve, they can cause an increase in the dead band and the dead time of the valve, thereby creating a potential for instability and cycling of the control loop. This test data can be called up via HART protocol and can be a vital part of the plant s asset management system (also discussed in Section 6.11). Some believe such data could also be transmitted to a computer via a fieldbus digital signal while the valve is actively controlling the process. Others argue that in order to measure the valve s performance, one has to provide an artificial offset to the control signal. Therefore, some plant operators are understandably reluctant to utilize this feature, because they fear that this offset to the control signal would create too much disturbance in the control loop. As a result, one typically has to wait until the valve is out of service (process is shut down) in order to perform functionality tests. It also has to be realized that continuous monitoring of valve performances can create an information overload for the engineer or technician who may be responsible for up to 600 control loops! It therefore may make more sense to monitor the quality of the control signal (against preset limit values) at the controller. Any significant deviation could then be used to trigger a check of the control valve and its positioner. Otherwise, such call-up of the valve s performance data, perhaps with a handheld HART, has to wait until it is time for the regular scheduled maintenance. Future Trends and Tasks Future uses of the intelligent positioner as data transmitter will no longer be restricted to the monitoring of the condition of the valve actuator but may include 1. Monitoring the leakage of the valve s packing box or stem bellows using the binary output of suitably placed pressure sensors, or by sniffing the ambient air around the valve stem to detect toxic substances by means of a miniaturized chemical analyzer. 2. Checking for fluid leakage between plug and seat by feeding the signal of a sound pressure transducer the sound frequency within the valve housing against a known frequency profile to the positioner. 3. Another method of indirectly discovering excessive seat leakage is by comparing the controller output signal at low flow conditions with the same signal, which was measured for the same low flow rate when the valve was new. Excessive leakage would result in such a signal change, which will show less travel (less opening) to compensate for the increase in leakage. An important feature of all positioners is the time they take to either evacuate all the air from the actuator (if the valve is air-to-open ), or to deliver the maximum air pressure to the actuator (if the valve is air-to-close ) whenever the control signal calls for the valve to be closed. When this occurs, in order to avoid seat leakage, the positioner output signal must change rapidly to apply the maximum actuator force onto the valve plug. Some digital positioners offer memory-embedded data such as serial number, date purchased, vendor, tag number, and valve size. These are all useful for asset management purposes and especially handy when the original valve serial plate is corroded and has become unreadable. Future intelligent positioners may be able to receive their control signal via radio transmission. The difficulties with such a schema include reliability, path loss, RF interference, multipath radio echoes, and the required transmission power. Most of all, these positioner will still require sufficient electrical power to operate their pneumatic and electronic components. Consequently, the valve positioners will never be truly wireless. CONTROLLING THE PROCESS Embedding the control algorithm (usually PID) inside the digital heart of the positioner transforms the control valve package into a dominant part of a distributed control system. Such a smart control valve requires only an input from a process sensor/transmitter and a supervisory (adaptive, feedforward, and so on)or set point signal from a central computer. Such imbedded algorithms initially tend to involve PID control only but may later include more advanced features to serve self-tuning or fuzzy logic functions. Changing the Valve s Characteristics Another common feature of digital positioners is their ability to modify the relationship between the controller s output signal and the pneumatic output signal to the valve actuator. The effect of this relationship is to alter the valve s inherent flow characteristic. Choices include quick opening, linear, equal percentage (Figure 6.7b), and a variety of userdefined characteristics that must be custom programmed. Such a modification of the valve s inherent characteristic is similar to placing a mechanical cam in the travel feedback linkage of a pneumatic positioner. For example, it can convert

152 1196 Control Valve Selection and Sizing the change in flow through the valve becomes proportional to the change in the controller s output signal. This way one would have not only a constant control valve gain but also a constant control valve time constant! Another problem caused by the changing of the flow characteristic has to do with the dead band of the valve. For example, if one changes the inherent linear valve characteristic to equal percentage, the valve travel for say the first 10% of signal range (0 10%) will be very small (typically 0 3%). This means that the apparent dead band of the valve stem will increase three times, and this can cause oscillation in the control loop. OPERATION OF SMART POSITIONERS FIG. 6.12b Typical intrinsically safe digital positioner that can provide singleor double-acting pneumatic output (note pneumatic pressure gauges). (Courtesy EMERSON Process Management Fisher.) the static flow characteristic of a butterfly valve, which has a nearly linear characteristic (C v vs. travel is linear), into an equalpercentage one (C v varies exponentionally with controller signal). The trouble is that the travel vs. flow coefficient (C v ) of a butterfly valve is still linear. The result is that the signal resolution at low flow becomes very poor, because with an equal percentage output signal at low flows, the valve travel is very small per given signal change. Therefore, in such cases the valve s dead band becomes high compared to its signal resolution. Valve Gain and Time Constant The purpose of changing the flow characteristic is to make the gain (rate of flow change vs. signal change) of the control valve such that the loop gain will be more constant, in order to aid in maintaining the stability of the loop control (Figure 6.7a). This can be accomplished, at least in theory, by the above-described modification of the control signal. In practice, this approach works only in applications where the time constants of the valve and of the process differ by at least a factor of five. The reason is that the static signal modification performed by the positioner results in vastly different stem travel speed requirements at low and at high flow rates. This in turn results in a continuous variation of the valve s time constant, 3 which makes controller tuning very difficult. It is hoped that future positioner designs will provide algorithms that will change the rate of output signal change in proportion to the rate of signal modification. This would mean that the travel speed of the actuator is changed so that From the outside there appears to be little difference between smart digital and analog electronic positioners (Figure 6.12b), other than a digital monitoring window on the outside of some smart positioners. The real difference lies inside. As can be seen in Figure 6.12c, all communication as well as power supply travel through a single pair of wires from the controller or computer to the printed wiring board containing all the functional electronics. This signal may be in the form of a 4 20 ma analog, or in a digital form according to the selected fieldbus protocol. The controlling electronic portion is usually encapsulated and field-replaceable. This printed wiring board receives feedback from a motion transmitter, which is sensing the position of the valve stem. Any deviation from the desired valve stem position will be detected, and a correcting signal is sent (in this case) to an I/P transducer that converts the electronic signal into a Auxiliary terminals Drive signal I/P converter Terminal box Printed wiring board Pneumatic relay 4 20 ma Input signal + Hart Output A Valve travel feedback Supply pressure Output B Valve and actuator FIG. 6.12c Schematic view of a digital valve positioner that is sending an electronic analog output signal into an I/P transducer, which in turn is generating a corresponding pneumatic output signal.

153 6.12 Intelligent Valves, Positioners, Accessories 1197 FIG. 6.12d Digital valve postioner with cover removed. It shows the adjustment knobs for zero, for span (travel), and for the type of action as direct: air to open or reverse air to close. Finally, there is a setting for the low voltage internal limit switch. (Courtesy of SAMSON A.G.) corresponding pneumatic one. This air signal is then amplified by a relay and fed to the valve s actuator. Maintenance and Calibration An additional pressure sensor (measuring the output of the air relay) can give additional feedback to the circuit board in order to aid in maintaining the stability of the loop. This pressure sensor also monitors the valve s performance by comparing the relationship between the air signal and travel with a stored relationship that was taken when the valve was new. As is illustrated in Figure 6.8c and discussed in Section 6.8, a major difference between the two relationships signals the need for maintenance. After the positioner is mounted on the valve, its initial calibration can be done by a remote device such as a computer via a HART protocol, or alternatively by using internal adjustment knobs as shown in Figure 6.12d. ACCESSORIES The most common built-in accessories available as options in intelligent positioners are micro-switches, which typically operate on low-voltage levels (24 V). These switches provide the capability of a safety shutdown of the valve, which is initiated independently from the control signal. Other accessory options include analog or digital valve travel position transmitters. In addition one also requires software, which is needed for the intelligent positioner to perform the desired diagnostic and remote calibration functions. Listed below are some of the capabilities that such a software package should support: Reading of valve diagnostic data Supporting the graphical diagnostic trends window displays Valve configuration, and valve configuration database Monitoring device variables Configuration of valve characteristic General process and valve database Security levels for users Diagnostic database FLOW CONTROL BY SMART VALVE It is possible to integrate a whole flow control loop into a control valve that is a self-contained part of a distributed control system. Such a device would include the valve, the flow sensor, the valve positioner. and the controller, all incorporated into a single entity. 4 As shown schematically in Figure 6.12e, it includes a temperature sensor and a flow detector, which consists of two piezoelectric pressure sensors embedded within the valve housing and located up- and downstream of the valve orifice. The mass flow through the valve is measured by detecting the differential pressure across the valve orifice and by applying a density correction to that differential on the basis of the detected temperature. This product is then multiplied by the known area of the vena contracta between orifice and valve plug at the prevailing inner valve travel position, and thereby

154 1198 Control Valve Selection and Sizing P 1 P 2 T1 To process m = Aur Desired mass flow Supervisory computer A = N 3 C v φ(ft 2 ); U = N 2 P 1 P 2 ( ft/s); r = N 1 γp 1 P 2 /T 1 (lbs/ft 3 ) A = Area of valve opening C v = Valve capacity coefficient m = Mass flow, lbs/s N 1, N 2, N 3 = Numerical constants P 1 = Absolute inlet pressure, lbs/ft 2 P 2 = Absolute outlet pressure, lbs/ft 2 FIG. 6.12e Schematic view of a complete flow control loop built around an intelligent control valve. 4 T 1 = Inlet temperature, R U = Orifical velocity, ft/s g = Specific weight, lbs/ft 70 F and 14.7 psia f = Fractional valve travel r = Density of flowing gas (lb/ft 3 ) the on-board computer calculates the mass flow passing through the valve. Flow control is provided by a PID algorithm that compares the measured flow with a desired set point and adjusts the valve opening as needed to meet that set point. Limitations In order to make this type of flow measurement reasonably acceptable, one must locate the pressure sensors in a long and narrow channel in the body cavity to eliminate the effects of flow turbulence disturbances. While not an accurate flow sensor, it can detect a deviation from a desired flow rate and correct the flowing quantity. Such a device can be used in temperature cascade loops, where this intelligent valve may control the flow rate of steam, while a temperature controlling cascade master adjusts the set point of this intelligent valve-based flow-controlling slave. Such an intelligent valve provides ease of installation, space savings, and the elimination of extra wiring. While the price is higher than that of a conventional control valve, it is less than the combined cost of the components of a mass flow control loop. References 1. Schneider, H.-J., Digital Feldgeraete: Vorteile, Probleme und Anforderungen aus Anwendersicht, ATP, Automatisierungstechnische Praxis, Germany, April 1995), pp Harris, J. Assessment of Control Loop Performance, Canadian Journal of Chemical Engineering, Vol. 67, pp , Baumann, H. D., Positioners Set the Course for Modern Control Valve Designs, I&CA, May 1997, pp Baumann, H. D., Control Valve Primer, 3rd edition, Research Triangle Park, NC: Instrumentation, Systems, and Automation Society, 1998, pp Bibliography Ali, R., Smart Positioners to the Rescue, InTech, June 2002, pp Ali, R. and Jero, L., Smart Positioners In Safety Instrumented Systems, Petroleum Technology Quarterly, Winter 2002/2003. Baumann, H. D., Interkama 1999: The Status and Future of Control Valves: A Global Review, ATP Automatesierungstechnische Praxis, Germany, Vol. 42, No. 3, pp , DeGaspari, J., A Tale of Two Buses, Mechanical Engineering, September Emerson Process Management, newproduct/newliter/pdfacilities.html. Kiesbauer, J., Detektion Der Inneren Leckage Bei Stellgeraeten, ATP, Automatesierungstechnische Praxis, Germany, Vol. 42, Heft 11, 2000, pp Kiesbauer, J. and Hoffmann, H., Anlagennahes Asset Management Bei Stellgeraeten-Eine Standortbestimmung, ATP, Automatisierungstechnische Praxis, Germany, Vol. 7, July 2002, pp Miller, L., Valve Diagnostic Past, Present, and Future, Fluid Handling Systems, November Render, N., Digitally Mastered, Control & Instrumentation, November Shinskey, F. G., How Good Are Our Controllers in Absolute Performance and Robustness? Measurement and Control, Vol. 23, No. 5, pp , 1990.

155 6.13 Miscellaneous Valve and Trim Designs C. S. BEARD (1970, 1985) B. G. LIPTÁK (1995) FY X J. P. WILSON (2005) Expansible tube or diaphragm valves Control pilot line Fluid interaction valves Flow sheet symbols Valve and Trim Designs: Available Range of Sizes: Design Pressure Limits: Maximum Operating Temperatures: A. Anticavitation Valves B. Dirty Service Valves C. Low-Noise Valves D. High-Capacity Valves E. Cryogenic Valves F. High-Temperature Valves G. Steam Conditioning Valves H. Tank Mounted Valves I. Expansible Tube or Diaphragm Valve J. Fluid Interaction Valves A. 1 to 36 in. (25 to 915 mm) B. 1 to 36 in. (25 to 915 mm) C. 1 to 48 in. (25 to 1220 mm) D. 12 to 72 in. (300 to 1830 mm) E. 1 to 36 in. (25 to 915 mm) F. 1 to 20 in. (25 to 508 mm) G. 6 to 48 in. (152 to 1220 mm) H. 4 to 24 in. (100 to 600 mm) I. 1 to 12 in. (25 to 300 mm) J. 1 / 2 to 4 in. (12.5 to 100 mm) A. Up to 8000 PSIG (550 bar) B. Up to 8000 PSIG (550 bar) C. Up to 4000 PSIG (275 bar) D. Up to 3000 PSIG (205 bar) E. Up to 1500 PSIG (100 bar) F. Up to 5000 PSIG (345 bar) G. Up to 5000 PSIG (345 bar) H. Up to 1500 PSIG (100 bar) I. Up to 1500 PSIG (100 bar) J. 100 PSIG (7 bar) or greater (no theoretical limit) A. Up to 1100 F (595 C) B. Up to 1100 F (595 C) C. Up to 1100 F (595 C) D. Up to 1100 F (595 C) E. 325 F ( 162 C) to 300 F (150 C) F. Up to 1800 F (980 C) G. Up to 1100 F (595 C) 1199

156 1200 Control Valve Selection and Sizing H. Up to 1100 F (595 C) I. Up to 150 F (66 C) J. Can handle molten metals Partial List of Suppliers: ABB Process Automation ( Emerson Process Management Daniel Division ( Emerson Process Management Fisher Valve Division ( Flowserve Corporation ( Masoneilan Division of Dresser Flow Control ( Siemens (www. siemens.com) INTRODUCTION This section consists of two distinct parts. In the first part, a number of control valve designs are described, which are different in that they are neither linear nor rotary in their operation. These miscellaneous valve designs utilize the energy content of the flowing fluid for their operation. They depend on fluid interaction or the static pressure of the process fluid and use flexible elements to throttle gas pressure regulators and other control valves. These valves are used in specialized services, such as in sensitive level control, in gas pressure or flow regulation, and in toxic services. In the second part of this section some of the special valve trim designs used in the traditional globe control valves are discussed in terms of their suitability for specialized applications. These applications include applications where noise; cavitation; flashing; high and low temperatures or pressures; or viscous, dirty, or slurry flows are involved. Such trim designs and applications have already been discussed in Sections 6.1, 6.7, 6.14, and MISCELLANEOUS VALVE DESIGNS Dynamically Balanced Plug Valves This family of control valves has been developed for pressure and flow control applications where no external power is available to operate the valve, and therefore the static pressure of the process fluid is utilized to achieve throttling. Figure 6.13a illustrates an installation for upstream (back) pressure control. Here, the upstream pressure is sent to the control pilot through port D. If the controlled upstream pressure drops, this lowers the pressure in the pilot chamber and the pilot spring moves the pilot poppet valve to the right. This opens port B and thereby upstream pressure is applied inside the plug chamber. This equalizes the pressure acting on the plug and allows the spring to move the plug to the left, to close the valve. These valves provide nearly linear characteristics (Figure 6.13b) and high flow capacities (C v = 30d 2 ). They are available in sizes up to 12 in. (300 mm), and special units D Controlled pressure E D A B C Hard faced plug A B C Pilot valve poppet assembly Plug guiding provides stable and consistent seating. Seat with tight-shut-off FIG. 6.13a Dynamically balanced valve plug. (Courtesy of Daniel Industries.) Plug chamber with low-force spring

157 6.13 Miscellaneous Valve and Trim Designs Closing port Opening port % Flow or Cv Expansible tube Dynamically balanced plug Plug % Stroke Cylinder Piston FIG. 6.13c Positioned plug valve in its fully open position. (Courtesy of Eisenwerk Heinrich Schilling.) FIG. 6.13b The characteristics of expansible tube and dynamically balanced plug valves are nearly linear. have been made in up to 24 in. (600 mm) sizes. The valves can be provided with up to ANSI Class 1500 rating and with pressure control settings up to 3000 PSIG (205 bar). The advantages of this design include its erosion and corrosion resistance, due to the hard facing of the plug; its good pressure recovery characteristics, due to the large and smooth annular flow area through the valve; its fast speed of response; and its bubble-tight seating. Positioned Plug In-Line Valves The positioned plug in-line valve, excluding control units, resembles a pipe spool. It is only necessary to inject pressure into its ports for positioning the valve plug. This simple design (Figure 6.13c) requires only three pressure seals. The plug is carried on a cylinder that also includes the piston. Pressure in one port causes closing, while the opposite port is used for opening. The valve has only one moving part and is available in sizes from 2 8 in. ( mm) for use to 350 PSIG at 400 F (2.4 MPa at 204 C). Control quality is dependent upon the pilot valves and auxiliary units employed. Another in-line valve available in small sizes (Figure 6.13c) carries the valve plug on a bridge in the operating cylinder, with the seat as part of a split body. A spring-loaded version of this design uses the beveled end of the moving cylinder to seat on a replaceable soft seat, retained in a dam, held in position by struts from the inside wall of the valve body. The spring loading may cause failclose or fail-open actions, as illustrated in Figure 6.13e. The unit can be powered with line fluid or by an external pressure source. A double-bleed feature can be incorporated to eliminate the possibility of actuation and line fluids combining if one of the dynamic seals should fail. All seats and seals are replaceable by separation at the body flange. The unit is particularly adaptable to fluids that are toxic or difficult to contain, such as nitrogen tetroxide, hydrogen, and others used in the aerospace industry. The unit is furnished in sizes from 1 1 / 2 18 in. ( mm), with ratings to 2500 PSIG (17.3 MPa). All types of end connections are available, and control is dependent upon the auxiliary control components selected for the application. An explosionproof limit switch can be furnished for position indication. Diaphragm-Operated Cylinder In-Line Valves An in-line valve using a low convolution diaphragm for positive sealing and long travel (Figure 6.13f) is designed particularly for gas regulation. The low level of vibration, turbulence, and noise of this in-line design makes it suitable for high-pressure gas service. Opening port Closing port FIG. 6.13d Positioned plug valve in closed position. (Courtesy of Control Air Inc.)

158 1202 Control Valve Selection and Sizing Opening port Closing port Opening port Closing port Fail closed design Fail open design FIG. 6.13e Spring-loaded positioned plug valve. Inlet pressures to 1400 PSIG (9.7 MPa) and outlet pressures to 600 PSIG (4 MPa) are possible in the 2 in. (50 mm) size. It is a high-capacity valve, as expressed by C v = 23 d 2. As a gas regulator the unit is supplied with a two-stage pilot to accept full line pressure. This pilot resists freeze-up and serves as a differential limiting valve. All portions of the pilot and line valve will withstand a full body rating of 600 PSIG (4 MPa). Expansible Valve Designs The expansible element in these valve designs can be a rubber cylinder, an expansible tube, or an expansible diaphragm. The common feature of all of these designs is that they utilize the process pressure to provide tight shut-off of gas flows. Expansible Element In-Line Valves Streamlined flow of gas occurs in a valve in which a solid rubber cylinder is expanded or contracted to change the area of an annular space (Figure 6.13g). A stationary inlet nose and discharge bullet allow hydraulic pressure to force a slave cylinder against the rubber cylinder to vary its expansion. Control is from a diaphragm actuator, with the diaphragm plate carrying a piston. The piston acts as a pump to supply hydraulic pressure to the slave cylinder. The rubber cylinder offers the seating ability of a soft seat valve. It has the capability of closing over foreign matter, and Rubber cylinder Hydraulic fluid Inlet pressure Pilot supply pressure Loading pressure Outlet pressure FIG. 6.13f Diaphragm-operated cylinder-type in-line valve used in high-pressure gas regulation services. FIG. 6.13g A rubber cylinder provides the soft seat in this low-noise, expandable element, pressure regulator in-line valve.

159 6.13 Miscellaneous Valve and Trim Designs 1203 Outlet slots Inlet slots Closed Throttling Open FIG. 6.13h An expansible tube valve utilizes the process pressure for its operation, while being controlled by a three-way pilot that determines if the valve is to be closed (left), throttling (center), or open (right). the design allows for the use of a restricted throat for reduced capacity. With this design, pressure drops as high as 1200 psid (8.3 MPa) have been handled with a low noise level. The valve may be utilized as a pressure reducer or for back-pressure control, depending upon the system requirements. Available sizes are 1 6 in. ( mm). The 1 in. (25 mm) valve can have screwed connections, while all sizes can be flanged. The body is steel with flange ratings to 600 PSIG (4 MPa). A valve positioner can be used, if the stem position is calibrated as a function of the annular space reduction. If such calibration is provided, a controller output can throttle the valve to obtain accurate flow control. Expansible Tube Valves Control of flow is obtained by use of an expansible tube that is slipped over a cylindrical metal core containing a series of longitudinal slots at each end and a separating barrier in between. The characteristics of such a valve were shown in Figure 6.13b. A cylindrical, in-line jacket surrounds the tube so that the process pressure can be introduced between the jacket and the sleeve to cause the sleeve to envelope the slots. This valve will open if the space between the jacket and the sleeve is connected to the downstream (lower) pressure and will open if that space in connected to the upstream (higher) pressure. With pressure connected to the downstream line (Figure 6.13h, right), the line pressure in the valve body will cause the valve to open fully. Control of the pressure on the sleeve creates a throttled flow condition by first uncovering the inlet slots and then progressively opening the outlet slots (Figure 6.13h, center). A continuous dynamic balance between fluid pressure on each side of the sleeve makes it possible to obtain wide rangeability between a no-flow and a full flow (fully open) condition. The basic operation of the valve can be accomplished with a three-way pilot valve positioned from a remote location. A variety of automatic pilots give versatility to the basic valve. For reduced pressure control, a pilot is used to modulate the jacket pressure in response to the sensed pressure in the downstream pipeline. As downstream pressure falls below the set point, the double-acting pilot positions itself to reduce the jacket pressure. This allows the valve to open to a throttling position. Therefore, downstream pressure increases to the set pressure, with attendant change in flow rate to maintain the set pressure. The static sensing line is separate from the pilot discharge line, in order to eliminate the pressure drop effect in the sensing line. Another form simulates a conventional regulator, in that system gas is bled into the jacket annular space through a fixed orifice and bled off through the pilot regulator. In this form, the static sensing line and pilot output are common. Doubleacting pilot systems use seven control ranges from 2 to 1200 PSIG ( MPa) with corresponding inlet pressures up to 1500 PSIG (10.4 MPa). The fixed orifice design is available for control from 2 oz 600 PSIG (0.86 kpa 4 MPa). Pilot Design Variations Back-pressure control and pressure relief are obtained in the same manner as pressure reduction is, except they sense the upstream pressure. By using a separate sensing and bleed port, a build-up from cracking to fully open can be varied from 3 to 14% of the set pressure. Return to normal operation causes the valve to create absolute shut-off. Emergency shut-off service may use an external pressure source piloted to obtain immediate shut-off upon abnormal conditions. A diaphragm-operated, three-way slide valve may also control jacket pressure by proportioning the inlet and outlet pressures. In this design, the controller output signal is sent to the diaphragm actuator of the three-way slide valve, which causes the sleeve valve to open proportionally, in a manner similar to that of a conventional diaphragm control valve. In this manner, the valve becomes a throttling control valve (Figure 6.13i). Differential pressure or flow control is accomplished by using a pilot valve in which the diaphragm is positioned by both upstream and downstream pressures. The differential

160 1204 Control Valve Selection and Sizing FIG. 6.13i The expansible tube valve becomes a throttling control valve if it is provided with a three-way slide-valve-type control pilot. static lines may be taken at the inlet and outlet of the line valve or, for flow control, across an orifice plate in the line. Pressure Boosting A practically unlimited variety of pressure and flow regulators can be configured with the range of pilots that are available. One important application is in pressure boosting in gas distribution systems. By these boosters, the line pressure losses due to increased consumption can be counteracted automatically by increasing the set pressure of the distribution control valve. In such a design, there can be as many as three pilots controlling the gas distribution control valve. In Figure 6.13j, Variable orifice Control signal Normal pressure control (pilot # 1) Boost pressure control (opens at high flow) Boost pressure control (pilot # 2) Flow element FIG. 6.13j Expansible tube-type gas distribution control valve can automatically boost the distribution pressure, when the demand for gas increases. the normal pressure is controlled by pilot #1. An increase in flow is sensed by the flow element (orifice, or flow tube), which opens a high-differential pilot to cut in a boost-pressure-control regulator (pilot #2), which has a higher set point than pilot #1. Return to normal flow cuts out the boost pressure control regulator and reinstates the normal control pilot. The expansible tube valve is made in sizes from 1 12 in. ( mm) with pressure rating from 200 PSIG (1.4 MPa) in iron to 1500 PSIG (10.3 MPa) in steel construction. It is made flanged or flangeless for insertion between line flanges. The flangeless body is cradled in the studs between the line flanges. Removal of the body is made easier by expanding the flanges about 1 / 8 in. (3 mm) using nuts on the studs inside the flanges. Tight shut-off or throttling requires a differential between the line pressure or external source used for closure and the downstream pressure sensed on the inside surface of the sleeve that is exposed to this pressure. This differential requirement for a special low-pressure 2 in. (50 mm) valve is 3.6 psid (24.8 kpa) and is 1.6 psid (11 kpa) for a 4 in. (100 mm) valve. The low-pressure series requires from 21 psid (145 kpa) for the 1 in. (25 mm) size to 4.6 psid (32 kpa) for the 10 and 12 in. sizes (250 and 300 mm). High-pressure models require from 58 psid (400 kpa) for the 1 in. (25 mm) size to 11 psid (76 kpa) for the 10 and 12 in. sizes (250 and 300 mm). The body design allows tight shut-off, even with comparatively large particles in the flow stream. Freezing of the pilot by hydrates is not common because the intermittent and small bleed occurs only to open the valve and as such is not conducive to freezing. The pilot may be heated or housed, or even located in a protected area close to the warm line. With its only moving part a flexible sleeve, this type of valve has no vibration to contribute to noise. The flow pattern also helps make this valve from 5 to 30 db more silent than most regulators. Flow capacities are comparable to those of single-seated as well as many double-seated regulators. Expansible Diaphragm Design In an expansible diaphragmtype regulator, an expansible element (Figure 6.13k) is stretched down over a dome-shaped grid causing shut-off of the valve when pressure above this resilient member overcomes the line pressure under the element. Line pressure is evenly directed over the expansible area by a series of pressure channels. Selection of the correct action on a pilot that supplies line or external pressure to the exterior of the expansible element causes the valve to control as a back-pressure or as a reducing control valve. For back-pressure control, or pressure relief, the static line is taken upstream of the valve. Increase in line pressure will increase the bleed from the annular space between the expansible element and the metal housing. Reducing regulation is accomplished by restricting the bleed upon increase in downstream pressure and increasing it upon decrease in downstream pressure.

161 6.13 Miscellaneous Valve and Trim Designs 1205 Splitter P L P R Closed Control ports closed FIG. 6.13l The Coanda effect is used in the flip-flop diversion of flow in the fluid interaction valves. FIG. 6.13k The expansible diaphragm valve can be used as either a pressure regulator or as a back-pressure relief valve. The valve is available in iron or steel with ratings to 600 PSIG (4 MPa). Models are available from F ( C), using a molded, Buna-N diaphragm. Relief valve pressures are from PSIG ( MPa), while reducing service varies from PSIG ( MPa). Capacity factors vary from C v = 11 d 2 in smaller sizes to C v = 14 d 2 in larger valves. Fluid Interaction Valves Wide open The Coanda effect, the basis of fluidics, is used in diverting valves from 1 / 2 4 in. sizes ( mm). The Coanda effect means the attachment of a fluid stream to a nearby side-wall of a flow passage. This effect can be used in a so-called flipflop valve for diverting a stream from one discharge port to another. Figure 6.13l shows the flow through the right-hand port due to both control ports being closed. Opening the right-hand port, to allow air or liquid to enter, will shift the flow. The industrial valve has rectangular diverting tubes, but the end connections may be circular. Control is maintained by opening or closing the control port or by injecting low-pressure air or liquid through a solenoid or other pilot valve. In this valve, with stream flowing in one diversion tube, the flow at the inlet contains some potential (pressure) energy and some kinetic (flowing) energy. Much of the potential energy is converted to kinetic energy at the nozzle. Up to 70% recovery of potential energy occurs in a diffuser section. Fifty percent recovery is guaranteed for commercial valves, and somewhat less for gases above critical flow and for viscous fluids ( P ~ P 1 /2). Installation must allow the recovered pressure to create the desired flow against friction effects of piping or fittings. An uninhibited flow will create some aspirating effects in the open outlet; restriction causes blocking, while excessive restriction will cause a leak or diversion to the open port. Fast Level Control Industrial valves, with their ability to divert in less than 100 ms, fill a wide variety of uses. The primary one is for level control in which the effluent not required for filling may be returned to storage. It is necessary only to use a dip tube set at the control point, as shown in Figure 6.13m. Lack of moving parts or of detrimental effects due to fast diversion action allows the system to provide close control. Valve inlet Dip tube Converging walls Side walls Control port Level fluctuates about end of dip tube FIG. 6.13m The response of level control with fluid interaction diverting valves can be very fast.

162 1206 Control Valve Selection and Sizing Numerous uses of diversion valves exist, such as tank filling, which is accomplished by using an external signal. Diversion of a process stream upon contamination, sensed by a ph or other analyzer, is important in paper mills and in chemical plants. The ability to divert rapidly makes this valve applicable to oscillating flows. The valve may be used for space conditioning, if the total bypassing of a heating or cooling medium is adequate for space temperature control. Fluidic valves can also be used for the diversion of engine exhaust gases from tailpipe propulsion nozzles to wingmounted lift fans with ambient control flow. Other unique applications include a four-ported fluidic valve, which has been used for direction control of a missile. It seems that all the potential uses of this valve design have not been exploited yet. SPECIAL VALVE APPLICATION In the following paragraphs, the capabilities of one control valve manufacturer (Fisher Controls) are discussed in handling such demanding services as cavitation and sludge. This is somewhat redundant, because Section 6.1 has already covered all the special control valve applications. These included the products of all valve manufacturers and covered noise, cavitation, flashing, high and low temperature or pressure services, and viscous, dirty, or slurry flows. Yet, the editor believes that this manufacturer s perspective that follows is still a valuable addition to the overall picture. The reader is reminded that Sections 6.14 and 6.15 provide in-depth information on control valve noise and on valve sizing aspects of the various special applications. For these reasons, the reader is advised to refer to the abovementioned sections for an in-depth treatment of the subject matters that are discussed below. Cavitation and Flashing As it was already discussed in Section 6.1, a fluid will flash when the downstream pressure is below the vapor pressure of the flowing process fluid. The vapor bubbles that form when the pressure falls below the vapor pressure continue to grow as long as the pressure keeps dropping, and eventually the liquid changes or flashes to a vapor. In connection with Figure 6.1w, cavitation has been also explained as it was also shown how vapor bubbles can be formed at the vena contracta and how these bubbles can implode and release powerful microjets that will damage any metallic surface, as the pressure rises downstream. In addition, the options available to the process control engineer to eliminate cavitation were also shown in connection with Figure 6.1x. There are a number of ways to protect against cavitation. As was shown in Figure 6.1aa, one of the most common is to expose the fluid to a series of restrictions as opposed to a single restriction. Each subsequent restriction dissipates a certain amount of the available energy and reduces the inlet pressure to the next stage. As was shown in Figure 6.1x, a well-designed pressurestaging device prevents cavitation by taking a large pressure differential and by maintaining the vena contracta pressure above the vapor pressure of the liquid, which prevents the liquid from cavitating. The expanding flow area concept of damage control is closely related to the pressure drop staging approach (Figure 6.13n). Figure 6.13n shows a pressure vs. distance curve for flow through a series of fixed restrictions where the area of each succeeding restriction is larger than the previous. Notice that the first restriction takes the bulk of the pressure drop, and the pressure drop through successive sections decreases. In the last restriction, where cavitation is most likely to occur, the pressure drop is only a small percentage of the total drop, and the pressure recovery is substantially lower. The expanding flow area concept requires fewer pressure drop stages to provide the same cavitation protection as does the concept that utilizes nearly identical areas for staging. The most common approach to cavitation protection employs a drilled-hole cage that incorporates both pressure staging and expanding flow area concepts. Each drilled hole has a significant impact on the overall pressure recovery of the valve. Drilled Trim and Multistage Designs Figure 6.13o shows the cross-section of three types of drilled holes that could be used in an anticavitation cage. The thin plate design is a very inefficient flow device, but it does provide a high F 2 L value and therefore a low pressure recovery. The thick plate design is not only more efficient, but it also provides a high pressure recovery as denoted by its low F 2 L value. The Cavitrol trim hole, designed by Fisher Controls, is a balance between the thick plate and the thin plate hole designs. It provides relatively high flow efficiency while still 2 maintaining a high F L value, which results in a low pressure recovery. This design represents the optimal choice between capacity and cavitation control. Figure 6.13p shows the cross-section of a three-stage, anticavitation trim. This particular design prevents the formation of damaging cavitation at pressure drops up to 3000 psid (207 bar) by utilizing a unique expanding flow area design, meaning that each stage has successively larger flow area. When a series of drilled holes are used to control cavitation, it is also easy to characterize the trim. In the trim illustrated in Figure 6.13p, as the valve plug travels through the cage, the cage design changes. It begins as a pressure-staging device and transitions to a straight-through, low-restriction flow design. Consequently, the cavitation control ability of

163 6.13 Miscellaneous Valve and Trim Designs 1207 P 1 Equal drop through six stages Pressure Cavitrol 4 trim Inlet pressure to final stage P 2 P V FIG. 6.13n Comparison of staged pressure drops to prevent cavitation. (Courtesy of Fisher Controls.) P VC Fluid travel through the valve stages Thin plate Thick plate Cavitrol hole Low C V High F L 2 High C V Low F L 2 High C V High F L 2 FIG. 6.13o Comparing various drilled-hole-type anticavitation trim designs. (Courtesy of Fisher Controls.) FIG. 6.13p The cross-section of a three-stage anticavitation trim. (Courtesy of Fisher Controls.)

164 1208 Control Valve Selection and Sizing this trim design is the greatest at low travels and decreases as the valve plug travel increases. In Figure 6.1y, a large number of anticavitation valve designs are shown. One disadvantage of most anticavitation trims is their potential for plugging. The flowing media often contains small particulate matter (e.g., sand) that can plug the passages, restricting or totally stopping flow through the valve. If this potential exists, the particles must be removed from the flow stream (e.g., by filtration), or an alternative approach to cavitation prevention should be taken. In situations where it is impossible to remove the particulates via filtration or separation, valve designs that can pass the particulate and still resist cavitation and erosion should be considered. The next paragraph describes some of the control valves that can be considered for sludge or slurry services. Dirty Process Services In Section 6.1, Figure 6.1ee shows a number of control valve designs that can be considered for sludge and slurry services. These applications involving entrained particulates are some of the most challenging and can be found in all industries where entrained particulates are present in the process stream, which can cause extensive erosion to the valve trim and valve body. Often these applications also involve high pressure drops that create the added potential for cavitation, flashing, and excessive noise and vibration. It is important to address all of these damage mechanisms in combination. The most common valve selection for process streams containing entrained particulates is a rugged valve design utilizing erosion-resistant body materials and internal valve body liners. This brute force approach often does not incorporate any method of staging the pressure drop and therefore can cause valve and pipe vibration, if cavitating conditions exist. Sweep-Angle and Rotary Ball Valves Figure 6.13q is an example of a valve that relies on brute force in being able to operate under dirty service conditions. This valve was initially designed for liquid asphalt service, but has also been successfully applied in separator let-down and slurry let-down applications. This valve utilizes a venturi-type throat design that directs the fluid into the center of the valve and downstream piping. This prevents impingement of the fluid on the valve body and downstream piping. The expanded valve outlet reduces the velocity of the exiting fluid, thereby reducing any associated erosion effects. This is especially important in applications that experience flashing or out-gassing, where there may be some areas of high velocity. The cylinder-guided valve plug ensures excellent controllability by providing plug stability, which results in better, more controlled flow of the fluid through the valve and reduced wear of the trim components. FIG. 6.13q Type 461 Sweep-Flo Valve by Fisher Controls. For high-temperature applications, up to 1100 F (593 C), the valve body can be specified with an extension bonnet to reduce conduction of heat to the valve packing and the actuator. Connections on the side of the bonnet are available to flush particulate out of the extension bonnet cavity and back into the flowstream as the valve opens. Another brute force selection for applications with lower pressure drops is to install a rotary ball valve. The plug in the rotary valve should be facing towards the downstream side of the valve body, so that the pressure drop will occur downstream of the valve body. However, with this approach it is important to provide either a durable downstream pipe liner or a sacrificial pipe section. As rugged as these sweep-angle and ball valve designs are, they still can be exposed to conditions that cause excessive noise and vibration. This is because the process fluid is not staged through the valve trim in a manner that prevents the onset of damaging cavitation and the resultant noise and vibration. Dirty Service Trim Another valve design that is available for dirty service applications (but not for heavy sludge, slurry services, or particle sizes exceeding 0.5 in., because of plugging) incorporates an axial flow path through a series of restrictions. These restrictions divide the pressure drop into stages, reducing the potential for cavitation and subsequent noise generation. Figure 6.13r shows one axial flow multistage trim design. The dirty service trim (DST) trim eliminates damaging cavitation and resultant noise and vibration by staging the pressure across a properly determined number of stages. The

165 6.13 Miscellaneous Valve and Trim Designs 1209 No significant pressure drop FIG. 6.13r Dirty service trim (DST) design of Fisher Controls. number of stages selected is dependent upon the pressure drop of the application and is designed in the same manner as was shown in Figure 6.13n. The higher the pressure drop, the more stages become necessary. This trim design can pass particulates up to 3 / 4 in. in diameter without plugging. The large open flow paths and expanded area staging design also compensate for volumetric expansion in flashing fluids, thus reducing velocities in the trim and the downstream piping. Also included in this DST design is a protected seat, because the shut-off function of the trim is separate from its throttling areas. This separation is accomplished by not allowing any significant pressure drop to be taken until the fluid is downstream of the seating surface. This type of design also ensures that all clearance flow is subjected to a staged pressure drop. Unlike in linear cage-style, anticavitation trim sets, here no segment of the process flow can drop directly from the upstream (P1) to the downstream (P2) pressure after the valve. High Noise More than 90% of pressure drop Very low inlet pressure to final stage Section 6.14 covers all aspects of control valve noise, and Figures 6.14o to 6.14s show some low noise control valve designs from a variety of suppliers. In many cases when the process fluid remains contained by the valve and piping, the noise generated becomes airborne FIG. 6.13s WhisperFlo noise abatement trim by Fisher Controls. only by its transmission through the valves and the adjacent piping. The sound field in the flow stream forces these solid boundaries to vibrate. These vibrations cause disturbances in the ambient atmosphere that are propagated as sound waves. There are several ways to eliminate excessive valve noise (see Section 6.14), but this section will focus mainly on source treatment methods, which are obtained through specially designed control valve trim. There are many types of control valve trims that were designed to reduce valve noise, but there are only two ways to accomplish this reduction. One method is referred to as frequency shifting, while the other is velocity control. The former approach, which has been used for many years, incorporates small, properly sized and spaced passages that raise the frequency of the fluid exiting the trim. The higher frequency reduces acoustic energy in the audible range by relying on the transmission loss of the piping. The latter approach (velocity control) can be accomplished in many ways, and the highest level of noise attenuation is usually accomplished by incorporating both methods. Figure 6.13s shows the cross-section of a multistage noise reduction trim that utilizes a combination of noise reduction strategies and reduces valve noise by up to 40 db. Some of the features of this design include Unique passage shape reduces the conversion of total stream power generated by the valve into noise power. Multistage pressure reduction divides the stream power between stages and further reduces the acoustic conversion efficiency. Frequency spectrum shifting reduces acoustic energy in the audible range by capitalizing on the transmission loss of the piping system. Exit jet independence avoids noise regeneration due to jet coalescence. Velocity management is accomplished with expanding areas to accommodate the expanding gas. Complementary body designs avoid flow impingement on the body wall and secondary noise sources. As explained in detail in Section 6.14, the amount of noise that will be generated by a control valve can be predicted quickly and reasonably by use of industry-standard methods. In order to obtain accurate noise predictions, it is important to utilize the standards described in Section 6.14.

166 1210 Control Valve Selection and Sizing High-Capacity Valves Globe-style valves larger than 12-in., ball valves over 24-in., and high-performance butterfly valves larger than 48 in. are considered to be special valves. As valve sizes increase, shutoff static pressure loads also increase. Consequently, shaft strength, bearing loads, unbalance forces, and available actuator thrusts all become more significant with increasing valve size. Noise levels must also be carefully considered in all largeflow installations because sound pressure levels increase in direct proportion to the magnitude of flow. To keep valve noise within tolerable limits, large cast or fabricated valve body designs have been developed. These bodies, which are normally cage-style constructions, use very long plug travel, a great number of small flow openings through the wall of the cage, and an expanded outlet line connection to minimize noise generation and to reduce fluid velocity. With the increase in valve plug travels, the selection of the actuator used becomes more important. Typically, longstroke, double-acting pneumatic piston actuators are selected. With these types of actuators, the accessories required to move the actuator also become more complex. Cryogenic Valves Cryogenic applications are those with temperatures that fall below 150 F ( 110 C). Globe-style valves in cryogenic services are used in both cold box applications and non-cold box applications. Cold boxes are commonly found in the air separation industry. Valves used in these applications feature bodies with welded extension necks and standard length bonnets to allow in-place trim maintenance from outside the cold box. For non-cold box applications, a cryogenic valve with an extension bonnet is used. For control valves in cryogenic service, the correct selection of plastic and elastomeric components are important, because they often cease to function properly at temperatures below 0 F ( 18 C). In these low-temperature ranges, components such as packing and plug seals require special consideration. For plug seals, a standard soft seal will become very hard and less pliable, thus not providing the shut-off that is required. Special elastomers have been applied in these temperatures, but require special loading to achieve a tight seal. Packing is a concern in cryogenic applications because of the frost that may form on valves. Moisture from the atmosphere condenses on surfaces where the temperature is below freezing and will freeze into a layer of frost. As this frost and ice forms on the bonnet and stem areas of control valves and as the stem is stroked by the actuator, this layer of frost is drawn through the packing, causing tears and a loss of seal. The solution is to use an extension bonnet (Figures 6.1t, 6.1u, 6.13t, 6.19p, and 6.19q) that allows the packing box area of the control valve to be warmed by ambient temperatures, thus preventing frost from forming on the stem and packing box areas. The length of the extension bonnet depends upon FIG. 6.13t The construction of a cryogenic valve with extended bonnet. (Courtesy of Fisher Controls.) the operating process temperature and on the insulation requirements. When testing cryogenic valves for shut-off or in hydrostatic tests, the use of water-based tests should be avoided. If water tests are conducted, it is possible that moisture can be trapped inside the body or extension bonnet, which could ultimately form ice in the valve after it is cooled down. For these types of applications, in order to prevent freezing, the proper test medium is usually helium. High-Temperature Valves Valves that operate at temperatures above 450 F (232 C) experience many of the same limitations as do cryogenic valves. At elevated temperatures, the standard materials of control valve construction might be inadequate. For instance, plastics, elastomers, and standard gaskets generally prove unsuitable and must be replaced by more durable materials. Metal-to-metal seating materials are always used. Semimetallic or laminated flexible graphite packing materials are commonly employed, and spiral-wound stainless steel and flexible graphite gaskets are necessary. It is important to select the trim materials and valve designs that will not experience sticking and gasket failure due to thermal expansion. If dissimilar trim materials are used, it is possible that one part will react to the high temperature faster than another, causing the components to gall. It is also important to allow the trim to grow axially in the valve body. Hanging the cage element from the top of the valve body does this. Some designs also incorporate internal springs or load rings to allow for the thermal expansion of the trim. Similar to cryogenic applications, extension bonnets are used to protect the packing box parts from extremely high

167 6.13 Miscellaneous Valve and Trim Designs 1211 temperatures. For the selection of metallic and packing materials (Figure 6.1o) and the use of jacketed valve designs (Figures 6.1p to 6.1r), refer to Section 6.1. Steam Conditioning Valves Steam conditioning applications are examples of services where valves are exposed to high temperatures. These valves serve the function of simultaneously reducing the steam pressure and temperature to the level required for a given application. Frequently, these applications deal with high inlet pressures and temperatures and require significant reduction of both. They are, therefore, best manufactured as a forged or a fabricated body that can better withstand steam loads at elevated pressures and temperatures. Forged materials permit higher design stresses, improve grain structure, and offer an inherent material integrity that is superior to cast valve bodies. The forged construction also allows the manufacturer to provide up to ANSI Class 4500, as well as intermediate and special class ratings, with greater ease vs. cast bodies. Due to frequent extreme changes in steam properties as a result of the temperature and pressure reduction, the forged and fabricated valve body design allows for the addition of an expanded outlet to control outlet steam velocity at the lower outlet pressure. Similarly, with reduced outlet pressure, the forged and fabricated design allows the manufacturer to provide different pressure class ratings for the inlet and outlet connections to more closely match the adjacent piping. Other advantages of combining the pressure reduction and desuperheating functions in the same valve versus two separate devices include: Improved spraywater mixing due to the optimum utilization of the turbulent expansion zone downstream of the pressure reduction element Improved rangeability Ease of installation and servicing of only one device The manifold steam conditioning valve design (Figure 6.13u) is the most common form of steam conditioning valve available. This valve design offers all of the benefits of a combined valve, but features the ability to provide multipoint water injection utilizing an externally mounted manifold around the valve outlet. With this manifold, large quantities of water can be injected with a homogeneous distribution throughout the steam outlet flow. Positioning of the valve plug within the control cage controls the steam pressure and flow. A signal from the pressure control loop to the valve actuator moves the valve plug within the control cage to increase or decrease the amount of free flow area. As the plug is lifted from the seat, steam passes into the center of the cage and out through the cage element. The outlet section is outfitted with a water supply manifold. The manifold provides cooling water to a number of FIG. 6.13u The design of a steam conditioning valve, which is provided with an integral cooling manifold. (Courtesy of Fisher Controls.) individual spray nozzles installed in the outlet section, which provide a fine spray mist that is injected radially into the highly turbulent stream of flowing steam. The combination of having a large surface area contacting the water and the high turbulence makes the mixing efficient and the vaporization rapid. For this control system an external water control valve is required, which is throttled by a downstream temperature controller to provide the required fine-tuning of the temperature control. These types of valves are most commonly used in power (utility, cogeneration, and industrial) plant applications, which include the start-up of steam turbines, and the bypassing, dumping, venting, and exporting of steam. Tank-Mounted Valves Tank-mounted control valves are commonly found in the chemical industry. These are reverse acting (push down to open) control valves designed for handling corrosive fluids. The throttling element of the valve is positioned at the valve outlet and is placed inside the receiving tank so that flow exits directly into the vessel. In flashing applications or those with entrained gas, the erosion potential is greatly reduced. Bibliography Ball, K. E., Final Elements: Final Frontier, InTech, November Carey, J. A., Control Valve Update, Instruments and Control Systems, January 1981.

168 1212 Control Valve Selection and Sizing Control Valves Globe, Plug, Pinch, Needle, Gate, Measurements and Control, February Control Valves, Regulators, Measurements and Control, June Control Valve Handbook, 4th edition, Fisher Controls, Inc., Control Valve Sourcebook, 3rd edition, Fisher Controls, Inc., Fernbaugh, A., Control Valves: A Decade of Change, Instruments and Control Systems, January O Connor, J., The Turbine Control Valve, Instrumentation Technology, December Rahmeyer, W., The Critical Flow Limit and Pressure Recovery Factor for Flow Control, InTech, November The Sourcebook of North American valve, Actuator and Control Manufacturers, 6th edition, 2004, Valve Manufacturers Association of America. Wilson, J., The Impact of Valve Outlet Velocity on Control Valve Noise and Piping Systems, Valve World, October Wilson, J., Tight Shutoff in Boiler Feedwater Control Valves, Valve World, April 2002.

169 6.14 Valves: Noise Calculation, Prediction, and Reduction H. D. BAUMANN (1970) J. B. ARANT (1985) B. G. LIPTÁK (1995) F. M. CAIN (2005) Valve Noise Types: Sizes: Design Pressure: Materials of Construction: Special Features: Cost: Partial List of Low-Noise Valve and Diffuser/Silencer Suppliers: Mechanical vibration (usually below 100 dba); hydrodynamic caused by liquid turbulence, cavitation, or flashing (usually below 110 dba); aerodynamic (can reach 150 dba) 1 to 24 in. (25 to 600 mm) in standard bodies; sizes above 24 in. in special castings or weldment fabrications Up to ANSI Class 2500 (PN 420) standard; above Class 2500 in special designs Any machinable wrought or cast metal for body and trim approved for use in valves or pressure vessels Balanced plugs, special seal designs, hard facings, piloted inner valve, characterized flow, dual (high/low) operating conditions, multistage trims Highly variable depending upon type of design, size, metallurgy, special features. Range may be from 2 to 10 times equivalent standard valve ABB Control Valves ( Control Components Inc. ( Dresser Flow Solutions ( Emerson Process Management ( Flo-Dyne Limited ( Flowserve Corporation ( GE-Nuovo Pignone ( Industrial & Marine Silencers Ltd. ( Koso Hammel Dahl ( McGuffy Systems, Inc. ( silenc.html) Metso Automation ( Samson AG ( SPX Valves and Controls ( Tyco Flow Control ( Weir Valves & Controls ( com) Welland & Tuxhorn ( INTRODUCTION This section begins with an overview of general noise principles, followed by a description of the types of noise produced by fluid flow through control valves. The discussion of control valve noise mitigation includes both the treatment of the noise source (modifying the valve) and the treatment of the noise path (providing downstream insulation or silencers). Other options include protection of the receiver (by personal protective equipment such as earplugs or earmuffs) or the removal of the receiver (by placing a barrier or distance between the noise source and personnel). The section ends with a discussion about recent improvements in predicting and calculating probable noise levels. Because most valve noise calculation standards avoid excessive detail, only the SI system of units will be used in this section. Users of U.S. Customary units should refer to Appendix A.1 and A.2 for the proper conversion factors, including gravitational units conversions (i.e., g c ) when necessary. 1213

170 1214 Control Valve Selection and Sizing SOUND AND NOISE A weed has been defined as an unwanted plant or flower. As an environmental analogy, noise may be considered as an unpleasant or unwanted sound. Sound, in the context of this discussion, is defined as pressure fluctuations generated in the air or other medium, which are capable of stimulating the physiological hearing response of the human ear and brain. For ease of understanding, we will hereafter refer to sound and noise as equivalent terms. Most common sounds are a complex mixture of many frequencies at varying magnitudes. Pure tones have discrete frequencies. It is customary to model sound as pressure waves with sinusoidal characteristics such as frequency (f ), magnitude (p), wavelength (λ), and speed (c). Of course, sound waves possess other more complex characteristics that are beyond the scope of this topic. Frequency is expressed in cycles per second (cps) or Hertz (Hz), where cps and Hz are equivalent units. The magnitude of sound pressure is measured in units of pressure (Pascal in the SI system). The range of sound pressures that humans can discern from the threshold of hearing to the threshold of pain spans over 12 orders of magnitude! Therefore, it is more convenient to use a logarithmic comparison of an actual sound pressure to a standard pressure reference at the threshold of hearing and to define this comparison as a sound pressure level, L p, expressed by Equation 6.14(1) in decibels (db). L p p p = p = 10 log10 20 log 10 p o (1) where p is the actual sound pressure, and the reference sound pressure, p o, is defined as Pascal ( microbar or psi.). Because the decibel is a logarithmic function, for every 10 db increase, there is a tenfold increase in sound intensity. Thus, a 100 db sound is 10 times as intense as 90 db and 100 times as intense as 80 db. However, the human ear perceives each 10 db increase as an approximate doubling of loudness. The sound pressure fluctuations must be generated by some energy source that transfers power into the air or other wave-conducting medium. (Sound waves cannot travel in a vacuum.) The total acoustic power created by the noise source is defined as sound power, W a, usually expressed in watts (W). The calculation of sound power will be used in this section to predict sound pressure levels in valve applications. It is worth remembering that, while sound is produced by a power source, it is sound pressure that the ear perceives. Sound power can also be presented as a sound power level, L w, in decibels by logarithmic comparison with the standard reference level, W o, of W. L w W a = 10 log10 W 6.14(2) Wavelength, λ, is the distance required for one complete pressure cycle. o o Speed of Sound The speed of sound, c, in any medium is a function of its mass density and elastic properties. For a solid: 6.14(3) where E is the elastic modulus and ρ is the mass density. For carbon steel pipe at 100 C, E = 198 GPa, ρ = 7.86 g/cm 3, and c = 5020 m/s. For CrMo steel alloy pipe at 100 C, E = 207 GPa, ρ = 7.84 g/cm3, and c = 5140 m/s. Austenitic stainless steel pipe (UNS S30400) at 100 C with E = 190 GPa and ρ = 8.03 g/cm 3 has c = 4860 m/s. So, for purposes of estimating the speed of sound in steel pipe, using 5000 m/s usually produces satisfactory results. 1 For a liquid: 6.14(4) where E s is the isentropic bulk modulus. It can be shown that at 20 C, speed of sound in fresh water is 1481 m/s, in seawater 1521 m/s, and in machine oil (sp. gr. = 0.90) 1297 m/s. For a gas or vapor: 6.14(5) where γ is the ratio of specific heats, R is the universal gas constant (8 314 J/kmol K), T is absolute temperature (Kelvin), and M is molecular mass of the fluid. Using this relationship, we find that the speed of sound in air at 0 C (273 K) is 331 m/s. Wavelength, frequency, and speed of sound are related as shown in Equation 6.14(6). THE HUMAN EAR c = E/ρ c = /ρ E s γ p c = = ρ c = λ f γ RT M 6.14(6) The human ear is an intricate acoustic instrument that is described here in only general terms. The anatomy of the ear is divided into three major regions, each with unique functions: the outer ear, middle ear, and inner ear. The outer ear consists of the pinna, ear canal, and outer layer of the eardrum. It channels sound waves to the eardrum, where sound pressure waves are converted into mechanical energy by vibrating the eardrum. 1 The most recent valve noise calculation standard and the field of acoustics in general use SI units. Users of U.S. Customary Units are cautioned to use the proper gravitational units conversions (i.e., g c ) when necessary. To avoid excessive detail, only the SI system of units will be used in this section.

171 6.14 Valves: Noise Calculation, Prediction, and Reduction 1215 The middle ear is an air-filled cavity containing the ossicles (bones) that connect to the oval window to the inner ear. The middle ear cavity is also connected to the Eustachian tube, which equalizes static pressure across the eardrum. The middle ear mechanism acts as an impedance-matching transformer. It is matching the impedance of the air in the ear canal to the impedance of the liquid of the inner ear. The inner ear vestibule leads to the semicircular canals (providing sense of balance) and the snail-shaped cochlea, where the final energy transformation occurs. In the cochlea, mechanical energy is conducted through a traveling wave pattern on the basilar membrane, causing a shearing of the cilia of the outer and inner hair cells of the Organ of Corti. The design and stiffness gradient of the basilar membrane allow more efficient response to higher frequencies at the basal end, and progressively lower frequencies are detected along the membrane toward its apex. The Organ of Corti is the sense organ that changes vibration energy into neural energy. This conversion takes place as shearing stress on hair cells induces a depolarization that generates neural impulses. The neural impulses are conducted by the auditory nerve to the brain, where they are processed and interpreted as sound. Damage to the hair cells connecting the basilar membrane and Organ of Corti usually produces permanent loss of hearing. Damage or deterioration can occur from sudden, loud noise (explosions), excessive exposure to moderately loud noise (industrial environments, loud music), physical injury (head trauma), advancing age, infections, or disease. Loudness Perception A healthy, young adult human is able to perceive sound over a wide range of frequencies from approximately 20 to 18,000 Hz, which is generally accepted as the audible range. The human ear, however, does not give equal weight (loudness perception) to the same sound pressure level across the frequency spectrum. Studies of apparent loudness by many human subjects over the frequency spectrum when compared to a pure tone of 1000 Hz frequency has resulted in mapping the ear response. The loudness level in phons represents the pressure level in db of a 1 khz tone that a typical hearer feels is as loud as the sound in question. Figure 6.14a shows the loudness level map as function of frequency. We can see from Figure 6.14a that a sound at 1000 Hz and 50 db sounds equally loud as 67 db at 100 Hz or 62 db at 10 khz. The resulting correction numbers, which are approximating the response of the human ear, are called A weighting. The corresponding decibel level is indicated as dba, as shown in Figure 6.14b. There are other weighting schemes for various purposes, but A weighting is used in governmental regulations on noise pollution. Hence, for the discussion of valve noise levels and environmental noise reduction, we will use the dba scale. Noise levels of some common environmental sounds are compared in Table 6.14c. Sound level (decibels) ,000 FIG. 6.14a Apparent loudness contours for human hearing. Limiting Valve Noise Feeling Loudness Levels-phons Sound frequency (CPS) There are several important reasons to limit the noise levels emitted by valves and piping. One of them is to prevent the harmful effects of environmental noise pollution, which includes hearing loss in people. As was noted earlier, we can tolerate much louder sounds at low and at very high frequencies than we can in the middle of the spectrum. This is represented in the A-weighting curve of Figure 6.14b. Note that in the Hz range, the human ear is most responsive, and this is the area where high noise level exposure can do the most damage. For this reason, the U.S. government enacted the Occupational Safety and Health Act of 1970 (amended in 1998), establishing the Occupational Safety and Health Administration (OSHA). OSHA regulations limit a Amplification db Internationally standardized A-weighting filter characteristic ,000 2,000 5,000 10,000 20,000 Frequency Hz FIG. 6.14b The A-weighting filter characteristic approximates the human ear s response to different sound frequencies.

172 1216 Control Valve Selection and Sizing TABLE 6.14c Approximate Sound Pressures Levels of Typical Sounds Source of Sound L p (dba) Near jet engine; artillery fire hp victory siren at 30 m; threshold of pain 130 Rock-and-roll band; threshold of feeling 120 Jet flying overhead at 300 m 110 Air chisel; high pressure gas leak 100 Motorcycle at 15 m; subway train at 6 m; symphony 90 Inside sports car (100 km/h) 80 Loud conversation; noisy business office 70 Normal conversation; light traffic at 30 m 60 Private business office; normal conversation 50 Quiet conversation 40 Quiet home at night; still forest; soft whisper 30 Empty theater; rustling leaves 20 Inside a soundproof room; quiet breathing 10 Sound pressure level decibels - RE microbar Pipe noise Maximum momentary level Valve noise Osha limits Allowable B HR level weighted 90 dba maximum level exposure to 8 hours per day. Table 6.14d below shows general exposure time limits established by OSHA. Figure 6.14e shows a typical frequency octave band noise level contour that will meet this limit. Note that if the predominant noise frequency exposure is in the critical middle frequency range of Hz, the allowable weighted noise level over 8 hours would be considerably less than 90 dba. VALVE NOISE While there are many noise sources in industrial and process plants, some of the main contributors can be control valves operating under conditions of high pressure drop. These are one of the few and sometimes the only sources of over 100 dba sound levels found in process plants. To gain some perspective of how loud 100 dba actually is, refer to Table 6.14c for a comparison of common environmental sounds. However, even if people are removed from areas with high noise levels, other hazards are still created by excessive TABLE 6.14d OSHA Exposure Time Limits for Various Noise Levels Hours per Day dba / / (max.) Over all Octave band center frequency Fig. 6.14e Frequency octave band noise level contours, which will result in the weighted average exposure to meet OSHA limits. noise. High intensity noise can produce vibrations in structures, which become magnified when the natural (resonant) frequencies within the structure are close to the dominant frequency of the noise. Even without resonance, studies by Fagerlund have shown that the sound power that is produced downstream of valves can cause fatigue failures in valves and piping systems. This can occur when noise levels outside the pipe (1 m downstream of a valve and 1 m away from the pipe wall) exceed dba depending on pipe size. Table 6.14f provides a summary of the likely causes of noise in valves and of their frequencies. The five major sources of noise generated by control valves are as follows: Mechanical vibration Control element instability Resonant vibration Hydrodynamic noise Aerodynamic noise Mechanical Vibration Mechanical vibration of valve internal parts is caused by unsteady flow and turbulence within the valve. It is usually unpredictable and is really a design

173 6.14 Valves: Noise Calculation, Prediction, and Reduction 1217 TABLE 6.14f Sound Frequencies and Sources in Valves Frequency (Hz) Octave Band Number problem for the manufacturer. Noise levels are typically low, usually well under 90 dba, and in the 50 and 1500 Hz frequency range. The problem is often not the noise, but the progressively worsening vibration as guides and parts wear. The solution can be in improving the valve design by adding heavy-duty stems and guides. Improvements in design may also include small changes in the flow path geometry of the trim, which can also eliminate some vibration problems. Control Element Instability Sound Description Typical Noise Source in Valves Rumble Vertical plug oscillation Cavitation* Rattle Howl Horizontal plug vibration Hiss Flowing gas Whistle Squeal Natural frequency vibration 20,000 and up Ultrasonic * Cavitation frequencies vary widely from about 100 Hz to 15 khz depending on valve and trim design. Control element instability is usually due to mass flow turbulence impingement on the valve plug. The relationship between velocity and static pressure forces acting across the plug or disc face and the actuator force balance varies over time. Without sufficient stiffness in the actuator, valve, and mechanical connections, fluid buffeting forces may produce vertical stem oscillations in linear valves and torsional shaft oscillations in rotary valves, resulting in low-level rattle noise usually under 100 Hz. This instability is detrimental to control. Correction requires changing the damping characteristics of the valve and actuator combination. This is done by providing a stiffer valve actuator or eliminating mechanical backlash. If the actuator is a spring-and-diaphragm type, then one can increase the nominal spring rate from kpa (3 15 PSIG) to kpa (6 30 PSIG). For single-acting piston actuators, the cushion air loading can be increased. If these changes do not solve the problem, then either actuator can be replaced with double-acting air piston actuators, which are generally stiffer and allow use of higher air pressures. In extreme cases, a hydraulic snubber, an all-hydraulic actuator, or electromechanical actuator may be required. Resonant Vibration Resonant noise is characterized by a discrete tone and possibly a few harmonic multiples. Resonance can involve merely an acoustic interaction within the valve and piping geometry, with certain frequencies of the flow turbulence. Resonant frequencies from 200 Hz to 10 khz can be excited acoustically by the flow in the same way that tones are produced in musical instruments. Localized metal fretting or wear is likely on internal valve parts. In some cases, turbulence and acoustic resonance excites mechanical or structural natural frequencies, producing severe vibration capable of causing damage to piping, equipment, and supporting structures. Resonant noise levels exceed calculated predictions based on current prediction standards and may be in the dba range. Work by Glenn refers to this discrete resonance as screech, and his research shows that screech is possible at pressure drops lower than required to produce sonic flow in gases or cavitation in liquids. It is possible for conditions to exist that produce screech-type resonance in virtually any valve type and brand. Glenn identifies several possible causes of screech in valves: Higher than expected velocities in the valve, due to uncertainties in the pressure recovery characteristics as a function of valve opening. (Refer to the discussion of the pressure recovery factor F L in Section 6.15.) Excitation of harmonic pipe modes of vibration Flow instabilities, due to Vortex shedding Tollmien-Schlichting waves in the laminar-toturbulent flow transition Bi-stable flow separation Unstable shock waves Unstable vapor-liquid interface Two approaches to solving these problems include 1) modifying the design of the flow path to change the characteristics of the turbulence, or 2) changing the stiffness and resonant frequencies of the valve trim. Valve trims can often be modified by a change in stem diameter, change in the plug mass, or method of guiding. Flow paths can be modified sometimes by reversal of flow direction through the valve or by minor design changes to seats, plugs, or cages. These changes shift the natural frequency of the plug and stem out of the excitation range of the flow turbulence, or vice versa. Valve manufacturers and users should collaborate to implement effective solutions in valves and piping when these problems arise. For example, one investigation identified the source of serious screech noise as a gap between valve and pipe flanges created by an oversized inside diameter of the gasket. Filling the gap with a properly sized gasket eliminated the problem.

174 1218 Control Valve Selection and Sizing Hydrodynamic Noise Hydrodynamic noise is generally less troublesome and less severe than aerodynamic noise and usually only becomes excessive when accompanied by cavitation or flashing (discussed in Sections 6.1 and 6.15). Severe cavitation can produce noise in the range of dba or higher. Problems with cavitation or flashing are usually avoided by use of a suitable trim or valve type with low-pressure recovery characteristics (high F L ). Noise caused by cavitation is the result of imploding vapor bubbles in the liquid stream. This noise can vary from a low-frequency rumble or rattling to a high-frequency squeal. This latter condition is due to acoustic or pipe resonance with cavitating fluid. In most cases, the problem is not so much with noise as it is the destruction of the valve trim and piping from erosion and pitting by the imploding vapor bubbles. Reducing or eliminating the cavitation and its damage also eliminates the noise. Single-stage multiorifice valves (see Figure 6.1z) and multistage valves (see Figures 6.1y and 6.1aa) are typical solutions to cavitation erosion and noise. The sizing of any liquid service control valve should include an evaluation of the cavitation potential, with emphasis on eliminating or mitigating the cavitation. Section 6.15 outlines methods for predicting the onset of cavitation. The standards VDMA and IEC include methods for calculating hydrodynamic noise, but these methods have been shown by Kiesbauer and Baumann to predict lower than actual noise in many cases. At the time of this writing, work is under way in the International Electrotechnical Commission (IEC) to improve the accuracy of hydrodynamic noise prediction. Flashing is rarely a significant source of valve noise, although it can cause valve trim erosion damage in some cases. Flashing produces increasing valve exit velocity and downstream piping velocity as a result of the higher specific volume of the two-phase flow. In cases where sonic flow and shock cells develop in downstream piping, excessive noise can result. Expanded outlet valves and larger downstream piping will be required under conditions where a large percentage of the liquid undergoes flashing. At this time there is not a standard method for predicting noise from flashing. Aerodynamic Noise In control valve design, aerodynamic noise can be a major problem. It is a category of valve noise capable of generating noise levels of 120 dba or greater. Noise produced by fluid turbulence in liquids is almost negligible as compared to the noise generated by the turbulence and shock cells due to the high velocity of gases and vapors passing through the valve orifice. The mechanisms of noise generation in valves and transmission through pipe walls are highly complex and are still not completely predictable. As a result of the many variables influencing noise generation and the need for simplifying assumptions in calculations, predicting the noise levels from valves or atmospheric exhaust vents is an inexact science. However, universities, manufacturers, and interested technical societies have made much progress, which has resulted in better noise prediction methods based on scientific fundamentals, which will be discussed in the section on Noise Calculations. Aerodynamic noise generation, in general, is a function of mass flow rate and the pressure ratio (p 1 /p 2 ) across the valve. The point at which sonic speed is reached in the valve vena contracta is a function of the valve design and its pressure recovery coefficient, F L, combined with the ratio of upstream to downstream absolute pressure (p 1 /p 2 ). For example, valves with F L values of 0.5 and 0.95 require pressure ratios of 1.15 and 1.80, respectively, to generate sonic flow in the valve. When sonic velocity is reached at the vena contracta, the valves are said to be choked, because their capacity does not increase if the pressure ratio is increased while the upstream pressure is kept constant. Generally, choked valves are the sources of the highest noise levels, but subsonic flows can also generate high noise levels. Valves that are not choked operate in a subsonic flow regime. For a given mass flow, they are less noisy than choked valves, but the noise level will increase as the pressure ratio approaches the sonic level. Velocity of the flow in downstream pipe can also generate significant noise starting at pipe velocities of about Mach 0.4 to Mach 1.0 (sonic). Noisy gas or vapor control valves can have acoustically induced and turbulence-induced vibration damage, trim wear, and control instabilities. Highintensity noise can produce vibration-related stresses at very high cycle rates (1,000 10,000 cps). Hence, noise-induced damage can drastically reduce valve service life, and in some cases, it can cause valve or piping failures in a matter of minutes or hours. CONTROLLING NOISE The transmission of a noise requires a source of sound, a medium through which the sound is transmitted, and a receiver. Each of these can be changed to reduce the noise level. In cases when the noise is from vibrating control valve components, the vibrations must be eliminated or they might result in valve failure. In cases when the source of noise is the hiss of a gas-reducing station, the acoustical treatment of the noise is sufficient. Depending upon the magnitude of the aerodynamic noise and assuming that massive valve damage is not a factor, valve noise treatment can be accomplished either by path treatment or source treatment. Valve damage can only be reduced or eliminated by source treatment, which minimizes or eliminates the damage mechanism. There is no absolute rule that will enable one to choose between path or source treatment. However, in general, if the noise is under 100 dba, then either a path or source treatment is a possible solution. Noise above 100 dba almost always

175 6.14 Valves: Noise Calculation, Prediction, and Reduction 1219 requires source treatment to successfully solve the noise problem. The proper choice of noise treatment method is not always easy to select, but with the help of improved noise predictions through frequency spectrum evaluation and with expertise based on experience it can be obtained. Conservative solutions are preferred, because reworking or retrofitting in case of poor design is often very expensive. Path Treatment Path treatment, as its name implies, does not focus on changing the noise source. The intent of path treatment is to attenuate the noise transmission from the source to the receiver (ear). There are several common path treatments: the use of heavy wall pipe; installation of diffusers, mufflers, or silencers; and application of acoustical insulation. Path treatment is not always a more economical solution than source treatment, and economics must be evaluated for individual applications. For existing installations, path treatment may be used, not because it is the best solution, but because it may be the only feasible one. Pipe Wall Thickness Heavy wall pipe reduces noise by increasing the transmission loss through the pipe wall. The amount of attenuation depends on the stiffness and mass of the pipe. The mechanisms are complex and beyond the scope of this text. However, as a simple rule for rough estimation, each doubling of pipe wall thickness results in approximately 6 dba more attenuation depending upon pipe size (attenuation increases with pipe size). Refer to the works of Fagerlund and Chow (1981) and Reethof and Ward (1986) for important foundation in calculating transmission losses through pipe walls. Sample calculations will be introduced in the Noise Prediction section. Insulation and Absorption Another method of increasing transmission loss at the pipe wall is the use of acoustic insulation. Even thermal insulation can add 3 5 dba attenuation. Proper selection and application of 1 2 in. (25 50 mm) of a good acoustic insulation can reduce the noise level by roughly 10 dba. Certain types of insulation are more effective at specific frequency bands, so this information is important for proper selection. Because sound travels down the pipeline with very little attenuation over long distances, increasing the pipe wall thickness or applying acoustical insulation can be a very expensive solution. This approach is most useful when downstream piping runs are short. The higher the frequency of vibration, the more effective are the commercially available sound absorption materials. Figure 6.14g gives an example of acoustical treatment for the outside of a pipe. It is often beneficial to cover the inside walls of the building with sound-absorbing materials to prevent the reflection Approximately 1'' Gustin bacon snap on insulation seal wrapped with glass cloth impregnated with resin FIG. 6.14g Acoustical treatment of pipe walls. Space necessary for flange bolting and radiation of the sound waves from process equipment. If a valve is installed close to a single reflective surface (e.g., a hard floor or wall), the apparent noise increases by 3 dba; with two reflective surfaces, noise increases by 6 dba; and for three nearby reflective surfaces, noise increases by 9 dba. A valve installed in a small room with all reflective surfaces can elevate noise levels by dba. When using walls as sound barriers, it is important to seal all openings. Isolation Locating a potentially noisy valve installation at a substantial distance from normal working areas may be effective and economical. If the valve can be located on top of a structure or pipe bridge, the distance attenuation can minimize the noise treatment required at the valve source and downstream pipe. For example, instead of a control valve with a noise specification of 85 dba, it may be possible to relax this to 90 or 95 dba, which can considerably reduce the cost of the valve or noise treatment system. It is important to note that very little noise actually radiates from the valve itself, due to generally heavy wall thickness and rigidity of most valve bodies; downstream piping radiates the great majority of noise produced in the valve to the surroundings. As a general rule, each doubling of a person s distance from the piping downstream of a valve will reduce the sound level by 3 dba in a nonreflective environment. For example, the sound level that a person hears at 8 m from a valve will be 9 dba quieter than the sound level at 1 m, and at 16 m it will be 12 dba quieter than at 1 m. Diffusers, Mufflers Diffusers located downstream of the valves (Figure 6.14h) can be helpful in both original installations and retrofit situations. These devices can aid in reducing exit flow turbulence or shock. Another important function of the multiple-hole design of diffusers depends on the fact that sound frequency increases as the size of the flow passage decreases. Using many small holes forces the dominant frequency of the turbulence into a higher range to which human hearing is less sensitive. Diffusers can be designed to serve as pressure drop devices to reduce the pressure drop across the control valve

176 1220 Control Valve Selection and Sizing Flange to customer specifications guidelines for application. Inlet velocity must be subsonic, and the silencer cannot be sized to serve as a pressure reducer. An inlet diffuser (as shown in Figure 6.14j) can be helpful, because it breaks up turbulence or shock cell oscillations that often occur in downstream sound fields and reduce the effectiveness of the unit. The outer shell should have a thick enough wall to prevent resonance. Materials of construction are selected to meet process conditions and to retain absorptive materials. Source Treatment Optional flange FIG. 6.14h Acoustical diffusers are used to reduce the exit turbulence downstream of the valve. (Courtesy of Emerson Process Management.) and thus reduce its noise generation. The valve and diffuser system works best in a situation when the flow rate is substantially constant or at least does not vary over a wide range. As a restrictor, the diffuser s effectiveness in generating backpressure on the valve decreases substantially as the flow rate drops. However, this shift of pressure drop back to the control valve does not necessarily increase noise, because when this occurs, the lower mass flow produces less noise. Figure 6.14i illustrates a silencer design that can be installed downstream of a gas-regulating valve. Due to the resulting acoustical attenuation, it can reduce the sound pressure by a factor of five (e.g., from 96 to 82 dba). Mufflers or silencers (Figure 6.14j) can be used for inline path treatment or for atmospheric vents. These are usually expensive devices, with the cost escalating dramatically with size. Dissipative or dissipative/reactive silencers are most commonly used, but a comprehensive discussion of these devices is beyond the scope of this text. However, there are some rough Source treatments reduce noise by limiting sound power generated at the source. In most cases, treatment consists of a special control valve and trim design, sometimes combined with a special diffuser or back-pressure element. While they may differ in concept, design, and manufacturing technologies, these special systems are designed for one or more of the following objectives: Reduce pressure drop in stages Limit fluid velocity to subsonic levels Reduce or eliminate the formation of high turbulence and shock cells Shift as much sound power as possible into higher frequency bands that have greater transmission losses in the pipe wall and have reduced response by human hearing Depending upon the particular design, noise can be reduced with relatively inexpensive, simple elements by 7 10 dba, whereas the more sophisticated valve designs or multielement systems can accomplish as much as dba attenuation from an untreated configuration. Specification of source treatment valves or systems is not a simple matter. There are a number of design considerations, including the following: Application: in-line or vent Noise reduction required or maximum SPL allowed (dba) 8'' butterfly regulator Gustin bacon fiberglass Perforated metal screen 10'' plug valve Flow 10'' pipe 10'' weld neck flange 10'' special flange Retainer plate Retainer plate FIG. 6.14i Silencer for gas regulating stations. (From King, C. F., Control Valve Noise, Emerson Process Management.)

177 6.14 Valves: Noise Calculation, Prediction, and Reduction 1221 Shell closure Inlet diffuser Acoustical absorption material Silencer core assembly Shell Nozzle FIG. 6.14j Silencers are installed in the flow path to dissipate the sound energy by absorbing it in an acoustical pack. They are designed to cause less than 1 PSI pressure drop. In-line silencers are often the most economical means of noise control in applications where the mass flow rate is high and the pressure drop is low. These units are normally installed immediately downstream of control valves, but in some cases they may also be required upstream of the valves. (Courtesy of Emerson Process Management.) Valve absolute pressure ratio: p 1 /p 2 or p/p 1 Pressure drop, p Fluid properties Temperature operating level and range Mass flow rate and turndown Metallurgy and mechanical design considerations Other potential velocity-induced problems Valve shut-off requirements Valve service life Valve location and orientation, piping arrangement, valve support, and maintenance access Actuation and control requirements Economics, including purchase, installation, and maintenance costs The importance of each factor is a matter of judgment and experience, understanding all aspects of the application and plant operation. Thus it is up to the user to carefully weigh and evaluate all vendor proposals. If a vendor is using one of the standardized noise prediction methods, the user should verify that accurate input values were specified and used for the calculations. Vendors that offer proprietary noise prediction calculations should also offer empirical justification to support their noise prediction. While initial cost is one factor, this is not the only consideration; it may be that the more expensive equipment is the most economical solution in the long run. Plant downtime and retrofit costs for deficient valve noise solutions are usually very expensive. With these caveats, there is much good equipment available, and vendor expertise and experience can be very valuable to the user with limited experience in controlling valve noise. Basically, source treatment control valve designs fall into three categories: multipath, multistage, and combination of multipath/multistage. These designs are listed in order of sophistication and capability for noise reduction under severe operating conditions. As might be expected, cost also increases. The cost of these special designs tends to range from 2 20 times the cost of a standard valve with the same flow capacity. Multipath Valves The multipath valve (Figure 6.14k) provides multiple orifices in parallel. A cylindrical plug to vary the flow rate through the valve uncovers these orifices. Although the path shape may vary by manufacturer, the principle consists of splitting the single-path flow into a large number of small paths. (See also the designs in Figures 6.1y and 6.1z.) The noise radiated outside the pipe from the combined flow through multiple small paths is much less than that from the same flow rate through a single path restriction. Typical attenuation levels are 7 10 dba but may reach dba in some applications. Variations of the multipath design are used for both hydrodynamic and aerodynamic valve noise with damage potential of low to moderate severity. Typically, compressible fluid pressure ratios (p 1 /p 2 ) of with valve exit velocities below Mach 0.33 are good candidates for this design. A A Section A A Flow FIG. 6.14k A multipath valve design, which can provide moderate noise reduction. (Courtesy of Emerson Process Management.)

178 1222 Control Valve Selection and Sizing Flow FIG. 6.14m Resistor element used for valve back-pressure and noise reduction. (Courtesy of Dresser Flow Solutions.) FIG. 6.14l Multistage step trim valve for use on compressible fluids. Outlet is expanded to compensate for volume change. (Courtesy of Dresser Flow Solutions.) Multistage Valves Pure multistage valves force the flow through a single path of two or more restrictions in series. (Figure 6.1y gives some examples.) An example of this design is shown in Figure 6.14l. The multiple orifices in series divide the total valve system pressure drop over several stages (typically three to nine). Thus, the reduced pressure drop per stage results in greater friction loss, reduced local velocities, and reduced noise. The shape of the trim element allows an increasing effective flow area between the inlet and the outlet to compensate for the change in gas density and increase in specific volume. Thus, the outlet flange size is often larger than the inlet to limit the exit velocity to a level that will not regenerate excessive noise. Typically, these valves can provide noise attenuation up to 25 dba, depending on pressure ratio and exit Mach number. Resistor Elements In addition to diffusers, other special designs of multiple orifice restrictors are available (Figure 6.14m). These devices are built in a wafer design for installing between flanges and can be used in single- or in multistage configurations. Such resistor elements can be installed in series, as shown in Figure 6.14n. These resistor plates are designed to work in series with the control valve to share the total pressure drop in a way that reduces pressure ratios on each element, thereby reducing the potential to generate noise. The design of some of these devices forces the fluid through multiple changes in direction, acting like friction elements with noise attenuation capability of several dba. These multiple orifice restrictors are very useful in valve noise control work, but like diffusers, they lose effectiveness with flow turndown. Combination Valves Combination multipath and multistage valves are usually required for the more severe and high noise producing services. These are the workhorse designs for the really tough applications, especially those that can cause extensive valve trim damage due to erosion or cavitation or noise levels in excess of 100 dba. Various manufacturers have taken different approaches to the design of this type of valve. Many are based upon multiple orifices in series and parallel with the flow controlled by a close-fitting cylindrical plug inside the cage for throttling (Figure 6.14o). A variation of this design also incorporates a secondary diffuser element built into the valve (Figure 6.14p). One manufacturer has designed a valve for moderate pressure drops using a standard valve trim assembly that eliminates the close-fitting cylindrical plug and uses a noise attenuator element ranging from one to seven stages (Figures 6.14q and 6.14r). Another design utilizes the pressure loss producing effects of a fluid passing through a series of sharp turns FIG. 6.14n Noise can be reduced by resistor elements that are installed in series. (Courtesy of Dresser Flow Solutions.)

179 6.14 Valves: Noise Calculation, Prediction, and Reduction 1223 Bonnet gasket Body Bonnet Bonnet flange bolting Bonnet flange Attenuator FIG. 6.14o Multipath and multistage valve with shaped first stage holes. (Courtesy of Emerson Process Management.) machined into a set of stacked disks (Figure 6.14s). Other similar designs are illustrated in Figure 6.1y and 6.1aa. Depending upon the manufacturer and design, multistage, multiorifice valves will typically have 2 7 stages, although some might use 20 or more. However, the number of stages required by any specific design for a given application depends on the design principles employed and their effectiveness in the application. In other words, having more stages does not necessarily make a valve quieter than another design with fewer stages. Inexperienced users are advised to Seat ring gasket Seat ring Plug FIG. 6.14q Two-stage noise attenuator in a valve, which was designed for use with a standard inner valve trim assembly. (Courtesy of Flowserve Corporation.) require some validation of manufacturers claims of noise reduction for these critical service valves in addition to their noise calculations. AERODYNAMIC NOISE PREDICTION Valve noise prediction is an inexact science because of the complex nature of noise generation by the control valve and the transmission of this noise through the pipe wall. So it is not surprising that a number of different prediction methods are used by manufacturers and others. What is surprising is that the various methods can give answers for the same application that differ by up to 20 dba. The subject of valve noise prediction is still subject to continuing research and evaluation. So what should we do? FIG. 6.14p Multipath and multistage valve with integral secondary diffuser element. (Courtesy of ABB Control Valves.) FIG. 6.14r Multistage noise attenuator (and detail of that attenuator) in a valve, which was designed for use with a standard trim assembly with pressure balance. (Courtesy of Flowserve Corporation.)

180 1224 Control Valve Selection and Sizing EDM disk stack Punched disk stack Disk stack configurations FIG. 6.14s Special noise element design using labyrinth passages incorporated on plates. (Courtesy of Control Components Inc.) Each manufacturer claims to be able to predict the valve noise level and provide a valve design solution. It finally falls upon the user to obtain the best possible process data, carefully evaluate all proposals, ask questions and resolve marked differences, and finally use good engineering judgment and experience in selecting the vendor for each application. It is wise to err on the conservative side when making a final selection, because the cost of mistakes and of the required retrofit may far outweigh valve cost differentials. Fortunately, the noise prediction standards of the various standards organizations have helped to make comparisons of noise predictions somewhat easier for users and manufacturers alike. Standards The past quarter-century has seen continuous improvements in the standardized methods and in the industrial acceptance of noise prediction standards for valves. In 1979 the Verband Deutcher Maschinen- und Anlagenbau e.v. (VDMA) published the first standardized method of calculating the sound level for valves as Standard VDMA 24422, which addressed both hydrodynamic and aerodynamic noise. VDMA revised in 1989 to include calculations of the frequency domain. The weakness of the VDMA method was that key valve noise parameters had to be determined experimentally; when testing was not practical, prediction accuracy was unsatisfactory. Meanwhile, other organizations were developing prediction methods based on free jet turbulence theories. The Instrumentation, Systems, and Automation Society (ISA) published standard ISA in 1989, and the International Electrotechnical Commission published the first edition of IEC in 1995 and the second edition IEC in The basic methods in these standards are essentially the same. They both are based on the published works by Lighthill, Powell, Fowcs and Hawkins, Reethof and Ward, Shea, Fagerlund, Baumann, and the contributions of many others. These organizations update their respective standards when new information is validated. The author has selected the more recent IEC Standard (2000) to demonstrate the basic calculation process. This standard and the field of acoustics in general use SI units. For conversion factors to U.S. Customary units, the reader should refer to Appendices A.1 and A.2 and is cautioned to use the proper gravitational unit conversions (i.e., g c ). To avoid excessive detail, only the SI system of units will be shown in this section. Calculations Acknowledgment: The author thanks the International Electrotechnical Commission for permission to reproduce information from its International Standard IEC All such extracts are copyright of IEC, Geneva, Switzerland. All rights reserved. Further information on the IEC is available from IEC has no responsibility for the placement and context in which the extracts and contents are reproduced by the author; nor is IEC in any way responsible for the other content or accuracy therein.

181 6.14 Valves: Noise Calculation, Prediction, and Reduction 1225 IEC The intent of this section is to familiarize the reader with the nomenclature and illustrate the basic procedure of the IEC aerodynamic noise prediction method. However, the standard covers additional special cases and details that are too extensive for coverage here. Like all theoretical methods, it is based on assumptions and limitations that must be applied with appropriate engineering skill and judgment to practical applications. Some of the stated assumptions and limitations of IEC include The valve is installed with steel or alloy steel piping upstream and downstream, possibly with pipe expanders, and that the downstream piping is straight for a length of at least 2 m from the point where the noise measurement is made. The method assumes that the fluid properties can be modeled on the perfect gas laws. The method can be used for most valve types. However, it is not applicable for full-bore ball valves where the product F p C (for the operating condition) exceeds 50% of the valve s rated flow coefficient. The method shown for standard single-stage trims in this section applies only for valve outlet and downstream pipe velocities with Mach numbers up to 0.3. Refer to IEC for multistage trims and higher velocities. Transmission loss of noise through the pipe wall is based on a simplified method due to the wide tolerances in pipe wall thickness for commercial steel pipe. The calculated sound pressure level assumes a location 1 m downstream from the valve or expander and 1 m from the outside of the pipe wall in an acoustic free field. The prediction does not guarantee actual results in the field. Validation tests were conducted with low pressure air and steam in laboratory tests, and the majority of test results were within 5 dba of the predicted sound pressure level. Nomenclature The following list of symbols, definitions, and units is a partial set of those used in IEC Symbol Description Unit A Area of a single flow passage m 2 A n C c vc c vcc Total flow area of last stage of multistage trim with n stages at given travel Flow coefficient (K v and C v ) (see IEC ) Speed of sound in the vena contracta at subsonic flow conditions Speed of sound in the vena contacta at critical flow conditions m 2 Various (see IEC ) m/s m/s c 2 Speed of sound at downstream conditions m/s D Valve outlet diameter m d Diameter of a circular flow passage m d H Hydraulic diameter of a single flow m passage d i Smaller of valve outlet or expander inlet m internal diameters D i Internal downstream pipe diameter m D j Jet diameter at the vena contracta m d o Diameter of a circular orifice, the area m of which equals the sum of areas of all flow passages at a given travel F d Valve style modifier = d H /d o Dimensionless F L Liquid pressure recovery factor of a valve Dimensionless without attached fittings (see note 4) F LP Combined liquid pressure recovery factor Dimensionless and piping geometry factor of a control valve with attached fittings (see note 4) F P Piping geometry factor Dimensionless f g External coincidence frequency Hz f o Internal coincidence pipe frequency Hz f p Generated peak frequency Hz f pr Generated peak frequency in valve outlet Hz or reduced diameter of expander f r Ring frequency Hz l Length of a radial flow passage m l w Wetted perimeter of a single flow passage m L g Correction for Mach number db (ref p o ) L pae A-weighted sound pressure level external dba (ref p o ) of pipe L pae,1m A-weighted sound pressure level 1 m from dba (ref p o ) pipe wall L pi Internal sound pressure level at pipe wall db (ref p o ) L wi Total internal sound power level db (ref W o ) M Molecular mass of flowing fluid kg/kmol M j Freely expanded jet Mach number in Dimensionless regimes II to IV M j5 Freely expanded jet Mach number in Dimensionless regime V M o Mach number at valve outlet Dimensionless M vc Mach number at the vena contracta Dimensionless M 2 Mach number in downstream pipe Dimensionless m Mass flow rate kg/s m s Mass flow rate at sonic velocity kg/s N Numerical constants (see Table 6.14v) Various N o Number of independent and identical flow Dimensionless passages in valve trim p a Actual atmospheric pressure outside pipe Pa (see note 3) p o Reference sound pressure = (see note 5) Pa

182 1226 Control Valve Selection and Sizing p s Standard atmospheric pressure (see note 1) Pa p vc Absolute vena contracta pressure at subsonic Pa flow conditions p vcc Absolute vena contracta pressure at critical Pa flow conditions p 1 Valve inlet absolute pressure Pa p 2 Valve outlet absolute pressure Pa p 2B Valve outlet absolute pressure at break Pa point p 2C Valve outlet absolute pressure at critical flow Pa conditions p 2CE Valve outlet absolute pressure where region Pa of constant acoustical efficiency begins R Universal gas constant = 8314 J/kmol K r w Acoustic power ratio (see Table 6.14w) Dimensionless T vc Vena contracta absolute temperature at K subsonic flow conditions T vcc Vena contracta absolute temperature at K critical flow conditions T 1 Inlet absolute temperature K T 2 Outlet absolute temperature K TL Transmission loss db t p Pipe wall thickness m U p Gas velocity in downstream pipe m/s U vc Vena contracta velocity at subsonic flow m/s conditions W a Sound power W W m Stream power of mass flow W W ms Stream power of mass flow rate at sonic W velocity W o Reference sound power = (see note 5) W α Recovery correction factor Dimensionless β Contraction coefficient for valve outlet or Dimensionless expander inlet γ Specific heat ratio Dimensionless η Acoustical efficiency factor (see note 2) Dimensionless ρ 1 Density of fluid at p 1 and T 1 kg/m 3 ρ 2 Density of fluid at p 2 and T 2 kg/m 3 Φ Relative flow coefficient Dimensionless Note 1. Standard atmospheric pressure is kpa or bar. Note 2. Subscripts 1, 2, 3, 4, and 5 denote regimes I, II, III, IV, and V, respectively. Note 3. 1 bar = 10 2 kpa = 10 5 Pa. Note 4. For the purpose of calculating the vena contracta pressure, and therefore velocity, in this standard, pressure recovery for gases is assumed to be identical to that of liquids. Note 5. Sound power and sound pressure are customarily expressed using the logarithmic scale known as the decibel scale. This scale relates the quantity logarithmically to some standard reference. This reference is Pa for sound pressure and W for sound power. Method Outline Numerous equations are involved in the calculation procedure, but a brief outline of the five general steps will ensure continuity in the procedures. 1. Gather the necessary input data Valve sizing data and dimensions of trim and body ports Configuration and dimensions of adjacent piping Service conditions and fluid properties 2. Calculate key pressures and pressure ratios, and determine the noise regime 3. Calculate the effective jet diameter 4. Calculate jet conditions, acoustic efficiency, sound power, peak frequency for the noise, and internal sound pressure level 5. Calculate pipe natural frequencies, pipe transmission loss, and external sound pressure level Noise Regimes There are five key noise regimes identified in IEC , which have different mechanisms governing the generation and transmission of sound. In order to determine which regime applies to a given set of conditions, several important pressures and pressure ratios must be calculated and compared with the actual downstream pressure, p 2. The vena contracta is the region of flow constriction with maximum velocity and minimum pressure, p vc, given in Equation 6.14(7). p = p p vc p F L 6.14(7) 2 If the valve has attached fittings, F L is replaced by the value of F LP / F P. At critical flow conditions, the vena contracta pressure is p vcc. p = 2 vcc p1 γ + 1 γ ( γ 1) 6.14(8) The downstream pressure at the critical pressure drop where sonic flow begins at the vena contracta is P 2C. 2 p = p F ( p p ) 2C (9) At the break point pressure, P 2B, shock cell turbulent interaction begins to dominate noise generation. For downstream pressures greater than p 2B, turbulent shear flow generates most of the sound power. p where α is a correction factor defined as 2B L p 1 1 = α γ 6.14(10) 2 Equations 6.14(7) (46) and 6.14(48) (53) are used here by permission of IEC. Copyright 2000, IEC, Geneva, Switzerland. 1 vcc γ ( γ 1)

183 6.14 Valves: Noise Calculation, Prediction, and Reduction 1227 α p 1 p 2C pvcc = p p 1 2 C p vcc 6.14(11) Regime V begins when downstream pressure drops to P 2CE and is where acoustical efficiency becomes constant. Further reductions in downstream pressure will not increase the sound pressure level. p = p1 2CE 6.14(12) 22α Table 6.14t summarizes the boundaries and characteristics of Regimes I through V. Calculate Jet Diameter Determining the jet diameter requires information about the valve trim dimensions in order to calculate the valve-style modifier F d, which is the ratio of hydraulic diameter, d H, of a single flow passage to the circle diameter, d o, corresponding to the total flow area. d o F d d H d = d H o A = 4 l w NA = 4 πo 6.14(13) 6.14(14) 6.14(15) TABLE 6.14t Characteristics of IEC Noise Regimes Regime Downstream Pressure Description I p 1 > p 2 p 2C Subsonic flow; isentropic recompression; turbulent shear noise II III IV p 2C > p 2 p vcc p vcc > p 2 p 2B p 2B > p 2 p 2CE Sonic flow at vena contracta; isentropic recompression; turbulent shear noise Supersonic flow past the vena contracta; no recompression; noise from shock turbulence and shear turbulence Sonic flow at vena contracta; supersonic Mach cone terminates in Mach disk at outlet; shock interaction dominates noise V p 2CE > p 2 Supersonic Mach cone reaches maximum Mach number; acoustical efficiency and noise level are constant Lacking specific dimensions, approximate values of F d are given in Table 6.14u. The jet diameter, D j, is calculated from Equation 6.14(16). Dj = N 14 Fd CFL 6.14(16) where units conversion factor N 14 is found in Table 6.14v and depends on whether the required flow coefficient, C, is given as C v or K v. (Refer to Section 6.15 in this chapter for information about flow coefficients.) TABLE 6.14u Typical Values of Valve-Style Modifier F d (Full-Size Trim) Relative Flow Coefficient Φ Valve Type Flow Direction Globe, parabolic plug To open To close Globe, 3 V-port plug Either* Globe, 4 V-port plug Either* Globe, 6 V-port plug Either* Globe, 60 equal diameter hole drilled cage Either* Globe, 120 equal diameter hole drilled cage Either* Butterfly, swing-through (centered shaft), to 70 Either Butterfly, fluted vane to 70 Either Butterfly, 60 flat disk Either 0.50 Eccentric rotary plug Either Segmented ball 90 Either NOTE: These values are typical only; actual values are stated by the manufacturer. * Limited p 1 p 2 may apply in flow-to-close direction. Copyright 2000, IEC, Geneva, Switzerland

184 1228 Control Valve Selection and Sizing TABLE 6.14v Numerical Constants N TABLE 6.14w Acoustic Power Ratio r w Flow Coefficient Valve or Fitting r w Constant K v N N Note: Unlisted numerical constants are not used in this standard Copyright 2000, IEC, Geneva, Switzerland. Regime I Calculations Calculate the following subsonic parameters for the vena contracta. Gas velocity: C v Globe, parabolic plug 0.25 Globe, 3 V-port plug 0.25 Globe, 4 V-port plug 0.25 Globe, 6 V-port plug 0.25 Globe, 60 equal diameter hole drilled cage 0.25 Globe, 120 equal diameter hole drilled cage 0.25 Butterfly, swing-through (centered shaft), to Butterfly, fluted vane, to Butterfly, 60 flat disk 0.5 Eccentric rotary plug 0.25 Segmented ball Expanders 1 U vc = Stream power: γ pvc 2 γ 1 1 p 1 ( γ 1) γ p 1 ρ (17) Copyright 2000, IEC, Geneva, Switzerland. Peak frequencies in Regimes I and II are based on Strouhal s equation with the Strouhal number = 0.2. W = mu ( ) vc m (18) f p U = 02 D. vc j 6.14(24) Absolute temperature: Speed of sound: T T p vc vc = 1 p 1 ( γ 1) γ 6.14(19) Common Calculations for Regimes II V For sonic conditions in the vena contracta, calculate the following parameters. Vena contracta temperature: T T = 2 1 vcc γ (25) Mach number: c vc RT = γ M vc 6.14(20) Velocity of sound: c vcc RT = γ M vcc 6.14(26) M vc U = c 6.14(21) With this information, calculate the acoustical efficiency factor, η 1, sound power, W a, and peak frequency, f p. vc vc 4 η 1 = 1 10 ( ) 36. M vc 6.14(22) Stream power: W ms mc = 2 2 vcc Mach number in the freely expanding jet: 6.14(27) W 2 = η 1 r W F a w m L 6.14(23) where r w is the acoustic power ratio taken from Table 6.14w. M j = 2 p 1 γ 1 αp 2 ( γ 1) γ (28)

185 6.14 Valves: Noise Calculation, Prediction, and Reduction 1229 Next, the acoustical efficiency factors, sound power, and peak frequency are calculated for the regime in question. The acoustical efficiency factor becomes constant: Regime II Acoustical efficiency factor: 2 4 M 5 η 5 = ( 1 10 ) j 2 ( 2 ) F L 6.14(38) Sound power: Peak frequency: Regime III Acoustical efficiency factor: Sound power: F η 2 = ( 1 10 ) M L j p p 1 2 Wa = η 2 rwwms p p f p 1 Mj c = 02 D. vcc j vcc F η 3 = ( 1 10 ) M L j 6.14(29) 6.14(30) 6.14(31) 6.14(32) Sound power generated in this regime that radiates into downstream pipe is Peak frequency: f p W a = η 5 rww ms 035. cvcc = D M 1 j 6.14(39) 6.14(40) Noise Calculations The following calculations are used for all regimes. Downstream mass density: j5 p 2 ρ2 = ρ1 p (41) W a = η 2 rww ms 6.14(33) Downstream sonic velocity: Peak frequency is calculated from Equation 6.14(31). Regime IV Acoustical efficiency factor: c 2 RT = γ M (42) Sound power: ( ) 4 η 4 = M j 2 ( 2 ) F L 6.14(34) where T 2 may be found from thermodynamic isenthalpic relationships. If fluid properties are not known, reasonable results can be obtained by assuming T 2 is approximately equal to T 1. Mach number at valve outlet: Peak frequency: f p W a = η 4 rww ms 035. cvcc = D M 1 Regime V Jet Mach number reaches it maximum: j j 6.14(35) 6.14(36) m M = 4 o 2 πd ρ c 6.14(43) If the outlet Mach number M o is above 0.3, accuracy of this method diminishes. IEC clause 7 provides further procedures for high Mach number applications, which is outside the discussion of this basic process. The internal sound pressure level, L pi, referenced to Pa is calculated in db from the following: 2 2 = 2 [( 1 22 ) γ 1 ] ( γ 1) γ M j5 6.14(37) L pi = 10 log 10 9 ( ) Waρ c Di 6.14(44)

186 1230 Control Valve Selection and Sizing Transmission through the Pipe The pipe wall must be made to vibrate in order for noise inside the pipe to radiate into the air outside the pipe. The mode of pipe vibration, for the purpose of this prediction method, is determined from the peak frequency of the noise source and the natural frequencies of the pipe. The assumption is made that the shape of the sound frequency spectrum is an arc or haystack -shaped curve that reaches a pronounced maximum level at peak frequency, f p. Although this is true for most valves, some configurations can possess flatter broadband spectra that could radiate more noise than the simplified model predicts. Pipe natural frequencies are functions of the pipe diameter, wall thickness, and density. The transmission loss model used by the IEC standard is based on the work of Fagerlund and Chow. The important characteristic frequencies are explained in detail by Singleton and are summarized below. Ring frequency, f r, has a wavelength exactly equal to the circumference of the pipe, which produces a resonant stress wave around the circumference. f r = 5000 π D 6.14(45) External coincidence frequency, f g, corresponds to the external acoustic wave speed that matches the speed of a flexural wave in pipe wall. Assuming the speed of sound in steel is 5000 m/s and 343 m/s in air, f g 2 3( 343) = πt ( 5000) p 6.14(46) First internal coincidence frequency, f o, is the lowest natural frequency of the pipe wall and produces a longitudinal flexural wave that spirals along the length of the pipe c fo = π D c = i o 6.14(47) Cutoff frequency, f c, though not part of the IEC standard, is significant because at the cutoff frequency and below, the wavelengths are too long to reflect off the internal pipe wall, making them incapable of vibrating the pipe. f c 6.14(48) The relationship of the peak frequency in the flow stream to the pipe natural frequencies is used to calculate the frequency factors used in the transmission loss calculation. Table 6.14x, taken from IEC , shows how frequency factors G x and G y are determined. i f r c c = D i TABLE 6.14x Frequency Factors G x and G y G G x y fo = f f p < f o G y = 1 for f o f g for f o < f g The transmission loss across the pipe wall is calculated from Equation 6.14(49). 6.14(49) Next, calculate the downstream pipe velocity correction factor, L g. where M 2 should not exceed 0.3 and is calculated by 6.14(50) 6.14(51) The A-weighted sound pressure level radiated from the outside surface of the pipe is given a 5 db correction to account for all frequency peaks and is calculated below. 6.14(52) Finally, a distance adjustment is made to calculate the sound pressure level in dba at 1 m from the pipe wall. Noise Calculation Example r fo = fg 23 4 f p fo f p f o G x = 1 for f p f r G y = 1 for f p f g for f p < f r for f p < f g Copyright 2000, IEC, Geneva, Switzerland, c 7 2 TL = 10 log 10 ( ) tpfp L pae,1m 6.14(53) These calculations are typically carried out with computer software and presented as part of the sizing calculations done by valve manufacturers. For a thorough understanding, a simple calculation example is tabulated below. G G x y fp = fr fp = fg L = 1 g 16log10 1 M M 4m = πd ρ c 2 2 i 2 2 L = 5 + pae L + pi TL + L 23 2 Gx ρ2c2 ( Gy ) Di + 2tp + 2 = LpAe 10 log 10 Di + 2tp 2 g p a ps

187 6.14 Valves: Noise Calculation, Prediction, and Reduction 1231 Example 1. Steam Valve Inputs Valve: NPS 4-in. (DN 100) Trim: Parabolic plug C = C v = 152 required (from sizing calculations) F L = 0.90 Φ = 0.60 F d = 0.31 Maximum allowable noise level: 90 dba Pipe: NPS 4-in. (DN 100) Schedule 40 carbon steel D i = m t p = m Fluid: Steam M = kg/kmol T 1 = 260 C = 533 K p 1 = 2.5 MPa p 2 = 1.7 MPa m = 9.10 kg/s γ = 1.32 From steam tables: ρ 1 = kg/m 3 T 2 = 247 C = 520 K ρ 2 = 7.58 kg/m 3 Preliminary Calculations Variable Equation Results p vc 6.14(7) 2.5 ( )/(0.9) 2 = MPa p vcc 6.14(8) 2.5[2/( )] 1.32/0.32 = MPa P 2C 6.14(9) 2.5 (0.9) 2 ( ) = MPa α 6.14(11) 1.355/1.573 = P 2B 6.14(10) (2.5/0.861) (1/1.32) 1.32/0.32 = MPa P 2CE 6.14(12) 2.5/[(22) (0.861)] = MPa Regime? Table 6.14t p 2 p 2C : Regime I D j 6.14(16) ( )(0.31)( ) 1/2 = m Regime I Calculations U vc 6.14(17) {2(1.32/0.32)[1(1.512 / 2.5) 0.32/1.32 ]( /11.17)} 1/2 = 460 m/s W m 6.14(18) (9.1)(460) 2 /2 = W T vc 6.14(19) (533)(1.512/2.5) 0.32/1.32 = 472 K c vc 6.14(20) [(1.32)(8314)(472)/(18.02)] 1/2 = 536 m/s M vc 6.14(21) 460/536 = Determine Internal Noise η (22) ( )(0.858) 3.6 = r w Table 6.14w 0.25 W a 6.14(23) ( )(0.25)( ) = 13.9 W f p 6.14(24) (0.2)(460)/(0.0167) = 5,509 Hz ρ (41) or steam tables 7.58 kg/m 3 c (42) [(1.32)(8314)(520)/(18.02)] 1/2 = 563 m/s M o 6.14(43) (4)(9.1)/[π(0.1016) 2 (7.58)(563)] = L pi 6.14(44) 10log[( )(13.9)(7.58)(563)/(0.102) 2 ] = db f r 6.14(45) (5000)/[π(0.102)] = 15.6 khz Determine Radiated Noise f g 6.14(46) (3) 1/2 (343) 2 /[π ( )(5000)] = 2155 Hz f o 6.14(47) (15600/4)(563/343) = 6401 G x Table 6.14x f p < f o : (6401/15,600) 2/3 (5509/6401) 4 = G y Table 6.14x f o f g : 1.0

188 1232 Control Valve Selection and Sizing TL 6.14(49) log ( ) () 1 ( )( 5509) ( 758. )( 563) = db + 1 ( 415)( 1) M (51) 4(9.1)/[π (0.102) 2 (7.58)(563)] = L g 6.14(50) 16log[1/( )] = 2.10 db L pae 6.14(52) = dba L pae,1m 6.14(53) log{[ ( ) + 2]/[ ( )] = 105 dba Conclusion: Noise level exceeds desired maximum; consider noise reduction trim; consult manufacturer. Applying Distance Corrections Placing extra distance between noisy equipment and people is sometimes a viable alternative to expensive noise reduction treatment, if there are no other detrimental effects of the noise at the source. If a noise source can be treated as a point in the acoustic far field, the sound radiates in a spherical pattern. Atmospheric vents can be treated this way. The reduced sound pressure level at some distance, r, from the center of a point source can be determined from a measured or calculated sound pressure level taken at a reference distance of r o from the center (typically r o = 1 m + pipe OD/2) using Equation 6.14(54) below. L pae,r = L pae,1m 20 log10 r 6.14(54) Because noise produced by valves radiates to the environment largely through the pipe for great distances downstream of a valve, this type of noise source is generally treated as a line source. Line sources radiate noise in a cylindrical pattern. The reduced sound-pressure level at distance, r, from a line source is L pae,r = L pae,1m 10 log (55) Example 2. Valve Noise at a Distance Problem: Using the same valve and conditions from Example 1, what would be the sound pressure level for a worker 30 m away from the downstream pipe (centerline)? Solution: From Example 1, L pae,1m = 105 dba, and pipe OD = m. Use Equation 6.14(55). r o r m L = 30 pae,30m log 10 1m m/ 2 = = 90dBA r o HYDRODYNAMIC NOISE PREDICTION Noise prediction for liquid flow through valves should consider three major flow regimes. 1) Turbulent flow, which, without cavitation, rarely produces noise levels high enough to create dangerous structural vibration or noise pollution. 2) Cavitating flow, which produces noise from vapor cavity formation and collapse as well as from turbulence, and it frequently causes excessive vibration and noise in addition to erosion of valve and piping materials. 3) Flashing of liquid into vapor across a valve sometimes causes high levels of noise and vibration, if vapor velocities in the downstream piping approach sonic velocities. Piping systems should be so sized as to avoid vapor or two-phase velocities, which are high enough to cause noise and erosion. Hydrodynamic noise prediction is currently in a state of development. Noise prediction Standards VDMA (1989) and IEC (1994) have been shown by Kiesbauer and Baumann to predict lower than actual noise in many cases. A more accurate method of hydrodynamic noise prediction has been proposed (Kiesbauer, J. and Baumann, H. D., Recent Developments in the Prediction of Hydrodynamic Noise of Control Valves, Valve World, February 2004), which is being considered by the IEC as a revision to Standard at the time of this writing. This method includes calculations for turbulent flow and cavitating flow regimes. There are no standards or generally accepted methods at this time for predicting noise under flashing conditions. For calculation of noise, the reader is advised to study the Kiesbauer-Baumann method or later revisions of IEC Standard Hydrodynamic noise predictions use a differential pressure ratio, x F, to identify noise regimes. x F p p = p p (56) The incipient cavitation index, x Fz, corresponds to the differential pressure ratio at which cavitation in a valve begins and should be determined from cavitation tests, although some of the methods include ways of estimating x Fz. This is v

189 6.14 Valves: Noise Calculation, Prediction, and Reduction 1233 the index that separates the turbulent flow regime from the cavitating flow regime. (If x F 1.0, the liquid is flashing.) Each of the methods discussed above follows a general process similar to that for aerodynamic noise prediction: 1. Gather the necessary input data Valve sizing data and dimensions of trim and body ports Configuration and dimensions of adjacent piping Service conditions and fluid properties 2. Calculate key pressures and pressure ratios, and determine the noise regime. 3. Calculate the effective jet diameter and stream power. 4. Calculate acoustic efficiency and internal sound-pressure level. 5. Calculate pipe natural frequencies, pipe transmission loss, and external sound-pressure level. Bibliography Arant, J. B., Special Control Valves Reduce Noise and Vibration, Chemical Engineering, March 6, Arant, J. B., How to Cope with Control Valve Noise, Instrument Technology, March Arant, J. B., Control Valve Noise, in Control Valves Practical Guides for Measurement and Control, G. Borden and P. G. Friedmann (eds.), Research Triangle Park, NC: Instrumentation, Systems, and Automation Society, Ball, K. E., Final Elements: Final Frontier, InTech, November Barnes, R. W., Understanding Automatic Control Valve Noise, ISA Transactions, Vol. 22, No. 1, Research Triangle Park, NC: Instrumentation, Systems, and Automation Society, Baumann, H. D., On the Prediction of Aerodynamically Created Sound Pressure Level of Control Valves, ASME Paper 70 WA/FE-28, December Baumann, H. D., How to Estimate Aerodynamic Valve Throttling Noise: A Fresh Look, ISA Paper #82-902, Baumann, H. D., A Method for Predicting Aerodynamic Valve Noise Based on Modified Free Jet Noise Theories, ASME Paper 87-WA/NCA-7, presented at ASME Winter Annual Meeting, Boston, MA, December Baumann, H. D., A Firm New Handle on Valve Noise, Chemical Engineering, December Baumann, H. D., Predicting Control Valve Noise at High Exit Velocities, InTech, February Baumann, H. D. and Hoffmann, H., Method for the estimation of frequencydependent sound pressures at the pipe exterior of throttling valves, Noise Control Engineering Journal, Vol. 47, No. 2, March April Baumann, H. D. and Page, G. W. Jr., A Method to Predict Sound Levels from Hydrodynamic Sources Associated with Flow Through Throttling Valves, Noise Control Engineering Journal, Vol. 43, No. 5, September October Boyle, S. J., Acoustical Design Considerations for Low-Noise Control Valves, Proceedings of Noise-Con 83, Institute of Noise Control Engineering, March Brockbank, G. and Glenn, A., Compressible Flow Noise, Valve Magazine, Fall CEI/IEC ( ), Prediction of Noise Generated by Hydrodynamic Flow, Geneva, Switzerland: International Electrotechnical Commission, Chow, G. C. and Reethof, G., A Study of Valve Noise Generation Processes for Compressible Fluids, ASME Paper 80 WA/NC-15, CEI/IEC ( ), Control Valve Aerodynamic Noise Prediction Method, 2nd edition, Geneva, Switzerland: International Electrotechnical Commission, Fagerlund, A. C, Use of Pipe Wall Vibrations to Measure Valve Noise, ISA Paper # Fagerlund, A. C., Recommended Maximum Valve Noise Levels, InTech, November Fagerlund, A. C. and Chow, D. C., Sound Transmission through a Cylindrical Pipe, ASME Journal of Engineering for Industry, Vol. 103, No. 4, November Glenn, A. H., Eliminating Screech-Type Noise in Control Valves, doctoral dissertation presented to the Department of Mechanical Engineering, Brigham Young University, Provo, UT, July Hynes, K. M., The Development of a Low Noise Constant Area Throttling Device, ISA #839-70, October ISA , Control Valve Aerodynamic Noise Prediction, Research Triangle Park, NC: Instrumentation, Systems, and Automation Society, Kiesbauer, J. and Baumann, H. D., Neues bei der Schallberechnung von flüssigketsdurchströmten Stellgeräten, Industriearmaturen, September Kiesbauer, J. and Baumann, H. D., Recent Developments in the Prediction of Hydrodynamic Noise of Control Valves, Valve World, February Kinsler, L. E., Frey, A. R., Coppens, A. B., and Sanders, J. V., Fundamentals of Acoustics, 3rd edition, New York: John Wiley & Sons, Inc., Lighthill, M. J., On Sound Generated Aerodynamically II, Turbulence as a Source of Sound, Proceedings of The Royal Society of London, Vol. 222, Series A, Lodge, J. M. and Turner, J. T., Aerodynamic Noise Generation in Control Valves An Experimental Study, Paper F1, International Conference on Developments in Valves and Actuators for Fluid Control, sponsored by BHRA and BVAMA, Oxford, England, September Ng, K. W., Control Valve Aerodynamic Noise Generation and Prediction, Proceedings of the National Noise and Vibration Control Conference, Chicago, IL: INCE, Ng, K. W., Control Valve Noise, ISA Transactions 33, 1994, pp Page, G. W. Jr., Predict Control Valve Noise A Graphical Method for Liquid- Flow Systems Eases the Task, Chemical Engineering, July Rahmeyer, W., The Critical Flow Limit and Pressure Recovery Factor for Flow Control, InTech, November Reed, C. M., Optimizing Valve Jet Size and Spacing Reduces Valve Noise, Control Engineering, September Reethof, G., Some Recent Developments in Valve Noise Prediction Methods, ASME Transactions of Winter Annual Meeting, November Reethof, G. and Ward, W. C., A Theoretically Based Valve Noise Prediction Method for Compressible Fluids, Transactions of the ASME Journal of Vibration, Acoustics, Stress, and Reliability in Design, Vol. 108, pp , July Saitta, G. N., A Realistic Look at Reducing Control Valve Noise Problems, Oil and Gas Journal, June 18, Sawley, R. J. and White, P. H., The Influence of Pressure Recovery on the Development of Valve Noise Description, Instrumentation, Systems, and Automation Society Paper #74-834, October Seebold, J. G., Reduce Noise in Process Piping, Hydrocarbon Processing, October Shea, A. K., A Comparative Study of Sound Level Prediction Methods for Control Valves, Proceedings of Noise-Con 83, Institute of Noise Control Engineering, March Singleton, E. W., The Impact of IEC on Control Valve Aerodynamic Noise Prediction, Measurement & Control, Vol. 32, pp , March VDMA , Richtlinien für die Geräuschberechnung Regel-und Absperrarmaturen, Verband Deutscher Maschinen- und Anlagenbaue. V., January 1989.

190 6.15 Sizing B. G. LIPTÁK (1970) C. G. LANGFORD (1985) F. M. CAIN (1995) H. D. BAUMANN (2005) Reviewed by F. M. Cain and B. G. Lipták INTRODUCTION The control valve is an important component of the total process control loop, and proper valve sizing is essential for this final control element to fulfill its role. There are two important prerequisites for the control valve to be able to fulfill its role: First is to have the correct process data. This includes the accurate knowledge of the maximum and the minimum flow conditions, the available pressure drops across the valve at the various flow conditions, the maximum inlet pressure and inlet temperature, the viscosity of the fluid, and whether or not the fluid is twophased. Correct selection of the valve s pressure drop, for example, depends on the head loss in the piping system or the characteristic of a pump, to name just two system variables (Figure 6.1b). As the reader will notice, just the selection of the process conditions involves considerable uncertainties. The second criterion is the selection of the correct valve sizing equation for the prevailing operating conditions. Therefore, before starting the process of sizing, the process control engineer has to determine which of the following sizing basis are applicable. For liquid flow sizing, they include 1) turbulent non-choked, 2) turbulent, choked, 3) saturated, 4) laminar (viscous), 5) non-newtonian, and 6) two-phase. For gas flows, the sizing conditions can be: 1) turbulent non-choked, 2) choked, and 3) laminar (small flow valves). The Standard IEC and the Instrumentation, Systems, and Automation Society equivalent Standard S have improved the accuracy of valve sizing over the years, lowering the sizing errors to less than 5% except for two-phase flow and for non-newtonian flow conditions. ABOUT THIS SECTION For the definition of all the terms used in this section, refer to the Nomenclature listing provided at the end of this section. For capacity testing procedures see Section 6.6. The subjects of control valve characteristics and rangeability are discussed in Section 6.7, and the subject of control valve noise and its calculation is covered in Section The engineering units used in this section are the U.S. Customary units. Use the numerical constants listed in Tables 6.15l and 6.15p for conversion to metric units. This section begins with a definition of the valve coefficients (C v and C d ) and with a brief description of fluid behavior during flashing, cavitation, and under turbulent or laminar flow conditions. After this general discussion the various liquid sizing factors (F L, F F, F P, F LP, and F R ) are defined, and then the liquid sizing equations are provided, as well as some examples. Some of the examples deal with viscous (laminar flow), sizing applications and pipe expander losses, and generalized approaches to liquid sizing. This is followed with a more detailed discussion of cavitation and flashing. The second half of this section begins with a discussion of valve sizing for gas and vapor services. After the sizing equations are described, an example outlines a generalized approach to gas/vapor sizing. The section is concluded with a description of sizing for two-phase flow applications. STANDARDS The study of control valve sizing is based on the science of fluid dynamics. The information in this chapter is an overview of current knowledge, which is continually growing. This know-how has been developed, in large measure, from the following two standards of the Instrumentation, Systems, and Automation Society (ISA): S , Control Valve Sizing Equations, and S , Control Valve Capacity Test Procedure. The author is grateful to ISA for its permission to copy and adopt portions of these standards for this section. The international equivalent of these U.S. standards are IEC , Industrial Process Control Valves, Part 2-1: Flow Capacity Sizing Equations for Fluid Flow under Installed Conditions, and IEC , Industrial Process Control Valves, Part 2-3, Flow Capacity Test Procedure. The equations in the international editions are essentially equivalent to those in the U.S. standards with the exception of using metric units. The user is cautioned that the calculations are not perfectly accurate for all conditions. Accuracy is best for water, air, or steam applications using conventional valve designs, which are installed in straight piping. For non-newtonian fluids or high-viscosity or low Reynolds number flows, these calculations are less accurate. In all cases, reliable calculations require reliable information about the characteristics of the valve and the process fluid, including the fluid s flow rates and pressures at normal and unusual operating conditions. The English units are used in the various sample calculations and also in the Nomenclature listing at the end of 1234

191 6.15 Sizing 1235 the section. While this was necessary because conversion constants are embedded in the working equations, other units can be substituted with appropriate conversions found in Tables 6.15i and 6.15n. The notations used in this section are those used in the ISA standards. General Principles The earliest control valves were manual globe valves. Sizing was easy: a 4 in. (100 mm) valve belonged in a 4 in. (100 mm) line. Later it became obvious that at higher pressure drops such a valve had too much capacity for good control. Thus evolved the next rule-of-thumb practice, which was to make the control valve one size smaller than line size. The limitations of such rule-of-thumb methods of valve sizing led to the development of a sizing coefficient based on Bernoulli s theorem for the conservation of energy, and the continuity equation for the conservation of mass. These equations can be combined to express the ideal flow rate through a pipe restriction such as a control valve: Q= A 2 2g( H) 1 ( A / A) (1) Calculating the actual flow rate requires the introduction of a discharge coefficient, C, which accounts for the real losses in the specific valve. The head loss ( H) may also be expressed as pressure drop divided by weight density (γ). Thus, Q= CA 2 2g( p/ γ ) 1 ( A / A) (2) By combining the area terms and discharge coefficient (C) with the other constants and expressing density in terms of specific gravity, G f, the flow equation can be written as a function of a capacity parameter, C v, which is a generally accepted valve capacity parameter called the valve coefficient. Q = C v p G 6.15(3) It should be noted that Equation 6.15(3) is only valid for turbulent and nonchoked (low pressure drop) liquid flows in sizing control valves. It is important to note that C v is not a dimensionless coefficient, and in the SI system K v is used. The conversion between engineering units is covered in later paragraphs of this section. Table 6.15i provides the conversion constants for liquid and Table 6.15n for gas and vapor sizing equations. The C v valve coefficient, which is based on English units, is widely used in the United States, while K v, based on the SI system, is the preferred unit in Europe. f The Flow Coefficient Prior to about 1946 control valves were sized using the flow area of the valve s orifice in square inches together with some, usually proprietary equations in order to estimate the amount of liquid or gas passing through the valve. Thereafter, a new coefficient was introduced called Cv, which started standardization in this field. This coefficient denotes the number of (U.S.) gallons of water that a valve would pass when the pressure drop across it is 1 lb/in. 2 Under turbulent flow conditions the water velocity is related to the square root of the pressure drop (inlet pressure minus outlet pressure). Therefore, one can multiply this quantity with the C v number of the valve to obtain the expected flow of cold water in gallons per minute. Later in this section it will be shown how this simple definition of the C v can be expanded to liquids having specific gravity other than cold water and with the use of conversion factors, for gases as well. Correction Factors Yet, the turbulent C v alone turned out to be inaccurate. Using it, occasionally there appeared to be severe sizing errors (up to 50%) that nobody could explain. This problem was resolved in 1963 by the introduction of a Critical Flow Factor. 1 This factor was later renamed the Liquid Pressure Recovery Factor and designated as F L. What this factor describes is the amount of pressure recovery, which is occurring downstream of the throttling orifice. A low F L factor, which is usually the case with streamlined valves, will promote vaporization in liquid streams or cause the development of sonic velocity in gas streams. These effects tend to choke the flow and consequently, the valve will pass less than would be predicted if the F L factor was not considered. In 1983 two additional factors were introduced. One is called Xt, the Pressure Differential Factor of a Control Valve at Choked Flow. 2 The other is called the Expansion Factor Y, which enables the calculation of the correct gas density at the valve s orifice. Yet another, albeit less serious, sizing error source has been noticed in applications where a valve, smaller than line size, was installed between reducers. The error was caused by the pressure drops across the reducers, which lowered the pressure drop available for the valve. In 1968, this problem was overcome by the introduction of a Piping Geometry Factor, 3 F p. One more factor was introduced for viscous fluid services, which is called the Reynolds Number Factor, 4 F R. In 1993, this factor was redefined and its accuracy was improved. 5 Finally, the Liquid Pressure Ratio Factor, F F (based on the original work of Allen 6 and others), was introduced for boiling or flashing liquid applications. Table 6.15a tabulates the various sizing equations for the overall orientation of the reader. The equations listed in this table will be derived in the following paragraphs. In this section, the equations are numbered consecutively, so the

192 1236 Control Valve Selection and Sizing TABLE 6.15a Orientation Table: Summary of Valve Sizing Equations (for U.S. Customary Units: gpm, SCFH, psi, F, lbm/hr, lb/ft 3, etc.) Fluid State Selection Basis Liquid Gas or Vapor Nonchoked, turbulent flow in liquids: Volumetric flow in gpm or SCFH 2 p < F ( P F P ) C L 1 F v Eq. 6.15(35) Eq. 6.15(78) v q = F P G f p C v Q = 1360 FPY P l GTZ g l x x p Y x Fx P F k = 1 ; = ; k = k T l Nonchoked vapors and gases: Mass flow in lbm/h Eq. 6.15(36) Eq. 6.15(77) p/p 1 < x T w w Cv = Cv = 63. 3FP ( p)γ FY xp γ P l l Choked flow due to cavitation or flashing liquids in liquids, or choked flow in gases Volumetric flow in gpm or SCFM Eq 6.15(31) Eq. 6.15(80) C v F q = F max LP G P F P 1 = ( P/ P) / 12 F v c f F v C v Qmax = 7320 FPY MT Z 1 Fx P l K T Choked flow due to sonic velocity in gases or vapors, or choked flow in liquids Piping effect for above equations Mass flow in lbm/h Eq. 6.15(32) Eq. 6.15(79) Not choked Eq. 6.15(22) wmax Cv= F ( P F P ) γ LP l F v l KCd FP = + 2 ( ) / C v wmax = FPY TZ 1 FxM P l K T ΣK See Eq. 6.15(21) Choked Eq. 6.15(23) F LP 2 1 KC i d = + 2 F 890 L 12 / K i See Eq. 6.15(20) Nonturbulent (viscous) flow Volumetric flow in gpm Eq. 6.15(63) q G f Cv = FR p where F R is from Figure 6.15m and is a function of Re v from Eq. 6.15(64 or 69) Mass flow in lbm/h Eq. 6.15(39) Laminar, i.e., nonturbulent, conditions generally do not occur in gases or vapors except for small flow valves. C v w = F ( p)γ R l reader can easily find the derivation of any of the equations listed in the table. Units In that part of the world where the SI system is used, the valve coefficient is called K v and it is defined as the flow of cold water in cubic meter per hour units, when the a pressure drop across the valve is one bar. The conversion between C v and K v is K v = 1.17C v 6.15(4) The flow coefficient can be converted into an area that is the area of the vena contracta, or the net flow area inside a

193 6.15 Sizing 1237 single stage valve orifice, in accordance with Equation 6.15(5): A v = C v F L in square inches 6.15(5) More refined equations are applicable for multistage valves but are not yet standardized. 7 Here the reader may refer to their respective vendors for assistance. There has been substantial progress in increasing the accuracy of valve sizing, yet additional research is needed in the areas of mixed phase and non-newtonian flows. LIQUID SIZING Fluid flowing through a control valve obeys the basic laws of the conservation of mass and energy. As the fluid stream approaches the valve restriction, fluid velocity increases in order to pass the same amount of flow through the restricted area. The restriction inside the valve is the result of moving a closure member (e.g., plug, disk, ball) closer to the valve seat. Energy to accelerate the fluid comes from a corresponding decrease in static pressure, as illustrated in Figure 6.15b. Maximum velocity and minimum static pressure occur immediately downstream from the minimum area of valve restriction (the throttling point), the most constricted area of the flow stream. This minimum pressure point downstream from the throttling point is known as the vena contracta. Figure 6.15b shows that as the velocity decreases after the vena contracta, some of the kinetic energy is converted back into static pressure. This conversion is called the pressure recovery of the valve, and its amount is equal to P 2 P vc. Valves with large pressure recovery relative to P 1 P 2 are called high recovery valves. Such valves include most rotary and gate valves. The net pressure loss (P 1 P 2 ) between the valve inlet and outlet is due to frictional effects (turbulence) and represents the permanent pressure loss. Relative Valve Capacity Coefficient (C d ) The valve capacity coefficient, C v, increases as valve size increases, but valve geometry is also a major factor in the magnitude of the pressure loss for a given flow rate. It is therefore useful to define the relative valve capacity (C d ) in order to compare the effects of geometry of different valve designs. C d is defined as 6.15(6) Representative values of C d for various valve styles (at full capacity) may be found in Table 6.15c. C d can also be useful for evaluating the relative magnitude of pressure recovery. Generally, the valve types with larger C d values are also more likely to have higher pressure recovery (P 2 P vc ); that is, they have a lower F L factor and therefore tend to choke or have liquids evaporate at lower pressure drops (P 1 P 2 ). The significance of this will become clearer in the discussions of sizing factors for choked flow. Caution: The published C v data from manufacturers is based on standard testing methods that include the friction losses associated with the test manifold pipe between the pressure taps (Section 6.6). When the C d is under 20, the friction loss from eight diameters of this pipe adjacent to the valve is relatively insignificant. However, as the C d increases above 20, the manifold pressure drop rises exponentially relative to the losses from the valve alone. Therefore, for valves with C d > 20, one must also consider the pressure drop in the valve test manifold piping to avoid significant errors. In the majority of valve applications the flow is turbulent, and the velocity profile across the restrictive (throttling) area of the valve is relatively uniform. Turbulent flow occurs when the Reynolds number is high (typically above 10,000). This occurs when the velocity is high and the viscosity is low [Re = (γud/µ)]. In the turbulent range, the incompressible fluid flow is proportional to the square root of the pressure drop across the valve, as was shown in Equation 6.15(3). This proportionality is shown by the straight line on the left side of Figure 6.15d. Factors F L, F F, F P, and F LP C = C / d d v 2 Inlet reducer effect P 1 P vc Increaser effect P 1 P 2 Pressure Recovery Factor (F L ) F L is the pressure recovery factor, which indicates the size of the pressure recovery relative to the valve pressure drop (Figure 6.15b). F L is defined as F = [( P P )/( P P )] / L vc (7) P 1 P 2 Pressure recovery (P 2 P vc ) Note: F L can also be used to calculate the pressure drop that causes sonic velocity in the valve orifice. As it has been discussed in Section 6.14, this is important for noise calculations, where P vc FIG. 6.15b Pressure profile through a valve. p sonc = 0.5 F L 2 P (8) The values of F L for various valve designs are listed in Table 6.15c.

194 1238 Control Valve Selection and Sizing TABLE 6.15c Representative Values of Relative Valve Capacity Coefficients (C d ) and of Other Sizing Factors for a Variety of Valve Designs. The Cd Values Listed are for Valves with Full Area Trims, When the Valve is Fully Open Valve Type Trim Type Flow Direction* X T F L F d F s C d = C v /d 2 ** K c GLOBE Single-port Ported plug, 4 port Either Contoured plug Open Close Characterized cage, 4 port Open Close Wing-guided, 3 wings Either Double-port Ported plug Either Contoured plug Either Wing-guided Either Rotary Eccentric spherical plug Open Close ANGLE Contoured plug Open Close Characterized cage, 4 port Open Close Venturi Close BALL Segmented (throttling) Open Standard port Either (diameter 0.8d) BUTTERFLY 60, no offset seat Either , offset seat Either , no offset seat Either *Flow direction tends to open or close the valve, i.e., push the closure member away from or toward the seat. **In this table, d may be taken as the nominal valve size, in inches. Liquid Critical Pressure Ratio Factor (F F ) The complex geometry of most valves makes experimental measurement of the pressure at the vena contracta (P vc ) impossible. The ISA sizing equations 9 use the liquid critical pressure ratio factor, F F, for approximating P vc used for saturated liquid flow conditions. where F F can be approximated by F P vc = F F P v ( P/ P) / 12 F v c 6.15(9) 6.15(10) A graph of this relationship is plotted in Figure 6.15e. Equation 6.15(10) is based on the assumption that the fluid is always in thermodynamic equilibrium. However, a liquid does not remain in thermodynamic equilibrium as it flashes across the valve restriction; therefore, the actual flow rate can be greater than that predicted by Equation 6.15(10). Estimating P vc by using F F is not fully agreed upon by experts but it is accepted by the valve sizing standard (ISA S75.01) and is in common use in the control valve industry. The pressure drop required to cause choked flow in liquids is given by the following equation: P choked = F L 2 (P 1 F F P v ) 6.15(11) Piping Geometry Factor (F P and F LP ) By convention, valve tests and calculations have included a portion of piping adjacent to the valve. P 1 is measured upstream of the pipe reducer,

195 6.15 Sizing 1239 Flow rate (q) P P ch q max the flow profile and cause some reduction in capacity. No procedure is available to predict these losses. The standard reducer and increaser fittings result in an abrupt change in size. As a result of the change in crosssection, first there is an irreversible energy loss due to turbulence (friction loss), and second there is the conversion between pressure energy and velocity energy. If the control valve is smaller than the line size, the pressure at the valve inlet is reduced by the friction loss and is also reduced because the smaller flow area requires acceleration to a higher velocity. If the downstream piping is the same size as the inlet piping, then the velocity-energy/pressure-energy exchange at the outlet is reversed and it cancels the effect at the inlet. Friction losses are always additives. For a reducer at the valve inlet, the ISA standard determines the inlet fitting friction loss coefficient: FIG. 6.15d Typical flow rate vs. p curve for liquid at constant upstream pressure and vapor pressure. The straight line segment on the left shows the behavior of turbulent liquid flow. As the pressure drop rises the pressure at the vena contracta drops and when it reached the vapor pressure of the liquid, vaporization starts. From this point on (right side of figure) the relationship is no longer linear, until at q max the choked flow condition is reached. if used, and P 2 is downstream of the increaser. Some manufacturers supply tables for C v for valves with standard pipe reducers. Where this data is based on actual test data, it should be used in preference to the approximations shown below. Elbows and block valves installed near the valve will upset F F = (P v /P c ) 1/2 Equation 6.15 (10) 1.0 K = 051.( (/ d D )) (12) For an increaser at the valve outlet the ISA standard defines the outlet fitting friction loss coefficient: K = ( 1 ( d/ D ) ) (13) Although not explicitly stated by ISA S75.01, an analysis of the references shows that for the unusual case where an increase might exist at the valve inlet, K 1 is calculated as K = ( 1 ( D/ d) ) 1 1 and for a reducer at the outlet, K 2 is found as K 6.15(14) 6.15(15) For the velocity-pressure exchange (Bernoulli effect) for the usual case: 2 2 = 051. ( ( D / d) ) reducer at inlet K = 1 ( d/ D ) 4 B (16) F F increaser at outlet K 4 = ( d/ D ) 1 B (17) It can also be shown from the same basic fluid mechanics for the unusual case: increaser at inlet K 4 = ( D / d) 1 B (18) 0.6 reducer at outlet K = 1 ( D / d) B (19) P v /P c Certain combinations of these coefficients are used in calculations: FIG. 6.15e Liquid critical pressure ratio factor F F, plotted as a function of the P v /P c ratio. Ki = K + K ( ) 1 B1 inlet combination K = K + K + K i 2 B2 6.15(20) 6.15(21)

196 1240 Control Valve Selection and Sizing F LP F P ΣK F P = [1 + (ΣK) (C d 2 )/890] 1/2 Ref (22) 1.0 C d F LP = [1/F L 2 + (K i ) (C d ) 2 /890] 1/2 Ref (23) F L K i F LP C d 0.9 Σ K F P F L 0.8 K i FIG. 6.15f Piping geometry factor as a function of C d and K. C d C d FIG. 6.15g Combined factor (F LP ) for liquids as a function of C d, K i, and F L. A piping geometry factor is defined as: F = [ 1 + ( K)( C )/ 890] P 2 1/ 2 d 6.15(22) Valve flow capacity is directly proportional to this F P factor (see Figure 6.15f ). Note that capacity is reduced for high values of ΣK combined with high C d. Because the inlet fittings also affect the valve inlet pressure, a combination of F L and F P is used to predict choking or cavitation (see Figure 6.15g): F = [/ 1 F + ( K )( C )/ 890] 2 2 1/ 2 LP L i d 6.15(23) Again note the large loss in capacity caused by reducers on valves with high C d. This equation provides only an estimate. The actual point where choking starts is preferred to be determined by test. 2 ( P P )( F / F ) = pcritical 1 vc LP P 6.15(24) For all practical purposes, the flow at rated valve capacity is always turbulent (otherwise the outlet velocity would be too high). This makes correcting for F p much simpler and avoids most iterations. Figure 6.15h provides the piping configuration factor (F p ) for two pipe-to-valve diameter ratios. The curve on the left is for a D/d = 2, and on the right for a D/d = 1.5. Installing a valve between reducers and expanders can lead to the distortion of the valve s inherent flow characteristic. This in turn can alter the gain of the valve (the percentage change in flow resulting from 1% change in control signal). F p Correction Shortcut The correction to the required C v for reducer and expander losses can also be obtained by the following steps: 1) Calculate the C v for nonchoked liquid or gas flows. 2) Estimate the valve size d (inches) by determining the required C d (dividing the calculated C v by d 2 ). 3) Make sure that this required C d is at least 40% smaller than the typical values listed in Table 6.15c. 4) Divide the pipe diameter by the chosen valve diameter to obtain the D/d ratio. 5) Select the applicable curve on Figure 6.15h. 6) Read F p from the applicable graph using the calculated C d (required C v /valve size squared). 7) Divide the required system C v by F p. This then gives the required and corrected valve C v. 8) Relationship F p v.s. required C d 1.0 F p for D/d = Required C d F p for D/d = 1.5 Relationship F p v.s. required C d FIG. 6.15h Correction factor for pipe reducers and expanders. The curve on the left is for a pipe to valve diameter ratio (D/d) of 2 and on the right, the curve is for D/d = Required C d 25

197 6.15 Sizing 1241 Check the manufacturer s literature to make sure that the C v of the chosen valve size is sufficient. EXAMPLE Assume that the required valve C v was calculated from the flow data as 70 and the pipe size is 4 in. If a line size valve is assumed, the corresponding C d = 70/4 2 = 4.4. This then could result in the selection of a 3 or 4-in. globe valve. However, one might consider a less costly 2-in. ball valve. In that case the required C d = 70/2 2 = Consulting Figure 6.15h for a D/d = 2 (curve on the left), the factor F p for a C d of 17.5 is Therefore, the required operating C v = 70/0.84 = For such a requirement, one needs to select a valve with a catalog C v of 100 or more. Based on Table 6.15c, a 2-in. segmented ball valve can provide that, so the choice is acceptable. Units Used in Valve Sizing USA Units The valve sizing equations described here are based on Instrumentation, Systems, and Automation Society Standard S75.01, Flow Equations for Sizing Control Valves. 9 For turbulent, nonvaporizing flow conditions, the flow rate in gallons per minute is calculated in accordance with Equation 6.15(25): The valve coefficient may be calculated as where P 1 and P 2 are in units of lb/in. 2 absolute. 6.15(25) 6.15(26) If flow rate, w, is given in lbm/hr and specific weight, γ 1, is in lbm/ft 3, then or if solved for C v : q = F C C v P q = F P v P1 P2 G 6.15(27) 6.15(28) Under choked flow conditions, the maximum flow rate is no longer dependent on the pressure drop but rather on the pressure drop at the onset of choking ( p ch in Figure 6.15d). See Equation 6.15(9). Therefore, the maximum volumetric flow rate, q, under choked conditions, in gallons/minute is f G f P P 1 2 w= FPCv ( P1 P2)γ 1 C v q w = F ( P1 P2)γ 1 max = F C L P v P1 FFPv G f 6.15(29) The maximum mass flow rate, w, in lbm/hr is 6.15(30) When reducers, increasers, or other fittings are installed between the pipe and the valve, a combined liquid pressure recovery factor, F LP, is used. See the discussion on liquid sizing factors in connection with Equation 6.15(23).The volumetric flow rate becomes P F P 1 F v qmax = F C LP v G f or if solved for C v : 6.15(31) q G max f Cv = F P F P The mass flow rate is calculated by or if solved for C v : wmax = FLCv ( P1 FFPv) γ 1 w max Cv = FLP ( P1 FFP v)γ (32) The pressure drop at the onset of choking ( p ch ) must always be calculated to check if choking will occur. It is convenient to write the above Equations 6.15(25 32) using the allowable sizing pressure drop, p a, which is defined as the smaller value of the actual pressure drop, P 1 P 2, and the pressure drop at the onset choking, p ch. The Reynolds number factor, F R, may also be included in the equation. Refer to the discussion on liquid sizing factors for nonturbulent in the later paragraphs discussing laminar and viscous flow. Thus, the following set of simplified equations can be written for volumetric flow rate (gpm): and if solved for C v : wmax = FLCv ( P1 FFPv) γ 1 LP q = F F C C v P R v q = FF R P p 1 F v 6.15(33) 6.15(34) Other Engineering Units The equations presented previously were derived using U.S. Customary Units. In these equations C v is not a dimensionless coefficient. Therefore, it is helpful to develop units conversion constants that will permit the use of these equations with other systems of units. ISA has chosen to write the equations using numerical constants (N 1, N 2, etc.) that take on unique values for specific G G f f p a a

198 1242 Control Valve Selection and Sizing TABLE 6.15i Numerical Constants for the Conversion of Liquid Flow Equations* Constant Units Used in Equations N w Q P, P d, D γ 1 ν (nu) N m 3 /h m 3 /h gpm KPa Bar psia N mm in. N 4 76,000 17,300 m 3 /h gpm mm in. Centistokes** Centistokes** N kg/h kg/h lb/h KPa Bar psia kg/m 3 kg/m 3 lb/ft 3 *Reprinted by permission. Copyright 1985, Instrumentation, Systems, and Automation Society. From ANSI/ISA-S75.01, Flow Equations for Sizing Control Valves. **To convert m 2 /s to centistokes, multiply m 2 /s by To convert centipoise to centistokes, divide centipoise by G f. combinations of units. Equations 6.15(35 41) and Table 6.15i are from ISA S75.01.* q N F C P P 1 2 = 1 P v G f or 6.15(35) q G f Cv = NF 1 p P1 P2 w = N FPCv P P 6 ( 1 2) γ 1 or w 6.15(36) Cv = NF 6 P ( P1 P2) γ 1 where: F P ( KC ) = 4 Nd 2 The corresponding nonturbulent equations * are 6.15(37) P P 1 2 q = N1FRCv G f or 6.15(38) q G f Cv = NF 1 R P1 P2 * Reprinted by permission. Copyright 1985, Instrumentation, Systems, and Automation Society. From ANSI/ISA-S75.01, Flow Equations for Sizing Control Valves. 2 v / w = N FRCv P P 6 ( 1 2) γ 1 or w 6.15(39) Cv = NF 6 R ( P1 P2) γ 1 The valve Reynolds number is defined* as NFq FC d L v Re v = + 12 / 12 / 4 νfl Cv Nd (40) The equations* for determining maximum flow rate of a liquid under choked conditions for valves in straight pipes of the same size as the valve are as follows: P F P 1 F v qmax = N1FLCv G f or 6.15(41) q G max f Cv = NF 1 L P1 FFPv Metric Capacity Coefficients, K v and A v Valve capacity coefficients, which have been derived in a similar manner as was the C v but which have their basis in the metric or SI system of units, can be easily converted to their equivalent C v values. If units for volumetric flow rate (q) are in cubic meters per hour (m 3 /h) and the pressure drop ( p), is given in bars (1 bar = 100 kpa), then the capacity coefficient K v is K = q( G / p) / v f / 6.15(42)

199 6.15 Sizing 1243 Simple units conversions lead to the following relationship between K v and C v : K v = 0.865C v 6.15(43) The derivation of the capacity coefficient, A v, yields an equivalent area in square meters (m 2 ) that is related to C v by the conversion A 6.15(44) It should be noted that valve capacity or sizing coefficients are defined in terms of a flow test with pressure taps located in straight pipe of the same nominal size as the valve and at specific distances upstream and downstream of the valve. The test configuration for determining C v is well defined in ISA Standard S and is discussed in Section 6.6. The above conversions are valid only if K v or A v are determined using the same test manifold configuration and method as those specified in ISA-S Sizing Example for Liquids = C The above working equations can be used in a logical sequence to determine the required valve capacity for a given application. The following example illustrates the basic steps. However, there are other considerations when selecting a valve, such as noise, cavitation, corrosion, and actuator sizing, that are discussed in other sections of this chapter. Example 1 Determine the required valve capacity coefficient (C v ) and valve size for the following application: Liquid: Water Critical pressure (P c ): psia Temperature: 250 F Upstream pressure (P 1 ): psia Downstream pressure (P 2 ): psia Flow direction (relative to valve plug): Flow-under-toopen Line size/class: 4 in., ANSI Class 600 Flow rate (q): 500 gpm Liquid vapor pressure (P v ): 30 psia Kinematic viscosity (v): centistokes Valve characteristics: Equal percentage Step 1: Calculate actual p. v p= P1 P2 = = 110 psi Step 2: Calculate choked pressure drop p ch and determine allowable sizing pressure drop p a. From Equation 6.15(10): F 12 / = ( P / P) 12 / = ( 30/ ) = 093. F v c v In order to determine the p ch, a preliminary valve-type selection must be made in order to establish the liquid pressure recovery coefficient F L to be used in the calculation. Assume the use of a single-ported globe valve, which has an F L equal to 0.9 (Table 6.15c). Solving for p ch from Equation 6.15(11): 2 2 p = F [ P F P] = 0. 9 [ ( 0. 93)( 30)] = psi ch L 1 F v The allowable sizing drop p a is the smaller of p and p ch. Therefore, we have determined that p a = 110 psi, and the flow is not choked. Step 3: Determine the specific gravity, G f. The Instrumentation, Systems, and Automation Society sizing equations are based on water at 60 F with a density of lb/ft 3. From water properties data we find the density of water at 250 F is 58.8 lb/ft 3. G f = / = Step 4: Calculate the approximate required C v assuming F P and F R equal 1.0. From Equation 6.15(34): C v q = FF R P G f p a 12 / = 500( 0. 94/ 110) = Step 5: Select the approximate body size based on the approximate C v from Step 4. From manufacturers catalogs it can be determined that the smallest globe valve body size that will accommodate C v 46.2 is a 2 in. size with a maximum valve capacity coefficient of C v 51. Step 6: Determine the piping geometry factor F P. If manufacturer s test data are not available for the selected valve type, F p may be estimated using Equations 6.15(12 22) or Figure 6.15f. d = 20., D = D = 40., and d/ D = Therefore, in Figure 6.15h, the left side (D/d = 2) curve should be used. For a reducer on the inlet and an increaser on the valve outlet the following values are calculated: K K K K 1 2 B1 B2 = = [ 5 1 (.)] = = 2 2 [ 1 ( 0. 5) ] = 1 ( 0. 5) = = ( 0. 5) 1 = K = = C d 2 = 46.2/( 2) = F = [ 1 + ( K)( C ) / 890] P 2 1/ 2 d 2 / = [ 1+ ( ) ( ) / 890] 1 2 = 0. 94

200 1244 Control Valve Selection and Sizing Step 7: Calculate the Reynolds number factor, F R. For most process applications, this step is not necessary, because in most cases turbulent flow (i.e., Re v > 10,000) conditions exist. The calculation is shown here for illustration purposes. From Equation 6.15(64): Re v NFq 2 d ( FC L d) 4 = + / / F C ν N L ( )( )( ) Re = [( 09. )( )] v ( ) ( 0. 9)( 46. 2) Re = Referring to Figure 6.15m, F R = 1. Step 8: Calculate the final required C v using Equation 6.15(34) and the values for F p and F R that were calculated above: C v v q = FF R P G Note: It was noted in Step 5 that the maximum C v for a 2-in. globe valve was found to be C v(max) = 51. The required C v(req) = 49.1 is too close to the maximum, because it exceeds 85% of it. Therefore, a 3-in. globe valve should be selected. Step 9: Calculate the fluid velocity at the 3-in. valve outlet using Equation 6.15(45) with appropriate units conversions. where U = velocity q = volumetric flow rate A o = area of the body outlet port 6.15(45) If q is given in gallons per minute (gpm) and A o is in square inches, then U can be calculated in feet per second: For this example, the outlet velocity is U = 0.321(500 gpm/7.07 sq. in.) = 22.7 fps 6.15(46) Evaluation: The valve sizing and outlet velocity of 22.7 fps is of consequence only under flashing (liquid/vapor) conditions, or where the fluid contains erosive particles. In such cases, the valve manufacturer should also be asked to doublecheck the sizing for the particular service. The Cavitation Phenomenon v f p a 2 Cavitation occurs when static pressure anywhere in the valve drops to or below the vapor pressure of the process liquid. 14 / / = ( 500)/[( 0. 94)( 1)]( 0. 94/ 110) 12 / = U = q/ A o U = q / A o ( fps ) ( gpm ) ( sq.in.) Vaporization begins around microscopic gaseous nuclei. The cavitation process includes the vapor cavity formation and sudden condensation (collapse) of the vapor bubble driven by pressure changes. The basic process of cavitation is related to the conservation of energy and Bernoulli s theorem, which describes the pressure profile of a liquid flowing through a restriction or orifice (Figure 6.15b). In order to accelerate the fluid through the restriction, some of the pressure head is converted into velocity head. This transfer of static energy is needed to maintain the same mass flow through the reduced passage. The fluid accelerates to its maximum velocity, which corresponds to the point of minimum pressure (vena contracta). The fluid velocity gradually slows down as it expands back to the full pipe area. The static pressure also recovers somewhat, but part of it is lost due to turbulence and friction. If the static pressure at any point drops below the liquid vapor pressure (P v ) corresponding to the process temperature, then vapor bubbles will form. If enough energy is imparted to the growing vapor bubble to overcome surface tension effects, the bubble will reach a critical diameter and expand rapidly. As the static pressure recovers to a point greater than the vapor pressure, the vapor will condense, causing the bubbles to collapse violently back into their liquid phase. The growth and collapse of the bubbles produce high-energy shock waves in the fluid. The collapse stage of the process (the bubble implosion) produces the more severe shock waves. Shock waves and liquid microjets radiate for short distances from imploding cavities and erode nearby surfaces. Cavitation can cause erosion, noise, and vibration in piping systems and therefore must be avoided. Extensive cavitation also causes choked flow conditions in the valve. Predicting and Mitigating Cavitation Sizing a valve in liquid service for choked flow allows one to determine its maximum flow capacity, but this is of limited value, because most liquid-service valves should not be operated under choked conditions. Special trim designs with multiple stages or multiple flow paths like those in Figure 6.1y are typically used to prevent severe cavitation and are better able to operate at or near choked conditions without damage. Metal erosion from cavitation damage has a very distinctive appearance, like that of cinder block or sandblasting. No known material will withstand continuous, severe cavitation without damage and eventual failure. The length of time it will take is a function of the fluid, metal type, and severity of the cavitation. Without special trim geometry, mild or intermittent cavitation can be mitigated in standard valves by the use of extremely hard trim materials or overlays. Sometimes it is feasible to increase the downstream backpressure, or limit the pressure drop by installing control valves in series to reduce the pressure drop in each valve. Another mitigating effect in some processes is a result of the fluid thermodynamic properties and operating conditions. As liquid operating temperature approaches its critical temperature, the

201 6.15 Sizing 1245 growth and collapse rates of cavities slow down, as heat transfer effects become increasingly significant relative to the dominant inertial effects. This can greatly reduce the energy of cavity collapse and surface damage to the valve parts. Some cryogenic and hydrocarbon applications are thought to behave this way, which may partly explain why cavitation damage in these cases is minimal or absent even when cavitation is present in the valve. Further information about cavitation and predicting its effects on valve performance can be found in the ISA Recommended Practice RP , Considerations for Evaluating Control Valve Cavitation. Various methods are in use to establish the pressure drop at which cavitation starts. Two of the common techniques are described here. The first is based on a cavitation index, K c, determined from flow capacity test data similar to that shown in Figure 6.15d. The second is based on a cavitation index, σ, and vibration test data. Flow Curve Cavitation Index (K c ) Figure 6.15j illustrates the effect that cavitation has on the linear relationship between flow rate and the square root of pressure drop. The inflection point, the point at which the linear slope of the C v curve marks the p, at which measurable amounts of vaporization exist is measured by a flow test. 10 The calculation of the index K c is based on the assumption that a valve may operate cavitation-free at any pressure drop that is less than that associated with K c, which is defined as 15 K c Pi = P P (47) where p i is the pressure drop at the inflection in the tested C v slope. K c can then be compared to the index describing the v actual service conditions of an application in question, K sc : K sc P = P P (48) If K sc for the service is less than the K c for the valve, the valve is assumed to be adequate for the service. Estimated K c values for various valve types are shown in Table 6.15c. These values of K c are only representative and should always be verified with the valve manufacturer. However, care must be exercised with this method, because it has not consistently indicated damaging cavitation for some valve designs. 2,15,18 Vibration Curve Cavitation Index ( ) Noise and vibration measurements have identified four recurring regimes of cavitation in many valves Figure 6.15k illustrates the relationship between acceleration measurements taken on the downstream pipe and a cavitation index called sigma (σ). There are several forms of sigma, all of which represent a ratio between the forces resisting cavitation and the forces promoting cavitation. The following forms of sigma have been used for valves. 6.15(49) 6.15(50) Upstream and downstream pressures should be corrected for piping and fitting losses to obtain net values of P 1 and P 2 at the valve for calculating σ when C v /d 2 > 20. The σ indexes v P P 1 v σ 1 = P1 P2 P P 2 v σ 2 = P1 P2 F L 2 C v = Q/ P/G f s mv s c K c Q (gpm) P ch Acceleration P i s i P/G f ( psi ) FIG. 6.15j The flow rate curve showing the effect of choking and the methods how F L and K c are determined. (Copyright 1991, Instrumentation, Systems, and Automation Society. From ISA Paper No , Solving the Problem of Cavitation in Control Valves. ) Supercavitation Full cavitation FIG. 6.15k Classical cavitation level plot (log-log). (Copyright 1991, Instrumentation, Systems, and Automation Society. From ISA Paper No , Solving the Problem of Cavitation in Control Valves. ) s Incipient cavitation No cavitation

202 1246 Control Valve Selection and Sizing can be related to the K-type indexes by using the relationship: 1 σ1 = σ2 + 1 = K 6.15(51) Although K c can be converted to a σ value for subjective comparison, the meaning of the value does not change. There is no proven correlation between specific points on the flow curve Figure 6.15j and the inflection points on the vibration curve in Figure 6.15k. Various cavitation regimes can be identified for many (but not all) valves with the type of plot shown in Figure 6.15k. 22 The ISA Recommended Practice RP uses σ defined by σ 1 in equation 6.15(49), so where σ is used in this section without a subscript, σ 1 is implied. The various regimes of cavitation identified in Figure 6.15k are determined in a test defined in ISA RP , which begins at a maximum attainable pressure drop in the test valve. The pressure drop is then reduced in small steps while maintaining a constant upstream pressure and temperature; vibration, pressures, and flow rates are measured at each step until the noise, vibration, and P indicate that a noncavitating flow regime has been attained. The results are plotted showing vibration as a function of σ. The noncavitating regime in Figure 6.15k is distinguished by turbulent flow noise and mild vibration. The incipient cavitation regime begins at a level called incipient cavitation and progresses to another level called constant 22 (or critical 18 ) cavitation. Incipient cavitation begins with intermittent cavity formation and collapse. The resulting noise may be scarcely audible above the background noise. Constant cavitation produces a steady process of cavity growth and collapse, so that the sound is no longer intermittent but is constant. Cavitation in the incipient regime is not generally harmful to valve hardware, 18 and limiting valve operation below a level of constant cavitation is a conservative application approach although sometimes overly conservative. In the full cavitation regime, vibration or sound intensity increases to the maximum vibration level. When damage testing can be done, the onset of cavitation damage, called incipient damage, is often found somewhere in this regime. 22 In many cases, the value of 1/K c plotted on this graph will also appear in this regime. Thus, K c is sometimes referred to as incipient choking. Noise and vibration in the super-cavitation regime 18 are less than at maximum vibration as a result of the way that large volumes of vapor in the flow affect the fluid density, compressibility, and vibration transmission properties. Super-cavitation may exhibit a vapor pocket extending from the valve downstream into the pipe, where it will collapse back into liquid as the static pressure recovers. This differs from true flashing, in which the downstream pressure remains less than the vapor pressure of the flowing fluid. Damage to downstream piping and fittings is common in this regime. The cavitation index σ describing the service conditions is useful if it can be compared with a valve cavitation index that represents the resistance of the particular valve to potential cavitation. The coefficient σ mr ( manufacturer s recommended σ ) is defined in ISA RP to designate the limit for σ above which the valve may be operated. If the operating pressures for the valve are very close to the pressures used to test the valve, a direct comparison could be made, such that σ > σ mr indicates safe operating conditions. However, research has shown that valve σ coefficients for incipient, constant, and incipient-damage cavitation do not remain constant for all pressures or in different sizes of geometrically similar valves. Scaling equations have been developed that take these effects into consideration, and they have been adopted in the ISA recommended practice. Pressure Scale Effect (PSE) The scaling factors PSE and size scale effect (SSE) are used to adjust the recommended valve coefficient, σ mr, when the reference size and reference pressure are known for which σ mr was determined. Equation 6.15(52) defines a scaled coefficient for the valve, σ v. σ = [ v σ ( ) ]( ) + mr SSE 1 PSE (52) σ v is then compared to the service σ calculated by Equation 6.15(49). The service σ σ v means that the valve will operate at a level of cavitation less severe than the level corresponding to the manufacturer s σ mr. The pressure scale effect, PSE, is P1 Pv PSE = ( P1 Pv) 6.15(53) The subscript R refers to reference pressures used in testing the valve. Typically, valves are tested at about (P 1 P v ) R = 100 psi; this value may be used as an estimate unless otherwise stated by the valve manufacturer. The exponent, a, is a constant determined for a specific valve. Table 6.15l shows representative values of a for different valve types and different levels of cavitation. 18,22 When the exponent a is 0, PSE reduces to a value of 1, indicating that there is no pressure scale effect for those conditions. Size Scale Effect The size scale effect, SSE, can be estimated from Equation 6.15(54). The exponent b was derived from limited testing. 18,20 It should be noted that testing under choked conditions showed no size scale effect; i.e., SSE for choked flow has a value of 1. SSE = d 6.15(54) 6.15(55) where d is the inside diameter of the valve port in inches d R refers to the reference (tested) valve port diameter in inches d R C v b = d b R a 14 /

203 6.15 Sizing 1247 TABLE 6.15l Exponent a for Calculating Pressure Scale Effect (PSE)* Valve Type Cavitation Level Exponent a Quarter-turn valves (i.e., ball, butterfly) Segmented ball and eccentric plug Single-stage globe Multistage globe Orifice Size scale equations are valid only when the valves are the same style and have approximately the same relative capacity; i.e., C v /d 2 must be about the same for both valves. There are additional considerations if the valve is larger or smaller than the pipe size with attached fittings (reducers, increasing). RP recommends additional calculations when C v /d 2 > 20. The following example demonstrates how this method can be applied. Example 2. Rotary Valve Service data: Fluid: Water T = 74 F Line size = 10 in. Sch. 40 P 1 = 82 psia Q = 3500 gpm G f = P v = 0.41 psia P 2 = 70 psia Incipient Constant Incipient damage Choking Incipient Constant Incipient damage Choking Incipient Constant Incipient damage Choking Incipient Constant Incipient damage Choking Incipient Constant Incipient damage Choking N/A N/A *Reprinted by permission. Copyright 1991, Instrumentation, Systems, and Automation Society. From ISA Paper No , Solving the Problem of Cavitation in Control Valves. Singlestage globe data modified per Tullis, Reference 18. Results of valve sizing: C v = 1010; 8 in., ANSI Class 150 throttling rotary disk valve; body outlet velocity = 22 fps; approximately 75% open with C v /d 2 = 1010/(8) 2 = 15.8, so effects of fittings may be ignored. Calculate σ using Equation 6.15(49) σ = = 68. Compare with manufacturer s cavitation data for the butterfly valve, which are given as: σ mr = 5.1 (P 1 P v ) R = 100 psi, a = 0.12 d R = 6 in. Calculate PSE, b, SSE from Equations 6.15(53 55). PSE = [( )/100] 0.12 = b = 0.068(15.8) 1/4 = SSE = (8/6) = Calculate σ v from Equation 6.15(52). σ v = [.( ) 1]( ) + 1 = 5. 2 Evaluation: Since σ is 6.8 and is greater than σ v of 5.2, the valve can operate without significant cavitation. Rule of Thumb for Incipient Cavitation One definition of the state of incipient cavitation is when the pressure drop increases to the point where it is first audible that some cavitation bubbles burst. This condition is usually not a cause for concern, although the pressure ratio factor X Fzi is important in noise calculations. 16 X Fzi = 0.9(P 1 ref /P 1actual ) /[1 + 3F d (C v /1.17F L ) 0.5 ] (56) In Equation 6.15(38), if P 1actual is in psia, the value of P 1 ref is 87 psia. If P 1actual is in bars, P 1actual is 6 bars absolute. Using this pressure ration factor X Fzi, the pressure drop causing incipient cavitation can be calculated using Equation 6.15(39): P inc = X Fzi (P 1 P v ) 6.15(57) As a rule of thumb that can be used as a guide only in an attempt to avoid damage or noise, one should make sure that the valve s pressure drop does not exceed the following limit: P limit = 0.5(X Fzi + F L 0.5 )(P 1 P v ) 6.15(58) For more accurate cavitation calculations one should use Equations 6.15(47 55), while for accurate noise calculations, should consult Section Example 3 Assume that a 2 in. globe valve was selected for a particular flow condition where the inlet pressure was 195 psia and the outlet pressure 87 psia. Let us also assume that using Equation 6.15(25), the required C v was calculated

204 1248 Control Valve Selection and Sizing to be 38, and that the supplier s catalog for the particular valve showed an F L of 0.87 and an F d of 0.4 at the given valve travel. Finally, assume that the vapor pressure of the flowing fluid is 8.5 psia. Using Equation 6.15(56), we can calculate the pressure differential ratio for incipient cavitation as X Fzi = 0.9 (87/195) /[ (38/ ) 0.5 ] 0.5 = Using this X Fzi value, the limiting pressure drop to avoid cavitation damage or noise, using Equation 6.15(58), is P limit = 0.5( )( ) = 96.8 psi. Because the actual pressure drop is only 80 psi, we should not be concerned. Flashing Flashing begins in the same manner as cavitation. However, in case of flashing, the pressure downstream of the cavity growth region remains at or below vapor pressure of the process fluid. This causes the vapor to stay in the vapor state as it leaves the valve and enters the downstream piping. The specific volume increases as liquid changes to vapor, which in turn causes an increase in the fluid velocity. If enough vapor is formed, the resulting high valve outlet, or pipe, the resulting velocities can erode metals. In many cases, flashing is a normal part of the process; it cannot be avoided, and special system and valve designs are required to accommodate it. The heat of vaporization comes from the process liquid, causing its temperature to decrease. The relative masses of liquid and vapor will thereby approach thermodynamic equilibrium. The amount of flashing can be calculated from an energy balance. Even small amounts of flashing (e.g., where the vapor equals 1 3% by weight) can significantly affect a valve s capacity, sizing, and selection; therefore, flashing should be stated in the valve specification data sheets. Large amounts of flashing (e.g., 10 15% by weight) require special valve designs, such as oversized outlets, and a larger downstream pipe. Note, choking can occur in the downstream pipe due to the large volume. This in turn will increase the back-pressure, P 2, which in turn then causes the valve to undergo cavitation again. When the valve outlet pressure, P 2, is less than or equal to the vapor pressure of the process liquid, some of the liquid flashes into vapor and stays in the vapor phase as it enters the downstream piping. The phase change from liquid to vapor may cause high velocities and erosion of metals at the outlet. Even small amounts of flashing can significantly affect valve sizing and selection. Large amounts of flashing (e.g., where the vapor equals 10 15% by weight) require special valve designs and materials. In order to select the right valve, it is necessary to know the fraction of the liquid that will flash to vapor and the flowing velocity of the resulting vaporized mixture. If the vapor content and therefore the velocity is high, the resulting back-pressure at the valve outlet can cause choked flow in the downstream pipe. This can cause cavitation in the valve and can cause severe erosion of the valve trim. Calculating the Flash Fraction (X) The method for calculating the vapor fraction is essentially the same for all liquids. Water is the most common liquid that is likely to flash, and data for water are readily available in the steam tables. Data for other fluids may be found in other references. 11 The vapor fraction by weight can be calculated from Equation 6.15(59). X = ( h h )/ h f1 f2 fg2 6.15(59) where h f1, h f2, and h fg2 are the enthalpy of saturated liquid upstream, the enthalpy of saturated liquid downstream, and the enthalpy of evaporation at downstream pressure, respectively. Calculating Velocity Velocity calculations of the liquidvapor mixture downstream of the valve assume that the mixture is in thermodynamic equilibrium. In most cases, this assumption is a good enough approximation when compared to the accuracy of other valve sizing factors. The velocity, U, for the mixture is U = ( w/ A)[( X) v + Xv ] 1 f2 g2 6.15(60) where all units are consistent and v f2 and v g2 are specific volumes of the saturated liquid and saturated vapor, respectively, at downstream conditions. If mass flow rate, w, is given in lb/hr, area A in in. 2, and specific volumes in ft 3 /lb, Equation 6.15(60) for the calculation of velocity, U, in fps becomes U = ( w/ A)[( 1 X) vf2 + Xvg2] 6.15(61) Similarly, if flow rate is given as q in gpm, then U in fps can be calculated by U = ( qgf/ A)[( 1 X) vf2 + Xvg2] 6.15(62) It has been the experience that velocities exceeding approximately ft/s have caused erosion in downstream pipe fittings in flashing water applications. Therefore, flashing water applications often require angle body valves and piping made out of chromium-molybdenum-steel alloys to resist erosion from the flashing water that is moving at high velocities. Sizing Example 4 A flashing application must always be sized on the basis of choked flow conditions. Piping downstream of a flashing valve should generally be larger than the valve size because it must accommodate the large increase in volume. Therefore, the corresponding combined

205 6.15 Sizing 1249 pressure recovery and piping geometry factor, F LP, must also be applied to the sizing equations. See Equation 6.15(23) and Figure 6.15g. Using the same configuration and conditions as in Example 1, except that P 2 in this case is 20 psia, determine the valve size and required C v. (The other conditions in Example 1 were: process fluid was water at 250 F, P v = 30 psia, G f = 0.94, P 1 = psia, and q = 500 gpm.) Step 1: Calculate actual p. p = = psi Step 2: Calculate P ch and determine P a. From Example 1, F F = Assuming that an angle valve with a characterized cage and flow over the seat (flow to close) is the preliminary selection for this flashing application, from Table 6.15c one can obtain the values F L = 0.80, F d = 1.0, and C v /d 2 = 12. From Equation 6.15(11) p = 2 FL P ch FFP = 2 ( = 1 v) 0. 8 [ ( 0. 93)( 30)] psi p p, therefore p = p ch < a ch = psi Step 3: Determine specific gravity of liquid at upstream conditions. This was determined in Example 1; G f = Step 4: Calculate the approximate required C v assuming F P and F R = 1. From Equation 6.15(34) C v q G f 12 / = = 500( 0. 94/ ) = FF p R P a Step 5: Select the approximate body size. Refer to manufacturers catalogs or, by using estimates from Table 6.15c, C v /d 2 = 12, in which case the minimum body size is 12 / d ( 35. 8/ 12) Thus, a 2-in. body size is selected. Step 6: Determine F LP ; d = 2.0 in.; D 1 = D 2 = 4.0 in.; d/d 1 = 0.5. From Equations 6.15(12 22) and Example 1: K 1 = ; K B1 = ; K i = = ; C d = 35.8/(2) 2 = Using Equation 6.15(23): F = 1/ F + K Cd / 890) 2 2 1/ 2 LP L i 2 2 1/ 2 = [ ( )( 89. 5) / 890] = Step 7. Calculate the final required C v from Equation 6.15(30). C = ( q / F ) G /( P F P) v max LP f 1 F v = 500/ /( ( 0. 93)( 30)) = Therefore, the 2-in. valve size is adequate and the capacity required for the application is within the estimated limit. Example 5 Given the flashing conditions described in Example 4, determine the mass fraction of vapor at the valve outlet. Step 1: Determine the enthalpies of saturated liquid at upstream conditions and of saturated vapor and liquid at downstream conditions using the steam tables. From saturated steam temperature tables at 250 F for upstream conditions: h f1 = BTU/lbm. From saturated steam pressure tables at 20 psia for downstream conditions: h f2 = BTU/lbm; h g2 = BTU/lbm; h hfg2 = BTU/lbm. Step 2: Using the assumption that the process is adiabatic and therefore the enthalpies immediately upstream and downstream of the valve are equal, Equation 6.15(59) can be used to calculate the vapor fraction, X, or steam quality downstream: X = ( hf1 hf2)/ hfg2 = ( )/ = Therefore, the amount of fluid mass flashing to steam is 2.3%. Example 6 Given the flashing case in Examples 4 and 5, estimate the velocity of the two-phase mixture at the valve outlet. Step1: From the steam tables, find the specific volumes of saturated liquid and vapor at downstream conditions. v 3 3 = ft /lbm; v = ft /lbm f12 g2 Step 2: Determine vapor fraction or steam quality at downstream conditions. For this example, the vapor fraction, X, was found in Example 5. X = Step 3: Find specific gravity (or density) of the mixture and the cross-sectional area at the valve outlet. From Example 4, G f = Flow area is the outlet area of a 2-in. ANSI Class 300 valve: A = π ( 200. ) / 4= 314. in. 2 2 Step 4: Calculate velocity using one of Equations 6.15(60 62). Selecting Equation 6.15(62), because the flow rate is in units of gpm and area in square inches, U = [ 20 qg / A]( 1 X) v + Xv ] f f2 g2 = [( 20)( 500)( 0. 94)/ 3. 14][( )( ) + ( ) ( )] = 1432 ft/s Evaluation: This is not a realistic velocity since the actual velocity cannot exceed the speed of sound (about 1300 ft/s). As a result of such high velocity, the pressure at the valve outlet will rise, which in turn will reduce the vapor content

206 1250 Control Valve Selection and Sizing until an equilibrium is established at sonic velocity. The result will be choked flow in the pipe. At that point, the back-pressure in the pipe can cause cavitation in the valve and in the pipe. Wet steam at 1300 ft/s can be erosive. An angle valve with hardened trim exhausting it into either a long, straight run of pipe or directly into a receiver vessel should be used in order to avoid erosion of downstream pipe and fittings. Another alternative would be to use an expanded outlet with a venturi outlet on the valve to reduce the velocity before it exits into the pipe. A venturi outlet with the same exit area as the 4-in. downstream pipe would reduce velocity to approximately 360 ft/s. These velocities are estimates only because the assumption of thermodynamic equilibrium was used to determine fluid properties. Note: The longer the downstream pipe and the more elbows there are in your piping system, the more the valve s back-pressure will increase, and with it cavitation will also rise. Laminar or Viscous Flow The proportionality between flow rate and the square root of pressure drop holds true only for fully turbulent Newtonian fluid behavior. Non-Newtonian fluids include most polymers and many other fluids. If no experimental sizing data exist, one might specify a line size valve and purchase a number of reduced C v trims for it, so that the final choice of trim would be determined by trial and error. Laminar or transitional flow regimes may result from very low flows, small valves or passages, low pressure differentials, and high viscosity. The valve Reynolds number, Re v, is calculated to determine the effect of laminar or transitional flow behavior on the required valve C v. Reynolds Number Factor (F R ) In the majority of control valve applications the flow velocities inside the valve are high, which in turn causes the Reynolds number to be high, and therefore turbulent conditions exist. Under turbulent conditions pressure drop relates to the square of flow. On the other hand, when the viscosity of the process fluid is high or when the size of the valve and the flow velocity through it are low, the Reynolds number will also be low and laminar flow conditions will exist. Under laminar conditions pressure drop linearly relates to flow. Therefore, the same increase in pressure drop results in a larger increase in flow than it would under turbulent conditions. Here the flow coefficient under laminar or transitional flow becomes C v = q/f R ( P/G f ) (63) The valve equivalent Reynolds number, Re v, is defined as Re v NFq 2 d ( FC L d) 4 = + / / νf C N L v 2 14 / 6.15(64) (See Table 6.15i for the values of constants N 2 and N 4 for various units.) F R = (Re v ) 0.5 /F L (C v /d 2 ) 6.15(65) When the valve operates in the laminar regime and the diameter (d) is in inches, F R = 1 + (0.06 (C v /d 2 ) 0.5 F L 0.5 ) log (Re v /10,000) 6.15(66) Note: When operating in the transitional regime, use the smaller of the two F R values calculated by using Equations 6.15(65) and 6.15(66). If the diameter (d) is given in millimeters, the F R numbers are determined by using Equations 6.15(67) and 6.15(68). For the laminar regime use Equation 6.15(67): F R = (Re v ) 0.5 /F L (C v /d 2 ) 6.15(67) For the transitional regime use Equation 6.15(68): F R = 1+ (1.53(C v /d 2 ) 0.5 F L 0.5 )log (Re v /10,000) 6.15(68) The curves provided in Figure 6.15m relate the valve Reynolds number (Re v ) to the Reynolds number factor (F R ). Example 7 Assuming that in a particular application C v = 10, d = 25 mm, F L = 0.9, Re v = 500, the F R in the transitional Valve Reynolds number factor-f R ,000 10,000 Valve Reynolds number-re v FIG. 6.15m The valve Reynolds number (Re v ) can be corrected as a function of the Reynolds number factor (F R ) and of the type of valve used (if the diameter d is in inches). 8 Curve No. 1 applicable for globe style valves C d 10. Curve No. 2 applicable for globe valves and eccentric rotary plug valves, C d = Curve No. 3 applicable for butterfly valves, C d = Curve No. 4, applicable for ball valves C d 25.

207 6.15 Sizing 1251 regime is calculated as follows: 1 + (1.53(10/625) ) log (500/10,000) = = In Figure 6.15m the straight diagonal lines extending downward from right to left, starting at an F R value of approximately 0.5, are in the laminar region and operate under the conditions where laminar flow exists. At a valve Reynolds number above 10,000, all three curves in Figure 6.15m reach an F R value of 1.0. At this number and at all higher values of Re v, fully turbulent flow conditions exist. Between the laminar region (straight diagonal lines in Figure 6.15m), and the turbulent region (over Re v = 10,000, where F R = 1.0), the flow regime is transitional (i.e., neither laminar nor turbulent). Equation 6.15(69) can be used for determining the valve Reynolds number Re v for globe valves and small flow valves. In these cases the Bernoulli correction for reducers (the bracketed term in Equation 6.15[64]) is not required. Re v = N 4 F d q/f L 0.5 C v (69) The valve-style modifier in Equation 6.15(69) is designated F d. Baumann 17 defined F d as the ratio of the equivalent hydraulic diameter(s) of a valve flow orifice to that of a circular one. For representative values please consult Table 6.15c. Valve Sizing for Viscous (Laminar) Services The following treatment is applicable to valves with or without attached fittings. The F p factor (Table 6.15c) is assumed to be 1.0, because of the absence of turbulence at the valve outlet. The first step in the valve sizing procedure is to calculate a pseudo valve flow coefficient C vt, assuming turbulent flow and using Equation 6.15(70): C vt = N (70) One may calculate the valve s Reynolds number by using either Equation 6.15(64) or 6.15(69) and then determine F R from Figure 6.15m. The required C v for the laminar or transitional flow conditions is then found by C vl = C vt /F R. The problem with this method is that it requires iteration, because one first has to guess the laminar valve C v in order to calculate the Reynolds number, and later one has to repeat this procedure if the actual C vl is too far off from the estimate. In the sizing approach below a simplified method is presented for two typical valve classes. This approach, while ignoring transitional flow (a minor error), saves the time associated with the above-mentioned iteration. Globe Valve (Example 8) For the calculations below, the values of F L = 0.9 and F d = 0.46 are assumed. In this simplified approach one would first calculate a pseudo Re vi by modifying Equation 6.15(69) by also using Equation 6.15(70) for C vt, in order to arrive at Equation 6.15(71) below: q P1 P2 G f Re vi = N4q/0.95νC vt (71) Next, the required C v is calculated under laminar flow conditions: C vl = C vt /0.022 (Re vi ) (72) Assuming that q = 4.13 gpm, ν = 31,729 cst, P = 8.41 psi, and G f = 0.977, one can, using Equation 6.15(70), calculate the C vt of Using that value in Equation 6.15(71), the Reynolds number comes out as 2. Based on Equation 6.15(72), the C v under laminar flow conditions is calculated as: C vl = 1.407/0.022 (2) = 40.3 This C v requirement can be met with a 2-in. globe-style control valve. Ball or Butterfly Valves (Example 9) For the calculations below, the values of F L = 0.6 and F d = 0.5 are assumed. Here the required C v under laminar flow conditions is C vl = C vt /0.04 (Re vi ) (73) Using the same data as in Example 8 above, but using Equation 6.15(73), we get C vl = 1.407/0.04(2) = 22. For this application, one could select a 1.5-in. ball valve instead of the 2-in. globe valve and thus obtain some cost saving. Note: If Re v is greater than 10,000, the flow may be taken as turbulent and it can therefore be assumed that F R = 1.0. In this case C vl = C vt. For butterfly valves refer to Section Small Flow Valves (Example 10) The sizing of small flow valves is more complicated, because their F d factor varies widely depending on the type of their trim. Laminar flow is likely to occur in small valves on liquid service, but at C v values under 0.1, it is also possible to have laminar or transitional flow with gases. The F d values drastically change with valve type. For V-notch-type plugs F d = 0.7, for flat seated trim F d = 0.3, and for tapered needle trim F d = 0.09 (C vl /d o ) 0.5, where d o = orifice diameter in inches. A simplified sizing equation for liquid service in small valves is given below: C vl = 0.192(N 11 G f νq/ PF d ) (74) Assuming the following process conditions and the use of a flat seated trim: q = gpm, G f = 0.87, P = 50 psi, ν = 16 cst, F d = 0.3 (flat seat). Having this information, one would first calculate the turbulent C v using C vt = q (G f / P) 0.5, which results in C vt =

208 1252 Control Valve Selection and Sizing Next C vl is calculated from 6.15(74) as C vl = ( /50 0.3) = Note that this C vl value is substantially higher than the turbulent C v of , hence it indicates that laminar flow will occur and that a valve trim with a rated C v that is above has to be selected. For small flow valves having tapered needle trims the sizing is even simpler. Here, C vl = 0.973(N 12 qg f νd o / P) (75) Using the previous example but choosing a needle trim with an orifice diameter of d o = in., C vl is calculated to be (a substantial difference from the with the flat seat trim). Sizing for Laminar Gas Flow In case of gases one should use absolute viscosity µ, because, in contrast with liquids, the kinematic viscosity varies with the absolute pressure of the gas. Sizing for the most common trim, the tapered needle, can be done as follows: One should first calculate the turbulent C v from the appropriate equation from the orientation table, Table 6.15a. Next, the laminar valve coefficient for gas service (C vlgas ) is to be determined by using the following equation, where d o, valve orifice diameter in inches, has to be obtained from the manufacturer: C vlgas = Cv T /0.0127(N g qg s /µd o ) (76) where µ = absolute viscosity, centipoises N l1 = a constant for liquids = 1 if q = gpm, P = psi for misc. trim 30.7 if q = m 3 /h, p = kpa N l2 = a constant for liquids = 1 if q = gpm, P = psi, and for tapered trim 1204 if q = m 3 /h, P = kpa, and d o = m, d o = inches N g = a constant for gas = 1 if q = gpm, and d o = inches for tapered trim 0.9 if q = m 3 /h, and d o = m Example 11 Assuming the following process conditions: q = 1.6 scfh and the gas has a specific gravity, G s, of The absolute viscosity µ is cp, P 1 = 190 psia, P 2 = 170 psia. Based on these conditions, the calculated C vt is found to be If one learns from the manufacturer that the orifice diameter d o = in., one can proceed to Equation 6.15(76) and calculate the required laminar C vl as: C vl = /0.0127( / ) 0.5 = This is 3.5 times the turbulent C v requirement! Note: One should always calculate C vt and C vl for small valves, especially when C v is below After having made both calculations, one should select the larger of the two values. GAS AND VAPOR SIZING * The flow rate of a compressible fluid varies as a function of the ratio of the pressure differential to the absolute inlet pressure ( p/p 1 ), designated by the symbol x. At values of x near 0, the equations in this section can be traced to the basic Bernoulli equation for Newtonian incompressible fluids. However, increasing values of x result in expansion and compressibility effects that require the use of appropriate correction factors. Equations for Turbulent Flow The flow rate of a gas or vapor through a valve or the required C v may be calculated by using any of the following equivalent forms of the equations, which differ only in the units used. If the flow is in mass units and the density is in specific weight units, w = N F C Y xp 6 P v 1γ 1 or w 6.15(77) Cv = NFY 6 P xp1γ 1 For volumetric flow and specific gravity units: x Q = N7FPCvY GTZ g 1 or 6.15(78) Q GTZ g 1 Cv = NFPY 7 P 1 x For mass flow and molecular weight units: Q N F C PY xm = 8 P v 1 TZ 1 or 6.15(79) w TZ 1 Cv = NFPY 8 P 1 xm For volumetric flow and molecular weight units: x Q= N9FPCvPY 1 MT1 Z or 6.15(80) q MT Z 1 Cv = NFPY 9 P 1 x Note that the numerical value of x used in these equations must not exceed the choking limit (F k x TP ), regardless of the actual value of x. * Reprinted by permission. Copyright 1985, Instrumentation, Systems, and Automation Society. From ANSI/ISA-S75.01, Flow Equations for Sizing Control Valves.

209 6.15 Sizing 1253 TABLE 6.15n Numerical Constants to be Used in Gas and Vapor Flow Equations, which Reflect the Units Being Used* Constant Units Used in Equations N W Q** P, P γ 1 T 1 d, D N N N N N kg/h kg/h lb/h kg/h kg/h lb/h m 3 /h m 3 /h scfh m 3 /h m 3 /h scfh kpa Bar psia kpa Bar psia kpa Bar psia kpa Bar psia kg/m 3 kg/m 3 lb/ft 3 K K R K K R K K R mm in. *Reprinted by permission. Copyright 1985, Instrumentation, Systems, and Automation Society. From ANSI/ISA-S75.01, Flow Equations for Sizing Control Valves. **Q is in cubic feet per hour measured at psia and 60 F, or cubic meters per hour measured at kpa and 15.6 C. Constants for Engineering Units The numerical constants, N, are chosen to suit the engineering units used in the equations. Values for N are listed in Table 6.15n. Expansion Factor (Y ) * The expansion factor Y accounts for the change in density of a compressible fluid as it passes from the valve inlet to the vena contracta and for the change in area of the vena contracta as the pressure drop is varied (Contraction coefficient). Theoretically, Y is affected by all of the following: 1. Ratio of port area to body inlet area 2. Internal geometry of the valve 3. Pressure drop ratio, x = p/p 1 4. Reynolds number 5. Ratio of specific heats, k The influence of items 1, 2, and 3 are defined by the factor x T. Test data indicate that Y may be taken as a linear function of x, as shown in the following equation (and in Figure 6.15n for a valve without attached reducers or fittings): x Y = 1 (lim its : 10. Y 067. ) 3Fx k T 6.15(81) * Reprinted by permission. Copyright 1985, Instrumentation, Systems, and Automation Society. From ANSI/ISA-S75.01, Flow Equations for Sizing Control Valves. For a valve with attached fittings, x TP should be substituted for x T. For the overwhelming majority of applications, Reynolds number effects may be disregarded when the process fluid is compressible except for small flow valves; i.e., if C v is less than The effect of the ratio of specific heats is discussed later. Choked Flow* If all inlet conditions are held constant and the differential pressure ratio (x) is increased by lowering the downstream pressure (P 2 ), the mass flow rate will increase till it reaches a maximum. Flow conditions where the value of x exceeds this limit are described as choked flow. Choking occurs when the jet stream across the crosssectional area of the vena contracta attains sonic velocity and when the vena contracta ceases to expand. Sonic velocity already occurs when P 1 /P 2 exceeds about 0.5 F L 2 P 1, while choked flow occurs at P 1 /P 2 ratios, which are higher than the ones causing sonic velocity. Therefore, x T is always higher than the sonic pressure ratio. The value of x at the inception of choked flow conditions (x T ) varies with the type of valve (Table 6.15c). It also varies with the piping geometry and with the thermodynamic properties of the flowing fluid. The factors involved are x T, x TP, and F k. Choking affects the use of Equations 6.15(77 81) in the following manner: The value of x used in the equations must not exceed F k x T or F k x TP, regardless of the actual value of x. The expansion factor Y at choked flow (x F k x TP ) is at its minimum value of 2 / 3.

210 1254 Control Valve Selection and Sizing Critical Pressure Drop Ratio Factor (x T ) For maximum accuracy, the pressure drop ratio factor, x T, must be established by using the test procedures specified in Reference 10. Representative x T values for valves are tabulated in Table 6.15c. These representative values are not to be taken as actual. Actual values must be obtained from the given valve manufacturer. When a valve is installed with reducers or other fittings, the critical pressure drop ratio factor of the assembly (x TP ) is different from that of the valve alone (x T ). For maximum accuracy, x TP must be determined by test. 10 When estimated values are permissible, the following equation (or Figure 6.15p) may be used to determine x TP : x TP x 2 T xtkc i v = F Nd P 6.15(82) In this equation, x T is the pressure drop ratio factor for a given valve installed without reducers or other fittings, and K i is the sum of the inlet velocity head coefficients (K 1 + K B1 ) of the reducer or other fittings attached to the valve inlet. Ratio of Specific Heats Factor (F k ) The ratio of specific heats of a compressible fluid affects the flow rate through a valve. The factor F k accounts for this effect. F k has a value of 1.0 for air at moderate temperatures and pressures, where its specific heat ratio is about Both theoretical and experimental evidence indicates that for valve sizing purposes, F k may be taken as having a linear relationship to k. Therefore, 6.15(83) Compressibility Factor (Z) Equations 6.15(78 80) do not contain a term for the actual specific weight of the fluid at upstream conditions. Instead, this term is inferred from the inlet pressure and temperature based on the laws of ideal gases. Under some conditions, real gas behavior can deviate markedly from the ideal. In these cases, the compressibility factor Z shall be introduced to compensate for the discrepancy. Z is a function of both reduced pressure and reduced temperature. In this section, reduced pressure P r is defined as the ratio of the actual inlet absolute pressure to the absolute thermodynamic critical pressure for the fluid in question: 5 k F = k 140. P r = P 1 / P c (84) Velocity of Compressible Fluids The velocity of sound in a compressible fluid is an important parameter. Pressure changes travel through an ideal gas at the velocity of sound. Most valve and piping systems are restricted to fluid velocities less than the speed of sound, due to the nature of the geometric discontinuities (supersonic wind tunnels being a notable exception). In valves, the throttling element and seat produce the vena contracta. As flow rate and fluid velocity increase, the effects of compressibility cause the relationship of mass flow (x) 1 / 2 to depart from a straight line. Once sonic velocity is attained in the vena contracta and the pressure ratio of x T is reached, choking occurs. Reducing the downstream pressure will not increase the flow rate or the velocity at the vena contracta. However, if local downstream geometry permits (e.g., similar to diverging nozzles), it is possible for velocity to increase downstream of the vena contracta to supersonic levels. Shock waves from supersonic velocities can produce undesirable and even dangerous vibration in process systems; use of supersonic velocities in system or valve applications should be avoided. The ratio of the actual fluid velocity, U, to the sonic velocity, c, is the Mach number, M N. For valves with standard trim M N should not exceed 0.3 at the outlet of the valve body. In order to avoid excessive noise. (See Section 6.14 for details.) M = N U / c 6.15(86) The sonic velocity can be determined from the first law of thermodynamics and the ideal gas law. Gas constants and ratios of specific heat are required in the calculation. c = gkrt 6.15(87) Flow rate for gases is usually given in terms of mass flow rate, w, or volumetric flow rate, Q. The velocity is 6.15(88) From these equations and from appropriate units conversions, the Mach number can be derived as follows: For gases (using molecular weight, M): M U = wv/ A= Q/ A M N For gases (using specific gravity, G g ): c Q = 5574 A 1 kta M 6.15(89) The reduced temperature is defined similarly. That is, T r = T 1 / T 6.15(85) Absolute thermodynamic critical pressures and temperatures for most fluids and curves from which the compressibility factor, Z may be determined can be found in many reference handbooks of physical data. 11 c For air: M M N N Q = 1035 A Q = 1225 A kta Gg T a 6.15(90) 6.15(91)

211 6.15 Sizing 1255 where For steam: M N wv = 1515 A Q = actual volumetric flow, ft 3 /hr (not scfh) A = flow area of the valve outlet, in. 2 T a = absolute temperature, R ( F + 460) w = mass flow rate, lbm/hr 6.15(92) v = specific volume, ft 3 /lbm G g = specific gravity of the fluid at standard conditions relative to air at standard conditions M = molecular weight k = ratio of specific heats The calculated exit velocity or Mach number should be compared to the limits set by the process or the valve manufacturer for the specific valve and application. Mach numbers exceeding 0.3 can cause significant noise problems. Sizing for Compressible Fluids (Example 12) The above working equations can be used in a logical sequence to determine the required valve capacity for a given application. The following example illustrates the basic steps. However, there are other considerations when selecting a valve, such as noise, erosion, corrosion, and actuator sizing, that are discussed in other sections of this chapter and in the references. For the example below, assume that the process conditions are the following: Fluid: Steam Temperature: 450 F Upstream pressure (P 1 ): 150 psia Downstream pressure (P 2 ): 65 psia Specific volume downstream (v): 7.98 ft 3 /lb Flow rate (w): 15,000 lb/hr Flow characteristic: Equal percentage Line size: 3 in., Class 600 Step 1: Select the appropriate equation based on the available process information and its units. As the information is given in mass flow units, Equation 6.15(79) can be used and N 8 can be selected from Table 6.15n as Thus, the working equation becomes T a For steam: Critical pressure (P c ): psia Critical temperature (T c ): R Molecular weight (M): Ratio of specific heats (k): 1.33 Step 2: Check for choked flow. Determine the critical pressure drop ratio, x T, from manufacturer s data or from the estimates in Table 6.16c. In order to do that, a preliminary valve-type selection is required. Assuming the use of a singleseated globe valve with contoured plug for this type of application and referring to Table 6.15c, x T = The ratio of specific heats factor, F k, must be calculated from Equation 6.15(83): F = k k/. 140 = /. = 095. Calculate the actual pressure drop ratio, x = p/p 1 : x = (150 65)/150 = Choked flow occurs when x equals or exceeds the value of F k x T or F k x TP (when the valve is installed with reducers or other fittings). The flow is not choked when x is less than F k x T or F k x TP, Fx k T = ( 0. 95)( 0. 72) = Note: If the value of x had been greater than F k x T, then the value of F k x T would replace the value of x in the calculation of Y and C v below. Step 3: Calculate the expansion factor, Y, from Equation 6.15(81) or obtain it from Figure 6.15o. Y = 1 x/( 3F x ) = 1 57/[( 3)( )] = k T Step 4: Determine the reduced pressure, P r, and the reduced temperature, T r using Equations 6.15(84 85). The compressibility factor, Z, is found from generalized compressibility charts available in reference books such as Perry and Chilton s Chemical Engineer s Handbook. 11 P = P/ P = 150 / = T r r 1 c = T/ T = 910 / = Z = 10. c Step 5: Calculate a preliminary C v using the equation identified in Step 1 and assuming a line size valve with F P = 1. C v = w FPY P 1 TZ 1 xm 6.15(93) 15, 000 C v = 19. 3( 1)( 150)( 0. 72) ( 910)( 1) ( 057. )( ) = In order to be able to use this equation, one must obtain other physical data for critical pressure and temperature, for molecular weight, and for the ratio of specific heats. Table 6.15q lists such data for a variety of gases and vapors. Step 6: Select the approximate body size based on the calculated preliminary C v. Manufacturers tables are the best source for C v values, but estimates can be made based on C d (C d = C v /d 2 ) values from Table 6.15c. Observing that

212 1256 Control Valve Selection and Sizing Y = 1 X/(3F K ) (X TP ) Ref. 6.15(81) X TP F P K i X T C d Y X TP = (X T /F 2 P )/[(X T K i C 2 d /1000) + 1] Ref. 6.15(82) 1.0 F k X X T F k F P K i X T X TP X T C d X = P/P 1 FIG. 6.15o Graph for determining the expansion factor Y. a single-ported, plug-type globe valve has a C d = 11, d = (67.7/11) = Thus, a 3-in. valve is selected for further evaluation. Step 7: Determine the effect of piping geometry. If the valve had been smaller than line size and installed with reducers, F p and x TP would need to be known or estimated. Where appropriate, F p can be estimated using Equations 6.15(12 22) or Figure 6.15f. The factor, x TP, can be estimated from Equation 6.15(82) or Figure 6.15p. Y and C v values may require recalculation using the value of F k x TP, if this product is less than x, to replace x in calculating Y and C v. In this case, the line size and the valve size are 3 in. NPS, so this step is not required. Step 8: Calculate the final required C v using the actual or estimated F P. In this example, F P = 1, so the required C v remains The body size must be verified if the C v changes at this step. Note: The Reynolds number factor, F R, is generally not considered when sizing for gases and vapors. However, in case of very small valves, with C v < 0.01, F R should be applied (when it becomes significant) in the same manner as for liquids. FIG. 6.15p Graph for determining the critical pressure drop ratio factor for the combined assembly of valve and fittings (x TP ). Step 9: Calculate valve exit velocity to evaluate the likelihood of velocity-related problems in the valve or in the downstream piping. Excessive velocity can result in noise, structural vibration, erosion, equipment damage, and personal injuries. Using Equation 6.15(92), with a flow area A = 7.07 in / M N = ( 15000)( 7. 98)/[( 1515)( 910) ( 7. 07)] = This is a high Mach number for the application; its acceptability should be verified with the manufacturer and with the process system engineer. Valve noise should be predicted and appropriate noise reduction measures taken to safeguard against hazards of excessive noise, as discussed in Section TWO-PHASE FLOW Valve sizing for two-phase flow is approximate at best, because the effects of thermal, mass, and energy transfer between phases of the mixture are not easily quantified. The methods commonly used by the control valve industry 12,13,14 will be explained here. More elaborate experimental and numerical methods may produce more accurate results for specific cases. Flow through a control valve is considered isenthalpic (adiabatic) overall and isentropic between the valve inlet and

213 6.15 Sizing 1257 TABLE 6.15q Physical Data for a Number of Common Gases and Vapors Gas Critical Pressure, P c (psia) Critical Temperature, T c ( R) Molecular Weight, M Ratio of Specific Heats, k Air Ammonia Argon Butane Carbon dioxide Carbon monoxide Chlorine Ethane Ethylene Helium Hydrogen Methane Natural gas Nitrogen Oxygen Propane Steam the vena contracta. It is also assumed for this analysis that the velocities of the liquid and gas or vapor are essentially equal at the throttling point or at the vena contracta. This discussion considers two cases of mixed-phase flow: (1) a liquid and a gas of different chemical species, and (2) a liquid and vapor of the same chemical species. Liquid-Gas Mixtures The procedure for sizing valves for liquid-gas mixtures applies to a homogeneous mixture of a nonvaporizing liquid and a noncondensible gas. As the mixture passes through the valve, the specific volume of the liquid remains essentially constant, but the volume of the gas expands, thereby increasing the effective specific volume, v e, of the mixture. 2 v = f v / Y + f v e g g f f 6.15(94) where f g and f f are the weight or mass fractions of gas and liquid, respectively, and Y is the expansion factor from Equation 6.15(81). Equation 6.15(94) expresses v e as the specific volume of an equivalent incompressible fluid, allowing use of v e in equations for incompressible fluids. The required valve capacity, F P C v, assuming F R = 1, can be obtained from rearranging Equation 6.15(36): w FC = P v N ( P P )γ (95) By substituting 1/v e for γ 1 and generalizing the equation for both choked and non-choked flow using p a as the smaller of P 1 P 2 and p ch (see Equation 6.15[11]), the valve capacity is FC w = N 6.15(96) These equations have provided the required C v values within an error of about ±10% based on limited testing on air and water mixtures in small globe valves in horizontal pipe. 15 Example 13 The goal in this example is to find the required valve capacity (F p C v ) for the conditions listed below: Air flow rate: 600 lb/hr Water flow rate: 26,000 lb/hr Upstream pressure, P 1 : 150 psia Pressure drop, p: 50 psi Temperature: 90 F (550 R) Line size: 3 in. schedule 40 p v 6 Step 1: Make a preliminary selection of valve type and determine the critical pressure drop ratio factor (x T ) for the valve. Assume a single-seated globe valve with a contoured plug with flow under the plug (to open). Using Table 6.15b, obtain an estimate of x T = (Manufacturers catalogs may also be used.) ve p a

214 1258 Control Valve Selection and Sizing Step 2: Determine the relative mass fractions of gas and liquid, f g and f f. The total mass flow rate is w = ,000 = 26,600 lb/hr. in Equations 6.15(94 96). For low vapor mixtures, while recognizing the possibility of large errors, use the effective specific volume at inlet conditions, v e, and Equation 6.15(98). f f g f = 600/ 26, 600 = = 26, 000/ 26, 600 = C v = w ve NFLP P ( F ) F 6.15(98) Step 3: Calculate the pressure drop ratio, (x), the ratio of specific heat factor (F K ), and the expansion factor (Y) x = p/p 1 = 50/150 = Because x < x T, and the gas flow is not choked, p a = p = 50 psi. From Equation 6.15(83), using k = 1.40 for air, F k = 1.40/1.40 = 1.0. Using Equation 6.15(81): x Y = 1 = /[( 3)( 1)( 0. 72)] = Fx Step 4: Determine the effective specific volume of the mixture at upstream conditions using Equation 6.15(94). The specific volume of the air can be calculated from the gas law equation of state: 6.15(97) From saturated steam temperature tables at 90 F, the saturated liquid specific volume is v f = ft 3 /lbm. The mixture effective specific volume from Equation 6.15(94) is v e = (0.0226)(1.357)/(0.846) 2 + (0.9774)( ) = ft 3 /lbm. Step 5: Calculate valve capacity from Equation 6.15(96). Liquid-Vapor Mixtures k T v RT MP = 1 g 1 ft lbf ( 1545 ) 550 lb mol R v g = lbm lbf ( 29 lb mol 2 FC P v ( R) ) ( ) ( ) = 2 in. 2 in. 3 ft lbm w ve 26, = = = p a Liquid-vapor mixtures are a more difficult, if not impossible, problem owing to the difficulty in determining the inlet vapor fraction and because of the nonequilibrium conditions prevailing as the fluid passes through the valve. Tests have shown that the flow rate of low-quality (vapor fraction X < 0.03) steam-water mixtures can be as much as ten times the amount predicted using calculations based on thermodynamic equilibrium, and about twice the flow predicted assuming no change of state between valve inlet and vena contracta. 15 Calculations produced better predictions for very highquality (vapor fraction X > 0.9) steam-water mixtures. For high vapor fraction mixtures, the same procedure is used as ft F LP is the combined piping and pressure recovery factor from Equation 6.15(23). Example 14 Let us assume that the following conditions are given for a liquid-vapor mixture in near-equilibrium conditions. Find the required valve capacity, F LP C v. Fluid: Ammonia Vapor flow rate: 200 lb/hr Liquid flow rate: 9000 lb/hr Upstream pressure, P 1 : 110 psia Pressure drop, p: 5 psi Temperature: 60 F Line size: 2 in. schedule 40 Step 1: Make a preliminary selection of valve type, and determine the valve x T. Assume a throttling (segmented) ball valve. Using Table 6.15c for an estimate, x T = Step 2: Determine the relative mass fraction of gas and liquid. The total mass flow rate is w = = 9200 lb/hr. f f g f = 200/ 900 = = 9000/ 9200 = Step 3: Calculate pressure drop ratio (x), ratio of specific heat factor (F K ), and expansion factor (Y) x = p/p 1 = 5/110 = Since x < x T, the vapor flow is not choked. From Equation 6.15(83) and Table 4.17q, using k = 1.40 for air and 1.31 for ammonia, F k = 1.31/1.40 = Using Equation 6.15(81): x Y = 1 = /[( 3)( )( 0. 25)] = Fx k T Step 4: Calculate F F from Equation 6.15(10). The critical pressure of ammonia from Table 6.15q is P c = psia. The vapor pressure from ammonia tables 11 at 60 F is P v = psia. F = ( P/ P) 12 F v c 12 = ( / 1636.) 1 = Step 5: Determine the effective upstream specific volume at upstream conditions using Equation 6.15(94). The specific volume of the ammonia vapor taken from thermodynamic

215 6.15 Sizing 1259 tables 11 is only approximate, because the vapor is not in true thermodynamic equilibrium. However, using the values for saturated ammonia provides sufficient accuracy due to the relative uncertainty of the capacity estimate. From saturated ammonia tables, 11 v g = ft 3 /lb and v f = ft 3 /lb. Therefore, one can calculate v e from Equation 6.15(94). F C LP Step 6: Calculate valve capacity from Equation 6.15(98). v CONCLUSIONS v e = ( )( )/( ) 3 + ( )( ) = ft /lb w ve 9200 = = P ( 1 F ) F Due to the uncertainties in correctly determining the process conditions, there is a tendency to be overcautious; i.e., to overestimate pressure losses in pipes and to increase the normal flow rates by safety factors. This tends to oversize the control valves. This in turn affects the installed gain of the valve and greatly reduces the turn-down, or rangeability of the valve. It is recommended that the pump heads be selected on the basis that the control valve will add no more than 10% to the pressure drop in the piping system at maximum flow. This will not only reduce the size of the pump, but will also save on the required operating energy. If this approach is taken to determine the required pump head, the maximum differential pressure across the control valve will be the pump head at maximum flow minus the sum of the pressure drop across the piping system and any altitude changes, plus 10% of the pipe friction losses. There are instances when the flow conditions are not accurately known. This might be the case if an old valve is to be replaced in a piping system and its serial number or source is not known. In such cases, the best guess approach would be to select the same type of valve having the same pressure rating as the surrounding pipe and having a flow coefficient (C v ) one size smaller than that listed for a full size orifice of a line size valve. Another important rule is to make sure that the maximum required flow capacity for the valve does not exceed 85% of the valve s maximum rated capacity. This is because there is a typical tolerance of ±10% between what is advertised as the maximum C v number and what actually is the C v of the particular valve. Similar rules apply to the minimum required flow coefficient, which is obtained by considering a process condition when the flow is the minimum and the valve pressure drop is the maximum. Once this C v(min) is determined, one has to make sure that the valve travel at the required minimum flow = ( ) coefficient is not less than 10% of the rated travel (or twice the expected dead band of the valve, whatever is higher) away from touching the valve seat. This safety margin is necessary to make sure that the plug does not touch the seat during any excursion caused by a disturbance in the control loop. If the minimum flow capacity of the larger valve is insufficient to provide the required rangeability for the loop, use two valves in parallel by split-ranging the control signal among them, as shown in Figure 6.1i. NOMENCLATURE A Area A 1 Cross-sectional area of pipe, in. 2 A 2 Area of valve opening, in. 2 a Pressure scale exponent A v Metric capacity coefficient, m 2 b C d Size scale exponent Relative valve capacity or valve discharge coefficients, C d = C v /d 2 (Table 6.15c) C v Valve capacity coefficient. Its value is 1.0 when the valve passes 1 gpm of cold water, having a specific gravity of 1.0, at a pressure drop of 1 psid. C vt Pseudo valve coefficient d Valve body port diameter, inches d/d Valve to pipe diameter ratio D 1 Upstream pipe inside diameter, inches D 2 Downstream pipe inside diameter, inches F d Valve-style modifier F F Liquid critical pressure ratio factor F f Mass fraction of liquid in a liquid-gas mixture F g Mass fraction of gas in a liquid-gas mixture F k Ratio of specific heats factor; k/1.4 F L Liquid pressure recovery factor of a valve without attached fittings F LP Combined liquid pressure recovery factor and piping geometry factor of a valve with attached fittings F P Piping geometry factor of a valve with attached fittings F R Reynolds number factor F s Laminar (streamline) flow factor g Gravitational acceleration, ft/s 2 g c Units conversion constant, (lbm-ft)(lbf-s 2 ) or 1.0 (kg-m)/(n-s 2 ) G f Specific gravity relative to water at 60 F G g Specific gravity relative to air 1 atm. and 60 F H Enthalpy K Ratio of specific heats for gases; 1.4 for air K B1 Inlet fitting Bernoulli coefficient (static pressure to velocity exchange) Outlet fitting Bernoulli coefficient K B2

216 1260 Control Valve Selection and Sizing K c Valve cavitation coefficient K I Inlet fitting combined loss coefficient K sc Cavitation index of service conditions K v Metric valve sizing coefficient, (m 3 /h)/(bar) 1/2 K 1 Inlet fitting friction loss coefficient K 2 Outlet fitting friction loss coefficient m Mass, lbm M Molecular weight M N Mach number N 1 to N 9 Numerical constants P Pressure, psia P 1 Valve inlet pressure, psia P 2 Valve outlet pressure, psia P c Critical pressure, psia PSE Pressure scale effect for cavitation P vc Pressure at the vena contracta, psia P v or P vp Vapor pressure at flowing conditions, psia P r Reduced pressure (used with compressibility curves); defined as P 1 /P c q Volumetric flow rate, gpm Q Volumetric flow rate, ft 3 /hr R Universal gas constant, 1545 (ft-lbf)/(lb-mol- R) or cal/(gm-mol- K) Re Pipeline Reynolds number Re v Valve Reynolds number SSE Size scale effect for cavitation T Temperature, F or C T a Absolute temperature, R T c Critical temperature, R or K T r Reduced temperature; defined as T 1 /T c T 1 Absolute temperature at valve inlet, R or K U Velocity, ft/s v Specific volume, ft 3 /lbm V Volume, ft 3 v e Effective specific volume of a liquid-gas mixture, ft 3 /lbm v f Specific volume of saturated liquid, ft 3 /lbm v g Specific volume of gas or saturated vapor, ft 3 /lbm W Mass flow, rate, lbm/hr X Vapor portion of a liquid-vapor mixture, related to steam quality X F zi Differential pressure ratio for incipient cavitation at actual inlet pressure. X T Critical pressure drop ratio factor (inception of choked flows) X TP Critical pressure drop ratio factor for a valve and fitting assembly Y Expansion factor for gases Z Compressibility factor H Head loss, ft. p Pressure difference (drop), psi p a Allowable sizing pressure drop (the smaller of P 1 P 2 and p ch ), psi p ch Pressure drop at onset of choked flow defined in terms of F L, psi p I Pressure drop at flow curve inflection point where C v begins to decrease due to cavitation γ (gamma) Weight density, lbm/ft 3 γ 1 Specific weight at upstream conditions, lbm/ft 3 µ (mu) Absolute viscosity, centipoise = 10 3 N sec/m 2 ν (nu) Kinematic viscosity, centistokes = 10 6 m 2 /sec σ (sigma) Cavitation index at flowing conditions σ c Valve coefficient for constant cavitation σ I Valve coefficient for incipient cavitation σ id Valve coefficient for incipient damage cavitation σ mv Valve coefficient at maximum vibration cavitation σ R Recommended operating cavitation coefficient for a valve at a reference pressure σ v Valve cavitation coefficient adjusted for pressure and size scale effects References 1. Baumann, H. D., The Introduction of a Critical Flow Factor for Valve Sizing, ISA Transactions, Vol. 2, Driskell, L. R., Control Valve Selection and Sizing, Research Triangle Park, NC: Instrumentation, Systems, and Automation Society, Baumann, H. D., Effect of Pipe Reducers on Control Valve Capacity, Instruments and Control Systems, December McCutcheon, E. B., A Reynolds Number for Control Valves, Symposium on Flow, Its Measurement and Control in Science and Industry, Vol. 1, Part 3, Baumann, H. D., A Unifying Method for Sizing Throttling Valves under Laminar or Transitional Flow Conditions, Journal of Fluids Engineering, Vol. 115, No. 1., pp , March Allen, W. F., Jr., Flow of a Flashing Mixture of Water and Steam through Pipes and Valves, Journal of Basic Engineering, April Singleton, E. W., Multistage Valve Sizing Made Easy, InTech, August 1997, pp George, J. A., Evolution and Status of Nonturbulent Flow Sizing for Control Valves, Research Triangle Park, NC: Instrumentation, Systems, and Automation Society, ANSI/ISA S75.01, Flow Equations for Sizing Control Valves, Research Triangle Park, NC: Instrumentation, Systems, and Automation Society, Revised ANSI/ISA S , Control Valve Capacity Test Procedure, Research Triangle Park, NC: Instrumentation, Systems, and Automation Society, Perry, R. H. and Chilton, C. H., Chemical Engineer s Handbook, 5th edition, New York: McGraw-Hill Book Company. 12. Masoneilan Handbook for Control Valve Sizing, Masoneilan-Dresser, Catalog 10, Technical Information, Fisher Controls Company. 14. Monsen, J. F., Spreadsheet Sizes Control Valves for Liquid/Gas Mixtures, InTech, December Hutchison, J. W. (ed.), ISA Handbook of Control Valves, 2nd edition, Research Triangle Park, NC: Instrumentation, Systems, and Automation Society, 1976.

217 6.15 Sizing Kiesbauer, J. and Baumann, H. D., Recent developments in the prediction of hydrodynamic noise of control valves, Valve-World, February Baumann, H. D., Determination of Peak Internal Sound Frequency Generated by Throttling Valves for the Calculation of Pipe Transmission Losses, Noise Control Engineering Journal, Vol. 22, No. 1, pp , March April, Cain, F. M. and Barnes, R. W., Testing for Cavitation in Low Pressure Recovery Control Valves, ISA, Volume 25, No. 2, Tullis, J. P., Hydraulics of Pipelines: Pumps, Valves, Cavitation, Transients, New York: John Wiley, Rahmeyer, W., Cavitation Testing of Control Valves, Instrumentation, Systems, and Automation Society, Paper No. C.I. 83-R931, presented at International Conference in Houston, TX, October Riveland, M. L., The Industrial Detection and Evaluation of Control Valve Cavitation, Instrumentation, Systems, and Automation Society, Paper No. C.I , presented at International Conference in Philadelphia, PA, October Cain, F. M., Solving the Problem of Cavitation in Control Valves, Advances in Instrumentation and Control Proceedings of the ISA/91 International Conference, Vol. 46, Part 2, 1991, Paper # , pp Bibliography Baumann, H. D., Control Valve Primer, 3rd edition, Research Triangle Park, NC: Instrumentation, Systems, and Automation Society, Cain, F. M., Solving the Problem of Cavitation in Control Valves, Advances in Instrumentation and Control Proceedings of the ISA/91 International Conference, Vol. 46, Part 2, 1991, Paper # , pp Catalog 10, Technical Information, Fisher Controls Company. ANSI/ISA S , Control Valve Capacity Test Procedure, Research Triangle Park, NC: Instrumentation, Systems, and Automation Society, Driskell, L., Control-Valve Selection and Sizing, Research Triangle Park, NC: Instrumentation, Systems, and Automation Society, Revised ANSI/ISA S75.01, Flow Equations for Sizing Control Valves, Research Triangle Park, NC: Instrumentation, Systems, and Automation Society, Handbook on Control Valve Sizing, Baumann Inc., Portsmouth, NH. Hutchison, J. W. (ed.), ISA Handbook of Control Valves, 2nd edition, Research Triangle Park, NC: Instrumentation, Systems, and Automation Society, ISA S20 Specification Forms for Process Measurement and Control Instrument, Primary Elements and Control Valves, Research Triangle Park, NC: Instrumentation, Systems, and Automation Society, IEC Standard , Flow Capacity: Sizing Equations for Fluid Flow under Installed Cconditions, Geneva, Switzerland: International Electronical Commission, September Masoneilan Handbook for Control Valve Sizing, 7th edition, Masoneilan- Dresser, Monsen, J. F., Spreadsheet Sizes Control Valves for Liquid/Gas Mixtures, InTech, December Morse, P. M. Thermal Physics, W.A. Benjamin, Inc., Perry, R. H. and Chilton, C. H., Chemical Engineer s Handbook, 5th edition, New York: McGraw-Hill Book Company, Rahmeyer, W., Cavitation Testing of Control Valves, Instrumentation, Systems, and Automation Society, Paper No. C.I. 83-R931, presented at International Conference in Houston, TX, October Rahmeyer, W. and Driskell, L., Control Valve Flow Coefficients, Journal of Transportation Engineering, Vol. 111, No. 4, ASCE Pipeline Division Paper No , Riveland, M. L., The Industrial Detection and Evaluation of Control Valve Cavitation, Instrumentation, Systems, and Automation Society, Paper No. C.I , presented at International Conference in Philadelphia, PA, October Shinskey, F. G., When You Have the Wrong Valve Characteristic, Instruments and Control Systems, October Technical paper No. 410, Crane Company, Tullis, J. P., Hydraulics of Pipelines: Pumps, Valves, Cavitation, Transients, New York: John Wiley, Valtek Control Valve Sizing and Selection, Flowserve Inc. (Valtek Inc.), 1992.

218 6.16 Valve Types: Ball Valves C. E. GAYLER F. D. MARTON (1970, 1985) B. G. LIPTÁK (1995) H. L. MILLER (2005) Standard ball control valve Three-way ball control valve Conventional Full-ported ball control Characterized Cage Flow sheet symbols Types of Ball Valves: Size and Design Pressure: Design Temperature: A. Conventional B. Characterized C. Cage A. 1 /2 to 42 in. (12 to 1180 mm) in ANSI Class 150; to 12 in. (300 mm) in ANSI Class 2500 B. Segmented ball 1 to 24 in. (25 to 600 mm) in ANSI Class 150; to 16 in. (400 mm) in ANSI Class 300; to 12 in. (300 mm) in ANSI Class 600 C. 1/4 to 14 in. (6 to 350 mm) C. Up to ANSI Class 2500 A. Varies with size and material, typically from 250 to 600 F ( 155 to 315 C), with special designs available from 300 to 1800 F ( 185 to 1020 C) B. From 50 to 300 F ( 45 to 150 C); special units available from cryogenic to 1000 F (540 C) C. From 425 F to 1800 F ( 255 C to 980 C) Rangeability: Refer to Section 6.7; generally claimed to be about 50:1 Characteristics: See Figure 6.16a Capacity: A and B. Standard ball: C v = 30 d 2 to C v = 45 d 2 ; segmented ball: C v = 24 d 2 to C v = 30 d 2 ; full bore ball: C v = 35 d 2 to C v = 100 d 2 ; see Table 4.3c for details C. C v = 20 d 2 (noncritical flow) Materials of Construction: A. Body: Cast or bar stock brass or bronze, carbon steel, stainless steel, ductile iron, aluminum, Monel, titanium, Hastelloy C, plastics, glass; also hafnium-free zirconium (for nuclear applications) and ceramic for abrasives. Ball: Forged naval bronze, carbon steel (also plated), stainless steel, plastics, glass, ceramics, Alloy 20, Monel, Hastelloy C, aluminum, titanium. Seats: Teflon, Kel-F (both tetrafluoroethylene), Delrin, buna-n, neoprene, Perbunan, Hypalon, natural rubber, graphite. 1262

219 6.16 Valve Types: Ball Valves 1263 B. Body, ball, seal ring, and shaft are available in 316 stainless steel. Chrome and tungsten carbide plating available for ball and carbon steel for valve bodies. Allceramic valves are also available. C. Stainless steel, other materials available. Leakage: A. ANSI V (see Section 6.1). B. Metal seats, ANSI IV to VI; composition seats, ANSI V or better. Costs: A. Low-cost 1 /2 to 1 in. PVC valves with on/off actuators can be obtained for $300 or with positioned throttling actuator for about $1200. For process control-quality valve costs, see Figure 6.16b below. Cost data are based upon ANSI Class 150 flanged bodies with single-acting, spring-return actuator for on/off service. Other operators can be furnished and positioners added for throttling service. Ball and stem materials are 316 SST for both carbon and stainless steel bodies with TFE seat seals. Other alloys are available at higher cost. B and C. See Figure 6.16b for on/off; modulating with metal seats for high temperature will be three to four times higher. Special Features: Partial List of Suppliers: A. Full-ported, three-way, split body, two-directional. B. Depending on contour edge of ball the flow characteristics vary slightly between suppliers. Slurry design provides for continuous purging of low-activity zone of valve to prevent build-up of solids, dewatering, or entrapment. C. Good resistance to cavitation and vibration. A Plus Development Co., Ltd. (A) ( Actuation Valve & Control Ltd. (A) ( Armstrong International Inc.(A) ( Assured Automation Inc. (A) ( Bardiani Valvole SpA (A) ( China Zhejiang Chaoda Valve Co. Ltd ( Circor Int., Inc. (Circle Seal controls) (A) ( Cole-Parmer Instrument Co. (A) ( Combraco Industries Inc. ( Control Components Inc. ( Cooke Vacuum Products Inc. (A) ( Cooper Cameron Valves (A) ( Crane Valve Group (Stockham, Xomox) (A) ( Cyclonic Valve Co. Inc. (A) ( Dresser (Masoneilan) ( Derex Company (C) ( Emerson Process Management (Fisher) ( Eurovalve s.r.l.(a) ( Flowdyne Controls, Inc. (A) ( Flowserve Corp (Anchor/Darling, Valtek, Worcester, McCanna,) (A, B) ( com) Fujikin of America (A) ( Hartmann KG (A) ( Hoke Inc. (A) ( Honeywell (A) ( ITT Industries, Engineered Valves (A) ( Kitz Corporation of America (A) ( Marpac Inc. (A) ( Metso Automation (Jamesbury) (A,B) ( Milwaukee Valve Co. (A) ( Mogas Industries Inc. (A, B) ( Nibco Inc. (A) ( Nihon Koso (A,B) ( Nordstrom Valves Inc. (A) ( Oliver Valves Inc. (A) ( Parker Hannifin Corp. (A) (

220 1264 Control Valve Selection and Sizing PBM Inc.(A) ( Plast-O-Matic Valves, Inc. (A) ( Power and Pumps, Inc. (A) ( Spirax Sarco Ltd. (A) ( SPX Valves (Daniel, Dezurik) (A,B) ( Tyco Flow Control (A) (MCF) ( Valvtechnologies Inc. (A) ( Velan Valve Corp. (A) ( Voss/Europower Inc. (A) ( Watts Regulator Co. (A) ( Wier Valves and Controls (A) ( Zurn Industries (A) ( INTRODUCTION The rotary ball, butterfly, and plug valves, which in the past were considered only as on/off shut-off valves, are extensively used today as control valves. Relative to the traditional globe valve, their advantages include their lower cost and weight and higher flow capacity (two to three times that of the globe valve, as listed in Table 6.16c). Other features, such as tight shutoff, fire-safe designs, and low stem leakage, make it easier to meet governmental regulatory requirements from OSHA and EPA in the United States and the Pressure Equipment Directive (PED) in the EEC. Some ball valve designs, such as the characterized ball valve, also provide a nearequal-percentage characteristic. For a comparison of ball and globe valves in terms of their features and performance, see Reference 1. Throttling Ball Valves When used for throttling service, some of their disadvantages are a direct consequence of the above-listed advantages. Their % flow or Cv Conventional Ball & cage (% rotation) V-notch U-notch Parabolicnotch Characterized ball valves FIG. 6.16a The characterized ball valve with a parabolic notch is near-equalpercentage, while the ball-and-cage valve characteristics are closer to linear, when used on water service. On gas service at critical velocities, the characterized ball valve s performance lines move closer to linear. $ 21,000 18,000 15,000 12, (25) 2 (50) 3 (75) 316 SST body Carbon steel body (100) (125) (150) (175) (200) (225) (250) (275) (300) Valve size-inches (mm) FIG. 6.16b The costs of conventional ball valves with on/off actuators are provided above. For characterized ball valves add 10%, and for cage-type ball valves add 20% to these costs. high capacity results in either using oversized valves or installing small valves in large pipes. This means a substantial waste of pumping energy caused by the reducer pressure drops. Also, the characteristic of a high-pressure recovery results in low vena contracta pressures. This in turn increases the probability of cavitation and noise. Reduction of these problems has been achieved by adding perforated parallel plates in the ball valve openings (see Figure 6.16o) or by adding other tortuous path features (see Figure 6.16p) to the initially fully open flow path. These design improvement have added measurably to the range of troublefree applications. In a ball valve, critical flow occurs when the pressure drop through the valve rises to about 15% of the inlet pressure to the valve. In operating rotary valves, the linear movement of cylinder- or spring/diaphragm-type actuators must be converted by linkages, which introduces hysteresis and dead play. In addition, a nonlinear relationship exists between actuator movement and the resulting rotation. These considerations make the use of positioners essential, which on fast processes can lower the quality of control.

221 6.16 Valve Types: Ball Valves 1265 TABLE 6.16c Valve Sizing Coefficients (C v ) of Full Ported and Reduced Ported Ball Valves* Full Bore Valves, Ball Opening (%) Valve Size / / / / ,150 1, ,188 1,890 2,700 4, ,100 1,500 2,200 3,500 5,000 7,600 10, ,280 1,760 2,400 3,520 5,600 8,000 12,160 16, ,080 1,920 2,640 3,600 5,280 8,400 12,000 18,240 24, ,413 2,512 3,454 4,710 6,908 10,990 15,700 23,860 31, ,935 3,440 4,730 6,450 9,460 15,050 21,500 32,680 43, ,140 2,565 4,560 6,270 8,550 12,540 19,950 28,500 43,320 57, ,460 3,285 5,840 8,030 10,950 16,060 25,550 36,500 55,480 73,000 Reduced Bore Valve, Ball Opening (%) Valve Size / / , ,190 1,700 2, ,298 2,065 2,950 4, ,200 1,760 2,800 4,000 6, ,320 1,800 2,640 4,200 6,000 9,120 12, ,120 1,540 2,100 3,080 4,900 7,000 10,640 14, ,440 1,980 2,700 3,960 6,300 9,000 13,680 18, ,760 2,420 3,300 4,840 7,700 11,000 16,720 22,000 *Courtesy of KTM Products, Inc. The torque characteristics of these valves are also highly nonlinear (Figure 6.4v), and because of the high breaktorque requirement, the actuator is usually oversized for the operation in the throttling range. Some pneumatic piston actuators have a torque characteristic that has a high torque delivery at the closed position, therefore allowing a better matching of the torque needs (see Figure 6.4w). The ball valve contains a spherical plug that controls the flow of fluid through the valve body. The three basic types of ball valves that are manufactured are (1) the conventional or quarter-turn pierced ball type, (2) the characterized type, and (3) the cage type. CONVENTIONAL BALL VALVES The quarter-turn (90 ) required to fully uncover or fully cover an opening in the valve body can be imparted to the ball either

222 1266 Control Valve Selection and Sizing Body Stem Body Seat Ball FIG. 6.16f The design of ball valve seats. configured as two-way, three-way, or split-body. Figure 6.16e illustrates some of the available multiport configurations. Ball FIG. 6.16d Top-entry pierced ball valve. Seal manually by turning a handle, or mechanically by an automatic valve actuator. Actuators used for ball valves may be the same as those used to control other valve types. As was discussed in Sections 6.3 and 6.4, they can be pneumatic, electric (including electronic or digital), hydraulic, or a combination. The latter types include electropneumatic, electrohydraulic, electromechanical, and pneumohydraulic actuations. Most of the ball valves on the market are available with builtin, or integral, valve actuators. The valves are designed so that they can be used with or without an actuator, and they can be fitted with other manufacturers actuators. The spherical plug lends itself not only to precise control of the flow through the valve body but also to tight shut-off. Thus the ball valve may assume the double role of control and block valve (Figure 6.16d). Special materials used for valve seats help achieve these functions. Ball valves are available with the features listed in the front of this section. Their tight shut-off characteristics correspond to ANSI Class IV and VI. The valve body can be The Valve Trim The ball in a ball valve is cradled by seats on the inlet and the outlet side. The seats are usually made of plastic and are identical on both sides, especially in double-acting valves. Tetrafluoroethylene materials are preferred for seat materials, because of their good resilience and low-friction properties (Figure 6.16f). In some valve designs the plastic seats are backed up by metallic seats in order to ensure tightness in the event that the soft seat gets damaged by high temperature, such as in a fire (Figure 6.16g). Such precautions are imperative on shipboard, for nuclear installations, and in cryogenic applications. The fire-safe design of ball valves can be certified to API- 607, which specifies the types of secondary seats that are acceptable to control the leaking of flammable fluids, when the primary seat (usually PTFE) sublimes during fire. The secondary metal seats of the fire-safe designs are also useful on erosive or abrasive services and on saturated steam service. Body Seat Ball Body seal Tee ports Angle ports Fire-safe insert FIG. 6.16e Porting arrangements of various multiport ball valve designs. FIG. 6.16g Fire-safe ball valve design.

223 6.16 Valve Types: Ball Valves 1267 Back-up ring Retainer Pressure allows seal to flex into ball Back-up ring Retainer Metal seal Ball Body Back-up ring Retainer Bi-directional flex-loc seal Metal seal Soft seal Body Ball Soft seals Retainer Metal seal Body Back-Up ring locks metal seal to ball Ball Soft seal Body FIG. 6.16h Ball valves can be sealed by using either flexible metal seals or soft seals. (Courtesy of Valtek Inc.) Some seat designs are such that the seat is always under compression, which allows the use of such nonresilient seat materials as carbon graphite. Seats and Their Maintenance Other seat designs utilize the flexing of metal seals or soft PTFE seats, usually backed up by stainless steel or Inconel metal seals (Figure 6.16h). Where fluids of high temperature are handled, graphite seats are recommended. They hold tight up to 1000 F (540 C). There have been significant advancements in the manufacturing processes that allow the use of primary metal seats in direct contact with the ball (see Figure 6.16i). Such designs Ball can deliver better than ANSI Class V performance in terms of their seat leakage. 1 The ball valves are designed so that lubrication is unnecessary and the torque required to turn the ball is negligible. Both upstream and downstream seats of the pierced ball can sometimes be freely rotated in order to reduce wear. In some designs the seats are forcibly rotated a fraction of a turn with each quarter turn of the ball. Thus seat wear, which is concentrated at the points where the flow begins or ends on opening or closing of the valve, is distributed over the periphery of the seat. To facilitate cleaning or replacing worn seats in some designs the whole seating assembly is made in the form of a tapered cartridge. If the valve has top-entry design, the cartridge can be removed without disturbing the valve arrangement. O-rings usually close off stem and seats, and thrust washers made of tetrafluoroethylene compensate for axial stem thrust due to line pressure and reduce stem friction to a minimum. Some seats are preloaded by springs or are made tapered for wear compensation and leak-tight closure. Ball and stem are often machined from one piece. Other designs use square ends on the stem to engage in square recesses of the ball. In this case, the ball is made floating in fixed seats, while other designs provide a fixed location of ball and stem through the application of top and bottom guiding and ball bearings. Balls are subject to wear by friction. Where long life and dead-tight closures are of paramount importance, the design is recommended that provides for lifting the ball off its seat before it is turned. This measure also prevents freezing or galling. Liftoff is achieved by mechanical means such as an eccentric cam. Valves of this design facilitate the handling of slurries and abrasive fluids, and they can be used for high pressures. The proper materials for body and trim depend on the application. For handling chemicals or corrosive fluids, all wetted parts will possibly require stainless steel, plastics, or glass (borosilicate glass is preferred for impact strength). Ball valves are made with the same connections as used in all other valve types. Where screwed connections must be used, valves with ends that take the place of unions are preferred. Flow Characteristics FIG. 6.16i Contoured metal ball seal. (Courtesy of Metso Automation.) The flow characteristics of a ball valve approximate those of an equal-percentage plug (Figure 6.16a). These characteristic curves compare favorably with those of other rotary-stem valves. The flow path through a ball valve includes two orifice restriction locations (Figure 6.16j). Balls characterized either by a notch or by a noncircular bore give somewhat better characteristics (Figure 6.16a). Critical flow in ball valves is encountered at P = 0.15P 1, far below the usual figure of 50% of absolute inlet pressure. Sizing of ball valves proceeds along the lines described in Section 6.15, with the possible exception that due to the

224 1268 Control Valve Selection and Sizing Orifice #2 essentially straight-through flow feature, a ball valve can be chosen whose size is equal to the nominal pipe size, which usually is an advantage on slurry service. This can only be done if an oversized valve can be tolerated, which also implies the acceptance of reduced sensitivity and increased cost. The low-pressure loss and high-pressure recovery of a ball valve must also be considered in the calculations. The application of ball valves for control requires some caution. Where a lot of noise is encountered some have found it necessary to bury the valve. Such may be the case with natural gas flowing at high speed. Valves handling liquids could cavitate and erode the pressure boundary as well as produce substantial noise. On the other hand, the use of ball valves to control liquid oxygen in the experimental X-15 air-space vehicle as early as 1961, and later in the Atlas rocket, attests to the precise controllability with ball valves. A special valve with a dual-ball design for mixing liquid oxygen with liquid ammonia in precise proportions has also been used in the Atlas rocket system. CHARACTERIZED BALL VALVES Orifice #1 FIG. 6.16j Two restricted orifice locations are formed when a ball valve is used for throttling the flow. The characterized ball valves can be V-notched, U-notched, parabolic, and anticavitation-antinoise designs. The notched trim valves and the partial-ball trim valves were introduced partially in an effort to solve the problem of valve clogging and dewatering in paper stock applications. Since then these valves have come into more widespread use as a result of increased valve rangeability and the shearing action at the sharp edges of the valve as it closes. In essentially all characterized ball valves, the ball has been modified so that only a portion of it is used (Figure 6.16k). The edge of the partial ball can be contoured or shaped to obtain the desired valve characteristics. The V-notching of the ball in Figure 6.16k serves this purpose as well as the purpose of shearing the process stream. This shape or contour of the valve s leading edge is the main difference between the various manufacturers products. The ball is usually closed as it is rotated from top to bottom, although this action can be reversed. FIG. 6.16k The open, throttling, and closed positions of the characterized ball valve. Construction Mechanically, the characterized ball valves are very similar to their ancestor, the conventional full ball designs. However, because of the asymmetrical design, the characterized ball valve has some design problems that are not significant with the conventional ball valves. A typical characterized ball valve is shown in Figure 6.16l, in its end and side views. The main parts of a characterized ball valve are described below. The controlling edge of the ball can be notched or contoured to produce the desired flow characteristics. Characterized ball valves are available as U-notched, as V-notched, and as a parabolic curved designs. Mechanically, this part can create problems by bending under pressure and thus introducing movement into the shaft seals. Also, the stub shafts can be distorted by the bending of the partial ball under operating loads. Early bodies were not designed for high-pressure services or for installations other than insertion between flanges. Today they are available with up to 12 in. (300 mm) flanges with up to ANSI Class 600 ratings. The seal ring and seal-retaining ring are usually held in place by companion flanges. Damage due to overtightening of flange bolts sometimes occurs. Figure 6.16m illustrates a special sealing arrangement useful in slurry applications, due to the purging effect created by the flow into the otherwise low-activity zone, through the indent in the ball plug. Stub shaft Bearings Open Throttling Closed Characterized ball Body Stub shaft Bearings Seal rings Body FIG. 6.16l The component parts of a characterized ball valve. Characterized ball

225 6.16 Valve Types: Ball Valves 1269 Body Shims Back up ring Seal ring Seal ring retainer Indent Characterized plug Note: Seal ring is initially preloaded against plug face. Service pressure tends to increase seal ring loading. FIG. 6.16m Special seal ring arrangement, where service pressure increases seal ring loading. Characteristics Inlet The flow characteristics are dependent upon the shape of the edge of the partial ball and on the installed flow direction. The shape of the V-notch at the edge of the valve varies from concave for small openings to convex for large openings. Figure 6.16n illustrates this characteristic together with the corresponding shapes for the parabolic ball valves. The flow characteristics for parabolic, U-notched, and V- notched valves are given in Figure 6.16a. These curves are based on water flow and are also applicable to compressible fluid flow at less than critical (choking) velocities. If the characteristics were evaluated using compressible fluids at critical velocities, these curves would be flatter, closer to linear. The characteristics of the conventional ball valves are also modified by the anticavitation and antinoise designs. One such design approach is illustrated in Figure 6.16o. Here the attenuator is placed inside the ball, so that when the valve is throttling, the fluid has to pass the attenuators, creating a number of pressure drop stages. The size, location, and distribution of V-notch Parabolic FIG. 6.16n Shapes of throttling areas of the V-notched and U-notched characterized ball valves. FIG. 6.16o Internal attenuation plates reduce noise/cavitation and can also modify the throttling characteristics of ball valves. 1 perforations on the attenuator plates can be modified to obtain changes in the valve characteristics. These valves can also handle impurities in the process fluid. Another approach to the anticavitation, antinoise designs is to add a multistep tortuous trim feature in the ball so that, because of the small openings, the fluid velocity is maintained below damaging levels. Such a design having 16 discrete stages of pressure drop is shown in Figure 6.16p. This design provides a rangeability of up to 300:1 and operating noise levels of less than 75 dba. The flow passages are designed to continually expand so that any solids that enter the trim will pass through and solids that are blocked will be swept through the valve at larger openings. The anticavitation and antinoise designs are particularly useful in eliminating the need for low-flow bypass systems, which have been used in the start-up and shut-down of many gas pressure regulation applications. BALL AND CAGE VALVES Positioning of a ball by a cage, in relation to a seat ring and discharge port, is also used for control (Figure 6.16q). This valve design consists of a venturi-ported body, two seat rings, a ball that causes closure, a cage that positions the ball, and a stem that positions the cage. Seat rings are installed in both inlet and discharge, but only the discharge ring is active. The body can be reversed for utilization of the spare ring. The cage rolls the ball out of the seat as it is lifted by the stem, positions it firmly during throttling, and lifts it out of the flow stream for full opening (Figure 6.16r). The cage is contoured for unobstructed flow in the open position. Cage design includes four inclined control surfaces. The two surfaces next to the downstream seat lift the ball out of the seat and roll it over the top edge of the seat ring as the valve is opened.

226 1270 Control Valve Selection and Sizing Flow FIG. 6.16q The ball is positioned by a cage in this valve design. Sizes and Other Features FIG. 6.16p Multistage low-flow control valve design with anticavitation and antinoise capability. (Courtesy of Control Components Inc.) As the valve opens farther, the ball rolls down the first two inclined surfaces to the center of the cage to rest on all four inclined surfaces. The Bernoulli effect of the flowing stream holds the ball cradled in this position throughout the rest of the stroke. A nonrotating slip stem is guided by a bushing at the bottom and by a gland at the top of the bonnet. A machined bevel near the base of the stem acts as a travel limit and allows for back-seating. Ball and cage valves are furnished in sizes from 1 /4 14 in. (6 350 mm), with ratings from PSIG (1 17 MPa). Reported flow coefficients (C v ) are consistently high. The flow characteristic reflects the increasing enlargement of the crescent between the surface of the ball and the discharge port (Figure 6.16a). With a flow characteristic starting at zero flow, the rangeability is very high, over 50:1, depending only upon the ability of the actuator to position the cage. Tight shut-off occurs over a long operating life due to the continual rotation of the ball at each operation, which offers a new seating surface each time it is closed. Closure is positive due to the wedging of the cage in addition to line pressure. Although tightly closed, the stem force for opening is approximately 25% of a single-seated globe valve due to the manner in which the inclined surfaces of the cage roll the ball away from the seat. Open Throttling Closed FIG. 6.16r The open, throttling, and closed positions of a cage-positioned ball valve.

227 6.16 Valve Types: Ball Valves 1271 to use a relatively small actuator. The design is conducive to minimizing cavitation effects because flow tends to follow the curve of the ball, thus reducing turbulence. Cavitation tends to occur in the venturi passage, not at the seat, allowing use of hardened or replaceable throats. The expanding venturi discharge assists in handling flashing liquids. The bonnet design lends itself to the adaptation of a variety of linear or rotary actuators. Because the ball must be moved completely out of the flow stream, stem travels are at least as much as the diameter of the valve throat. Ball Unseated by Stem FIG. 6.16s Valve design in which the ball is unseated by the valve stem. (Courtesy of Powers Process Control.) Opening and closing force factors have been determined for all sizes of the valve. The low opening force requirement (9520 lb for a 4 in. valve at 2000 PSIG, or 42,400 N for a 100 mm valve at 13.8 MPa) is beneficial in being able The ball cage has also been used for regulators. The ball is cradled in the cage with the valve installed in the vertical position (Figure 6.16s). The stem of the regulator, coming from below the ball, forces the ball away from the seat. Flow is around the ball through the annular space, similar to the flow in a single-seated valve. Ball Gripped by Cage A variation of the ball-and-cage design is used for emergency closure (Figure 6.16t). Separate springs and ejection pistons allow high and low limit settings. Pressure above the high setting pushes the piston down to eject the ball from the holder into the seat. Low pressure allows the low-pressure spring to push the piston down. The ball is held firmly on the seat by the differential pressure. An internal bypass is opened to equalize the system pressures. Rotation of the bypass handwheel moves the ball back into the holder, the reset rod is retracted, and the bypass valve closed for normal operation. References 1. Bruckent, F. W., Using Ball Valves in Control Applications, Hydrocarbon Processing, August Husu, M., Trimming Control Valve Noise/Cavitation, InTech, December Bibliography FIG. 6.16t Ball-and-cage valve design, which is used to provide emergency closure. Baumann, H. D., Trends in Control Valves and Actuators, Instruments and Control Systems, November Carey, J. A., Control Valve Update, Instruments and Control Systems, January Cox, J., All Thanks to Metallic Sealing Systems: Ball Valves Withstand the Toughest Operating Conditions, Valve World 98 Conference, KCI Publishing BV, The Netherlands, Dobrowolski, M., Guide to Selecting Rotary Control Valves, Instrumentation Technology, December Fernbaugh, A., Control Valves: A Decade of Change, Instruments and Control Systems, January 1980.

228 1272 Control Valve Selection and Sizing Hammitt, D., Rotary Valves for Throttling, Instruments and Control Systems, July Holton, A. D., Control Valve Update, Instruments and Control Systems, January Lucitti, A. Choosing Valves for Isolation: The Case for Ball Valves," Valve Magazine, Winter Monsen, J. F., Valve Wars Rising Stem vs. Rotary, Plant Services, January Pyotsia, J., A Mathematical Model of a Control Valve, 1992 Instrumentation, Systems, and Automation Society Conference, Houston, October Ytzen, G. R., Ball Valves for Throttling Control, 19th ISA Conference, Preprint No

229 6.17 Valve Types: Butterfly Valves C. E. GAYLER (1970) B. G. LIPTÁK (1985) J. B. ARANT (1995, 2005) Flow sheet symbol Types of Designs: Sizes: A. General-purpose, aligned shaft B. High-performance, offset (eccentric) shaft A. 2 to 48 in. (51 mm to 1.22 m) are typical, but units have been made in sizes from 0.75 to 200 in. (19 mm to 5 m) B. 4 to 16 in. (0.1 to 0.4 m) are common, but units are available from 2 to 80 in. (50 mm to 2 m) Design Pressures: A. Most are available through ANSI Class 300 ratings and for up to 200 psid (1.4 MPa) pressure drop. Special units have been designed for up to 6000 PSIG design pressure B. For installation purposes most are available through ANSI 600 ratings and for up to 720 psid (5 MPa) pressure drop Design Temperature: Body/Disc Materials: A. 450 to 1000 F ( 268 to 538 C). Special refractory lined units have been made for up to 2200 F (1204 C) B. 320 to 450 F ( 196 to 232 C) for Teflon-seated valves; 1200 F (649 C) for metal-seated ones. Special units are available up to 1700 F (927 C) A. Iron, ductile iron, carbon or alloy steels, stainless steel ( ), aluminum bronze, Alloy 20, Monel, Hastelloy C, titanium, chrome plating, nickel plating, Kynar, Nordel, Viton, EPDM, Buna-N, neoprene elastomer lining, TFE encapsulation B. Steel, 316 stainless steel, alloy steel, Durimet 20, aluminum bronze, Alloy 20, Monel, Hastelloy C, titanium, tungsten titanium carbide (TTC) coating Seal Materials: A and B. TFE, Kel-F, EPT, polyethylene, PTFE with titanium, Inconel, or 316 stainless steel or other metals Characteristics: See Figure 6.17a Rangeability: Generally claimed as 50:1 Leakage: Capacity: Special Features: Cost: See Figure 6.17b Partial List of Suppliers: A. Unlined, 2 to 5%; lined, ANSI V (see Table 6.1gg) B. Metal seat, ANSI IV; soft (toggle) seat, ANSI VI A. With 60 rotation, C v = (17 to 20)d 2. Typical for throttling with 75 rotation, C v = (25 to 30)d 2 ; with 90 rotation, C v = (35 to 40)d 2 B. C v = (20 to 25)d 2 ; see Figure 6.17c Reduced torque disc designs, fire-tested seals, reduced noise disc, special disc seal designs ABB Kent Inc. ( AMRI, Inc ( Bray Controls ( Cashco Inc. ( Circle Seal Controls ( DeZurik/SPX Valves & Controls ( Fisher Controls International Inc. ( Flowserve, Flow Control Div. Valtek ( FMC Blending & Transfer ( 1273

230 1274 Control Valve Selection and Sizing Foxboro-Invensys ( George Fischer Inc. ( com) Halliburton Energy Services ( Honeywell Industry Solutions ( ITT Industries, Engineered Valves ( Keystone International ( Love Controls Corp. ( Metso Automation ( MKS Instruments Inc. ( Nibco Inc. ( North American Mfg. Co. ( Tyco Flow Control ( Ultraflo Corp. Xomox ( com) INTRODUCTION The orientation table (Table 6.1a) compares the main features of butterfly valves to other control valve designs. The rotary valves such as butterfly, ball, and plug valves were once considered to be only on/off valves. In recent decades the rotary valves in general and the butterfly design in particular have been used more and more as throttling control valves. Relative to the traditional globe control valve, the butterfly valves have the advantages of lower cost and weight, two to three times the flow capacity of globe valves (as shown in Figure 6.17c), fire-safe designs, and low stem leakage, which enables them to more easily meet the Federal Clean Act. Their leakage is high, unless special soft-seated configurations or high-performance designs are used. Some of them can also be provided with near-equal-percentage characteristics and tight shut-off. When used for throttling service, some of their disadvantages are a direct consequence of the above advantages. Their high-capacity design results in either using oversized valves or having small valves mounted in large pipes. If small valves are used, this means substantial waste of pumping energy caused by the reducer pressure drops. Also, their high-pressure recovery nature results in low vena contract pressures, which in turn increase the probability of cavitation and noise. Reduction of these problems has been attempted by adding flutes to the butterfly disc, but the problems have not been fully resolved. In operating rotary valves, the linear movement of cylinder or spring/diaphragm actuators must be converted by linkages, which introduces hysteresis and dead 24,000 21,000 % Flow or C v inch (0.61 m) high-performance Typical general purpose 2 inch (50 mm) high-performance $ 18,000 15,000 12, SST body Carbon steel body % Rotation (of 90 ) FIG. 6.17a The flow characteristics of butterfly valves are affected by the location of the shaft (aligned or eccentric) and by the relative size of the shaft compared to the valve. The characteristics of high-performance designs are also slightly affected if the shaft is moved from the upstream to the downstream side of the disc. For throttling purposes the rotation of the valve is usually limited to move between the 0 and 60 positions (75) (100) (150)(200) (250) (300) (350) (400) (450) (500) (550) (600) Valve size, inches (mm) FIG. 6.17b Approximate costs of high-performance butterfly valves provided with throttling actuators and positioners. Cost data are based upon standard high-performance, eccentric disc, soft-seat valves with doubleacting piston operator and positioner. Carbon steel bodies up through 8 in. (200 mm) size have 316 SST disc and 17-4PH shaft. Above 8 in. size, discs are chrome-plated carbon steel. Stainless steel bodies have 316 SST disc with 17-4PH shaft in all sizes. Other operators can be furnished. Other alloys are available at higher cost.

231 6.17 Valve Types: Butterfly Valves 1275 C v at degrees open Size Flow Shaft upstream C v at degrees open Size Flow Shaft downstream FIG. 6.17c Listed in the tables are the valve capacity coefficients of high-performance butterfly valves at various degrees of opening. The C v is different if the shaft of the disc is upstream (top) or downstream of the disc (bottom). (Courtesy of Valtek, Inc.) play. In addition, a nonlinear relationship exists between actuator movement and the resulting rotation. These considerations make the use of positioners essential, which on fast processes can lower the quality of control. The torque characteristics of these valves are also highly nonlinear (Figure 6.4v), and because of the high break-torque requirement, the actuator is oversized when operated in the throttling range. The characteristics of butterfly valves (Figure 6.17a) are somewhere between linear and quick-opening. CONVENTIONAL BUTTERFLY VALVES The butterfly valve is one of the oldest types of valves still in use. The dictionary defines the butterfly valve as a damper or throttle valve in a pipe consisting of a disc turning on a diametral axis (Figure 6.17d). Butterfly valves are not only used in industry, but variations are found in consumer products such as furnace dampers, automobile carburetors, and shower heads. Closed (Damper perpendicular to flow) Throttling Open (Damper parallel to flow) FIG. 6.17d The vane positions of butterfly valve when closed, throttling, or open.

232 1276 Control Valve Selection and Sizing Wide use of butterfly valves dates back only to the 1920s, when improved designs resulted in their acceptance for public waterworks applications. The valve was particularly applicable to the low-pressure on/off service usually encountered in waterworks applications. Today s modern butterfly valve designs are suitable for a wider variety of fluid applications, including those with higher pressure drops, tight shut-off, and corrosive characteristics. The straight-through design has a high capacity and has advantages when erosion is a consideration. Especially since the development of the eccentric shaft and disc designs known as the high-performance butterfly valve (HPBV), it is one of the fastest-growing segments of the valve industry. Lubrication fitting Packing Bearings Packing follower Shaft Solid ring type body Operation Disc Butterfly valve operation is basically simple, because it involves only rotating the vane, disc, louver, or flapper by means of the shaft to which it is fastened. This may be done manually by a lever handle on smaller valves or by a handwheel and rotary gear box on larger sizes. Automatic operation may be accomplished by pneumatic, hydraulic, or electrical motor drives attached to the shaft by various methods. Unfortunately, some of these methods of attachment do not provide good valve control, and the design of the operator-to-shaft connections must be closely examined. For the connection, some manufacturers just use the same square shaft end and clamp with set-screw that is used on manual valves. This type of connection is very prone to wear play, which is not acceptable for automatic control, where dynamically responsive valve operation is needed. The proper way to make a suitable connection is by using a valve shaft with a spline end and a corresponding mating connection at the operator shaft. Linkages should be the self aligning ball end type for an overall best connection. As the disc moves through a 90 rotation, the valve moves from fully closed to fully open (Figure 6.17d). The area open to flow increases as the disc rotates from closed to open, and this variation is used for throttling. The characteristic curve, which is a plot of the free area vs. percentage vane rotation, is shown in Figure 6.17a for a general-purpose butterfly valve and for some high performance designs. However, modern valve operator positioners are available either with cams or signal conversion units that can be programmed to give almost any valve characteristic that is desired. For a detailed discussion of intelligent positioners, refer to Section Construction Mechanically, butterfly valves vary widely in their construction features. However, common to all are the valve body, the disc and shaft, shaft support bushings or bearings, shaft packing, and a means of attaching an operator to the shaft. Butterfly valves also fall into two basic categories, swingthrough and shut-off designs. FIG. 6.17e The design of a swing-through butterfly valve. (Courtesy of Emerson Process Management, Fisher Controls Company.) Most swing-through designs (Figure 6.17e) have a symmetrical disc and shaft design with a certain clearance required between disc and body. The body is usually the solid ring type, which is mounted between pipe flanges. It can be either the wafer type or the single flange lug pattern, where the flange bolting also goes through the valve body. Discs are cast in one piece. The thickness of the disc and hub along with the diameter of the shaft is a function of the maximum pressure drop and torque required. Careful alignment of the body, bushings, shaft, and disc eliminates binding. Hard facing materials can be applied to the disc edge and body bore where erosive fluids such as steam are involved. Refractory-type linings are also available for the body. The operating temperature ranges of the various materials used in the construction of butterfly valves are given in Figure 6.17f. The swing-through butterfly valve designs are available with a variety of disc shapes that serve to reduce the required torque and to increase throttling angle range (Figure 6.17g). Swing-through butterfly valves are normally limited to a maximum throttling of angle about 70 open for the standard patterns and 60 open for the heavy patterns, due to their larger diameter shafts (Figure 6.17h). This is because the disc profile projection tends to disappear into the shaft area as the valve opens. HIGH-PERFORMANCE BUTTERFLY VALVES The most significant design advance in butterfly valves was the development of the high-performance butterfly valve. This design concept combined the tight shut-off of the lined valves, the reduced operating torque and excellent throttling capabilities of the swing-through disc shapes, and the ability to operate with relatively high pressure drops.

233 Packing & gaskets Shaft & pins Asbestos Grafoil Inconel 600 Al-bronze Inconel 718 Inconel X750 Monel K ASTM 453 Gr. 660 Nitronic PH AISI 316 AISI Body & disc Inconel 600 Hastelloy C Monel Al-bronze ASTM A296CA6MN ASTM A351CF8C ASTM A351CF8M ASTM A351CF8 ASTM A352LC3 ASTM A352LCB ASTM A217WC9 ASTM A217WC6 ASTM A216WCB 350 F 212 C F 50 F 20 F 212 F 600 F 700 F 800 F 101 C 45 C 29 C 100 C 315 C 371 C 427 C Permissible but not recommended for prolonged use ref.ansi B Manufacturer does not recommend long exposure at these temperatures as it may cause embrittlement. ANSI B16.34 does not recommend a flanged valve in class 150 LB. FIG. 6.17f Operating temperature ranges of various materials used in the construction of high-performance butterfly valves F 1000 F 1100 F 1200 F 510 C 538 C 593 C 649 C 6.17 Valve Types: Butterfly Valves 1277

234 1278 Control Valve Selection and Sizing Conventional Cambered Fishtail FIG. 6.17g Cambered and fishtail disc shapes are used to reduce the required torque and to increase the throttling angle range Light pattern Heavy pattern FIG. 6.17h Effect of design pattern and shaft diameter on the flow area of butterfly valves. The compact size, reduced weight, and lower cost have made the HPBV a formidable competitor to other control valve designs in sizes 3 in. (75 mm) and larger. There are many designs available, which usually all have the characteristics of 1) a separable seat ring contained in the body and 2) an eccentric cammed disc (Figure 6.17i). This camming action enables the disc to back out of and into the seat before and after the disc rotation when throttling. This is accomplished by having the shaft offset from both the centerlines of the disc and the valve body. Tight Shut-off Designs Special shut-off seals, such as piston rings for high temperatures to 1500 F (816 C) and T-ring seals for tight shut-off, are available (Figure 6.17j). These seals are not as popular as they used to be before the development of the HPBV valves and other designs with improved high-temperature seals. Butterfly valves designed for tight shut-off fall into two categories. One is the valve that is provided with an elastomer or plastic liner. In this configuration, the disc is also encapsulated in some cases (Figure 6.17k). The other tight shut-off design is the HPBV with the cammed disc and a separate seal ring clamped into the body (Figure 6.17l). In addition, there are some special designs with laminated seal rings located on the disc edge that wedge into a conical seat in the valve body (Figure 6.17m). These laminated seal designs are especially suitable for high pressure and temperature shut-off. Some time back, the lined butterfly valves were the only butterfly valves designed for tight shut-off. For their lining, Travel arc of spherical diameter of disc about stem and pivot CL Seat Spherical diameter of disc shown in closed position Disc lifting off seat Stem and pivot CL Spherical diameter of disc shown in open position C Lof disc (closed) Stem and pivot CL Rotation to open C Lof disc (open) Closed 2 Degrees rotation 4 Degrees rotation 6 Degrees rotation FIG. 6.17i The high-performance butterfly valve with eccentric shaft and cam action disc operation is shown on the left. On the right it is illustrated how the disc is lifting off the seat as the valve begins to open. (Courtesy of Flowserve-Valtek, Inc.)

235 6.17 Valve Types: Butterfly Valves 1279 FIG. 6.17l High-performance butterfly valve design provided with cammed disc. (Courtesy of Flowserve-Valtek, Inc.) FIG. 6.17j Special butterfly seal designs for tight shut-off. (Courtesy of Fisher Controls Company.) various elastomer materials were used, with a more rigid backup ring, which completely lined the bore of the valve and the valve gasket face area. Some liners can be bonded to the body and others are removable. Plastic liners, such as Teflon, are suitable for corrosive applications. In some cases the disc is also encapsulated in the elastomer or Teflon. Sealing in these valves is usually accomplished by a wedging action of the disc edge into the elastomeric or plastic seat. The discs may be symmetrical on the shaft (similar to swing-through), offset from the shaft, or canted on the shaft. The objective of the latter two designs is to give a 360 seal contact on the disc edge. Caution must be exercised in the selection of elastomers because they may be subject to attack by the process fluid. This attack may result in softening, swelling, cracking, or other effects. Plastic liners are not immune to these problems if the process fluid seeps past liner shaft seals and attacks the backing material or the body metal. By their very nature, these materials are also temperature limiting and can seldom exceed 350 F (177 C). Leakage Ratings FIG. 6.17k Lined butterfly valve design. (Courtesy of Keystone International, Inc.) The seat retainer rings are fastened to the body by various means. The most common method is by countersunk screws or bolts. Usually these fasteners are within the gasket area, and it is advisable to evaluate this interference in the gasket area for the intended service. Normally, this does not pose a problem, but it may require special gaskets or a change in gasket type to effect a proper seal. Other methods, such as a snap ring, friction fit, or retaining pins, do not intrude into the gasket area and may offer a better choice in some applications. Seats are commonly made of plastic materials such as Teflon or various elastomers. Each manufacturer has a specific idea of how this seat seal should be designed and configured. These designs range from very simple to very complex shapes, and in some cases may incorporate elastomers behind the Teflon to effect a pressure-energized seal.

236 1280 Control Valve Selection and Sizing Shaft Stuffing box packing Body Retainer Seal ring Gasket Bearings Key disc Taper pin Weld in seat Thrust bearing device FIG. 6.17m Eccentric disc butterfly with laminated disc seal ring. (Courtesy of Tyco-Vanessa, Inc.) Basically, these various seat designs are all classified as ANSI Class VI, which is commonly construed to be bubble-tight. Where temperatures are too high for the soft seat materials, metal seats are available that can also provide excellent shut-off equivalent to ANSI Class IV. Some metal seat designs can even approach ANSI Class V. Table 6.17n gives the seat leakage options for a particular HPBV design. Fire-Safe Designs A special design variation known as a fire-safe seat is also available from many manufacturers. This combines the soft seat discussed above with a backup metal seat. If the fire destroys the soft seat, the metal seat will serve to minimize leakage (Figure 6.17o). These designs also incorporate special fire-resistant gaskets and stem packing to minimize external leakage during and after a fire. TABLE 6.17n Seat Leakage Choices in a High-Performance Butterfly Valve* Type of Seat Metal seat Jam-lever toggle soft seat Flow ring Dual seat *Courtesy of Flowserve-Valtek, Inc. Leakage ANSI Class IV ANSI Class VI 2 % of rated C v ANSI Class IV TORQUE CHARACTERISTICS Operating torque requirements of butterfly valves require more careful consideration than do any other types of control valves (Figure 6.4v). The disc acts much like an airfoil or the wing of an aircraft. However, the special disc shapes and the HPBV designs already discussed have much lower torque requirements due to the shape effects, much like spoilers on an airfoil. The conventional symmetrical disc behavior as an airfoil results in different pressure distributions around the face of the disc, producing a torque that tends to close the valve. Only at 0 and 90 are the pressures equal on both sides of the disc. Between 0 and 90 the thrust load of the disc wing turned toward the upstream side is larger than that on the downstream side. This is called the unbalanced, or hydraulic, torque, and its magnitude is a function of the pressure drop and disc diameter. The valve operator must have enough power to overcome the unbalanced torque and, in addition, overcome the friction of the bearings and packing on the valve shaft. The total is known as combined torque. The combined torque required to open the valve is larger than that required to close the valve because the unbalanced torque helps to close the valve. This difference is known as torque hysteresis. In fact, when the symmetrical disc closes, the torque may become negative somewhere around from the closed position, and the valve will tend to close itself. The torque characteristics also indicate that at about a opening, the torques for both opening and closing maximize. Above this rotation angle, the torque falls quickly to zero. Thus, the

237 6.17 Valve Types: Butterfly Valves 1281 Face retainer ring graphite gasket Valve body Back-up ring Teflon insert soft seat Face retainer ring graphite gasket Valve body Three-point sealing Disc System pressure Constant metal-to-metal sealing System pressure Disc Before fire After fire FIG. 6.17o One type of fire-safe disc seal seat design Conventional 4" (100 mm) swing thru disc shape Combined torque in lbs/psi* Low torque cambered disc shape 5 4" (100 mm) Fluted disc Vane position-degrees from vertical *See section A.I for SI units 90 FIG. 6.17p Combined torque requirements (sum of shaft friction and unbalanced torque) of a number of butterfly disc shapes at various degrees of rotation. (Courtesy of H. D. Baumann Co.) FIG. 6.17q Substantial noise reduction can be obtained by the fluted butterfly disc design.

238 1282 Control Valve Selection and Sizing torque characteristics are highly nonlinear. They pose a considerable burden on the valve automatic operator, because it must cope with sharp increases and decreases in torque as well as positive and negative forces. Typical torque curves for conventional symmetrical, special shape, and fluted discs are shown in Figure 6.17p. It should be noted that while HPBV disc designs can be considered a reduced torque disc, the amount of torque and its characteristic behavior curve is a function of whether the valve is installed with the shaft upstream or downstream. Flow tends to open the valve with the shaft downstream and tends to close the valve with the shaft upstream. However, with the shaft downstream, dynamic torques with the disc open are much lower. On gas service, the shaft can be in either location, but on liquid service the shaft should be downstream only because of the effects of liquid inertial forces. Lined butterfly valves are not only subject to the above torque considerations but also encounter the additional torque required to seat and unseat the disc into the liner. The manufacturers usually rate the torque requirements of the various butterfly valve designs, because they are subject to so many variables. The manufacturer s recommendations should be followed for selecting the appropriate operator size required to operate the particular valve. It is best to be conservative when sizing butterfly valve operators, because operating and seating torques can often be greater than predicted. Sound level/dba Fluted disc Conventional globe valve Medium, saturated steam AP/P 1 = 0.5 to 0.61 Valve C v FIG. 6.17r Noise reduction produced by a fluted disc in comparison to the noise level produced by a conventional globe valve. NOISE SUPPRESSION Butterfly valves will generate noise, as will any other valve when throttled at high flow rate and pressure drop, as is discussed in full detail in Section The noise characteristics of some special disc designs, such as the cambered and the fishtail designs shown in Figure 6.17g, are improved in comparison to the conventional swing-through disc shape designs. However, with higher mass flows and pressure drops the generated noise can still be substantial. A newer design development for the butterfly is the fluted disc shown in Figure 6.17q. For compressible fluid applications, the fluted disc design is capable of delivering noise reductions of up to 10 dbs on the A-weighted scale (dba), as shown in Figure 6.17r. In addition, the airfoil spoiler effect of the flutes enables this disc to have the lowest operating torque requirements of any disc design. It can be provided for valve sizes 2 in. (25 mm) through 16 in. (400 mm). The fluted disc design is more expensive than others but is a very useful alternate where needed. The cost of manufacturing the noise-reduction spoilers has been somewhat reduced by relocating the flutes, as shown in Figure 6.17s. The addition of swing-through vanes to the butterfly disc not only reduce noise emissions and cavitation, but it also increases valve rangeability and improves valve characteristics (Figure 6.17t). The valve characteristics can be changed by varying the area, spacing, and number of the flow orifices. It is possible to provide such spacing of the flow orifices that the two semispherical contours will generate near-equal-percentage characteristics. The addition of a diffuser pack (shown in Figure 6.17t) brings the characteristics closer to linear. FIG. 6.17s Flutes can be added to the butterfly disc to act as noise-reduction spoilers. (Courtesy of Flowserve-Valtek, Inc.)

239 6.17 Valve Types: Butterfly Valves 1283 Flow % of maximum With optional integral diffuser pack Without diffuser Equal % Swing through vanes Diffuser pack % Valve openings FIG. 6.17t The addition of perforated swing-through vanes not only reduces noise and cavitation but also improves valve characteristics. (Courtesy of ABB Kent.) Bibliography Ball, K. E., Final Elements: Final Frontier, InTech, November Baumann, H. D., The Case for Butterfly Valves in Throttling Applications, Instruments and Control Systems, May Baumann, H. D., Trend in Control Valves and Actuators, Instruments and Control Systems, November Boger, H. W., Low-Torque Butterfly Valve Design, Instrumentation Technology, September Cain, F. M., Solving the Problem of Cavitation in Control Valves, Paper # , 1991 ISA Conference, Anaheim, CA, October Carey, J. A., Control Valve Update, Instruments and Control Systems, January Daneher, J. R., Sizing Butterfly Valves, Water & Wastes Engineering, July and September Dobrowolski, M., Guide to Selecting Rotary Control Valves, Instrumentation Technology, December Fernbaugh, A., Control Valves: A Decade of Change, Instruments and Control Systems, January Hammitt, D., Rotary Valves for Throttling, Instruments and Control Systems, July Page, G. W., Predict Control Valve Noise, Chem. Eng., New York 107(9): August Passage, D., Butterfly Control Valves, Instruments and Control Systems, March Pyotsia, J., A Mathematical Model of a Control Valve, 1992 ISA Conference, Houston, TX, October Rahmeyer, W., The Critical Flow Limit and Pressure Recovery Factor for Flow Control, InTech, November 1986.

240 6.18 Valve Types: Digital Valves Digital D. R. A. JONES (1985) B. G. LIPTÁK (1995, 2005) Flow sheet symbol Size: 3 /4 to 10 in. (19 to 250 mm) in-line and angle pattern Design Temperature Range: Design Pressure Limits: Cryogenic to 1250 F (677 C) Up to 10,000 PSIG (690 bars) Capacity: C v = 13 d 2 Applications: Where high speed is required, as in surge control Where accurate flow control is needed, such as in provers or in the blending of expensive ingredients Where great rangeability is needed, such as in ph control Where flow is to be both controlled and accurately measured, such as in natural gas regulator stations at high-pressure drops or in power plants Where tight shutoff is needed Rangeability: No. of bits Resolution 255:1 1023:1 4095:1 16,383:1 65,535:1 Speed: Characteristics: Leakage: Materials of Construction: Cost: Partial List of Suppliers: 25 to 100 ms Unlimited ANSI V Body Aluminum, carbon steel, stainless steel, titanium Seals Buna, rubber, Viton, TFE, Kel-F, Derlin, Hypalon, and graphite Typical prices for ANSI Class 600 carbon steel body, 12-bit valves are $5000 for 3 / 4 in. (20 mm), $5000 to $10,000 for 1, 1 1 /2, and 2 in. (25, 32, 40, and 50 mm), $12,500 for 3 in. (75 mm), $15,000 for 4 in. (100 mm), $20,000 for 6 in. (150 mm), and $30,000 for 8 in. (200 mm). For 316 SST bodies, add $1000 for 3 /4 in., $2000 for 1, 1 1 /2, and 2 in., $3000 for 3 in., $5000 for 4 in., $7000 for 6 in., and $14,000 for 8 in. ABB Kent, Introl Valves Div. ( Emco-Digital Valve ( com/digitalvalve.htm) Emerson-Daniel ( Herion Inc. ( Hoke Inc. ( Instrutech Inc. (www. instrutechinc.com) INTRODUCTION These days, when one refers to digital control valves, people think of intelligent control valves that are provided with fieldbus interaction capability (Figure 6.18a). Fieldbus interaction is not the topic of this section; it is covered in Section The digital control valves discussed in this section (while they can also be operated by digital networks) are multiported valves, with the number of ports ranging from 8 to 16. A digital valve contains a group of valve elements assembled into a common manifold. The elements have a binary relationship to each other; i.e., starting with the smallest, each 1284

241 6.18 Valve Types: Digital Valves 1285 Operation Engineering Maintenance HSE fieldbus Firewall LIMS Execution Business I. S. barrier H1 fieldbus devices Conventional devices & HART FIG. 6.18a Networked system architecture. increasing size element is twice as large as its next smallest neighbor (Figure 6.18b). Each element is controlled by an individual electric or electronic signal. Thus, an 8-bit digital valve requires 8 parallel, on/off electric (or electronic) signals, a 12-bit digital valve will require 12 parallel signals, and a 16-bit digital valve will require 16 parallel signals. The main advantages of digital control valves are their speed, high precision, and practically unlimited rangeability. Their main disadvantages are their high cost and their suitability for only clean services, because the smaller ports plug very easily. HISTORY In the early digital control valves, the ports were distributed circumferentially, as shown in Figure 6.18c. These designs were difficult to manufacture. In the later designs the circumferential distribution of ports was replaced by top-entry, vertical ports. Balanced Piston Digital Control Balanced piston type designs (Figure 6.18d) are also called digital control valves, because their piston-operated main inner valve is positioned by the opening or closing of two solenoid pilots. Their advantages include their positive shutoff and fail-safe nature, as they close on loss of electrical failure. These valves provide precise flow rate and batch controls, including preprogrammed low flow start-up, highrate delivery, and low-flow shutdown. Top-Entry Design An 8-bit digital valve throttles the flow by controlling the openings of eight flow elements in the valve body (Figure 8.18e) and provides a flow rangeability of 255:1. This body, an in-line, top-entry design, includes an air reservoir to ensure the fail-safe operation of the cylinder actuators. The design includes adequate manifold area to ensure consistent performance of the individual elements. The manifolds are large enough to minimize the possibility of cavitation and resulting erosion. Each element in the array is on/off. Flow throttling is accomplished by opening enough ports to provide the exact flow area required by the controller output signal. There is a 1:1 relationship between the binary weighted signal and the binary weighted flow area. Figure 6.18b illustrates schematically the size relationship between binary elements in a digital valve. Applications The main applications of digital valves are ones where speed, rangeability, and precision are critical. Such applications include the accurate blending and batching of both gases and liquids. The high-speed operation of compressor surge controls and all flow control applications, where the process fluid is clean and the rangeability of the Bit No FIG. 6.18b An 8-bit digital valve is provided with a balanced distribution of ports and guarantees a resolution of 255:1, which corresponds to a flow control accuracy within 0.39% of the total valve C v. (Courtesy of Emco-Digital Valve.) 8

242 1286 Control Valve Selection and Sizing Control orifices Inlet torpedo Actuator assembly Cage Solenoid Three-way valve assembly Inlet manifold plug Outlet manifold Center line for six parallel binary controlled elements at 0, 60, 120, 180, 240, and 300 degrees Actuation gas manifold Center line for six parallel binary controlled elements at 30, 90, 150, 210, 270, and 330 degrees FIG. 6.18c Early digital control valve design with circumferentially distributed ports. 1 flow exceeds the capabilities of conventional valves. Such applications include but are not limited to environmental chamber controls, aircraft pressure cycling, and the operation of both liquid and gas provers. instead of one 50% element for the largest bit. Thus, an 8- bit digital valve would have nine elements three 25%, one 12.5%, one 6.25%, one 3.125%, one 1.56%, one 0.78%, and one 0.39% element, as was shown in Figure 6.18b. Resolution and Throttling Because a binary array of 12 bits will provide resolution of 1 in more than 4000 (the area of the smallest flow element in a 12-bit digital valve is less than 0.025% of the total flow area), it is essential that the digital valve be leak-tight (Figure 6.18f). Leakage rates that are acceptable in standard globe valves would make the abovementioned resolution impossible to achieve. It is for these reasons that digital valves are commonly manufactured to extremely high leak-tight standards. In order to smoothly change the valve opening, say between 49% and 50%, it is usual to use two 25% elements Air reservoir Electric or electronic signal Check valve Air inlet Air pilot Air cyclinder Orifice or sonic nozzle Flow outlet Flow inlet In line base FIG. 6.18d Balanced piston operated digital control valve. (Courtesy of Emerson Process Management Daniel.) FIG. 6.18e An 8-bit, 3 inch, carbon steel digital control valve designed for high pressure service (3000 ANSI).

243 6.18 Valve Types: Digital Valves 1287 FIG. 6.18f 1n a digital valve, the area of each valve port is half that of the previous port. In higher-resolution digital valves, multiple elements are common; i.e., a 12-bit digital valve could have 16 elements, and a 14-bit digital valve could have 18 elements. The largest element then would be 12.5%. Each such valve would have seven 12.5% elements. The 6 in. (150 mm) cast steel body digital valve (Figure 6.18g) uses 12 elements to provide 8-bit performance. The binary series of control element capacities is 1, 2, 4, 8, 16, 32-32, 32 32, In this way no individual element handles more than 12.5% of the total flow. The largest bit (50%) is handled by four elements spread out around the body. This arrangement improves redundancy in operation and uniformity in individual element sizing. Noise generation is reduced by breaking the flow up into many flow paths. The 8-bit computer word remains the same but operates 12 control elements. This method of breaking the flow into many streams has a number of advantages. Each element actuator (electric, hydraulic, or pneumatic) is small and very quick. The multiple element arrangement provides redundancy, usually permitting operations to continue even if one or more elements are disabled. Flow calibration requirements are affected, because the calibration unit need be only 1 / 8 as large to qualify 100% capacity. Leakage, Ratings, and Speed Flow difference between individual elements is determined by the control orifice or the nozzle size. Each element consists of a plunger and a seat. The plunger is operated by a solenoid or solenoid-piloted A A A Air in Electric (Electronic) signals Air cyclinder operator Carrier Air reservoir Nozzle or orifice Inlet Outlet Cast body in-line base FIG. 6.18g A 6 inch, 8-bit, explosion-proof cast steel (900 ANSI) digital valve with 12 elements. (Courtesy of Emco-Digital Valve.) A

244 1288 Control Valve Selection and Sizing cylinder (pneumatic or hydraulic). Elements are usually dynamically balanced for low actuating force and are designed to assure tight shut-off. Because of the necessary leak-tight construction, digital valves of very small capacity coefficients (C v ) are practical. Standard digital valve body sizes range from 3 / 4 in. NPT (19 mm) through 10 in. (254 mm) pipe size. Pressure ratings up to 10,000 PSIG (68, 788 kpa) have been supplied. Digital valves have been used effectively in cryogenic service, and special high-temperature models are under development (to 1250 F, or 677 C). Fluid temperatures above F ( C) require special seals and seal design. Digital valves provide exactly repeatable performance, because each digital command causes the opening of a precisely defined port area. Nominal transfer time from any one position to any other position is usually under 100 ms, and the transfer time is uniform from one position to another. The binary relationship between elements provides a linear increase of port area with a linearly increasing digital signal command. Any desired valve characteristic can be obtained by the correct programming of the digital control command. FLOW METERING A valve-flowmeter is a version of the digital valve that is equipped with flow sensing nozzles or orifices in each element (flow port). Gas Flow When gas is flowing through a sonic venturi and the inlet pressure is high enough to induce sonic flow at the vena contracta (throat) of the venturi, that condition is called choked flow. Under such conditions, the rate of gas flow will only be dependent on the inlet conditions (i.e., absolute temperature and absolute pressure) of the flowing gas. When a sonic venturi is installed in each of the ports of a digital valve (Figure 6.18h), the choked flow at each port is independent of the downstream pressure. FIG. 6.18i Extremely wide rangeability can be provided by using sonic venturitype digital flowmeters. As long as the downstream pressure is less than the critical pressure at the vena contracta, choked flow can usually be maintained with a pressure drop across the venturi of 15% of the inlet absolute pressure. The divergent portion of the sonic venturi recovers part of the velocity head. Thus, variations in the downstream pressure do not affect the flow rate. The flow measurement is so accurate under these conditions that a digital valve-based flowmeter can be used as a transfer standard for calibrating other flow meters. An example of such installations is an 8 in. (200 mm) digital valveflowmeter equipped with sonic venturis that has been installed in a power plant fuel line to act as a combination pressure regulator and fuel flowmeter (Figure 6.18i). Liquid Flow An orifice can be installed in each element, and the differential pressure across the digital valve can be used as a measure of the fluid flow through the open port area. Digital valves can be used as wide-range flowmeters by controlling the valve opening to maintain a constant pressure drop across the valve. Thus, a low-pressure drop-based measurement can provide a very wide flow range (up to 16,000:1 or more). CONCLUSIONS FIG. 6.18h The sonic venturi element. In summation, digital valves can provide high resolution, very fast response, exact repeatability, and very wide range. They can also be equipped with flow elements to provide both measurement and control of flow. Recommended applications are flow blending, compressor surge control, gas meter-regulators, transfer-standard flow provers, and precise liquid flow rate measurement and control. In fact they are

245 6.18 Valve Types: Digital Valves 1289 appropriate any place where speed, accuracy, and high resolution are needed and the process fluid is clean. Reference 1. Langill, A. W., New Control Valve Accepts Digital Signals, Control Engineering, August Bibliography Alspach, W. J. and Maurer, G., Consider Digital Valves, Hydrocarbon Processing, December ANSI/ISA , Test Procedure for Control Valve Response Measurement from Step Inputs, Research Triangle Park, NC: Instrumentation, Systems, and Automation Society, Ball, K. E., Final Elements: Final Frontier, InTech, November Beeson, J. and Alspach, W. J., How Sonic Nozzle Proving Works, Gas Industries, February Britton, C. L., Sonic Nozzles, Appalachian Gas Measurement Short Course, August Calibrating Control Valve Accuracy, Mech Eng, 122 (6): 33-33, June Corey, J. A., Control Valve Update, Instrument and Control Systems, January Clark, H. L., Turbine Meter Testing, AGA Transmission Measurement Conference, May Fernbaugh, A., Control Valves: A Decade of Change, Instrument and Control Systems, January Grumstrup, B., Digital Valve Mounted Instrumentation Opening a Window to the Process, ISA/94 Technical Conference, Anaheim, CA, Harrold, D., Calibrating Control Valves, Control Engineering Europe, Vol. 3, Issue 3, p. 43, June July Jones, D. R. A., Digital Valves, AGA Transmission Measurement Conference, May Langford, C. G., A Method to Determine Control Valve Dynamic Requirements, Instrumentation, Systems, and Automation Society, Miller, C. E. and Alspach, W. J., Digital Valves and Digital Flow Technology, Proceedings of the Instrumentation Symposium for the Process Industry, Texas A&M University, Miller, L., Valve Diagnostic Past, Present, and Future, Fluid Handling Systems, November Morris, W., Digital Valve as a Transfer Standard, Gas Magazine, September Page, G W., Predict Control Valve Noise, Chem Eng-New York 107 (9): 23 26, August 2000 (Reprinted, July 1997). Pyotsia, J., A Mathematical Model of a Control Valve, 1992 ISA Conference, Houston, TX, October Rahmeyer, W., The Critical Flow Limit and Pressure Recovery Factor for Flow Control, InTech, November Smart Actuator Incorporates all the Controls in One Package, Chemical Engineering Magazine, January 1, 2003.

246 6.19 Valve Types: Globe Valves H. D. BAUMANN (1970) J. B. ARANT (1985) B. G. LIPTÁK (1995) F. M. CAIN (2005) FC Three-way Note: The letter S if marked inside the valve symbol, refers to split body and the letter C to cage design. Flow sheet symbols FC Fail closed globe valve FO Fail open angle valve Types: Sizes: Design Pressure Ratings: Maximum Pressure Drop: Design Temperature: Materials of Construction: A. Single-ported with characterized plug B. Single-ported, cage-guided C. Single-ported, split body D. Double-ported, top-bottom-guided or skirt-guided plug E. Angle F. Y-type G. Three-way H. Eccentric plug, rotary globe A. Typically NPS 1 /2 to 16 in. (DN 15 to DN 400); available up to NPS 48 in. (DN 1200) B. Typically NPS 1 /2 to 16 in. (DN 15 to DN 400); available up to NPS 48 in. (DN 1200) C. NPS 1 /2 to 10 in. (DN 15 to 250) D. NPS 1 /2 to 16 in. (DN 15 to DN 400) E. Typically NPS 1 /2 to 16 in. (DN 15 to DN 400); available up to NPS 48 in. (DN 1200) F. NPS 1 to 16-in. (DN 25 to DN 400) G. Typically NPS 1 /2 to 6 in. (DN 15 to DN 150); available up to NPS 24 in. (DN 600) H. NPS 1 to 12 in. (DN 25 to DN 300) DN = diameter, nominal mm assumed. Typically all ratings are available from ANSI Class 150 (PN 20) to Class 2500 (PN 420) with special designs up to Class 4500; types C and H are limited to ANSI Class 600 (PN 100) PN = Pressure, nominal bar assumed. Up to maximum allowed by body pressure rating depending on limitations of actuator size and trim design and materials Depends on material properties. Generally from 20 to 1200 F ( 29 to 538 C). Cryogenic designs for temperatures down to 423 F ( 253 C). Special valves have been designed for operation up to 1600 F (871 C). Body and bonnet materials: Most cast and forged grades of carbon steel, low-alloy steel, stainless steel, Alloy 20, duplex stainless steel, nickel and nickel alloys, bronze, titanium, and zirconium. See Table 6.19ww. Fluoropolymer lining also available for corrosion protection. Trim materials: Generally available in stainless steel, nickel, nickel alloys, bronze, titanium, and zirconium. Hard facing is available for erosive applications. See Table 6.19j. Seal and soft seat materials: FEP, PFA, PTFE, PCTFE, ETFE, EPT/EPDM, Fluoroelastomers, Nitrile, polyethylene, polyurethane, UHMWPE, compressed graphite, and soft metals. 1290

247 6.19 Valve Types: Globe Valves 1291 Leakage: Characteristics: (See Table 6.1gg for FCI leakage classes.) Metal seats in double-ported designs are Class II, while in single-seated designs they can meet Class IV or Class V. Soft seats in double-ported designs can meet ANSI Class IV or V, while in single-seated globe valves they can give Class VI performance. Refer to Section 6.7 for details; see Figure 6.19a. Rangeability: Based on Instrumentation, Systems, and Automation Society or IEC , it seldom exceeds 30:1. Special designs can achieve 50:1 or higher by increasing precision of control at small valve openings. See the discussion under Rangeability below and Section 6.7 for details. Capacity: Cost: Partial List of Suppliers: (Includes both manual and control valves) C v /d 2 = 10 to 15 with single-ported designs closer to the bottom of the range and with double-ported and eccentric disc designs closer to the top of the range (see Table 6.19c). See Figure 6.19b. ABB Control Valves ( American Valve, Inc. ( ARI-Armaturen Richter ( Asahi-America ( Cashco Inc. ( Collins Instrument Co. ( Control Components Inc. ( Conval Inc. ( Crane Valves ( Curtis Wright Flow Control ( Dresser Flow Solutions ( Emerson Process Management ( Flowserve Corporation ( GE-Nuovo Pignone ( index.htm) Invalco ( Kitz Corp. ( Koso Hammel Dahl ( Metso Automation ( Milwaukee Valve Co. ( Nibco Inc. ( Powell Valves ( Richards Industries Valve Group, Inc. ( Samson AG ( Severn Glocon Ltd. ( Spirax Sarco, Inc. ( SPX Valves and Controls ( Tyco Flow Control ( Velan Valve Corp. ( Warren Controls Corporation ( Weir Valves & Controls ( Welland & Tuxhorn ( Yamatake Corp. ( VALVE TRENDS When this handbook was first published some 35 years ago, the overwhelming majority of throttling control valves were the globe types, characterized by linear plug movements and actuated by spring-and-diaphragm operators. At that time, the rotary valves were considered to be on/off shut-off devices. Globe valves are still widely used, but their dominance in throttling control applications has been diminished by the less expensive rotary (ball, butterfly, and plug) valves as a result of improvements in rotary valve and actuator designs. Generally, globe valves use a linear-motion stem connected to a plug head that controls the flow area through a stationary seat ring. One exception to this is the rotary globe valve, which rotates an eccentric spherical plug into the seat ring; this type of rotary stem valve will be discussed later. Unless otherwise noted, the discussion of globe valve characteristics will apply to the linear-stem globe valve.

248 1292 Control Valve Selection and Sizing % Flow Theoretical quick opening Installed linear Installed equal percent Theoretical linear Theoretical equal percent % Lift FIG. 6.19a The theoretical valve characteristics shift as a function of installation. The dotted lines reflect such a shift in a mostly friction process where at 100% flow, 20% of the pressure drop was assigned to the control valve. The main advantages of the traditional globe design include 1. The simplicity of the pneumatic actuator designs 2. The availability of a wide range of valve characteristics 3. The relatively low likelihood of cavitation and noise 4. The availability of a wide variety of specialized designs for corrosive, abrasive, and high-temperature or high- pressure applications 5. Relatively small amounts of dead band and hysteresis The main reason for the increasing popularity of rotary valves is their lower manufacturing cost and higher capacity (C v /d 2 = for rotary vs. C v /d 2 = for globe). They generally weigh less, some designs can act as both control and shut-off valves, and they can be easier to seal at the stem to meet clean air requirements. The limitations of globe valves, in addition to their higher cost per unit C v, include their greater weight and dimensional envelope relative to their flow capacity. For the valve coefficients of globe valves, refer to Table 6.19c. One major disadvantage of rotary valves is their higher tendency to cavitate and produce excessive amounts of noise (Section 6.14). They are also more likely, due to their smaller size per unit C v, to have larger pipe reducers with the associated waste of pressure drop and distortion of characteristics. Their control quality can suffer from the linkages, which can introduce substantial hysteresis and dead band. As a result of advances in distributed control system (DCS) technologies, both rotary valves and globe-style (linear) valves generally benefit from the use of positioners with either single-acting or double-acting pneumatic actuators Budgetary price (US$) Valve size - inches FIG. 6.19b Approximate cost data are based upon typical globe control valves with Class 300 flanged bodies, double-acting piston actuator, and positioner with I/P transducer. As an example, Figure 6.19b would estimate the average cost of a 4 in. globe valve with a steel body and cage-type trim to be about $6000. Other actuators such as single-acting piston, spring-and-diaphragm, and electric motor drive are available. Bodies and trim are available in numerous metal alloys but usually at higher prices. (diaphragm or piston types). Previous problems with positioners in fast processes are largely relics of old-fashioned control systems, but much misinformation has perpetuated an aversion to positioner use on fast processes, even though DCS technology overcame these issues decades ago. For more information about control theory in general and controller tuning in particular, consult Chapter 2, and about DCS systems refer to Chapter 4. For conventional and for intelligent positioners, refer to Sections 6.2 and 6.12 respectively, in this chapter. TRIM DESIGNS Approximate cost of globe control valve Carbon steel body Stainless steel body The valve trim consists of the internal parts contained within the body and wetted by the process fluid. The main components are the plug and stem and the seat ring(s). Some globe valve body designs also incorporate other parts such as cages or seat retainers, spacers, guide bushings, and special elements. The trim parts create the flow restriction or throttling action responsible for most of the pressure loss dissipated in the valve. The trim design also serves to determine the inherent flow characteristics of the valve. The various aspects of trim design, construction, and selection will be discussed.

249 6.19 Valve Types: Globe Valves 1293 TABLE 6.19c Valve Coefficients (C v ) for Single-Ported, Equal-Percentage, Unbalanced Globe Valves with Flow under the Plug. * Valve Size (in.) Trim Size (in.) Stroke (in.) C v At Percent Open A B A * For each valve size, the values given in the first line correspond to full-area trim; the values for reduced trims follow in descending order. Courtesy of Flowserve Corporation.

250 1294 Control Valve Selection and Sizing Contoured plug (top guided) V- port (top and bottom guided) Balanced Piston plug (cage guided) Unbalanced FIG. 6.19d Valve with plain bonnet and separable bonnet flange design. (Courtesy of Flowserve Corporation.) Linear Equal percent Typical components that make up a globe valve are shown in Figure 6.19d. Globe valve trims are available in many design variations, depending upon manufacturer and the intended application of a particular valve. A complete discussion of every trim design variation is impossible, but those with the most general application use will be covered, including special trims for severe services such as noise, cavitation, and erosion. See Sections 6.1 and 6.14 for discussions of low-noise control valves. Trim Flow Characteristics All control valves are pressure-reducing devices; in other words, they have to throttle the flowing fluid in order to achieve control. The most widely used form of throttling is with a single-stage orifice and plug assembly. Multiple-stage orifice elements are usually found in trim designs for combating noise, erosion, and cavitation (Figure 6.19g). In all cases, the valve trim is the heart of the valve and operates to give a specific relationship between flow capacity and valve plug lift. This relationship is known as the valve flow characteristic and is achieved by different cage orifice patterns (Figure 6.19e) or valve plug contours (Figure 6.19f). The term flow characteristic usually refers to the inherent characteristic, which is a function of a number of valve design and manufacturing parameters. The inherent characteristic is determined by testing the valve flow vs. valve lift using a constant differential pressure across the valve throughout the test. These types of tests are standardized by ANSI/ISA and IEC : Therefore, the manufacturers trim characteristic curves or tables should not be confused with the installed flow FIG 6.19e Types of valve plug configurations for various valve designs. (Courtesy of ITT Conoflow.) characteristic in the actual process fluid flow loop. In actual service, the differential pressure across the valve varies throughout the valve lift and flow range as a function of the system characteristics. This variation is due to such factors as pump head changes with flow, piping friction losses, and the hydrostatic resistance of pipe fittings, block valves, flow measurement devices, heat exchanges, and other system elements. Equal percentage Linear Quick opening FIG. 6.19f Valve plug shapes to produce the three common flow characteristics: equal percentage, linear, and quick opening.

251 6.19 Valve Types: Globe Valves 1295 P 1 B P 2 (Outlet pressure) P vc (pressure at vena contracta) A = P F L = (Valve pressure drop) FIG. 6.19g Pressure profile of a single-seat valve that is experiencing cavitation. Control valve inherent characteristic data are expressed in graphs or tables, such as shown in Table 6.19c, where a flow coefficient (C v = K v /1.17) is expressed as function of the percentage of valve opening. It is important to understand the meaning of these flow coefficients. A detailed discussion of control valve sizing is provided in Section 6.15, which discusses control valve sizing. In its simplest form, the valve capacity coefficient, C v (or K v ), for liquids can be expressed as C K Q G v( or v) = p P v (vapor pressure) 6.19(1) where Q is volumetric flow rate in gpm (or m 3 /h for K v ), G f is specific gravity relative to water, and p is the pressure differential (the lesser of the actual p or the choked p) across the valve in lb/in. 2 (or bar for K v ). Important observations should be made from this simplified expression. 1. C v and K v are not dimensionless coefficients. They have units of (volume/time) (area/force) 1/2 = (length) 4 / [(time)(force) 1/2 ]. 2. Valve manufacturers publish valve inherent characteristic in terms of C v (or K v ) vs. lift (or percentage open). 3. Valve users need to know flow rate vs. lift for the installed characteristic. They must determine the system pressure differential allocated to the valve for the full range of flow rates in order to calculate the flow rate vs. lift. Control valve manufacturers commonly furnish three types of inherent characteristic valve trims along with some minor variations (Figure 6.19a). These are idealized curves and do not accurately reflect the actual characteristic as determined by test. Examination of actual test data will show deviations in lift vs. flow of 10% or more, slope variations, and other distortions from the ideal curve. f A B This is due to a number of factors; in order of their significance, they include a) the trim type and design, b) valve body geometry effects, c) test variations and repeatability, and d) manufacturing variations. For practical purposes, these distortions, if kept within reasonable limits, do not materially affect the valve in actual service. Allowable limits on variations in flow characteristic are established in industry standards ANSI/ISA and IEC The typical inherent characteristic (i.e., C v or K v vs. lift) test data and pressure loss vs. flow rate data for the static elements of the process system can be used to approximate the valve characteristic behavior in the installed system. This can be used to select the best valve trim for the controlled process, which will keep the control loop gain constant or optimized for process control (see Figure 6.7a in Section 6.7). A traditional rule of thumb is to use a linear trim if the control valve pressure drop is relatively constant (such as in pure pressure reducing). Where there is significant system and valve pressure drop variation as flow changes, the equalpercentage trim is recommended. Rangeability Rangeability can be expressed in different ways with different meanings. Inherent rangeability is the ratio of the largest controllable flow coefficient (C v or K v ) to the smallest controllable flow coefficient within specific deviation allowances. 1 Valve flow rangeability, referred to as turndown, is the ratio between the valve s maximum and minimum controllable flow rate at stated operating pressures. Generally, the minimum controllable flow is considered to be about twice the minimum clearance flow as the plug lifts off the seat. For single-seated contoured plug control valves, manufacturers often state inherent rangeability from 30:1 to 50:1 based on C v vs. lift tests. In practice, these numbers are merely benchmarks. When applied to processes with high system pressure losses, the installed flow rangeability of the control valve is more likely to be in the order of 7:1 to 15:1. This is usually enough because most processes do not operate much over 5:1 turndown. Some processes can require flow rangeability beyond the capability of one valve and will require parallel valves with split-ranging. For more details, see Figures 6.1i and 6.1j. While linear and equal-percentage trims are designed to throttle essentially over the full valve travel, the quick-opening characteristic is designed to act more like a bath stopper plug. The flow characteristic develops approximately 80% of its C v capacity in a nearly linear manner over the initial 20 30% of valve lift. The remaining capacity is added over the balance of the lift. This plug can be used for on/off service, for short-stroke valves such as self-contained pressure regulators, or in some process applications where a decreasing valve gain is required as the load increases. 1 See ANSI/ISA and IEC : 1989.

252 1296 Control Valve Selection and Sizing Standard Trim Configurations The valve plug configurations used on modern control valves include the contoured plug, ported plug, or piston plug (see Figure 6.19e). The turned or contoured plug is probably the most common, followed by the piston and ported plugs. The contoured plug is simple to machine out of bar stock from stainless steels and special alloys and can be hard-faced easily for erosive services. The contoured plug is usually used with single-seat valves, but it is also available in double-seat designs. The ported plugs usually have cast or forged stainless steel or special alloy heads mechanically attached to a similar stem material. They are more difficult to hard-face than designs with uninterrupted control surfaces. This plug is most common in double-seat valves, but it is also available in single-seat designs. Both the contoured and ported plugs are shaped to obtain the desired flow characteristic as they move in and out of the seat ring. The piston plug is relatively easy to fabricate and is usually made from a hardenable type of stainless steel such as 410, 440, or 17-4PH. Other materials such as nickel alloy 718, alloy K-500, and austenitic stainless steel with hardfacing are also available. Plug heads are usually a hardened material, or hard-facing must be done, because these plugs are normally guided in a cage assembly. The flow characteristic for this trim design is incorporated into the cage. Special Trim Configurations There are severe applications that may require special trim configurations. These applications usually involve noise, cavitation, erosion, or combinations of these problems. The special trim designs for all of these are often similar in concept although they will differ in design detail. Because noise is an especially difficult and complex problem to deal with, it is covered in depth in Section The following discussion will touch upon the noise reduction trims, but will primarily concentrate on cavitation and erosion services. Cavitation Liquid cavitation is a complex fluid dynamic reaction to pressure change, which is also discussed in Sections 6.1, 6.14, and It has many aspects that have not been successfully explained, but the fundamentals are relatively simple. The basic process of cavitation is related to the conservation of energy and Bernoulli s theorem, which describe the pressure profile of a liquid flowing through a restriction or orifice (Figure 6.19g). In order to accelerate the fluid through the restriction, some of the pressure head is converted into velocity head. This transfer of static energy is needed to push the same mass flow through the smaller passage. The fluid accelerates to its maximum velocity, which is also the point of minimum pressure (vena contracta). The fluid velocity gradually slows down as it again expands back to the full pipe area. The static pressure also recovers, but part of it is lost due to turbulence and friction. If the static pressure at any point drops below the liquid vapor pressure (P v ) for that temperature, then vapor bubbles will form. As the static pressure recovers to a point greater than the vapor pressure, the vapor bubbles collapse back into their liquid phase. The cavitation process includes the vapor cavity formation and sudden condensation (collapse) driven by pressure changes. The growth and collapse of the bubbles produce high-energy shock waves in the fluid. The collapse stage of the process (the bubble implosion) produces more severe shock waves. These implosions generate noise, fluid shock cells, and possible microjets that impinge upon the trim parts. This generates highly concentrated impact forces that cause surface fatigue and localized fractures that destroy the metal. This erosion process gives cavitation damage a very distinctive appearance, like that of cinder block or sandblasting. No known material will withstand continuous, severe cavitation without damage and eventual failure. The length of time it will take is a function of the fluid, metal type, and severity of the cavitation. Without special trim geometry, some of the possible mitigating actions include the use of extremely hard trim materials or overlays, increasing the downstream back-pressure, or limiting the pressure drop by installing control valves in series to reduce the pressure drop in each valve. Another mitigating effect in some processes is a result of the fluid thermodynamic properties and operating conditions. As liquid operating temperature approaches its critical temperature, heat transfer effects become increasingly significant relative to the dominant inertial effects, which causes the growth and collapse rates of cavities to slow down. This can greatly reduce impingement stresses on the valve parts. Some cryogenic and hydrocarbon applications are thought to behave this way, which may partly explain why cavitation damage in these cases is minimal or absent even when cavitation is present in the valve. Further information about cavitation and predicting its effects on valve performance can be found in Section 6.15 of this text and in the Instrumentation, Systems, and Automation Society (ISA) Recommended Practice RP , Considerations for Evaluating Control Valve Cavitation. Some of the special trim designs for combating cavitation are the Drag, Cavitrol, Turbo-Cascade, VRT, Hush, ChannelStream, and various other staged or step-type plugs and orifices, which are shown in Figures 6.1y, 6.1z, and 6.1aa. The multihole Cavitrol-style trims are designed to break the flow into multiple fluid jets and force the jets to impinge upon themselves with extreme turbulence. This turbulence converts part of the upstream energy (static pressure) into heat energy. Bubbles from the many small fluid jets are small and tend to implode within the turbulent fluid core away from the internal surfaces, which greatly reduces trim damage. However, these trims are suitable for only moderate cavitation and moderate pressure drops. The balance of the trims listed can be considered variations of staged trims. These trims reduce the total valve

253 6.19 Valve Types: Globe Valves 1297 pressure drop in multiple steps such that the vapor pressure of the fluid is not reached in any stage. In some cases, the fluid is also forced to undergo multiple changes in direction to promote gradual head loss and energy conversion. It should be pointed out that virtually all anticavitation trims are designed on the basis of data from tests with water. Therefore, actual performance with different fluids cannot be absolutely extrapolated. However, because water is one of the most destructive fluids under high pressure drop conditions, it is likely that a particular valve trim that is tested with water will be reasonably effective with most other fluids. In most applications, actual field experience is necessary to demonstrate the adequacy of a particular trim design. Erosion Erosion of the valve trim can also be caused by highvelocity liquid impingement, abrasive particles, and erosivecorrosive combination action. Erosion damage is roughly proportional to some power (e.g., 1.4 n 6) of velocity, which depends on the erosive environment and the boundary material. As in cavitation, one major key to the solution is to reduce the velocity through the trim. It is no surprise that the valve and trim designs discussed above are also useful on most erosion problems, although there may be some variations of design and materials of construction needed to better cope with a particular problem. A properly chosen and specified control valve and trim type can be one of the most reliable pieces of equipment in process service. Indeed, many companies no longer use control valve bypasses, except in some unique situations. However, it is a requirement that there be close consultation with the control valve manufacturer in specification of special trims. High-velocity liquid impingement erosion is usually associated with high pressure drop coupled with undesirable valve geometry. High-velocity fluid jets developed through the seat area will often result in erratic flow patterns that allow the liquid to impinge directly on the valve trim and body. Such damage is often confined to specific areas in the valve. Liquid droplets in a vapor stream can also cause impingement erosion, but it is generally spread over a greater area. Impingement damage is characterized by relatively smooth grooves and pockets worn into the metal. Abrasive erosion occurs when the fluid stream contains solid particles that are harder than the trim surface and are traveling at sufficient velocity. This erosion can be likened to a type of scouring action that wears away metal, similar to a file or grinder. Solutions to the problem involve the use of harder trim materials, streamlining the flow pattern, and reducing velocity. However, abrasive erosion can only be reduced in magnitude and not entirely eliminated. Good valve and trim service life can be obtained in some cases, but in severe problems other alternatives should be considered. If the fluid and operating conditions are compatible with elastomers, it might be better to consider pinch valves (see Section 6.20). With high velocities or corrosive fluids, ceramic-lined valves or chokes can be used. Erosion-Corrosion Metals in most ambient and process environments resist corrosion by means of a protective metaloxide film. Rust is a form of protective film on iron and steel, even though it has some undesirable characteristics. If the protective film is damaged, worn, or dissolved, the base metal is exposed to further corrosive action and a new oxide film is formed at the expense of the base metal. Protective films can be damaged by particle abrasion, mechanical wear, cavitation, chemical attack, and fluid velocity or turbulence. Common rust is a relatively weak film that is easily disturbed chemically or mechanically. Small additions of alloys, such as copper, to the steel can increase the stability of the iron oxide against atmospheric corrosion, such as in weathering steel. Substantial additions of chromium to steel create a stainless steel that forms a relatively strong protective chromium oxide film instead of iron oxide, which protects against further attack in a wide range of chemical environments and is referred to as the passive layer or film. The flow of fluid through a piping system and especially through valves and fittings can have both electrochemical and mechanical effects on protective films. Velocity, turbulence, and impingement can increase the polarization rates of the oxidation and reduction reactions at the metal-electrolyte interface, which can weaken or dissolve metal-oxide films. The mechanical effects of fluid velocity can more easily remove a weakened or thinned protective film, leaving the base material exposed to further corrosion. When these effects are combined in a way that accelerates the rate of corrosion from a static state, it is called erosion-corrosion or flow-accelerated corrosion. This is a complex phenomenon, and actual service experience may be needed to determine which alloys and trim configuration will give the best service. Depending on the actual metal, chemical, and velocity characteristics of a specific application, erosion-corrosion may appear in different ways. In some cases, the eroded surface may appear like a sandy beach with wave-like ripples or ridges. When there is more direct impingement or severe turbulence, the surface might have deep gouges, undercuts, or gullies. Some cases exhibit a pattern of elongated pits. Erosion-corrosion has been observed in the following applications. Deoxygenated water (condensate and boiler feedwater) or wet steam in the temperature range of F ( C) in carbon steel valves and fittings of fossil-fueled and nuclear power plants Polluted or silty, salt or brackish waters with dissolved or entrained gases with low levels of sulfur compounds in copper, bronze, or brass systems of oilfields and wastewater applications Slurry flow or cavitation in stainless steel systems Users should consult experienced corrosion specialists or metallurgists regarding alloy selection, system design, and process chemistry (or water treatment) to prevent or correct these situations.

254 1298 Control Valve Selection and Sizing Trim Materials The most popular general service trim material is austenitic type 316 stainless steel, which is commonly used with and without hard-facing up to about 800 F (427 C) and in special cases up to 1200 F (649 C). Other harder materials are frequently required in special trims, higher temperatures, and trim parts that might gall because of close tolerance metalto-metal sliding action (cage guiding and guide bushings). Among these materials are 17-4PH, 410, 416, and 440C stainless steels; hardenable Ni-Cr-Fe-Mo alloys (e.g., Inconel 718); cobalt-chromium alloys (e.g., Stellite ); nickel-boron alloys (e.g., Colmonoy and Deloro hard-facing); tungsten carbide; and ceramics. For very corrosive services, more noble or high alloy metals are used to advantage. Among these are Alloy 20 stainless steel, nickel, titanium, tantalum, zirconium, 70Ni- 30Cu alloys (e.g., Monel 400, Monel K-500), Ni-Cr-Fe-Mo alloys (e.g., Inconel), Ni-Cr-Fe alloys (e.g., Incoloy ), Ni- Mo alloys (e.g., Hastelloy -B/B2), and Ni-Cr-Mo alloys (e.g., Hastelloy-C/C276, -C22). It is difficult to generalize on recommended materials or material combinations for valve trims because of the wide range of valve designs and process application requirements. The specifying engineer should utilize not only his or her own knowledge, but also enlist the experience and expertise of the manufacturer and material specialists and metallurgists when needed. Fortunately, many applications are reasonably straightforward, and standard trim material combinations set forth by the manufacturer can be used. Some general guidelines can be given for specifying trim. When specifying hard-faced plugs and seats, the plug can be supplied with hard-facing alloy on the seat surface only (Figure 6.19h). This may be sufficient if the valve is subject to high pressure drop primarily during shut-off. However, for continuous high pressure drop throttling, the full face or contour of the plug should be completely overlaid with hardfacing alloy, unless the base material is already hardened. If the plug is stem-guided or post-guided, the lower guide area should also be hard-faced for high pressure drop throttling or if the fluid temperature is above 750 F (400 C). As with the plug, the seat ring can be hard-faced only on the seating surface or over the entire bore surfaces, depending on the severity of the pressure drop and temperature. A variety of cobalt-chromium and nickel-chromium-boron alloys are available for hard-facing. In some cases, coating techniques such as flame or plasma spraying will be used. The valve manufacturer s recommendations are valuable guidance. Commonly used hardenable alloys for trim materials include 410 and 416 stainless steels hardened up to about 38 Rockwell C (HRC) 440C stainless steel with hardness up to 60 HRC 17-4PH stainless steel, a precipitation-hardened material combining good corrosion resistance with a range Full contour FIG. 6.19h Hard-facing of plug and seat rings typically used in erosive services. (Courtesy of Flowserve Corporation.) of tensile properties and hardness between 28 and 42 HRC depending on heat treatment The 400-series (martensitic) stainless steels have limited corrosion resistance suitable for most water and steam service up to about 700 F (371 C). They are generally used for trim parts that can be made from bar stock or forgings, but some types are available as castings. Stainless steel 17-4PH is available in cast and wrought forms and is a usable generalpurpose alloy up to 750 F (400 C). It is not as corrosion resistant as austenitic 316 stainless steel, but it is significantly better than most common 400-series types. Where 316 or other soft alloys are indicated as the proper alloy for a cageguided trim, hard-facing will be required to minimize metalto-metal galling problems. Leakage Seat surface Stellite variations seat Seat surface Stellite variations plug Full bore Lower guide area Control valves have varying degrees of shut-off capability, depending upon the valve and internal trim design, material, and manufacturing methods. Tight shut-off is not always a requirement, especially for throttling control valves. However, as reliability of control valves has increased, more applications are requiring tight shut-off for control valves in order to minimize the number of isolation valves and bypass loops. Another consideration should be the cost of lost or contaminated product and wasted energy resulting from leakage through valves. Users should consider all of these factors along with the cost and operating requirements of the valve before specifying the shut-off requirements. Shut-off requirements are usually specified as an allowable volumetric leakage rate measured in a standardized test.

255 6.19 Valve Types: Globe Valves 1299 Plug Teflon insert FIG. 6.19i Typical soft-seat insert designs. (Courtesy of Dresser Flow Solutions.) While it is permissible for a user to specify any value for the allowable leakage at a specified pressure drop, it is more common to specify one of the standard leakage classes defined by industry standards. The two most commonly applied standards, FCI 70-2 (formerly ASME B16.105) and IEC , define several classes, with Classes II, III, IV, V, and VI being most commonly used for globe valves. Class V and Class VI represent the smallest allowable leakage depending on the pressure drop and test method specified. Refer to Section 6.1 for information about determining allowable leakage in each class. Good single-seated valve trims can give Class IV or V shut-off in nonbalanced plug designs. Leakage Classes III or IV are typical for balanced plugs, but elastomer seals and special pilot-operated balanced trims have been used to achieve Class V shut-off. Figure 6.1gg gives the seat leakage class tabulation. Typically, Class VI shut-off is specified for soft seat inserts such as polytetrafluoroethylene (PTFE) or Teflon, ETFE (Tefzel ), or other plastics. Figure 6.19i shows two typical configurations where the soft insert can be located in either the plug or the seat ring. Class VI leakage is also achievable with special lapped-in or precision-fit metal plugs and seats. Note that operating service temperatures, high or low, will limit the use of soft seat materials. The pressure drop across the valve is another limiting factor, although some protected insert designs (Figure 6.19j) will operate at very high pressure drops. Seat leakage tests and classes are defined only for new valves, and it should not be assumed that the same level of tightness can be maintained in service. When a valve is placed in service, the seating surfaces can be worn or damaged by high velocity, pipe scale, process solids, corrosion, or vibration, which cause leakage to increase over time. Wear-resistant materials and careful installation and start-up practices can minimize wear and damage to seating surfaces. Plug Stems The valve stem connects the plug head to the actuator stem or coupling. It has several requirements as part of a carefully balanced mechanical system consisting of the plug head and stem, seat ring, cage, bonnet, stem bushings, stem packing, and actuator. It must be strong enough to transfer the load Seat ring housing Sliding collar FIG. 6.19j Protected soft seat insert for high pressure drop and tight shut-off. (Courtesy of Dresser Flow Solutions.) from the actuator to plug and bear the seat loading forces without cyclic fatigue. It must be stiff enough under maximum actuator loading to prevent buckling or significant deflections that promote packing wear and leakage. Yet it cannot be so big as to generate excessive packing friction or unnecessary static pressure forces that make actuation and control difficult. The proper function of the valve also requires precise alignment of the plug and stem with the seat, bonnet, and actuator in order to prevent excessive friction and binding and to ensure minimal leakage at the seat or through the packing. Materials, design, and surface finish of the stem and bushings must prevent galling and minimize friction and packing wear. Material selection of the stem includes consideration of the process environment inside the valve, the range of environmental conditions outside the valve, and a mixture of those environments in the packing interface. In some unusual corrosive services, moisture or chemicals in the atmosphere can react with process chemicals to create a more corrosive mixture in the stem-packing interface. This situation may require a stem material with more corrosion resistance than the plug head that is exposed only to the process fluid. Valve stems are normally the same material as the plug head, but they can be different based on the design criteria discussed above. Depending upon the type of valve design, the valve stem may be integral (one piece) with the plug, or it can be threaded into the plug and then pinned to prevent unscrewing. Because the manufacturer does a number of things to insure proper alignment and a solid, vibration-resistant threaded connection, good maintenance practice may dictate replacement as a unit rather than separate pieces. Stem guides in one form or another are an integral part of the trim assembly. Metal stem guide or bushing material is selected to minimize metal-to-metal wear and galling

256 1300 Control Valve Selection and Sizing TABLE 6.19k Common Trim Material Characteristics Trim Material Hardness Rockwell C Impact Strength Corrosion Resistance Recommended Max. Temperature Erosion Resistance 316 stainless steel (without hard-facing) 316 stainless steel (with hard-facing) Rockwell B Excellent Excellent 600 F 316 C Varies with hardfacing Excellent Excellent 1200 F 649 C 410/416 stainless steel Good Fair 700 F 371 C 420 stainless steel 50 Good Fair 800 F 427 C 17-4PH H Good Good to excellent 800 F 427 C 440C stainless steel Fair Fair 800 F 427 C Ni-Cu Alloy K Good Good to excellent 600 F 316 C Tungsten carbide ~72 Fair to good Good (with nickel alloy binders) Co-Cr-W Alloy 6 (hard-facing, cast, or wrought) 1200 F 649 C 42 Fair Good 1500 F 816 C Ni-Cr-B No. 5 (hard-facing) Fair Good 1200 F 649 C Fair Excellent Good Excellent Good Excellent Fair to good Excellent Good Good Courtesy of Flowserve Corporation. against the stem. For lower temperatures and light duty, nonmetallic guide bushing materials, such as reinforced PTFE fluoroplastic or compressed graphite, are common choices. Metal guides may be of such materials as 17-4PH, 440-C, Stellite, or hard-chrome plated or nitrided stainless steel, bronze, and aluminum-bronze. For the main characteristics of the common trim materials refer to Table 6.19k. BONNET DESIGNS The valve bonnet is the top closure assembly for the globe valve, as well as for several other valve body design types. In addition to closing the valve body, the bonnet also provides the means for mounting the actuator assembly to the valve body and sealing the valve stem against process fluid leakage. The various bonnet designs will be discussed along with the subject of stem sealing utilizing packing materials, lubricants, and special seal designs. In addition to considerations of pressure containment, manufacturers design bonnets to provide features that predominate in their particular philosophy of valve design. The depth of the stuffing box and surface finish, provision of guides, method of operator attachment, packing design flexibility, and packing follower design are all details that vary from manufacturer to manufacturer or even among various body designs offered by one manufacturer. Some low-pressure valves, especially in sizes below 2 in. (DN 50), can be provided with a threaded bonnet. This reduces the weight and is more economical in first cost than the flanged and bolted design. Depending on the materials of construction, it can be difficult to remove a threaded bonnet after extended service in high temperature or corrosive service, making this type more costly to maintain in these types of applications. Economic justification for using a valve with a threaded bonnet is sometimes based on a life-cost model of discard and replace, whereas larger valves, bolted bonnet valves, and control valves typically require a maintain and repair cost analysis. In special cases the bonnet-to-body joint can be seal welded to prevent leakage of highly reactive, lethal, or radioactive fluids. Bolted Bonnets The most common valve bonnet design is the bolted bonnet, like the one shown in Figure 6.19l. It is usually fastened to the valve body by high-strength stud bolts and heavy nuts. Removal of the bonnet gives complete access to the valve trim for maintenance purposes. Because the bonnet is a pressure-retaining part of the overall fluid containment system,

257 6.19 Valve Types: Globe Valves 1301 FIG. 6.19l Bolted bonnet joint showing retained gasket design. Bridgeman gasket (Figure 6.19m). These latter designs are metal gaskets. The flat gasket is usually either graphite-based or a reinforced PTFE. The spiral-wound gasket usually consists of graphite ribbon or PTFE wound between thin metal strips. The metal windings of a spiral-wound gasket are usually stainless steel, but they are available in a variety of corrosion resistant and high-temperature alloys. Asbestos fillers were commonly used in gaskets until the late 1970s, largely due to the lack of effective substitutes for many high-temperature or corrosive sealing applications. Since then, sealing materials and technologies have improved to the point where there should be no further requirement for using this potentially hazardous material in gaskets. Proper handling and disposal are required if very old valves are encountered in order to reduce the risks of asbestos exposure. Pressure Seal Bonnets the design and materials are determined in accordance with an applicable pressure vessel standard. For example, the ASME Pressure Vessel Codes, ASME B16.34, API 6D, AD Merkblätter, European norms, ISO standards, and numerous individual nations pressure vessel standards give material requirements and design criteria for flanges, wall thickness, and flange bolting. Bonnet Gaskets The seal between the valve bonnet and body can take several forms depending upon the valve design and application range, including containment pressure and fluid temperature. Fluid corrosion is also a factor that must be considered. The most common seal is a contained gasket, either a flat or spiral-wound composite design (Figure 6.19l). Other designs that may be found are the API ring joint (oval or octagonal cross-section), lens type, delta gasket, and Another typical bonnet seal is called the pressure seal, and it is used in high-pressure valves Class 900 (PN 150) and higher to minimize size and weight of the bonnet-to-body connection. The pressure-sealed bonnet does not rely on heavy bolting and a thick flange to retain pressure, like the bolted bonnet. Instead, a wedge-shaped graphite-composite gasket is installed on top of a bonnet sealing lip and is compressed between the bonnet and a set of antiextrusion or spacer rings and heavy retaining ring segments that are indexed into the body. During assembly, the bonnet is pulled up tightly against the seal, spacer ring, and gasket retainer segments by the pull-up bolting and bonnet retainer to create a radial seal between the body neck and bonnet, as shown in Figure 6.19n. When the valve is pressurized in service, the upward force from pressure under the bonnet further energizes the radial sealing. The design and materials of pressure-sealed API Ring joint gaskets Lens type gasket Delta gasket FIG. 6.19m Special bonnet gaskets for high pressure and temperature. Bridgeman gasket

258 1302 Control Valve Selection and Sizing from 20 to 600 F ( 30 to 315 C). Above 450 F (230 C) graphite-based packing or extended bonnets are recommended. The reason for setting a temperature limit for standard-length bonnets is to limit the temperatures to which the packing and actuator are exposed. Over 90% of all control valve applications can be handled by the plain bonnet design. It may or may not incorporate stem guides or bushings and may have very broad or very limited packing configurations available, depending upon the specific manufacturer and valve design. FIG. 6.19n Pressure seal bonnet configuration showing composite wedge gasket. (Courtesy of Flowserve Corporation.) gasket joints are especially important. If the gasket material bonds so tightly in service to the inside bore of the body that it becomes difficult to remove for maintenance, it may be time-consuming or impossible to free the bonnet without damage to the body. Fortunately, recent design and material improvements have minimized these types of maintenance problems. Pressure-sealed bonnet joints have advantages in high-pressure, high-temperature applications, because they permit more uniform body wall thickness at the body neck opening, which minimizes thermal stresses in the body during temperature transients. The valves are frequently lighter in weight than comparable bolted bonnet valves, and they are relatively easy to assemble. Extended Bonnet The extended bonnet (Figure 6.19o) is usually required when the fluid temperature is outside the plain limitations of the standard bonnet temperature. Even when normal process temperatures are within the plain bonnet limits, it may be necessary to use the extended bonnet to protect the packing and actuator against temperature excursions during occasional process upsets. Originally, the extended bonnet used different designs for hot and cold service. The hot service extended bonnet was provided with cooling fins, while the cold service extended bonnet was a plain casting without fins. Over many years it was demonstrated that fins on the bonnet added only marginal capability to heat dissipation in the packing area. It made a costly and complex casting and has been largely abandoned in favor of the plain extension. In most modern control valve designs, the bonnet extension is similar for hot or cold service, except where deep cryogenic temperatures under 150 F ( 100 C) are encountered. Some manufacturers offer two standard bonnet extension Bonnet Classification Bonnets fall into basically three classifications. These are standard, extended for hot or cold service, and special designs such as cryogenic extensions and bellows seals. These classifications along with the stem seal systems, guides, and bushings will be discussed in more detail later in this section. Standard Bonnet The standard or plain bonnet (Figure 6.19d) is the normal bonnet design furnished on most valves. It covers the range of pressures and temperatures compatible with standard seal gaskets and stem packing materials. Generally, this includes valves designed for ANSI Pressure Classes 150 (PN 20) through 2500 (PN 420) and temperatures FIG. 6.19o Standard high-temperature extension bonnet. (Courtesy of Flowserve Corporation.)

259 6.19 Valve Types: Globe Valves 1303 FIG. 6.19p Standard cold extension bonnet. (Courtesy of Flowserve Corporation.) FIG. 6.19q Typical cryogenic cold-box valve. (Courtesy of Flowserve Corporation.) lengths (other than cryogenic) depending upon the operating temperature. In general, the standard extension bonnet is suitable from 20 to 800 F ( 30 to 425 C) in carbon steel construction and from 150 to 1500 F ( 100 to 815 C) in austenitic stainless steel construction. Cryogenic Bonnet The cryogenic bonnet or cold-box bonnet (Figure 6.19p) is a special design adaptation of the extended bonnet. Depending upon a manufacturer s valve design, this style bonnet may be required with operating temperatures ranging from 150 to 300 F ( 100 to 185 C) down to 425 F ( 255 C) using a bolted bonnet. The main purpose for extending the bonnet for cold service is to keep the packing and upper plug stem warm enough to prevent icing or frost around the packing area and stem. Ice crystals on the stem can be pulled into the packing, damage the seal, and create a leak path. The bonnet length is selected for the application based on valve body size, piping requirements, and operating temperature needs; it will generally range from 12 in. (300 mm) to 36 in. (900 mm). The standard cryogenic bonnets are distinctly different from cold-box designs or extended-neck designs (Figure 6.19q). The standard design is usually similar to the standard extension bonnet, except much longer. It can be cast or fabricated from multiple pieces by welding. For extreme temperatures near 454 F ( 270 C), the body neck is often extended to remove the bonnet gasket away from the extremely low temperature and to make the bonnet joint accessible outside of the cold-box. The extended neck of the cold-box body is usually fabricated from thinwalled stainless steel tubing (to reduce cool-down weight and heat leakage into the cryogenic process). The extension is welded directly to the body casting or forging (Figure 6.19q). At the top of the extension is the flange for connecting with the bonnet. In general, these applications are limited to ANSI Pressure Class 600 or below. Special designs are also available for high-pressure service up to ANSI Class 4500 and above. Because the extreme cold requires good impact resistance, materials of construction are limited to the austenitic stainless steels (Types 304L or 316L) and bronze. These valves incorporate a stem seal system at the plug end of the bonnet to keep the cryogenic liquid out of the bonnet and packing area. This seal may be vented or nonvented, but it must allow a pocket of vapor to exist below the bonnet as insulation against severe convective heat loss from the warmer bonnet area if exposed to cryogenic liquid. The seal design must also allow any build-up of gas pressure in the warmer bonnet area to relieve back into the valve body. Vented designs are typically used if the valve is installed with the neck and stem oriented vertically up to within 30 of horizontal to prevent liquid from getting to the packing area. Nonvented designs incorporate a unidirectional seal and are used when the valve neck is installed at or near horizontal. In some cases, where additional insulation is needed to reduce outside heat flow, the bonnet can be fitted with a vacuum jacket. Bellows Seal Bonnet When no stem leakage can be permitted, many globe valve manufacturers provide extended bonnet designs that incorporate a bellows seal around the stem (Figure 6.19r). Bellows seals are justified in applications involving toxic or radioactive fluids, where leakage to the

260 1304 Control Valve Selection and Sizing for size 1 in. (DN 25) and smaller valves to 8000 cycles for sizes 3 6 in. (DN 75 DN 150). In some cases, cycle life can be improved by reducing operating pressures or by using special short stroke valve plugs. Operating pressures can be increased to as much as 2900 PSIG (20,000 kpa) and temperatures up to 1100 F (590 C) by multiple or heavy-wall bellows and selection of a highstrength metal alloy. However, this reduces cycle life considerably. As a result of these various factors, metal bellows bonnet seals are selected for relatively few applications. Improved environmental packing systems are specified for a majority of hazardous fluids, except for dangerous applications where the risks justify the additional costs of purchase and maintenance for metal bellows seals. Bonnet Packing FIG. 6.19r Typical bellows seal bonnet with formed bellows. (Courtesy of Flowserve Corporation.) outside would pose personnel safety hazards. The bellows is usually made of stainless steel or other corrosion-resistant nickel alloys such as Hastelloy C-276 and Inconel 625. They can be hydraulically or mechanically formed (as shown in Figure 6.19r), or they can be made by stacking many individual leaf segments that are welded together at their outer edges and are known as welded or nested bellows. The bellows is attached to one end of the stem by welding; the other end is welding to a clamped-in fitting with an antirotation device. The antirotation device prevents the bellows from being twisted during assembly and disassembly or by vibration of the plug in service. Bellows seals are usually leak-tested from atmospheric pressure to vacuum using a mass spectrometer to detect helium leakage rates below cc/sec. The service life of a metal bellows depends on the design, material, manufacturing processes, service pressure and temperature, corrosion effects, and stem cycle history. Failures of metal bellows usually occur by cyclic fatigue. Bellows manufacturers that have tightly controlled production processes have cycle-tested their designs and are able to make reasonable predictions of expected cycle life. Bellows-sealed bonnets are backed up with a standard stem packing set and a leakage monitoring port between the bellows and the packing in order to prevent catastrophic release of hazardous fluid in the event of a bellows leak. Metal bellows seals have pressure and temperature limitations. Ratings of about 150 PSIG (1030 kpa) at 100 F (40 C) or 90 PSIG (620 kpa) at 600 F (315 C) are typical. The average full stroke cycle life can vary from 50,000 cycles In order to seal the valve stem against leakage of process fluid to the atmosphere, the upper part of the bonnet contains a section called the stuffing or packing box. This assembly consists of a gland flange, packing follower, packing spacer or lantern ring, lower packing retainer, and a number of packing rings. The surface finish of the stem should be very fine, on the order of 8 Ra (micro-inch roughness average), and the internal finish of the bonnet stuffing box should be 16 Ra or better. Various valve packing materials are available, but for control valves use three general groups of materials: fluoroplastics (PTFE, FEP, PCTFE), carbon graphite (compressed rings and braided yarn), and synthetic polymer fibers (PBI and aramid). Control valve packing must be compatible with the process fluid, seal the stem and bonnet, produce minimum starting and sliding friction, and give long service life in modulating service. The most popular material that meets all of these conditions over the broadest range of fluid applications is PTFE, used as V-rings or braided filament. As a result, the majority of control valve manufacturers provide a variety of packing configurations with this material. Fluoroplastic Packing Polytetrafluoroethylene (or Teflon) is the most common packing material in the family of fluoroplastics. It is normally formed or machined as chevron or V-rings from virgin (not reprocessed) material. For special needs, a shape variation known as cup-and-cone is available. Note that in V-ring and cup-and-cone configurations, the top and bottom rings are adapter rings with one flat surface. Braided packing is also formed from PTFE filament. For higher temperature or pressure applications, PTFE can be reinforced or filled with up to about 25% by volume of glass fibers, silica, carbon, or graphite and other fillers to add strength and stiffness and to improve its resistance to coldflow or creep. PTFE braided packing can be reinforced by using PBI or aramid fiber in the corner braids. As noted earlier, PTFE is limited to a maximum exposure temperature

261 6.19 Valve Types: Globe Valves 1305 Pressure (bar) Temperature (F ) Braided carbon PTFE vee Filled PTFE vee & braided PTFE Braided graphite Combination graphite Temperature (C ) 0 Carbon and graphite packings PTFE packings Pressure (psig) FIG. 6.19s Recommended temperature-pressure limits for some common packing materials. (Courtesy of Flowserve Corporation) of about 450 F (230 C) in a plain bonnet, depending on the service pressure. When used in extended bonnets, some users have successfully used reinforced PTFE packing with process temperatures up to about 850 F (455 C), providing the packing box temperature remains below 450 F (230 C). Although PTFE packing has useable properties down to 320 F ( 195 C), its practical lower limit for packing in a standard length bonnet ranges from 50 F ( 46 C) to 100 F ( 73 C), depending on bonnet design. See Figure 6.19s for pressure-temperature limits for PTFE and other packing in standard bonnets. In order to prevent ice damage to packing at lower temperatures, extended bonnets or cold-box-style bodies are used (Figures 6.19p and 6.19q). Contrary to some claims, PTFE packing does not require live-loading springs to be effective as a stem seal. Normal packing follower loading and adjustment is all that is required, especially with the V-ring or chevron shape (Figure 6.19t). The cup-and-cone style requires a higher loading to effectively energize the seal. Live-loading springs are beneficial in special applications, such as in temperature cycling duty and for some types of environmental seals to reduce volatile organic chemicals (VOCs) and toxic emissions. The packing rings and stuffing box dimensions should be very accurate for proper contact with the sealing lips of the V-rings. Dimensional tolerances are even more critical for the PTFE cup-and-cone rings because of the high degree of stiffness and packing loading required. PTFE packing has extremely low friction characteristics and does not require supplemental lubrication. However, excessive overtightening of the gland bolting can transform the V-ring structure into a solid compression packing capable of transferring high radial forces and excessive friction to the stem. Asbestos Packing Asbestos is one of the oldest packing materials and was in wide use until the late 1970s, even though the health hazards of airborne fibers were known before then. Health risks were generally insignificant when handling new packing with binders to keep the fibers from fraying. However, in high-temperature service, the binders disappear, and removal of old packing from stuffing boxes Teflon or graphite lined upper guide bushing and packing follower Packing spacer Lower guide bushing and packing follower FIG. 6.19t Typical Teflon V-ring arrangement. Teflon v-rings Bonnet Teflon v-rings

262 1306 Control Valve Selection and Sizing had the potential of creating airborne fibers unless special dust mitigation procedures were employed. The lack of effective asbestos substitutes for many high-temperature or corrosive sealing applications made it difficult to abandon its use. Since then, sealing materials and packing technologies have virtually eliminated further need for using this potentially hazardous material. In order to reduce the risks of asbestos exposure, proper handling and disposal are required if very old valves are encountered. Graphite Packing The improvement and optimization of graphite foil and yarn for use as valve packing came about largely out of the need to find substitutes for asbestos packing. Flexible graphite packing is available in formed rings of laminated graphite foil. Rings are formed in square, wedged, and cup-and-cone cross-sections. A wide variety of braided constructions are available. Braided graphite can include dry lubricants to reduce friction. Braided graphite can be reinforced for high-pressure service using synthetic fibers like PBI or aramid. Reinforcing of the graphite braid with an Inconel wire core or fine wire-encapsulated yarn is common for high-temperature, high-pressure service. Several typical graphite packing configurations consist of a combination of formed laminated rings with braided rings above and below the laminated set to act as wiper rings. Flexible graphite packing can be used in many fluids including reducing and mildly oxidizing acids (not strong oxidizers), caustics, hydrocarbons, solvents, water, steam, and gases. It has high-temperature capability up to 1200 F (649 C) in steam; and in inert, nonoxidizing media (i.e., oxygen-free environments like nitrogen or carbon dioxide) pure graphite it is useable up to 4500 F (2500 C). In atmosphere and oxidizing media, the high-temperature limit can vary between 650 F (343 C) and 850 F (454 C) depending on the specific materials; consult with packing manufacturer for specific applications. If an extended bonnet is used, the upper temperature can be extended to 1200 F (650 C) on both oxidizing and nonoxidizing service. The low temperature limit for pure graphite braided packing can be as low as 400 F ( 240 C). Packing and valve manufacturers have undertaken considerable developmental work in an attempt to overcome or mitigate past problems with graphite. Among these shortcomings are the following: Relatively high stem friction Difficulty in energizing the packing to give an effective stem seal Low cycle life without leakage Electrolytic pitting of stainless steel stems in conductive or high-temperature services Shortened packing life due to graphite plateout on the stem There are several methods that can be used to extend the life and improve the performance of graphite packing. The packing or valve manufacturer should be consulted for proper methods for their specific sealing system. Some general things that have helped to improve graphite packing performance include the following: 1. Use combination packing assemblies consisting of laminated graphite and braided graphite fiber rings. The braided graphite rings help as antiextrusion or wiping rings and prevent graphite plateout on the stem. 2. Use sacrificial zinc washers where possible or hard chrome plating or stem materials resistant to pitting attack. Some packing include a zinc powder as anodic protection for the stem. 3. Carefully torque the packing flange nuts to the minimum torque recommended by the valve manufacturer. Overtorque will result in excessive valve stem friction and may even lock the stem. 4. Remove the packing while the valve is in storage or out of service for extended time periods. 5. Clean graphite plating from the stem before installing new packing. 6. Avoid trapping air between the rings during installation. Leave the ring level with the chamfer of the stuffing box cavity and install the next ring on top. 7. Breaking in new packing by tightening the gland in gradual steps while cycling the valve at least 10 times may help. A break-in lubricant applied to the packing rings and stem (e.g., nickel antiseize or silicone grease) is also recommended. One of the oldest methods for reducing stem friction with graphite or any braided packing (such as asbestos) involved a lubricator fitting that was added to the bonnet. In this design, a compatible grease compound is injected into the lantern ring area by a packing lubricator assembly (Figure 6.19u). A loading Note: Lubricator optional FIG. 6.19u Stuffing box assembly with external lubricator.

263 6.19 Valve Types: Globe Valves 1307 bolt is turned to force the grease sticks into the packing. An isolating valve should be used for safety. External lubrication works reasonably well, although it is cumbersome and requires constant maintenance for checking and reloading of the lubricator. Also, it is sometimes difficult to find lubricants compatible with the process fluid. FIG. 6.19w Typical twin- or double-packing arrangements. (Courtesy of Flowserve Corporation.) Packing Arrangements There are a number of packing arrangement systems in use to suit various types of packing and fluid containment problems (Figures 6.19v). In general, the arrangements are equivalent for either Teflon V-rings or square rings (braided and laminated), although the number of rings will vary for each packing type and manufacturer s sealing system. In a standard packing arrangement, a lower set of rings serves to minimize ingress of solids from the process into the stuffing box and acts as a stem wiper. The upper set provides the primary seal and consists of several rings. For harder-to-hold fluids or when a leak detection port, bonnet purge port, or lubricator is used, a packing arrangement known as the twin-seal is often used (Figures 6.19w). A variation of the twin seal arrangement is also recommended for vacuum service. In the twin seal, there are two full sets of packing installed. This system requires a bonnet with deeper stuffing box, because space is needed for spacers or lantern rings and a lower guide bushing as well as the packprimary stem seal. However, in vacuum service, the upper set is inverted with the V-opening toward the atmospheric side as the primary seal, because positive pressure is from atmosphere to the negative pressure inside of the valve. For toxic or radioactive fluid services, the bonnet can be tapped at the lantern ring area, and this connection can be used in three different ways, depending upon specific design requirements. 1) The tap can be used for a leakage monitor connection using a pressure gauge or switch. 2) It can be used for sampling the process. 3) More commonly, the tap is used as a leak-off connection, whereby any process leakage is piped to a disposal header or nonpressurized waste container. In cases where the fluid is a slurry or tends to solidify, the tap can be used as a purge connection to keep process media out of the bonnet. Here, an inert gas or liquid (depending upon the process) is introduced at a pressure well above the highest expected process pressure. This purge material provides an additional pressure seal, and any packing leakage will be in the direction of purge into the process or from the purge to atmosphere. FIG. 6.19v Typical standard packing arrangements. (Courtesy of Flowserve Corporation.) Environmental Packing The need for more aggressive protection of the environment against volatile organic chemicals and clean air regulations in Europe (e.g., TA-Luft standards), United States (e.g., EPA standards), and most other countries have motivated the development of special packing configurations and

264 1308 Control Valve Selection and Sizing in service can cause excessive O-ring wear or sticking in dry service. A standard stem seal packing ring system is usually preferred in sliding stem valves. FIG. 6.19x Typical configurations of environmental (fugitive emissions) packing. (Courtesy of Flowserve Corporation.) materials. These regulations generally establish compliance below benchmark levels varying from parts per million (ppm) of VOC. Figure 6.19x shows only two of many different environmental packing system designs. Environmental packing systems often consist of a combination of ring designs and different materials. For low to moderate temperatures, a common packing set can have unfilled PTFE or chemically inert elastomer V-rings stacked between stiffer thermoplastic or reinforced PTFE chevrons for a tighter and more durable seal. For higher temperatures, various combinations of graphite braided rings with laminated graphite rings (square, wedged, or cup-and-cone) are used. Any of these configurations can include live-loading spring arrangements to help maintain constant packing loads and make up for consolidation or cold-flow of the packing over time. Antiextrusion rings are also used in some systems to promote longer packing life. Most valve manufacturers offer a variety of environmental packing styles to suit a wide range of applications. Other Packing Materials Elastomeric O-rings are used for stem sealing in some globe valve designs, usually for less low-pressure, low-temperature utility services such as water. O-ring stem packing tends to be more common for rotary valves than for sliding stem valves. Proper selection of an elastomer that is compatible with the process fluid chemistry and temperature can also be a problem. In gas service or with liquids containing gases, many elastomers will absorb some gas. Depressurizing the valve will result in sudden expansion of this gas. If the gas cannot diffuse rapidly out of the elastomer, the resulting internal pressure can rupture or split the O-ring. This effect is called explosive decompression (ED). Most O-ring manufacturers now offer special grades of some elastomers that are resistant to ED failures. The O-ring gland design requires a fully retained groove and probably backup rings to minimize extrusion or blowout. Installation of O-rings requires great care to avoid cutting, twisting, or other damage. Lubrication may be needed to overcome sliding friction, and gradual loss of this lubrication Packing Temperature The relationship between the process and the packing temperature is a function not only of the type of bonnet used but also of the valve metallurgy and the valve and bonnet physical relationship. Heat is transmitted to the packing area by conduction through the metal, by convection via the process fluid, and by the relative heat radiation balance with the ambient environment. Stainless steel, for example, has a much lower heat conductivity coefficient than carbon steel and thus is about 20 30% less efficient in conducting heat into the packing area. This does not mean that the packing temperature rating can be increased, but it may serve to reduce some heat load and increase packing life. (See Figure 6.1o in Section 6.1 for further discussion and illustration of packing temperature determination.) In some cases, the valve can be installed upside down so that the bonnet is below the valve body. In liquid service, this reduces convection, and heat is transferred to the bonnet mainly by conduction. In vapor service (with or without superheat), it may be possible in the stem-down orientation to condense sufficient vapor in the bonnet to create a condensate seal and lower the packing temperature to a suitable level. The potential corrosive effects of alternate wetting and drying must be considered with this approach. However, this approach to controlling packing temperature has serious disadvantages where maintenance is concerned, and the benefits generally do not justify the potential maintenance difficulties except in unusual cases. BODY FORMS The actual pressure containment and fluid conduit portion of a control valve is called the valve body assembly. This assembly consists of the body, a bonnet or top closure, sometimes a bottom flange closure, and the internal elements known as the trim. The body can have flanged, threaded, or welded end (butt weld or socket weld) connections for installation into the piping. The trim consists of such elements as the plug and stem, guide bushings, seat rings, cages or seat retainers, and stuffing box lantern ring. The body configuration can be in-line, angle, offset inline, Y-type, and three-way. Some of these will be discussed in more detail. The shape and style of the valve body assembly is usually determined by the type of trim elements it contains, piping requirements, and the function of the valve in the process system. There are a large number of body designs on the market, including a number of special-purpose designs. The end result is a device that can be fitted with an actuator and used to modulate the flow of process fluid to regulate such things as pressure, flow, temperature, liquid

265 6.19 Valve Types: Globe Valves 1309 level, or any other variable in a process system. Examples of some of the more widely used body configurations are shown in Figure 6.19y. Double-Ported Valves The double-ported (double-seated) balanced valve (Figure 6.19z) was one of the first globe valves developed during the early 20th century. It is still available today, but has been replaced in most applications by single-seated globe valves. Size for size, it is much larger and heavier than its single-seated counterpart. Shut-off is poor because it is not practical to have both plugs in tight contact with the seats at the same time, but the valve was intended for throttling control rather than for tight shutoff. Some special seat designs have been developed to help overcome this, but application is limited. The double-ported valve is considered semibalanced; i.e., the hydrostatic forces acting on the upper plug partially cancel out the forces acting on the lower plug. The result is less actuator force requirement, and a smaller actuator can be used. However, there is always an unbalanced force due to the difference between the upper and lower plug diameters required for assembly. In addition, unbalance forces are generated by the effect of dynamic fluid forces acting on the respective throttling areas of each plug. Such forces can be quite high, particularly with the smooth contoured plugs; these can reach as much as 40% of the forces of an equivalent unbalanced single-seated valve plug. Double-ported valves have been built in sizes up to 24 in. (DN 600), although most manufacturers now limit them to 12 in. (DN 300) as a maximum. Figure 6.19z shows that the valve can be converted from the push-down-to-close configuration shown to a push-downto-open design. This is done by removing the bottom closure flange, bonnet, and stem, and by inverting the entire assembly and reinstalling the stem, flange, and bonnet. This, coupled with the use of direct-acting (air pushes the stem down) and reverse-acting (air pushes the stem up) actuators, gives full flexibility to provide the required valve failure mode. Single-Seated Valves Single-seated valves are the most widely used of the globe body patterns. There are good reasons for this. They are available in a wide variety of configurations, including special-purpose trims. They have good seating shut-off capability, are less subject to vibration due to reduced plug mass, and are generally easy to maintain. There are three general types of seat construction. 1) The floating seat ring sits in machined bore in the body with a gasket to seal the joint between the body and seat ring. It must be retained in the body by a cage or seat retainer to maintain gasket tightness and concentricity with the plug. 2) The screwed-in seat ring is threaded into matching body threads with a special tool; a gasket may or may not be required. A separate seat retainer is not required, but some designs use a cage for guiding the plug or characterizing the flow. 3) The seating surface can be machined directly into the body; this is called an integral seat. The floating and screwed-in seat rings can be replaced after they wear out, but the integral seat requires resurfacing or machining of the body to repair wear or damage. Single-seated valve plugs are guided in one of four ways: post-guided (Figure 6.19aa), top-and-bottom-guided (Figure 6.19bb), stem-guided (Figure 6.19cc), and cage-guided (Figure 6.19dd). The most popular globe valves are stem-, cage-, or post-guided types, which require only one body opening for the bonnet and have one less closure gasket subject to leakage than the top-and-bottom-guided configuration. The stem-guided and post-guided designs provide more streamlined flow and are less subject to fouling in dirty service. The stem-guided valve minimizes stagnant fluid cavities and may be a better selection than the post-guided and cage-guided valves when dealing with fluids containing solids, sticky or viscous fluids, or highly corrosive fluids. The top-and-bottom single-seated valve (Figure 6.19bb), like its double-seated counterpart (Figure 6.19z), has similar limitations, but some users still prefer this design where the plug is held by two, widely spaced bushings. The top-entry single-seat globe valve is most commonly used in sizes from 1 in. (DN 25) through 12 in. (DN 300) from most manufacturers. Some top-entry designs are manufactured with bodies suitable for slip-on flanges (Figure 6.19y) rather than with integral cast flanges. This type of flange construction is discussed in more detail below, under Split-Body Valves and Valve Connections. Cage Valves The cage valve is a variant of the singleseated valve and is the most popular design used in the process industries. The top-entry bonnet and trim design makes it extremely easy to change the trim or to do maintenance work. Cages are used with floating seat rings and with screwed-in seat rings. The design is very flexible in that it allows a variety of trim types to be installed in the body. This includes such variations as reduced trim, anticavitation, and low-noise trims (Sections 6.1 and 6.14). The overall design is very rugged, and with proper specification of trim type and materials, cage valves provide relatively trouble-free service for extended time periods. These valves may eliminate the need for block and bypass valves in some cases, because their service life can be as good as or better than most other components in the process that require periodic maintenance. There are two basic design configurations available for cage valves. One type uses the cage solely as a seat retainer to clamp a floating seat ring into the valve body (Figure 6.19cc). This design is usually stem-guided or post-guided, and the valve plug is characterized and does not guide or control flow through the cage. The other type uses the cage to guide the plug head, and the cage openings are shaped to provide the desired flow characteristic as the valve plug exposes the ports (Figure 6.19dd), and it is called a cage-guided valve. In the cage-guided valve,

266 1310 Control Valve Selection and Sizing O S FIG. 6.19y Typical globe body configurations. (Courtesy of Flowserve Corporation.)

267 6.19 Valve Types: Globe Valves 1311 Interchangeability: Class 600 body used for class 150, 300 and 600 ratings. Very heavy guiding: Two widely-spaced guides on heavy stem. Plug does not guide in retainer. Clamped-in seat ring: Valve may be disassembled quickly, easily by removing four bonnet bolts. FIG. 6.19z Top- and bottom-guided invertible double-seated globe valve. FIG. 6.19cc Cage valve with clamped-in seat ring and characterized plug. (Courtesy of Flowserve Corporation.) FIG. 6.19aa Top-entry, post-guided plug, threaded single-seated globe valve. the cage may be used to clamp a floating seat ring, or it may be used with a screwed-in seat where the cage is part of the bonnet assembly. Both of the above cage designs are usually provided in smaller valve sizes with unbalanced plugs for best shut-off. The stem-guided design (Figure 6.19cc) has advantages when handling fluids with solids, sticky fluids, fluids that coat or plate out onto the trim, and corrosive fluids. This is because the large plug-to-cage clearances in the stem-guided design are not sensitive to debris or build-up, and corrosionresistant trim materials (316 stainless steel and nickel alloys) are more practical to use without special treatments to prevent metal-to-metal galling. FIG. 6.19bb Top- and bottom-guided invertible single-seated globe valve. FIG. 6.19dd Cage valve with unbalanced plug and characterized cage ports. (Courtesy of Emerson Flow Management.)

268 1312 Control Valve Selection and Sizing FIG. 6.19ee Cage valve with balanced and characterized plug. (Courtesy of Flowserve Corporation.) Tight clearances are required in the cage-guided version, which requires that metal surfaces must be hardened by heat treatment, hard-faced with Co-Cr or Ni-Cr-B alloys, plated with hard chrome or electroless nickel, or surface treated (e.g., nitriding) to eliminate metal-to-metal galling. Pressure-Balanced Valves Pressure-balanced trims are also available in stem-guided (Figure 6.19ee) and cage-guided (Figure 6.19ff) versions. Pressure-balanced designs help provide better control of high-pressure drops, reduce the magnitude of unbalanced plug forces, and help to reduce actuator size. The balanced plug designs can be provided with a variety of balance seal styles and materials to meet service conditions. These may FIG. 6.19ff Cage valve with balanced plug and cage port characterization. (Courtesy of Dresser Flow Solutions.) FIG. 6.19gg Cage valve with pilot-balanced construction and characterized cage ports. (Courtesy of Dresser Flow Solutions.) be plastics such as PTFE, elastomeric O-rings, and metal piston rings. Generally, the use of balanced plugs will degrade the valve shut-off capability to some degree. For example, an unbalanced valve rated for Class V shut-off may drop to Class IV with balanced trim. This is due to leakage past the balance seal. (Refer to the discussions on leakage in this section and in Section 6.1.) One variation of the pressure-balanced plug that has improved shut-off capability is shown in Figure 6.19gg. This is called the pilot-balanced plug. This design is particularly helpful when dealing with high-temperature, high-pressure drop situations when tight shut-off is difficult to achieve without the use of elastomeric seals. When the valve is shut off, the actuator has compressed the pilot spring and closed the pilot. There is provision for a pressure path from the upstream pressure side of the plug to the area above the pilot plug, allowing upstream pressure on the upper plug area to assist the actuator in achieving a tight shut-off. In the shut-off position, there is no pressure drop across the conventional balance seals. When the actuator begins to open the valve, the stem first lifts the small pilot plug off its seat. This allows the pressure above the plug to vent to the downstream side, and the valve operates as a conventional pressure-balanced valve. The spring shown in this particular design holds the pilot valve off the pilot seat during normal operation. When the valve begins to close for shut-off, the plug first seats off on the valve seat, and a small amount of additional stem travel compresses the spring and seats the pilot valve. Special Designs Many customized, special-purpose designs are possible to meet unique service and operating requirements. The example in Figure 6.19hh shows a cage and seat that are installed from the bottom of the valve and supported by a bottom flange. This design was intended as a flangeless

269 6.19 Valve Types: Globe Valves 1313 FIG. 6.19hh Quick-change trim valve designed for bottom trim removal. (Courtesy of Emerson Process Management.) valve for small sizes of forged or bar stock corrosion-resistant alloys such as Alloy 20, nickel, and Hastelloy C-276 for corrosive service. It allows for a bottom drain and quick inspection of the valve trim without removal of the actuator and bonnet. Similar bodies can be machined in custom configurations and may be more economical in small sizes for special alloys than other globe designs. Split-Body Valves Another type of single-seated globe body is the split-body valve (Figure 6.19ii). This design was the original stem-guided chemical service valve intended for hard-to-handle services involving slurries, sticky fluids, and corrosive services. The seat ring is clamped between the body halves, and the body is easily disassembled for replacement of the seat and plug. Another feature is the adaptability for building the body to use slip-on flanges. This results in cost savings when corrosionresistant alloy castings or forgings are required for the wetted FIG. 6.19ii Split-body valve with separable end flanges. (Courtesy of Dresser Flow Solutions.) body. Because the flanges are not normally subject to wetting by the process fluid, they can be of carbon steel or lower alloys than the body. In small sizes, it is economical to cast or forge the body for a standard ANSI Class 600 rating and install Class 150, 300, or 600 slip-on flanges as needed. It is also possible to rotate the lower body 90 degrees to the line axis and eliminate a pipe elbow, although this is rarely done. The design of the split body makes it possible to be molded in structural plastics for low-pressure corrosive applications. Refer to the discussion on lined and thermoplastic valves below. While still popular for some applications, split-body valves have limitations on installing special trim options. The bolted body joint may leak if exposed to large piping stresses from severe thermal cycling, and it should not be welded into the piping because maintenance requires body separation. Split body valves are available with flanged ends up to size 10 in. (DN 250), but separable (slip-on) flanges are only available up to size 4 in. (DN 100), and threaded ends are available up to size 2 in. (DN 50). Availability is generally limited to small sizes in high-pressure classes. ANSI Class 900, 1500, and 2500 ratings (PN 150, PN 250, and PN 420) are in sizes 2 in. (DN 50) and smaller. Angle Valves Angle valves (Figure 6.19y) are often used in a flow-to-close direction for high-pressure drop service. This is favorable to the valve body and trim but requires careful design of the downstream piping to avoid erosion problems in high-velocity liquid or two-phase flow. Depending on actual downstream velocities, these applications can require a larger pipe size than the valve and up to 20 diameters of straight pipe before the first elbow. Angle valves are also used to accommodate special piping arrangements to aid drainage, on erosive services to minimize solids impingement problems, and on other special applications such as coking hydrocarbons. Figure 6.19jj shows a coking valve design with a streamlined, sweep-angle flow path and a replaceable venturi outlet sleeve. This body flow path is designed to reduce flow velocity in the body to minimize erosion. Trim materials such as tungsten carbide and ceramics can be selected to resist erosion due to the higher trim velocities. The venturi-style outlet has high-pressure recovery and low vena contracta pressures, which makes the trim and downstream pipe highly susceptible to cavitation on liquid service, even with moderate pressure drops. In order to avoid damage to piping, these valves can be installed directly on vessel inlets or with a larger downstream pipe to mitigate the effects of cavitation or flashing. Y-Type Valves The Y-type valve has application in several special areas. Among these applications are those where good drainage of the body passages or high flow capacity is required, such as in controlling molten metals or polymers, cryogenic fluids, and liquid slurries. Cast or forged Y-type valves are available up to size 16 in. (DN 400) and with pressure ratings up to Class 4500 (PN 760). The valve can be installed in

270 1314 Control Valve Selection and Sizing FIG. 6.19jj Streamlined (sweep) angle valve with lined venturi outlet. horizontal, vertical, or angled piping to suit the application requirements. Because of the simple body shape and bonnet design, they are easy to fit with thermal or vacuum jackets. Figure 6.19kk shows a Y-type valve with a vacuum jacket that has been specially designed for low-pressure cryogenic liquid service. The vacuum jacket provides maximum thermal insulation for such applications as liquid hydrogen. This body is designed for minimum and uniform wall thickness, which enables rapid cool-down rates. The single-seated design allows good shut-off and can be provided with soft inserts for exceptionally tight shut-off. Note that the vacuum jacket is designed with a metal bellows to allow for mechanical tolerances or dimensional changes during temperature transients and has a provision for welding to a matching jacket around the adjacent piping. FIG. 6.19kk Y-pattern valve fitted with vacuum jacket for cryogenic service. FIG. 6.19ll Rotary globe with eccentric spherical segment plug. (Courtesy of Dresser Flow Solutions.) Eccentric Plug Rotary Globe Type The rotary globe valve (Figure 6.19ll) is mentioned here, because its design and performance combine features of conventional rotary valves with conventional globe valves. Compared to most rotary valves, the eccentric spherical segment plug valve has the advantage of lower torque requirements. The seating surface of the plug has the form of a spherical segment. The plug design makes exaggerated use of the offset center to obtain contact at closure without rubbing on the seat, like ball valves. It is capable of substantial seat contact loading for tight shut-off approaching that of conventional globe valves. The flow characteristic approaches linear (Figure 6.19a). Changes in control characteristic can be accomplished with a cam in the positioner or by modifying the controller output signal. Capacity (C v ) is between that of a double-ported globe and a high-performance, double-offset butterfly valve. Highflow capacity is achieved with only moderate pressure recovery in the body, so the critical flow coefficient (F L ) is higher than that of a butterfly valve throughout its throttling range and therefore less likely to cavitate. The valve is made in sizes from 1 16 in. (DN ) in pressure Classes 150, 300, and 600 (PN 20, PN 50, PN 100). Flanged and flangeless versions are available in most sizes. Allowable operating pressure drops depend on the manufacturers designs, materials, and actuators. In most cases, the valves are rated for operating pressure differentials significantly less than the body s maximum pressure rating, so it is necessary to consult the manufacturer s data for specific limitations. C v rangeability data are stated in ranges from 100:1 up to 160:1. However, these ratios are achievable in service only if the actuator and control system can control in the range of 1 5% open, which is possible but often not practical. From a practical control standpoint, assuming a control range from 5 95% open, the C v rangeability for eccentric rotary plugs is about 35:1 or lower, which is comparable with conventional globe valves.

271 6.19 Valve Types: Globe Valves 1315 FIG. 6.19mm Three-way valve for mixing service. Three-Way Valves Three-way valves are a specialized double-seated globe valve configuration. There are two basic types. One is for mixing service, i.e., the combination of two fluid streams passing to a common outlet port (Figure 6.19mm). The other is for diverting service, i.e., taking a common stream and splitting it into two outlet ports (Figure 6.19nn). The three-way design shown in Figure 6.19y uses an adapter FIG. 6.19nn Three-way valve for diverting service. to convert a conventional single-seated globe into mixing or diverting configurations. A typical mixing valve application would be the blending of two different fluids to produce a specific outlet end product. A diverting valve might be used for switching a stream from one vessel to another vessel or for temperature control on a heat exchanger. In the latter, the valve would direct one portion of the flow through the exchanger, and the balance of the flow would bypass the exchanger. The relative split would provide the heat balance needed for temperature control. The forces acting on the double-seated, three-way plug do not balance in the same way that a double-seated, two-way plug is balanced. This is because different pressure levels exist in each of the three flow channels. Also, there are different dynamic forces acting on each plug head. Therefore, these valves are not normally used in high-pressure drop service. The valve plugs are usually seat-guided and post- or stem-guided to maintain stability under fluctuating dynamic forces. Due to the larger size and piping complications, some users prefer to use two opposite-acting, two-way globe valves operating from one controller to do the same job as a three-way valve. Lined and Thermoplastic Valves For extremely corrosive services, high alloy metals such as Alloy 20, nickel, nickel alloys, titanium, and zirconium are often required where austenitic stainless steels are inadequate. However, special alloy valves are very expensive and the service life may still be limited by corrosion. An alternative to special alloys could be the lined globe valve (Figure 6.19oo). The lined valve uses a carbon steel or stainless steel shell for retaining pressure, but the inside surfaces of the body and bonnet are bonded with a chemically inert or corrosion-resistant fluoroplastic, such as PFA or PTFE. These valves are commonly for 1 in. (DN 25) through 2 in. (DN 50) sizes with capacity from C v = 0.33 to 45. For very low C v flow capacities, the design shown in Figure 6.19pp is available in 1 in. size (25 mm) with trim selections covering a C v range from to 1.0. These valve designs are limited to operating temperatures below F ( C) and pressure drops less than PSI ( bar). Lined valves are also available in various rotary valve types and pinch valves (see Section 6.20). Conventional globe patterns and split-body designs can be molded entirely from structural thermoplastics for corrosive applications with low to moderate pressures and temperatures. Bodies can be molded in polyvinylidene fluoride (PVDF), polyvinyl chloride (PVC), polypropylene (PP), and other plastics. Allowable pressures and temperatures vary with material and body design. Typically, PVDF is limited to 300 PSIG (21 bar) at 73 F (23 C) and 150 PSIG (10.3 bar) at 225 F (106 C). PVC has a limit of 300 PSIG (21 bar) at 73 F (23 C) and 150 PSIG (10.3 bar) at 140 F (60 C). Polypropylene can operate up to 300 PSIG (21 bar) at 73 F (23 C) and 150 PSIG (10.3 bar) at 175 F (79.4 C). End connections are available in Class 150 flanged,

272 1316 Control Valve Selection and Sizing Packing follower Upper body half Packing box plug Packing Liner Guide bushing Lantern ring Teflon packing secondary seal Seat ring shims Seat ring gasket Seat ring Valve plug and stem assembly Lower body half FIG. 6.19oo Standard Teflon-lined valve with Teflon-encapsulated trim. (Courtesy of Emerson Process Management.) 316 stainless steel body and bonnet Teflon body liner Teflon packing primary seal Tantulum plug flangeless (for clamping between pipe flanges with throughbolting), or threaded ends. Packing and gaskets are generally made of PTFE or perfluoroelastomer (e.g., Kalrez ) materials. VALVE CONNECTIONS Valve connections to adjacent piping can be categorized into four general types: 1) flanged, 2) welded, 3) threaded, and 4) clamped. Each style incorporates elements that contain pressure, bear piping loads, and seal the joint between the valve and piping. As a general rule, valves smaller than 2 in. (DN 50) can use threaded connections, while sizes 2 in. (DN 50) and larger use flanged connections. In a few process systems where fugitive emissions or process leakage is not a problem (such as water systems), threaded connections up to size 4 in. (DN 100) have been used. Most applications require both ends of the valve to have identical connections, but on some vent and drain valves, the upstream port and the downstream port may require different sizes or types of connections. Typical end connections are discussed below. Flanged Ends Flanges are the most common valve connection to piping. Most countries and piping codes accept the flange design and FIG. 6.19pp Small-flow Teflon-lined valve with tantalum plug. (Courtesy of Emerson Process Management.) ratings from the standard ASME B16.5, Pipe Flanges and Flanged Fittings commonly referred to as ANSI flanges (because the earlier standards were published by the American National Standards Insitutue). Some countries and applications use other flange standards, such as American Petroleum Institute (API) standard for oil production equipment and the International Standards Organization (ISO) ISO-7005 flanges (adopted from the German DIN standards). These rating systems are different and are not interchangeable; they use different rating systems, bolting, gaskets, and flange dimensions. The ASME (or ANSI) flanges use a Pressure Class rating system (e.g., steel flanges are Class 150, Class 300, Class 400, Class 600, Class 900, Class 1500, and Class 2500). API flanges are designated by a nominal pressure system (e.g., 2,000 PSI, 3,000 PSI, 5,000 PSI, 10,000 PSI, 15,000 PSI, and 20,000 PSI). The ISO (and DIN) standards also use a nominal pressure or PN numbering with typical ratings PN 10, PN 16, PN 20, PN 40, PN 64, PN 100, PN 160, PN 250, PN 320, and PN 400, which indicate maximum nominal pressure in bar (1 bar = PSI = 100 kpa). Flanges may be flat face, raised face, ring-type joint (RTJ), tongue and groove, male and female, or other configuration to suit the application. Cast iron, ductile iron, and

273 6.19 Valve Types: Globe Valves 1317 bronze are usually flat face; carbon steel and alloys are usually raised face; and above ANSI 600, the RTJ is fairly common. Figure 6.19qq shows three common configurations of flanges: separable flange-raised face, integral flange-raised face, and integral flange-rtj. Raised face and flat face flanges use gaskets from sheet stock, such as graphite or PTFE composites, or a spiral wound thin metal strip with graphite, PTFE, or other mineral filler between each metal winding. RTJ gaskets are oval or octagonal in cross-section made of any suitable metal softer than the flange. Valves with separable (slip-on) flanges retain the flange on the valve body ends with two circular half-rings held in the body grooves. Separable flanges are used mainly on small sizes from NPS 1 / 2 in. up to 4 in. With separable end flanges, the body can be designed for an ANSI Class 600 rating and then adapted as needed with ANSI Class 150, 300, or 600 flanges. Separable flanges are not wetted by the process, which permits less expensive carbon steel or stainless steel flanges to be used on high alloy valves. Separable flanged valves are easy to install with the mating piping, because the flanges can be rotated to fit mismatched line flange hole patterns. However, care must be taken during installation to prevent the valve from rotating in the line until the flange bolts are properly tightened. Flangeless bodies (clamped between pipe flanges) are sometimes used in the small bar-stock bodies. While they permit lower cost where expensive alloys are involved, they require care with bolting, gasket, and piping alignment. The tie-rod bolting should be high tensile strength material, and the valve must be carefully centered to permit proper gasket sealing and loading. The longer the bolt studs are, the more they are affected by longitudinal thermal expansion differences with the valve body and piping, which can lead to gasket leakage in thermal shock or severe temperature cycling. Welded Ends Welded ends are not common in the chemical process industries, but they are generally recommended in power generation and other applications where high piping stresses or thermal shock conditions exist. They are also used in hazardous services when no leakage is allowed from gasketed joints. Socket-weld ends (Figure 6.19rr) are easy to align and may be used in small sizes of 2 in. (25 mm) or under. FIG. 6.19rr Socket-weld end valve and pipe joint with fillet weld. (Courtesy of Flowserve Corporation.) However, most valves are installed with butt-weld ends where maximum joint strength is achieved by full penetration of the weld (Figure 6.19ss). Butt-weld joints are commonly checked by radiographic inspection (X-ray) to confirm full penetration and lack of defects such as cracks, voids, and slag. The valve body material should be selected so that it is suitable for welding to the adjoining piping. In many cases the valve and adjacent pipe require a post-weld, stress-relieving heat treatment. The mating ends of the body and pipe are usually machined at an angle to form a V-joint for ease of welding. The design of joint (end-preparation) depends on the thickness of the mating parts. ASME B16.25 is the most common standard for design of the butt-weld end preparation, but some users and engineering contractors choose their own dimensions for the end preparation to suit their particular fabrication practices. Threaded Ends Threaded connections with NPT (tapered pipe threads per ASME B1.20.1) threads are common on valve sizes 1 in. (DN 25) and smaller and are sometimes used on valves up to 2 in. (DN 50). Threaded connections are usually used in pressure ratings up to ANSI Class 600 (PN 100). The threaded connection normally consists of the valve body with female NPT threads; then the pipe with male NPT threads is screwed into the body end. Materials commonly used in FIG. 6.19qq Flange configurations: separable raised-face, integral raised-face, ring-type joint (RTJ) integral flange. (Courtesy of Flowserve Corporation.) FIG. 6.19ss Butt weld end valve and pipe joint showing full penetration weld.

274 1318 Control Valve Selection and Sizing threaded end valves are brass, bronze, and carbon steel. Stainless steel bodies usually do not have NPT threads, because they gall easily with stainless steel pipe. An antigalling thread sealant compound should be applied to the threads during installation to prevent leakage and seizing and to allow for future disassembly. Threaded connections are not normally used with process valves because they can leak more easily, they are more difficult to make up in the piping, and they may corrode or gall together, making disassembly difficult. Threaded connections do not work well on corrosive fluids, because the corrosion attack is generally more severe in the threaded crevice, causing the threads to corrode together and making it difficult to disassemble. Accelerated corrosion in the threads also promotes leakage. Threaded joints also tend to loosen in services where the temperature fluctuates. Threaded ends are not recommended for services with severe thermal shock or thermal cycling that occurs above 500 F (260 C) or below 50 F ( 46 C). However, some small valves under size 2 in. (DN 50) are furnished with threaded end connections for utility services such as low-pressure steam, water, and gas. Where it is necessary, threaded end valves can be converted to flanges by welding a flange to a pipe nipple, threading it into the valve, and seal-welding the nipple to the body. Special End Fittings In very high pressures, usually above 5000 PSIG (345 MPa), some operating companies use proprietary fitting and flange designs. One widely used high-pressure connection design is the lens-ring-type fitting shown in Figure 6.19tt. This is a self-energizing type of seal; i.e., the lens-ring deforms to give a tighter seal with increase in-line pressure. (Further details are given in Section 6.1.) In other high-pressure applications, special clamped fittings are very effective. One proprietary clamped fitting called FIG. 6.19tt High-pressure lens-ring-type joint. the Grayloc fitting (developed by the Gray Tool Company) greatly reduces the bulk of the valve end as compared to the large high-pressure, bolted flanges. The Grayloc fitting utilizes a lens-ring-type seal, similar to one shown in Figure 6.19tt, but the special hub design is fastened together with a bolted clamp-type fitting instead of flanges. See Figure 6.19uu. These fittings are also available in 10,000 and 15,000 PSIG (69 and 104 MPa) designs. The ASME Boiler and Pressure Vessel code permits the use of clamped connections and includes design rules in Section VIII, Division 1, Appendix 24. For standards on globe valve end connections, see Table 6.19vv. MATERIALS OF CONSTRUCTION A valve body assembly is a pressure containment vessel. As such, design and material selection must follow guidelines of a pressure vessel code that is recognized by the country or state in which it will be operated. Some of the FIG. 6.19uu Grayloc end connection detail and comparison with high-pressure flange connection. (Courtesy of Grayloc Products.)

275 6.19 Valve Types: Globe Valves 1319 TABLE 6.19vv End Connection Standards used for Globe Valves Type Standards Organization Designation* Title Flanged ASME International American Petroleum Institute (API) European Committee for Standardization (CEN) International Standards Organization (ISO) Manufacturers Standardization Society (MSS) ASME B16.5 Pipe Flanges and Flanged Fittings (NPS 1 / 2 through NPS 24) ASME B16.47 Large Diameter Steel Flanges (NPS 26 through NPS 60) ASME B16.1 Cast Iron Pipe Flanges and Flanged Fittings ASME B16.42 Ductile Iron Pipe Flanges and Flanged Fittings, Class 150 and 300 ASME B16.24 Cast Copper Alloy Pipe Flanges and Flanged Fittings, Class 150, 300, 400, 600, 900, 1500, and API Spec 6A Specification for Wellhead and Christmas Tree Equipment EN EN PREN ISO ISO ISO ISO MSS SP-6 MSS SP-44 ASME International ASME B16.25 Butt Welding Ends Flanges and Their Joints: Circular Flanges for Pipes, Valves, Fittings, and Accessories, Part 1: Steel Flanges PN Designated Flanges and Their Joints: Circular Flanges for Pipes, Valves, Fittings, and Accessories, Part 2: Cast Iron Flanges PN Designated Flanges and Their Joints: Circular Flanges for Pipes, Valves, Fittings, and Accessories, Class Designated Part 1, Steel Flanges NPS 1 / 2 to 24 Metallic Flanges Part 1: Steel Flanges Metallic Flanges Part 2: Cast Iron Flanges Metallic Flanges Part 3: Copper Alloy and Composite Flanges Petroleum and Natural Gas Industries Drilling and Production Equipment Wellhead and Christmas Tree Equipment Standard Finishes for Contact Faces of Pipe Flanges and Connecting End Flanges of Valves and Fittings Steel Pipe Line Flanges Welded European Committee for ASME B16.11 Forged Fittings, Socket-Welding and Threaded Standardization (CEN) EN Industrial Valves Butt Welding Ends for Steal Valves EN Valves Socket Welding Ends for Steel Valves Threaded ASME International ASME B Pipe Threads, General Purpose (Inch) AMSE B Dryseal Pipe Threads (Inch) ASME B16.11 Forged Fittings, Socket-Welding and Threaded Clamped ASME International ASME B&PVC Sec.VII-Div. 1- Appendix 24 Design Rules for Clamp Connections * Always use the current edition of any standard, unless governmental or construction specifications require otherwise. prominent standards include the ASME Unfired Pressure Vessel Code (Section VIII and Section III), ASME B31.1, Power Piping, and ASME B31.3, Process Piping. Most other countries have similar standards, and the European Community has moved to harmonize and unify some of these standards under a European Pressure Equipment Directive (PED) 97/23/EC. Most of these codes recognize ASME B16.34, Valves Flanged, Threaded, and Welding End, as an accepted standard for design, materials, and testing of valves. The selection of proper materials for valves requires consideration of all factors related to the process fluid and operating conditions. This can include forces applied from actuators and piping, (wind and seismic loads, hoists and rigging, accidental loads), changes in process and ambient temperatures, and the erosive and corrosive potential of both the process fluid and the outside environment. The complex scope of material selection is outside the purpose of this text, but a few of the common globe valve materials are listed in Table 6.19ww with typical ASTM specifications and process applications.

276 1320 Control Valve Selection and Sizing TABLE 6.19ww Common Valve Body Materials, Specifications, and Applications Common Material Designation Body Material Specification (Castings) Body Material Specification (Forgings) Max. ANSI Pressure Class Rating (ASME B16.34a-1998) Temperature Limit F ( C)* Min. Max. Typical Service/Applications 304 SS A351-CF8 A182-F304 Class ( 268) 1000 (538) General corrosion, organic acids, nitric acid, hydroxides, hydrocarbons, high temperature; poor resistance to halogens (chlorides, fluorides, etc.), poor resistance to pitting and crevice attack 304L SS A351-CF3 A182-F304L Class ( 268) 850 (454) Same as 304, low to moderate temperature 316 SS A351-CF8M A182-F316 Class ( 254) 1000 (538) Same as 304, high temperature; fair resistance to pitting and crevice attack 316L SS A351-CF3M A182-F316L Class ( 254) 850 (454) Same as 316, low to moderate temperature 316H SS A351-CF10M A182-F316H Class ( 198) 1500 (816) Same as 316, high temperature, creep resistant Alloy 20 A351-CN7M A182-F20 Class ( 198) 600 (316) Sulfuric acid, organic acids, hydroxides, nonhalogenated organic chemicals Bronze B61 (C92200) n/a B16.24, Class ( 198) 400 (204) Steam, fresh and chloride water, oxygen Aluminum- Bronze B148-C95200 n/a B16.24, Class ( 254) 600 (316) Brine; fresh, brackish, and salt water; chlorides Carbon Steel A216-WCB A105 Class ( 28) 800 (427) Neutral and alkaline waters, steam, dilute caustic, hydrocarbons Carbon Steel A216-WCC A350-LF3 Class ( 28) 800 (427) Neutral and alkaline waters, steam, dilute caustic, hydrocarbons Cast Iron A126-A n/a B16.1, 400 PSIG 20 ( 28) 406 (208) Neutral and alkaline waters, lowpressure steam, dilute caustic Chrome-Moly WC9 Chrome-Moly C12A Duplex SS 22% Cr Duplex SS 25% Cr Hastelloy B/B2 Hastelloy C/C-276 A217-WC9 A182-F22 Cl.3 Class ( 28) 1200 (649) Mildly corrosive, high temperature, resists erosion by steam and flashing water A217-C12A A182-F91 Class ( 28) 1200 (649) Mildly corrosive, high temperature, resists erosion by steam and flashing water A351/A995- CD3MN (J92205) A182-F51 Class ( 28) 600 (316) Corrosive, brine, salt water, polluted water, acid-chlorides, good against pitting and crevice attack A351-CD4MCu A182-F61 Class ( 28) 600 (316) Corrosive, brine, salt water, polluted water, acid-chlorides, better against pitting and crevice attack A494-N-7M B335-N10001/ N10665 Class ( 198) 800 (427) Superior in hydrochloric acid up to boiling, hydrofluoric acid, strong reducing chemicals A494-CW-6M B564-N10276 Class ( 198) 1250 (677) Oxidizing and reducing acids, hypochlorite, chlorine, seawater, acidchlorides, brines, excellent against pitting and crevice attack (continued)

277 6.19 Valve Types: Globe Valves 1321 TABLE 6.19ww (Continued) Common Material Designation Inconel 600 Body Material Specification (Castings) A494-CY-40 Class 2 Body Material Specification (Forgings) Max. ANSI Pressure Class Rating (ASME B16.34a-1998) Temperature Limit F ( C)* Min. Typical Service/Applications B564-N06600 Class ( 198) 1200 (649) Caustics with chlorides or sulfides, mild oxidizers, excellent resistance to chloride SCC Max. Low Temp CS A352-LCB A350-LF3 Class 4500 LCB: 50 ( 46) 650 (343) Neutral and alkaline waters, steam, dilute caustic, hydrocarbons, lowtemperature impact strength Low Temp CS A352-LCC A350-LF3 Class 4500 LCC: 50 ( 46) 650 (343) Neutral and alkaline waters, steam, dilute caustic, hydrocarbons, lowtemperature impact strength Monel A494-M35-1 B564-N04400 Class (482) Hydrofluoric acid, caustic, seawater, brine; not for oxidizing service Nickel A494-CZ-100 B160/B564- N02200 Class (316) Hot concentrated caustics, SCC resistant in chlorinated organics; not for oxidizing service Titanium B367-C-3 B381-F-3 Class (316) Better than Hastelloys for pitting and crevice attack, wet chlorine, dilute HCl, bleaches, brines; not for dry chlorine or fluorides Zirconium 705 B C/ Flowtherm HT *Temperature limits based on ASME B16.34 and ASME B31.3. Courtesy of Flowserve Corporation B550-R60705 Class (371) Dry chlorine, hot hydrochloric, - sulfuric, nitric, and acetic acids; not for fluorides or wet chlorine Trademarks 17-4PH is a trademark of Armco, Inc. Cavitrol is a registered trademark of Emerson Process Management. ChannelStream is a registered trademark of Flowserve Corporation. Colmonoy is a registered trademark of the Wall Colmonoy Corporation. Drag is a registered trademark of Control Components, Inc. Grayloc is a registered trademark of Grayloc Products. Kalrez is a registered trademark of DuPont Dow Elastomers. VRT is a registered trademark of Dresser Industries, Inc. Hastelloy and C-22 are registered trademarks of Haynes International, Inc. Hush is a registered trademark of DeZurik/Copes Vulcan, a unit of SPX Corporation. Inconel, Incoloy, and Monel are trademarks of the Special Metals Corporation group of companies. PBI is a registered trademark of Celanese AG. Stellite and Deloro are registered trademarks of Deloro Stellite, Inc. Teflon and Tefzel are registered trademarks of the DuPont Company. Turbo-Cascade is a registered trademark of Yarway Corporation, a unit of Tyco-Flow Control. Reference 1. See ANSI/ISA and IEC : Bibliography Adams, M., Control Valve Dynamics, InTech, July ANSI/ISA , Control Valve Capacity Test Procedures, Research Triangle Park, NC: Instrumentation, Systems, and Automation Society, ANSI/ISA , Flow Equations for Sizing Control Valves, Research Triangle Park, NC: Instrumentation, Systems, and Automation Society, Ball, K. E., Final Elements: Final Frontier, InTech, November Baumann, H. D., How to Assign Pressure Drop Across Control Valves for Liquid Pumping Services, Proceedings of the 29th Symposium on Instrumentation for the Process Industries, Texas A&M Univ., 1974.

278 1322 Control Valve Selection and Sizing Borden, G. and Friedmann, P. G. (eds.), Control Valves Practical Guides for Measurement and Control, Research Triangle Park, NC: Instrumentation, Systems, and Automation Society, Borden, G. and Zinck, L., Control Valve Seat Leakage, InTech, November Buckley, P. S., A Control Engineer Looks at Control Valves, Proceedings of the 1st ISA Final Control Elements Symposium, Wilmington, DE, Carey, J. A., Control Valve Update, Instruments and Control Systems, January Control Valve Handbook, 3rd edition, Marshalltown, IA: Fisher Controls International Inc., Control Valve Seat Leakage, FCI , Cleveland, OH: Fluid Controls Institute, Inc., Control Valves: Globe, Plug, Pinch, Needle, Gate, Measurements and Control, February Davis, J.R. (ed.), Corrosion Metals Handbook, Ninth Edition, Volume 13, Metals Park, OH: ASM International, Cunningham, E. R., Solutions to Valve Operating Problems, Plant Engineering, September 4, Engel, H. O., Control Valves for Process Automation, Maintal, Germany: Honeywell Regelsysteme GmbH, Fernbaugh, A., Control Valves: A Decade of Change, Instruments and Control Systems, January Fontana, M. G. and Greene, N. D., Corrosion Engineering, 3rd ed., New York: McGraw-Hill, George, J. A., Seat Leakage Standard Revised, InTech, October Hutchison, J. W. (ed.), ISA Handbook of Control Valves, 2nd ed., Research Triangle Park, NC: Instrumentation, Systems, and Automation Society, IEC : 1989, Inherent Flow Characteristic and Rangeability, Geneva, Switzerland: Research Triangle Park, NC, International Electrotechnical Commission, Inherent Flow Characteristic and Rangeability in Control Valves, ANSI/ISA , Research Triangle Park, NC, Instrumentation, Systems, and Automation Society, ISA-RP , Considerations for Evaluating Control Valve Cavitation, Research Triangle Park, NC: Instrumentation, Systems, and Automation Society, Lyons, J. L., Lyons Valve Designers Handbook, New York: Van Nostrand Reinhold, Pyotsia, J., A Mathematical Model of a Control Valve, 1992 ISA Conference, Houston, TX, October Rahmeyer, W., The Critical Flow Limit and Pressure Recovery Factor for Flow Control, InTech, November Shinskey, F. G., Control Valves and Motors, Foxboro Publication No Whatever Happened To... A Guide to Industrial Valve and Actuator Companies - Past and Present, Washington, D.C.: Valve Manufacturers Association, Wolter, D. G., Control Valve Selection, InTech, October 1977.

279 6.20 Valve Types: Pinch Valves FY X C. S. BEARD (1975) J. B. ARANT (1985) B. G. LIPTÁK (1995) J. B. ARANT, D. GARDELLIN, B. G. LIPTÁK (2005) Pneumatic pinch valve Flow sheet symbols Mechanical pinch valve Applications: Sizes: Design Pressures: Maximum Pressure Drop: Design Temperatures: Materials of Construction: Characteristics: Abrasives, minerals in suspensions, hydraulically transported solids, slurries, viscous, or food industry products Generally 1 to 24 in. (25 mm to 0.61 m); special units from 0.1 to 72 in. (2.5 mm to 1.8 m) Generally up to ANSI Class 150, with special units available with up to ANSI Class 300 ratings. See Figure 6.20a Generally about 10 to 200 psid (69 to 138 kpa) 20 to 350 F ( 29 to 204 C) Sleeve materials: Buna-N, Chlorobutyl, EPDM, Hypalon, Neoprene, PGR, Polyurethane, Teflon, Viton Bodies: Aluminum, cast grey iron, cast ductile iron, cast steel, stainless steel, Polyamide blend plastic For the characteristics of reduced port and Clarkson pinch valves, refer to Figures 6.20v and 6.20w; for others, refer to Figure 6.20b Costs: See Figure 6.20c Capacity: C v = 60 d 2 for full-ported pinch valves and C v = 20 d 2 for reduced-ported pinch valves; see Tables 6.20e and 6.20f Rangeability: From 5:1 to 10:1 Leakage: Partial List of Suppliers: Generally ANSI Class IV or Class V (see Table 6.1gg in Section 6.1for definitions) Clarkson Co. ( Elasto-Valve Rubber Products Inc. ( Ever-Flex ( Larox (Finland USA) ( Onyx Valve Co. ( Red Valve Co. ( Richway Industries ( INTRODUCTION These valves are called either pinch or clamp valves, depending upon the configuration of the flexible tube and on the means used to compress the tube. The compression can be done by mechanical clamping mechanisms or by external pneumatic or hydraulic power within a metal jacket enclosure. Pinch valves have been improved due to introduction of plastic tubes, elastomers, and reinforcing fabrics. Tubes can be fabricated from pure gum rubber or from a variety of rubber-like elastomers such as Buna-N, butyl, neoprene, Nordel, hypalon, Viton, silicone, polyurethane, polypropylene, white butyl, and odorless and tasteless white neoprene. The latter two materials are often used in the food and allied industries. Reinforcing fabrics may include some of the materials used in automobile tire fabrication, such as cotton duck, rayon, nylon, fiberglass, and Kevlar, which is a new arimid polymer material that is as strong as steel at 1323

280 1324 Control Valve Selection and Sizing Press. psi Pinch valve P, max Shell & Tube Mech Custom design Size, inches FIG. 6.20a Pinch valve pressure rating drops as size increases. The bottom curve gives the rating of shell and tube, the center of mechanical and the top of custom designs. Price, $US Pinch valves Hand wheel operated 8000 pinch valve Shell & Tube Size, inches FIG. 6.20c The shell and tube-type pinch valves are less expensive than cylinder or electric-operated valves and generally cost less than hand-operated valves. Cost data are based upon standard flanged cast iron housing, pure gum rubber sleeve, and on/off operation. Numerous other elastomers are available for the sleeve, a range of material for the body, and a variety of actuators and drives. one sixth the weight. Teflon plastic tubes are uniquely capable of handling highly corrosive or sticky fluids. As was shown in Figure 6.20b, pinching a standard full round sleeve from 100% to 50% of travel has little effect on the flow through the valve. The advantage of the prepinched tube design is that this dead part of the valve control is nearly eliminated. As can be seen from Table 6.20e and Table 6.20f, fluid capacity is sacrificed, because the shape of a full open % Flow or C v Full round sleeve Standard or double walled sleeve Cone sleeve Variable orifice sleeve Teflon tube clamp Reduced port sleeeve Shutter closure % Travel FIG. 6.20b The characteristics of a full round sleeve is nearly quick opening. Double-walled or cone-shaped sleeve cross-sections bring the characteristics closer to linear. $12,000 $10,000 $8,000 $6,000 $4,000 $2,000 0 Electric Price Pneumatic Size FIG. 6.20d Price comparison pinch valves provided with electric or pneumatic actuators.

281 6.20 Valve Types: Pinch Valves 1325 TABLE 6.20e Valve Coefficients (C v ) of Standard and Reduced-Port Pinch Valves during Throttling (% of Travel)* Standard and Double-Wall Sleeve Flow Chart C v Valve Size, in. Valve Opening % of Total Travel /2 3 / / / , , ,092 1, ,176 1, ,216 1, ,237 1, ,251 1, ,259 1, ,034 1,381 1, ,531 2,496 3,335 4,355 1,843 2,883 4,701 6,280 8,202 2,466 3,856 6,288 8,400 10,971 2,706 4,233 6,902 9,224 12,047 2,827 4,420 7,207 9,624 12,569 2,913 4,552 7,422 9,911 12,944 2,961 4,629 7,548 10,083 13,170 2,998 4,686 7,641 10,209 13,333 3,014 4,710 7,680 10,260 13, ,302 3,405 4,215 5,558 8,223 10,180 10,467 15,486 19,171 14,000 20,713 25,642 15,373 22,745 28,157 16,040 23,731 29,378 16,519 24,440 30,255 16,806 24,865 30,781 17,015 25,174 31,164 17,100 25,300 31,320 Reduced Port Sleeve Flow Chart C v Valve Size, in. Port Size, in. Valve Opening % of Total Travel /2 1 /2 3 /4 3 /4 1 1/4 3 /8 3 /8 1 /2 1 / /2 1 1 / /4 3 / / /2 2 1 / / / (continued)

282 1326 Control Valve Selection and Sizing TABLE 6.20e (Continued) Valve Coefficient (Cv) for Standard and Reduced-Port Pinch Valves* Reduced Port Sleeve Flow Chart C v Valve Size Port Size Valve Opening % of Total Travel ,211 1,255 1,438 2, ,346 1,325 1,808 1,375 2,413 1,844 3,125 2,189 3, , ,545 1,209 2,242 1,805 3,041 2,421 4,060 3,247 5,258 3,854 5, ,375 1,105 1,516 1,799 3,220 2,687 4,144 3,603 5,532 4,832 7,164 5,736 8, ,780 1,374 2,726 2,238 3,992 3,342 5,364 4,482 7,161 6,010 9,273 7,134 9, ,107 9,01 2,175 1,650 3,331 2,689 4,878 4,015 6,555 5,384 8,751 7,220 1,1332 8,570 11, ,250 1,036 2,472 1,898 3,736 3,091 5,544 4,616 7,450 6,190 9,946 8,301 12,880 9,853 13, ,372 1,062 2,490 1,945 3,313 3,169 5,584 4,732 7,504 6,345 10,018 8,509 12,973 10,100 14, ,491 1,14 2,521 2,041 3,861 3,324 5,654 4,964 7,597 6,656 10,143 8,925 13,183 10,594 14, ,502 1,123 2,570 2,057 3,936 3,351 5,764 5,004 7,745 6,710 10,344 91,25 13,394 10,821 15,120 *Courtesy of Red Valve Co. prepinched tube has approximately 80% of the area of a full open standard round tube. Pinch valves can be mechanically pinched or be of the shell and tube type. The mechanically pinched design can have a frame that is open or enclosed and can be provided with handwheel, pneumatic, or electric operators. If a pneumatic actuator is selected for a mechanically pinched valve, that design is available with either an open or a closed failure position. The shell and tube designs can be pinched pneumatically or hydraulically and can only fail open. A summary of pinch valve features is given in Figure 6.20g. THE SLEEVE A pinch valve consists of a short section of flexible hose combined with a mechanism for throttling or completely closing the hose. The hose is made of a composite matrix of elastomer and a fabric carcass. The fabrication and materials of construction are similar to automobile tire manufacturing, but inside out, in the sense that the tire has the wear surface on the outside, a pinch valve has the wear surface on the inside of the tube. Much of the art of making pinch valves lies in the design and manufacture of the elastomer sleeve. By balancing the number of fabric layers, the type and orientation of the fabric, and the ratio of rubber to fabric, the sleeves can be made soft and pliable for low-pressure or tough and rigid for highpressure applications. Table 6.20h lists the properties of a variety of elastomers used as pinch valve sleeves. The tensile strength of the sleeve is of particular importance because it limits the ability to resist tearing and delamination of the rubber and fabric. As can be seen from Table 6.20h, natural gum rubber is superior to all the synthetic compounds for abrasion resistance, tensile strength, and elongation. It is also the least expensive. This makes it the material of choice for most pinch valve sleeves. In addition to the elastomers listed in Table 6.20h, most manufacturers offer rubber compounds in FDA White Odorless and Tasteless versions for food and cosmetic applications. The white pigments in these compounds compromise temperature resistance and tensile strength. Pinch valves are available with conductive gum rubber sleeves that safely dissipate static electricity. These are routinely used in the manufacture of nitro-glycerin, gunpowder, and ordinance materials. Some manufacturers offer proprietary extreme abrasion resistant compounds such as Linatex Red Crepe and Onyx Blue. Urethane compounds are available on special order; they offer good abrasion resistance but are plagued by delamination and splitting at the pinch points. Kalrez has been debated for years as a pinch valve material but its extreme cost has made it prohibitive even for small pinch valves. Reinforcing fabrics for pinch valves borrow heavily from the tire industry and include polyester, nylon, cotton, rayon,

283 6.20 Valve Types: Pinch Valves 1327 TABLE 6.20f The Valve Capacity Coefficients (C v s) are Listed for Full and Reduced Ported Pinch Valves between the Valve Sizes from 0.5 to 30 in. Valve Port C v Valve Port Prepinch Full Full Round C v Valve Port Prepinch C v Full Round C v Full 12 Full Full 8,550 10,260 1/ ,764 6, ,351 4,021 Full ,837 2,204 3/ , Full 42 Full Full 11,200 13, ,745 9, ,004 6, ,002 3, ,711 2,053 Full 86 Full Full 14,300 17, ,344 12, ,710 8, ,500 5, ,790 3, ,632 1,958 Full 170 Full Full 17,600 21, ,400 16, ,125 10, ,000 7, ,166 4, ,650 3,180 Full 346 Full Full 26,100 31, ,120 18, ,830 12, ,536 9, ,167 6, ,780 4, ,475 2,970 Full 576 Full ,680 Full 42,500 51, , ,250 25, , ,000 14, , ,900 10, ,461 7,753 Kevlar, and occasionally fiberglass. They can be built up in either a bias-ply or radial pattern, or a hybrid combination where different layers in the same sleeve have different orientation. The rubber sleeve is the most critical component of the pinch valve and it determines the maximum time between service intervals. Sleeve quality varies considerably between manufacturers, and price tends to reflect quality. There are two methods for making pinch valve sleeves. One is the vulcanization process, where they are held together with heat tape, or they can be compression molded.

284 1328 Control Valve Selection and Sizing Pinch valves Shell & Tube Mechanically pinched Open frame Enclosed Hydraulic Pneumatic Pneumatic Electric Hand wheel Pneumatic Electric Hand wheel FIG. 6.20g Pinch valves design variations and feature options. Fail open Fail open Fail open Fail closed Fail open Fail closed An important feature of the sleeve design is the positive opening feature (POF), which is illustrated in Figure 6.20i. This feature is provided by a pair of rubber and fabric tabs, which are molded into the trunk of the pinch valve sleeve. These tabs are clamped to the pinch bars. They force the sleeve to open properly when there is no process pressure on the inside to open it, which is of value if the sleeve has been closed for an extended period of time. PINCH VALVE TYPES Pressure Limitations As was shown in Figure 6.20a, shell and tube-type pinch valves have the lowest pressure rating. Their main limitation is the available pressure of the compressed air used to close the valve. Standard mechanically pinched valves can operate against line pressure in the range of 75 to 200 PSIG, depending on size. The limitation is the strength of the fabric reinforcement in the rubber sleeve. Special-order pinch valves can be fabricated for high-pressure service. Manufacturers use exotic fabrics including Kevlar and the latest engineering software to create hybrid sleeves built up with a mix of bias ply and radial layers to increase working pressure. Shell and Tube Design The simplest pinch valve is the shell and tube type, which is also referred to as a direct operated or as a jacketed pinch valve. This design consists of a cast iron, aluminum, or plastic shell with a rubber sleeve inside. For a cost comparison between shell and tube-type and handwheel-operated pinch valves, refer to Figure 6.20c. As shown in Figure 6.20j, injecting compressed air into the housing causes the rubber sleeve to collapse, thereby blocking flow of the process fluid through the valve. The valve housing is usually configured to coax the sleeve into a straight-line closure to operate with the lowest possible differential pressure. Flanged bodies are available in 1 through 72 in. sizes. Body material for flanged valves is cast iron, ductile iron, TABLE 6.20h Comparative Properties of Elastomers Used to Make Pinch Valve Sleeves ANSI/ASTM D NR/IR CR NBR CHR CSM EPDM FKM AFMU SI Trade Names Poly Isoprene Pure Gum Neoprene Nitrile Buna-N Chloro Butyl Hypalon Nordel Viton Fluorel Teflon Silicone Durometer-A (typical) Abrasion Resistance D Outstanding Excellent Good Good Good Very good Fair Poor Poor Tensile Strength, psi Elongation% T range F Acid Fair to good Excellent Good Excellent Excellent Excellent Excellent Best Excellent Resistance Diluted Acid Fair to good Good Good Good Very good Excellent Excellent Best Fair Resistance Concentrated Oil Tolerance Poor Good Excellent Poor Good Poor to good Excellent Excellent Fair

285 6.20 Valve Types: Pinch Valves 1329 threads or grooves required. These are generally made in cast aluminum. Advantages 1. Simple operation, sleeve is the only moving part 2. Light weight 3. Compact 4. Inexpensive 5. Closes around suspended particles, rags, powder, pellets, and other debris without jamming or leaking Caveats FIG. 6.20i Pinch valve provided with the positive opening feature (POF). (Courtesy Larox Co.) cast aluminum, and fabricated carbon steel. Small valves in the range of 1 /8 through 3 in. are available with screw-on pipe connections with steel, stainless steel, and plastic bodies. Valves in the 1 /2 through 4 in. size are available with slip-on connections, where the mating pipe is simply inserted into the valve body and is retained solely by friction with no 1. Limited line pressure capability. This is because a higher pressure is needed on the outer surface of the sleeve to force it closed. Typically a 35 to 40 psi differential between internal and external pressure is required to close the sleeve. In a compressed air system operating at 100 psi, this limits the process pressure to 60 psi. 2. The valve must fail open (FO). There is no practical way to put a spring or other energy storage device inside the valve so it is not available in a true failclosed version. Auxiliary devices can be added to the valve; for example, a solenoid valve can be configured for fail closed on loss of electric power. Likewise, designers add air reserve tanks and other peripheral devices to close the valve on compressed air failure. None of these devices are functional when the sleeve wears through; at that point there is no way to sustain air pressure in the valve housing and the valve opens. 3. Low-pressure and vacuum systems can inhibit full opening or cause valve to flutter, thereby compromising flow capacity. There are ways to force the valve open using vacuum jet pumps but they add complexity. 4. Disappointing performance in throttling applications. Because these valves have no valve stem, there is no place to attach a positioner feedback lever. The alternative is to use the valve in concert with an I/P transducer coupled to a pneumatic amplifier. This converts a Compressed air Iron housing Rubber sleeve Compressed air Iron housing Rubber sleeve Compressed air Iron housing Rubber sleeve Closed FIG. 6.20j The throttling of shell and tube-type pinch valves. Partially closed Valve open

286 1330 Control Valve Selection and Sizing 4'' 3'' Reducer flange 4'' Valve body 3'' Pipe FIG. 6.20k The design of a double walled, 3 in. size, shell and tube pinch valve madc electronic signal to a pneumatic PSIG signal that is introduced into the valve shell. This varies the opening of the rubber sleeve, but without any feedback in the valve. The only means of closing the loop is through the measuring element and controller. This makes for sluggish response, erratic stability, and shallow turndown. Modulating shell and tube valves are generally limited to slow, linear applications like level control on large vessels. Exception to the rule: For ph control, use a shell and tube pinch valve with a solenoid valve instead of an I/P transducer. Configure the controller for pulse width modulation (PWM) with a s-l-o-w time base around 10 sec. This introduces a response time lag but cranks the turndown up to 100:1 or better with zero sleeve abrasion. Want well all-around modulating control? Read on and see how well mechanically pinched valves work in throttling applications. 5. Limited sleeve life. In mechanically operated pinch valves, designers add extra fabric for ample reserve strength. Designing a shell and tube valve does not offer this luxury. Additional fabric adds stiffness and increases the differential pressure needed to close the valve. This drives the sleeve design to the minimum number of fabric layers. The result is that the fabric operates right up to its yield point. After a certain number of open-close cycles the fabric fails through fatigue and the sleeve delaminates. Sleeve life in a shell and tube pinch valve operating at its maximum rated pressure is usually around a mean of 75,000 cycles with a variation of ±25,000. In comparison, mechanical pinch valves operate through practically an unlimited number of cycles because the fabric is operating at reduced stress. Mechanical pinch valves operating for millions of cycles over a period of 20 years are not uncommon. One method to increase the sleeve life is to regulate the compressed air pressure to precisely 40 psi over process pressure. For example, a 4 in. valve might be rated for 65 psi max process pressure with 105 psi max compressed air pressure. However, if the actual process pressure is 25 PSIG, reducing the air pressure from 105 down to 65 PSIG greatly reduces stress in the reinforcing fabric. This increases sleeve life by about 50%. The additional cost of an air pressure regulator pays for itself in improved sleeve life. Another method of improving sleeve life is to double the sleeve s wall thickness. As illustrated in Figure 6.20k, this can be done by increasing the valve size, underboring the hole through the sleeve, and doubling the thickness of the rubber sleeve. This requires the use of reducer flanges to make the connection between the pipe and valve. The Clarkson-C Valve The Clarkson division of Tyco makes a unique variation of the shell and tube valve (Figures 6.20l) Elastomer sleeve Muscle FIG. 6.20l The Clarkson-C valve. Housing Retainer

287 6.20 Valve Types: Pinch Valves 1331 throughout the usable throttling range. Compared to conventional shell and tube pinch valves, this valve can handle higher pressure and more abrasive fluids, and it offers more precise throttling. Naturally, the use of hydraulics adds another level of complexity and, therefore, both maintenance and first-investment costs. FIG. 6.20m The Clarkson C-Valve showing the air-powered hydraulic pump, amplifier, tank, positioner, check valve, pressure regulator, and gauges. and 6.20m. It is called the Clarkson-C or shutter valve. This rather unusual valve has been around since the late 1960s. Instead of compressed air, this valve uses hydraulic fluid, which is pressurized to several hundred PSIG, in order to be able to collapse the rubber sleeve. The assembly includes an airpowered hydraulic pump, valves, positioner, air filter-regulator, and gauges piggybacked onto the valve. The hydraulic fluid is separated from the sleeve by a thick-walled rubber cylinder called the muscle. To throttle the slurry, the muscle is compressed inward by hydraulic pressure, reducing the orifice size of the sleeve in a 360 squeeze. As it constricts, a round concentric aperture is maintained Opening Sleeve Pinch bar FIG. 6.20n Full ported and fully round single pinch design. Mechanically Pinched Valves The alternative to the shell and tube-type pinch valve described above is the mechanically pinched valve. In this design, steel pinch bars clamp the rubber sleeve closed. A handwheel, lever, electric actuator, or pneumatic cylinder actuates the pinch bars. The mechanically pinched valve has ample energy available to effect closure so the rubber sleeve can incorporate more reinforcing fabric. This enables the valve to handle higher process pressure and extends potential sleeve life to millions of cycles. A pneumatic actuator can be a double-acting cylinder, which, if you combine it with a two-coil solenoid valve, is inherently a fail-in-last position device. Pneumatic actuators are available for fail-open and fail-closed operation, using mechanical springs or air reservoirs to store the energy required to operate in the absence of compressed air pressure. As a general rule, pneumatic actuators are the leastexpensive for pinch valves up to 6 in. size, electric actuators have the advantage for valves over 10 in., and in between it is roughly a draw. The majority of mechanically pinched valves use flanged connections, although there are some designs that use a slipon connection with a hose clamp. 125/150 lb flanges are the most common, with 250/300 lb flanges also available. A word of explanation on the difference between 125 and 150 lb flanges: The bolt circles are identical between 125 and 150 lb flanges. The difference is the housing material: If the valve housing is cast grey iron or aluminum, ANSI 125 lb specifications apply; if the housing is steel, ductile iron, or stainless steel, then ANSI 150 lb specifications apply. The same relationship holds for 250 and 300 lb flanges. Single Pinch Design As illustrated in Figure 6.20n, the simplest mechanical pinch valve design consists of a rubber tube and a single steel bar that pinches the tube from above. When the valve is open, the inside is a round circle with its diameter of opening equal to the diameter of the adjacent pipe. Therefore, a full open 2 in. valve can pass a 2-in.- diameter sphere. This design offers good abrasion resistance, allows the pipe to drain completely, and can be cleaned with a pipe pig. This design is also a simple and economical mechanism. The main limitation of this design is valve size: It works for pinch valve sizes up to and including 3 in. Another problem is due to the fact that as the steel bar pinches the tube closed, the rubber and reinforcing fabric have to stretch. On small valves, this is not a problem. But on larger valves, the change in length exceeds the allowable elongation of the rubber and fabric matrix.

288 1332 Control Valve Selection and Sizing Pinch bar Sleeve Pinch bar Sleeve Opening Opening FIG. 6.20o Prepinched design. There are three possible solutions to this dilemma: 1. Make the valve longer. This reduces the elongation on the sleeve. This solution is controversial, because longer valves take up more space, and in today s competitive environment, plant space is a valuable commodity. 2. Prepinch the valve. 3. Use two moving pinch bars. Prepinched Design The valve in Figure 6.20o is a prepinched design. The term pre-pinched refers to the iron housing. The valve itself is full ported, because the sleeve has a full round shape before insertion into the housing. The iron housing has a weir cast into it that pinches the valve to centerline, resulting in a D shaped opening. This valve can pass a sphere equal to half nominal line size. If you want to pass a 4-in.diameter grapefruit through a 4 in. valve, you have to cut it in half and push the pieces though flat side down. Advantages Simple operation. Reduced cost (compared to dual pinch valves). Shorter actuator stroke, reducing the cost of the associated actuator. Good throttling characteristic. Disadvantages: Iron housing Pre-pinched design Largest particle that the valve can pass is half nominal valve size. Reduces valve capacity by 20%. With valve full open, adjacent pipe will not drain below centerline. You cannot pig out the line. Accelerated sleeve wear in high-velocity, high-abrasion applications FIG. 6.20p Fully round and full ported dual pinch mechanism. Dual-Pinch Mechanism As shown in Figure 6.20p, an alternate design is the dual-pinch mechanism, where two separate pinch bars squeeze the rubber tube closed. Here, there are two moving pinch bars that meet at the centerline of the valve to pinch off the rubber sleeve. Full round refers to the mechanical parts of the valve; full port refers to the sleeve design. Advantages Excellent abrasion resistance. Minimal friction. Valve can pass particles equal to line size. Adjacent pipe can drain completely. Valve can pass a cleaning pig. Disadvantages Mechanism is more complex. More expensive than prepinched. Actuator moves when stroking valve. Requires longer stroke actuator compared to prepinched design. Open and Enclosed Designs Most mechanical pinch valves have a housing that surrounds the rubber sleeve. As shown in Figure 6.20q, this housing connects the flanges, supports the handwheel or actuator, and protects the rubber sleeve from the environment. This housing also serves to contain the process fluid when the rubber sleeve wears out. An alternative is the open frame design. This pinch valve consists of a rubber sleeve, a clamping mechanism, and little else. Its light weight makes it popular for use in nonmetallic piping systems. It is also popular in mining operations as a throw-away valve. Advantages Easy visual inspection of the rubber sleeve Economy Lightweight

289 6.20 Valve Types: Pinch Valves 1333 Gearbox and Handwheel Pinch valves require extraordinary amounts of thrust to operate. The pinch valve requires brute force and lots of it to close the sleeve against the pressurized process fluid, which is a similar task as trying to crush a pressurized truck tire. For example, a 10 in. handoperated pinch valve closing against, say, 75 PSIG process pressure requires about 5 tons of force at the seat. Using a normal acme-threaded stem, this works out to 150 ft lb f of torque. If one uses an unassisted handwheel, one will require a 6-foot diameter handwheel to reduce the rim pull down to 50 ft lb f. Therefore, once the valve size exceeds 6 or 8 in. and depending on the process pressure, consider the use of a gear assist for manual valve operation. This is particularly true in case of chain wheel operators. Cutting corners by eliminating the gearbox or using a smalldiameter chain wheel can cause serious problems, because once one exceeds the requirement of 150 ft lb f rim pull, the plant operator might not be able to provide it. FIG. 6.20q Enclosed pinch valve. Part of the housing has been cut away to show the operation of the guide rods. (Photo courtesy Larox Company.) Caveats The rubber sleeve is a sacrificial part. When the sleeve wears out, whatever is being pumped just sprays all over the place. These valves are well suited for sand mining and stormwater applications, because if one inadvertently dumps a ton or two of sand on the ground, one just sends a crew to clean it up, but no serious harm is done. On the other hand, always use a fully enclosed pinch valve when the process fluid is scalding hot, acidic, caustic, toxic, or flammable. Always consider the consequences of leakage, and never compromise safety. Actuators Mechanical pinch valves can be actuated by: Handwheel Handwheel and gearbox Pneumatic actuators Electric actuators Handwheel When a pinch valve is to be buried or installed in a pit, it can be equipped with a torque tube extension and a handwheel operator. From an esthetic point of view, the occasional handwheels protruding from the ground are acceptable, but the cost can be substantial when one has to dig up the valve for servicing. Pneumatic Actuators Pneumatic actuators work particularly well with pinch valves. They are simple, economical, and dependable. One should use pressure regulator controls so that the closing thrust can be set precisely. This is necessary, on the one hand, to provide sufficient thrust to close the valve drop tight, and on the other hand, to avoid overpinching to the point where one might damage the rubber sleeve. Also, as the sleeve wears from erosion, a pneumatic actuator has reserve travel to compensate for the lost material. The most common pneumatic actuator used on pinch valves is the conventional double-acting air cylinder. When the air supply fails, a double-acting air cylinder fails in its last position. Using an Air Cushion The pneumatic cylinder can be supplied with a spring or air cushion on either top or bottom for fail-closed or fail-open operation. Air cushions are cheaper than springs but are less reliable, because a minor leak can void the actuator s ability to execute the proper fail action. The operation of the air cushion is shown in Figure 6.20r, where compressed air is supplied through tube a to pressure regulator b, which reduces the pressure to 45 PSIG. Air flows through check valve c and into cylinder head d and is trapped in the cylinder head by check valve c at a pressure of 45 PSIG. If the air supply fails and, therefore, the pressure below the piston drops, this drives the piston cylinder rod e down, closing the valve. To retract the cylinder rod, one needs both the air supply to be on and the three-way solenoid (f) to be energized. When this is the case, 90 PSIG air is injected underneath the piston, which lifts it up, compressing the air in the cylinder head (d) to about 80 PSIG. If the solenoid valve is de-energized or the compressed air supply fails, the 80 PSIG air that is trapped air above the piston will drive the piston down, thereby closing the valve. As the air cushion expands, its 80 PSIG pressure falls back

290 1334 Control Valve Selection and Sizing Gauges b Regulator c Check valve Air cushion a Piston d Solenoid valve f Compressed air FIG. 6.20r Double-acting cylinder can be configured to fail closed on air failure, when it is provided with an air cushion. to 45 PSIG. Therefore, the cylinder has to be sized to operate at this reduced pressure. Suspension Bag One pinch valve manufacturer has devised a particularly clever pneumatic actuator, where instead of a conventional air cylinder, a rubber pneumatic suspension bag is used as the actuator (Figure 6.20s). One advantage of this actuator on modulating applications is that it has a nonlinear e FIG. 6.20t 1 in. mechanically operated pinch valves, with pneumatic actuator on the left and electric on the right. As can be seen in Figure 6.20d, the cost of the electric valve in this size range is higher. For its cost, the user can purchase the pneumatic valve, plus a 2 hp compressor, and have $700 left over. thrust output. The reason for this is that as the suspension bag expands vertically, it necks in around its circumference. This eliminates breakaway friction and overtravel, enabling them to position a valve with 1/1,000th in. precision. Pneumatic suspension bags are normally used in trucks and buses as shock absorbers. They are visible on tractortrailers between the frame and axel assembly. Air bags have a decades-long track record of reliable operation in temperature extremes, rain, snow, sleet, dust, oil, vibration, and shock. They tolerate contamination of the compressed air system without adverse effects. They require no lubrication and operate virtually friction free. Electric Electric actuators also work well with pinch valves (Figure 6.20t). They avoid problems associated with compressed air including, compressor maintenance, condensation freezing, and the need to run airlines throughout the plant. In smaller valves, pneumatic actuators are more economical. As it was shown in Figure 6.20d, at around 6 or 8 in., the balance of economy tips towards electric actuators. Throttling Characteristics FIG. 6.20s Suspension bag actuator with its cover removed to clearly show its components. (Courtesy Onyx Valve Company.) The performance of the first throttling pinch valves was very poor, because they were grossly oversized. Until 1985, when David Gardellin ran the first capacity tests on pinch valves, nobody had a clue as to what their C v and F L values were. As it turns out, pinch valves offer higher C v values per valve size than any other valve style (see Tables 6.20e and 6.20f). This is not unexpected, because they have a smooth streamlined internal configuration with no disk, seals, gate tracks, packing cavity, or other sources of friction or turbulence.

291 6.20 Valve Types: Pinch Valves 1335 Pinch bar Pinch bar Sleeve Opening Pinch bar Full round reduced port Opening Iron housing Pre-pinched design Pre-pinched reduced port FIG. 6.20u Pinch valves can be provided with reduced-port sleeves in two ways. The full round reduced port version is shown on the left and the prepinched reduced port version is shown on the right. The reason for the poor performance of the full-ported modulating pinch valves is their excess capacity. The consequence of it being oversized is that the automatic controller tends to operate it nearly fully closed, down near the seat. This precipitates two problems: One is accelerated sleeve erosion caused by the high velocity flow. Pumping abrasive slurry through a pinch valve running just off the seat is like firing a shotgun through a mail slot. The sleeve doesn t last very long. The other problem is poor control quality. This was because if a 6 in. valve is used, but it needs to open only 1 /4 in. to pass the required maximum flow, throttling over such narrow range is next to impossible, and no positioner can correct that. Reduced Port and Clarkson Sleeve One way to improve the control performance of pinch valves is to use reduced port sleeves. As shown in Figure 6.20u, the reduced port can be provided either in a full round form or in a prepinched form. On the left of Figure 6.20v, the inherent throttling characteristics of a full round reduced port valve is described. It is not exactly linear, but using a characterizing positioner with the X 2 function gets it close enough to provide a nearly linear and stable control in most situations. On the right of Figure 6.20v, the inherent throttling characteristics of a prepinched reduced port pinch valve is shown. It is prepinched by the weir in the cast iron housing. Because the sleeve is prepinched, the opening is D shaped. It is reduced port because of the taper molded into the rubber sleeve. The prepinched design offers considerable cost savings and a more linear inherent throttling characteristic (on the left of Figure 6.20v) than a full round valve. Figure 6.20w illustrates the throttling characteristics of the Clarkson C-1 pinch valve, which is between linear and equal percentage. C v Full round valve capacity Valve Valve with X 2 positioner characteristic Opening C v Pre-pinched valve capacity Valve characteristic Valve with X 2 positioner characteristic Opening FIG. 6.20v Characteristics of reduced port pinch valves: On the left the characteristics of the full round, on the right the prepinched design is shown.

292 1336 Control Valve Selection and Sizing Capacity Clarkson-C1'' can just change the rubber sleeve to a 6 in. 4 in. to double the C v to 680. If more capacity is required, a 6 in. 5 in. sleeve will give a C v = 1723 and a full 6-in. sleeve will provide a C v = 2200, which is a sixfold increase in capacity without changing the valve or the fittings. C v Opening FIG. 6.20w Characteristic curve of a 1 in. Clarkson C valve. Example When compared to ball, butterfly, or plug valves, the reduced port and Clarkson pinch valves have higher valve coefficients (C v and F L ) and better throttling characteristics. In many installations, they are also more cost effective. As an example, let us assume that we have an application that requires a maximum C v of 336, and the pipe size is 6 in. From both Tables 6.20e and 6.20f, we find that the C v of a full ported 3 in. valve is the most suitable (C v = 570). This application, therefore, works out to require a 6 in. 3 in. pinch valve. There are three ways to design this system: Option 1 Buy a 6 in. 3 in. reduced port pinch valve that (according to Table 6.20f) has a C v measured from 6 in. flange to 6 in. flange of 336. This option (shown on the left of Figure 6.20x) is the most expensive solution, but it also offers the most flexibility, because if the plant capacity increases, one Option 2 The second option (shown on the right side of Figure 6.20x) is to purchase a 3 in. full ported valve and install it within 6 in. 3 in. reducer fittings. This installation will also provide a C v (measured from 6 in. flange to 6 in. flange) of 336. This second option is the least expensive, but it also offers the least flexibility in terms of accommodating future increases in plant capacity. Option 3 One can compromise between the above options by installing a 4 in. 3 in. pinch valve inside two 6 in. 4 in. pipe reducers. This solution might be the best of both worlds. It offers better economy than buying a 6 in. reduced ported pinch valve, but it does have flexibility for increased capacity by changing out the 4 in. 3 in. reduced port rubber sleeve to a 4 in. full ported one. This change would increase the C v from 336 to 680: double the original capacity without having to change the valve, actuator, or adjacent fittings. Such options do not exist with ball, butterfly, or plug valves. APPLICATIONS Pinch valves are used in a surprising range of industries and applications. Their primary application is in controlling the flow of abrasive slurry streams, but they are also used elsewhere. Some examples will be listed below. Wastewater Pinch valves can control sewerage and stormwater streams, which might contain oil, paper, rags, sticks, disposable razors, garbage, grit, and other contaminates. 6'' Pipe 6 3 Pinch valve 6'' Pipe 6'' Pipe 6 3 Reducer 6 3 Reducer 6'' Pipe C v = 336 C v = 336 3'' Pinch valve FIG. 6.20x The same valve coefficient (C v = 336) can be obtained by installing a 6 in. 3 in. reduced ported pinch valve (left) or by using a 3 in. full ported pinch valve inside 6 in. 3 in. reducers (right).

293 6.20 Valve Types: Pinch Valves 1337 Flue-Gas Desulfurization Pinch valves are suited for lime slurry service because they offer a straight-through design without any crevices or cavities for material to accumulate. Closing a pinch valve causes the rubber sleeve to stretch. As the rubber sleeve stretches, any scale that has built up on the surface will flake off. The fluid acceleration that occurs as the valve is closing helps to wash these flakes off the rubber surface, so the interior of the valve is self-cleaning. Unlike ball and plug valves, there is no dead space in the valve where a plug of slurry can solidify and block the line. Mine Slurries Pinch valves have numerous applications in mining applications, including: Paper and Tile Manufacturing Pinch valves can handle countless fluids in paper manufacturing, including bleaching and coating applications. As shown in Figure 6.20y, pinch valves can control the levels in head boxes in ceiling tile manufacturing applications, where the thickness of the finished tiles is within 1/16th in. of the specifications. Toxic Gas Applications Pinch valves are frequently used to handle carbon monoxide, methane gas from sewerage and landfill applications, and so on. Abrasion is not an issue; the pinch valve is preferred because there is NO PACKING BOX that could leak toxic or flammable gasses into the atmosphere. Consequently, fugitive emissions are not a problem with pinch valves. Pigments, Paint, and Ink Pinch valves routinely handle thick viscous fluids that would cement out in other valves after prolonged closure. Pinch valves never jam or clog. Glue Sand mining Copper tailings Gold slime Taconite slurry Coal and water slurry Coal and oil fuel Borax Phosphate slurry Diatomaceous earth Tar sands Mercury Molybdenum Kaolin and other clays Pinch valves are used to modulate the pressure in a closedloop system that sprays the glue onto plywood layers prior to lamination. A pinch valve can provide both throttling control and reliable drop-tight closure even on wood glue. FIG. 6.20y A 10 in. modulating pinch valve used for throttling pulp stock and chopped glass fibers to make acoustic ceiling tiles. (Courtesy Onyx Valve Co.) Food Pinch valves are used to convey tomatoes, chili peppers, live shrimp, chicken feet (for export to China), mustard, chicken bones and entrails (pet food stock), pig and cattle hooves (for glue), and grains. Powders and Grinding Compounds Pinch valves are routinely used to control air-conveyed cement, dry lime, dust, detergent powder (they fill containers very precisely), and fertilizer. They are also used to transport grinding compounds including diatomaceous earth, aluminum oxide, and garnet slurry used to polish teeth, TV screens, auto bodies, and the kitchen sink. Chemicals Pinch valves can handle a variety of chemicals without corrosion and with zero leakage because there is no packing box in their design. CAVITATION The phenomenon of cavitation has been discussed under the subjects of control valve applications (Section 6.1), control valve noise (Section 6.14), and control valve sizing

294 1338 Control Valve Selection and Sizing (Section 6.15). For that reason, its discussion here will be brief, and the reader is referred to the noted sections for a detailed treatment. The Phenomenon FL The pressure recovery factor (F L ) relates to the ratio between the pressure drop across the valve and the pressure difference between the inlet pressure and the vena contracta pressure (P vc ). F L = [(P 1 P 2 )/(P 1 P vc )] 1/2 6.20(1) The higher the pressure recovery factor (F L ), the better the cavitation resistance of a particular valve design (see Figure 6.1v for a range of valve designs). The cavitation coefficient (K c ) is the ratio of the difference between the inand outlet pressures and the difference between the inlet and the vapor pressure of the flowing fluid (P v ). K c = (P 1 P 2 )/(P 1 P v ) 6.20(2) If the pressure at the vena contracta falls below the vapor pressure of the liquid, gas pockets form in the flowing liquid. The process reverses itself when the liquid emerges from the restriction, because as the fluid decelerates its pressure recovers. Pressure does not recover to its original magnitude, but if it recovers beyond the vapor pressure, the cavities collapse and the vapors reliquefy. This is cavitation. Figure 6.1w illustrates a pressure profile where cavitation takes place as vapor bubbles form at the vena contracta and then implode as the pressure recovers and exceeds the vapor pressure. The microjets generated by these implosions cause the damage to metallic surfaces in the area. The resulting micro shockwaves generate localized impact pressures over 200,000 PSIG, which no material can withstand. The location where the cavitation damage occurs is where the bubbles start to collapse, which is downstream of the vena contracta. Figure 6.15d shows that the formation of these bubbles reduces the process flow until it reaches the choked flow condition, when an increase in inlet pressure to the valve will not increase the flow through it. The pressure drop, which will cause choking ( p choked ), can be approximately calculated as The Pinch Valves p choked = F L 2 (P P v ) 6.20(3) Pinch valves are one of the high recovery valve designs with a pressure recovery factor around F L = Reduced port valves have higher F L coefficients, and the more severe reduction in port size, the higher it will be. Figure 6.20z shows the F L coefficient of a reduced port pinch valve as a function of its percentage opening. Some manufacturers suggest that one way to avoid cavitation is to size the control valve, not for the available total pressure drop (P 1 P 2 ), but the pressure drop that would FL Valve % open FIG. 6.20z The recovery coefficient F L as a function of the opening of a reduced ported pinch valve. cause choking, or some fraction of that (0.9 p choked ). This approach will redistribute the total system drop and eliminate cavitation in the valve (see C in Figure 6.2aa), but it does it by oversizing the valve, which will also deteriorate the quality of control. Metal valves and rubber-lined valves respond differently to cavitation. Metal valves begin to deteriorate at a point between incipient cavitation and choked flow. Ironically, rubber fares better than metal under cavitation conditions. Rubber absorbs much of the shock of the imploding bubbles, so pinch valves can wade far into the chocked flow zone without any adverse effects. Limiting or Eliminating the Damage Figure 6.20aa illustrates five ways how cavitation can be eliminated by making changes in the installation or in the valve design. One pinch valve manufacturer provides a calculated approximation based on the available total pressure drop (P 1 P 2 ) and the pressure drop that would just begin to cause choking ( p choked ). They argue that the difference between these two pressure drops is the energy source that drives cavitation (the formation and collapsing of bubbles), and therefore a cavitation damage factor (G) can be calculated by using this difference: G = [(P 1 P 2 ) ( p choked )](Q/17.44d) 6.20(4) It is claimed that if G is calculated by using Q in gpm units of ambient water and d as the diameter of a rubber pinch valve, Table 6.20bb will give acceptable selection guidance for cavitation. Using a Choke Fitting Cavitation does not start at the throat of the valve, but develops past the throat where pressure

295 6.20 Valve Types: Pinch Valves 1339 P P 1 P P 1 P P 1 P P 2 Valve with less recovery higher F L & K C P v P 2 Two valves in series P v P 2 Reduce valve P P v P vc P vc (A) (B) (C) P vc P 1 P P 1 P P 2 Move valve closer to pump or to lower elevation P v P 2 Lower the temperature P v P vc P vc P v (D) (E) FIG. 6.20aa Ways to eliminate cavitation: A) use valve with higher Kc and F L, B) use two valves, C) reduce valve pressure drop, increase valve size, D) move valve to lower elevation or closer to pump, and E) reduce the temperature of the flowing process fluid. begins to recover. As a result, cavitation damage occurs downstream of the valve throat and destroys the pinch valve s sleeve between the pinch point and the valve exit. Modulating pinch valves are usually supplied with a reduced port sleeve. The sleeve is molded with a venturi shape that tapers down to a small diameter in the center of the valve. The sleeve is normally symmetrical, tapered on both inlet and outlet so the valve can be installed for either flow direction. This design is vulnerable to cavitation damage. Revising the sleeve design to an asymmetrical shape will enable the valve to withstand high levels of cavitation (Figure 6.20cc). This design is called a trumpet mouth or cone sleeve. The thicker rubber at the valve outlet can absorb more cavitation damage. There are several ways to exploit this feature to tolerate even higher levels of cavitation with minimum sleeve damage. One method is to install the valve at the end of the piping run, allowing the pressure to recover as the fluid emerges from the piping system. Here the cavities form after they have emerged from the choke fitting, where they cascade harmlessly into the open tank. If it is not practical to install the valve at the end of the pipe run, some pinch valve manufacturers recommend the use of a trumpet mouth-type valve, followed by a short spool piece of pipe having a diameter that equals the valve s port size. The minimum spool piece length is typically 10 times the nominal diameter of the valve. The spool is connected to TABLE 6.20bb The Predicted Degree of Cavitation Damage Based on the Size of the G Factor 4'' Pipe 2'' Nozzle Pressure recovery Cavitation Damage Factor (G) Predicted Degree of Cavitation Damage 0 G 100 Cavitation damage will be undetectable. 100 G 300 Some damage will be observable after 1 year of operation. 300 G 500 Sleeve damage will be observable in 6 months. 500 G 750 Frequent maintenance and sleeve replacement is likely. 750 G Consult factory. FIG. 6.20cc If the valve can be located at the end of a pipeline, the addition of a choke fitting at the end of an asymmetric sleeve can move the zone of pressure recovery (cavitation) away from the valve.

296 1340 Control Valve Selection and Sizing the downstream full-size piping by a reducer flange, which creates a sharp, sudden expansion. This design is not desirable because it does not eliminate the cavitation, but only moves its location downstream of the valve. Yet, some manufacturers argue that if the expansion downstream the spool is made with a rubber hose-type pipe joint, it can absorb a fair amount of wear and can be replaced once or twice a year at relatively low costs. CONCLUSIONS The phenomenon of cavitation is discussed in more detail in connection with control valve applications (Section 6.1), control valve noise (Section 6.14), and control valve sizing (Section 6.15). In general, one should note that the difference between the total available pressure drop (P 1 P 2 ) and the pressure drop that would cause choking ( p choked ) is the energy that drives cavitation (the formation and collapsing of bubbles). Therefore, instead of trying to find ways of harmlessly wasting this excess and unnecessary energy, one should concentrate on finding ways to NOT introduce it in the first place. Bibliography Carey, J. A., Control Valve Update, Instruments and Control Systems, January Control Valves Globe, Plug, Pinch, Needle, Gate, Measurements and Controls, February Dobrowolski, M., Guide to Selecting Rotary Control Valves, Instrumentation Technology, December Fernbaugh, A., Control Valves: A Decade of Change, Instruments and Control Systems, January Gardellin, D., Valve Cavitation, in Encyclopedia of Chemical Processing and Design, Vol. 61, Marcel Dekker, 1997.

297 6.21 Valve Types: Plug Valves C. S. BEARD (1970) R. D. BUCHANAN (1985) B. G. LIPÁTK (1995, 2005) Flow sheet symbol Sizes: 1 /2 to 36 in. (12.5 mm to 0.96 m) Types: V-ported, three-way, four-way, five-way, fire-sealed Design Pressure: Typically from ANSI Class 125 to ANSI Class 300 ratings and up to 720 PSIG (5 MPa) pressure, with special units available for ANSI Class The retractable seat design is suited for 10,000 PSIG (69 MPa) service. Design Temperature: Typically from 100 to 400 F ( 73 to 204 C), with special units available from 250 to 600 F ( 157 to 315 C) Rangeability: Refer to Section 6.7; generally 20:1 Characteristics: See Figure 6.21a Capacity: C v = (25 to 35)d 2 ; see Tables 6.1a and 6.21b Leakage: Materials of Construction: Costs: Partial List of Suppliers: Metal seats ANSI Class IV, composition seats ANSI Class V; see Table 6.1gg for definitions Iron, forged and alloy steel, chrome plating, 302 through 316 stainless steel, Alloy 20, Ni-resist, Monel, nickel, Hastelloy B and C, and zirconium, plus rubber or plastic, including PTFE linings Costs vary drastically with design and accessories. In general, the cost of conventional plug-type control valves is about half that of globe valves, while the cost of eccentric rotating plug valves is about the same as that of globe valves of the same size and materials. (For the costs of carbon steel and stainless steel globe valves, refer to Figure 6.19b.) ABB Inc. ( Anchor/Darling Valve Co. ( Cashco Inc. ( Circle Seal Controls Inc. ( Dezurik/ SPX Valves & Controls ( Emerson Process Management (www. emersonprocess.com) FMC Fluid Control Div. ( Halliburton Services ( Honeywell Industrial Controls ( Hydril Co. ( Jordan Valve ( Mar-In Controls ( Nordstrom Valves Inc. ( Offshore Technology ( Spirax Sarco Inc. ( Xomox/Tufline ( 1341

298 1342 Control Valve Selection and Sizing % Flow or C v Modified linear V-ported Modified parabolic Equal percentage FIG. 6.21c The self-cleaning nature of the eccentric rotating plug valve makes it a good option for slurry service applications FIG. 6.21a Plug valve characteristics are a function of the type of the particular V-port or of the shape of the throttling plate used. GENERAL CHARACTERISICS % Rotation The rotary plug valves (similar to ball and butterfly valves) used to be considered only as on/off shutoff valves, but today they are also used as control valves. Table 6.1a shows how they compare in their characteristics and applications to some of the other control valve designs. Plug valves are well suited for corrosive, viscous, dirty, fibrous, or slurry services, while they are generally not recommended for applications where cavitation or flashing is expected. Relative to the traditional globe valve, the advantages of conventional plug valves include their lower cost and weight, higher flow capacity, which can be two to three times that of the globe valve, if the plug is not characterized. In addition, they provide tight shutoff, fire-safe designs, and low stem leakage, which meets OSHA and EPA requirements. The designs using characterized or eccentric rotating plugs provide good control performance and a self-cleaning flow pattern, which also reduces noise and cavitation. Actually, the performance of the rotating spherical segment-type valve is just as good as that of a globe valve, and for that reason some refer to it as a globe valve (Figure 6.21c). In addition to their superior stem-sealing capability, plug valves are also suited for such corrosive applications as chlorine, phosgene, hydrofluoric acid, and hydrochloric acid. Plug valves are widely used on lethal and toxic services and can be made fire-safe by the use of Grafoil packing and can meet the external leakage requirement limits of API 607. TABLE 6.21b Valve Coefficient (C v ) Values of Standard and Characterized Plug Valve Designs* Valve Size-inch (mm) 1 /2 (12) 3 /4 (19) 1 (25) 1 1 /2 (38) 2 (51) 3 (76) 4 (102) 6 (153) 8 (224) 10 (254) 12 (305) Standard Modified parabolic plate Modified linear plate Three-way Equal-percentage cage Modified parabolic cage Modified linear cage V-ported *Courtesy of Xomox Corp./Tufline and DeZurik.

299 6.21 Valve Types: Plug Valves 1343 Plug Valve Features The plug valve is a type of quarter-turn valve that is among the oldest designs known in engineering. Wooden plug valves were used in the water distribution systems of ancient Rome and probably predate the butterfly valve. Although no longer as popular as ball or butterfly valves, they lend themselves to special designs that work very well in specific control applications. Conventional plug valves are usually lower cost and lighter weight than comparable gate or globe valves. Plug valves afford quick opening or closing with tight, leakproof closures under conditions ranging from vacuum to pressures as high as 10,000 PSIG (69 MPa). Some, including the various characterized and Y-ported or diamond design, can be used for throttling, while others, like the multiport, are used for diverting and bypass applications. Plug valves are used on gas, liquid, and nonabrasive slurry services. Lubricated plug valves can also be used for abrasive slurries, and eccentric plugs are also used on applications involving sticky fluids. Plug valves are also used for applications requiring the contamination-free handling of foods and pharmaceuticals. In general, plug valves can handle applications with the following requirements: High flows at low pressure drops Low flow control Flow diversion High- or low-temperature applications Vibration-free operation Throttling control, only with eccentric and V-ported characterized designs. The conventional plug valves are generally undesirable for the following types of applications: In operating rotary valves, the linear movement of cylinder or spring-and-diaphragm actuators must be converted by linkages, which introduces hysteresis and dead play. In addition, a nonlinear relationship exists between actuator movement and the resulting rotation. These considerations make the use of positioners essential, which on fast processes can lower the quality of control. The torque characteristics of these valves are nonlinear (Figure 6.4v), and because of the high break-torque requirements, the actuators can be oversized relative to the torque requirements in the throttling range. DESIGN VARIATIONS The first plug valves consisted of a tapered or straight vertical cylinder containing a horizontal opening or flow-way inserted into the cavity of the valve body (Figure 6.21d). They have developed through time into numerous shapes and patterns, depending on the application, but almost all are adaptations of the cylindrical or tapered plug. Within that plug, however, the ports may be round, oval, rectangular, V-, or diamondshaped, and can be the flow-through type two-way valves or multiport. These make up the special designs described in subsequent paragraphs. Plug valve designs can be categorized as lubricated or nonlubricated. In the lubricated type, the thin film of lubricant serves not only to reduce friction between the plug and the body, but also to form an incompressible seal to prevent gas or liquid leakage. Because the seating surfaces are not exposed in the open position, gritty slurries may be handled. The lubricant hydraulically lifts the plug against the resilient packing to prevent sticking. A special lubricant must be injected periodically while the valve is either fully open or fully closed. Flow modulation or continuous, exact flow throttling Maintenance-free operation (occasional lubrication is usually required and plugs may wear) Throttling and Actuator Considerations Typical actuator Lubricant fitting Plug stem When used for throttling service some of the above-mentioned advantages of rotary valves, such as their high-capacity, can become disadvantages. Their high flow capacity can result in installations where small valves are mounted in large pipes. This results in a substantial waste of pumping energy, as the pump has to overcome the reducer pressure drops. Also, the high-pressure recovery provided by most plug valve designs results in low vena contracta pressures, which in turn increase the probability of cavitation and noise. These problems, which are even more pronounced with pinch, butterfly, and ball valves, have been reduced by various means. In case of ball valves, perforated parallel plates have been inserted into the ball valve openings (Figure 6.16o) or flutes have been added to the butterfly disc (Figure 6.17s); however, the cavitation problems associated with high recovery valve designs have not been fully resolved. Packing Tapered plug FIG. 6.21d Conventional, lubricated plug valve with tapered plug. Lubricant check valves Thrust bearing with adjustment screw Lubricant grooves

300 1344 Control Valve Selection and Sizing Plug Cage O-rings Cage TFE sleeve Throttling plate Body FIG. 6.21e Throttling plates and characterizing cages can both be used to modify the inherent characteristics of plug valves. (Courtesy of Xomox/Tufline.) The plugs of nonlubricated plug valves are treated with coatings such as Teflon or are specially heat-hardened and polished to prevent sticking. Often they are constructed so the tapered plug may be lifted mechanically from the seat for easier operation. Characterized Plug Valves Plug valves can be characterized by the use of characterizing cage or plate inserts (Figure 6.21e). The resulting characteristics are a function of the shape of the opening on the cage or plate. Some examples of available plug valve characteristics are illustrated in Figure 6.21f. Rotation of the plug is inside a TFE sleeve, which is locked into the body in such a way that recessed areas minimized. Although a rangeability of 20:1 is claimed, this is made possible only if the valve can be fully FIG. 6.21g The design of a V-ported plug valve. open in order to provide full flow. The valve is available in 1/2 12 in. ( mm) sizes and up to 600 PSIG ANSI (4.1 MPa) rating for use up to 400 F (204 C). V-Ported Design The V-ported plug valve (Figure 6.21g) is used for both on/off and throttling control of slurries and fluids containing solid concentrations in suspensions greater than 2%. These applications occur principally in the chemical and pulp and paper industries. A diamond-shaped opening is created by matching a V-shaped plug with a V-notched body. Straight-through flow occurs on 90 rotation, when the plug is swung out of the flow stream. Shearing action and a pocketless body make the valve applicable for use on fibrous or viscous materials. The opening develops a modified linear flow characteristic with C v capacities approximating 17d 2. Valves are flanged from 3 16 in. ( mm) in bronze, corrosion-resistant bronze, or stainless steel. The body may be rubber-lined with a rubber-coated plug. A cylinder actuator and valve positioner are used for throttling control. A variation is the true V-port opening (Figure 6.21h). It is obtained by a rotating segment that is closing against a Modified linear Modified parabolic Equal percentage FIG. 6.21f Plug valve characteristics can be modified to linear, parabolic, equal-percentage, and so on.

301 6.21 Valve Types: Plug Valves 1345 Open There are no sharp corners or narrow openings to pack with stock. Large port area and clean interior design assure high flow capacity. FIG. 6.21h Illustration of how throttling is provided by a V-ported plug valve. Throttling Close plug-to-seat clearance remains constant. The V-orifice retains its shape throughout the cycle. Closed Flow continues until the V-orifice in the leading edge of the plug rotates past the seat, throttling is smooth down to the shutoff position. straight edge. The valve can be smoothly throttled on thick stock flows without the stock packing or interfering. The valve is available in sizes from 4 20 in. ( mm) and with C v stated as more than 20d 2. The valve is available in many body and trim materials for use in the chemical or pulp and paper industries. A cylinder-operated rack and pinion is used for on/off service with the addition of a valve positioner for throttling services. Adjustable Cylinder Type In another form of quarter-turn valve, flow is varied by rotating the core and by raising or lowering a curtain with an adjusting knob (Figure 6.21i). Proportional opening at any curtain position is made with the control handle, which may be attached to an actuator for automatic control. Various openings are obtained by these manipulations. The valve is used for combustion control and for mixing applications, in which case valves are stacked on a common shaft or operated by linkages from the same actuator. To obtain linear flow with constant pressure drop, a port adjustment technique is used. After installation, the curtain is closed until the pressure drop across the valve is one sixth of the total pressure drop of the system using the control handle in wide open position. This provides a flow characteristic approximating linear without decreasing sensitivity by limiting valve stroke. The percentage of flow is equalized by manipulation of the linkage to the actuator. Semispherical Plugs for Tight Closure Various designs have been developed to obtain tight closure while eliminating the continuous friction of the seals during rotation, as with most ball valves. The valve design illustrated in Figure 6.21j uses an eccentric ball. In the closed position, the rectangular end of the stem protrudes into the ball and the closure face is wedged toward the seating surface. As the stem is rotated to open the valve, the closure surfaces separate and the pin moves into a vertical slot so that rotation occurs. A nonlubricated seal can be provided with a primary Teflon seal enclosed in a body seat retainer ring. The valve is adapted for automatic operation by connecting the stem to a diaphragm actuator. Expanding Seat Plate Design As shown in Figure 6.21k, metal-to-metal or resilient seats can be provided in two seating segments. These segments are Bonnet gasket Stem packing chamber Pins Curtain ½ open, core full open Curtain ¾ open, core rotated ¼ Curtain ¾ open, core rotated ½ FIG. 6.21i Illustrations of both the curtain and the core openings of an adjustable cylinder-type plug valve. Seating core Bushing Hard facing Teflon insert Retainer ring FIG. 6.21j The semispherical plug in this plug valve design provides tight closure. (Courtesy of Offshore Technology, formerly Orbit Valve Co.)

302 1346 Control Valve Selection and Sizing at which point additional rotation of the drive creates a camming action to compress the packing ring. When the valve is being opened, the packing ring is released before the rotation of the plug occurs. Although also applicable to low-pressure service, this valve is particularly useful up to API 5000 PSIG (34.5 MPa) rating for 10,000 PSIG (69 MPa) service. It is used mostly in oil fields. Side view Closed position carried on a rail that is tapered so that downward stem movement forces the plates against the inlet and outlet ports. When the valve is being opened, the first few turns of the actuator cause the retraction of the plates and then plug rotation proceeds. These plates can be removed by merely removing the bottom plate of the valve V-body. Retractable Seat Type Side view Turning position FIG. 6.21k Plug valve with expanding seat plate design. End view Open position Positioning a movable seal after a spherical plug is in the closed position creates tight closure with sliding friction. In the retractable seat plug valve, a trunion-mounted partial sphere is operated by spur or worm gears (Figures 6.21l). The gear system rotates the plug until it is in a closed position, Overtravel Seating Design By fabricating a cylindrical flow passage within a tapered cylindrical plug, a plug valve can be built at least up to the 60 in. size, without undue weight and attendant inertia to rotation. In the overtravel seating design shown in Figure 6.21m, the rotation of the inner valve is caused by a rod operated by a piston pushing down on a rotator. Continuous movement of the rod closes the plug, but leaves a small crescent, because the plug is slightly raised from seating. For tight closure, the rod contacts a seating adjustment to force the plug into a tapered seating surface. In this design, the restriction of the flow area is rapid; 30% rod travel causes about 65% closure. Complete rotation of the plug occurs at about 65% stroke, with the additional stroke being utilized for seating. If rapid closure to about 20% opening does not create serious surge pressures, this valve can be used for emergency closure without much consideration of the piston speed. The rangeability of the valve exceeds 50:1 for proportional control. Free rotation and relatively low plug weight contributes to lower power requirements. Spherical plug Movable seal FIG. 6.21l Spherical plug and retractable seat for tight shut-off. (Courtesy of Hydril Co.)

303 6.21 Valve Types: Plug Valves 1347 Multiport Design FIG. 6.21m Two views of a plug valve of the overtravel seating design, shown in a throttling position. As shown in Figure 6.21n, a three-way plug valve is obtained by providing the plug with an extra port at 90 from the inlet, so that flow can be directed in either of two destinations. A multitude of directions can be achieved by nesting combinations of the simple multiport valves or by using more complex designs. These include a multistoried arrangement with the plug extending upward to connect to a series of tiered outlets. In such multistoried configurations, the plug has a long, vertical passageway connecting the horizontal ports. Another method of increasing the number of flow-directions is to design the plug with a diameter that is sufficiently larger than the ports so that intermediate ports can be placed at 45 or even 30 and 60. In that case, the actuators can be programmed to serve a variety of process applications. Bibliography Plug positioned for flow-through Plug positioned for discharge at 90 degrees FIG. 6.21n The design of a three-way plug valve. Ball, K. E., Final Elements: Final Frontier, InTech, November Baumann, H. D., Trends in Control Valves and Actuators, Instruments and Control Systems, November Bialkowski, B., Coughran, M., Beall, J., Control Valve Performance Update The Last 10 Years, Pulp Pap-Canada, (102):21 22, November Bishop, T., Chapeaux, M., Jaffer, L., et al., Ease Control Valve Selection, Chem Eng Prog, 98 (11): 52 56, November Borden, G. and Friedmann, P. G. (Eds.), Control Valves Practical Guides for Measurement and Control, Research Triangle Park, NC: Instrumentation, Systems, and Automation Society, Cain, F. M., Solving the Problem of Cavitation in Control Valves, Paper # , 1991 ISA, Anaheim, CA, October Control Valve Handbook, 3rd edition, Marshalltown, IA: Fisher Controls International Inc., Control Valve Seat Leakage, FCI , Cleveland, OH: Fluid Controls Institute, Inc., Control Valves Globe, Plug, Pinch, Needle, Gate, Measurements and Control, February Dobrowolski, M., Guide to Selecting Rotary Control Valves, Instrumentation Technology, December Hammitt, D., Rotary Valves for Throttling, Instruments and Control Systems, July Hutchison, J. W. (ed.), ISA Handbook of Control Valves, Research Triangle Park, NC: Instrumentation, Systems, and Automation Society, Kawamura, H., Selecting Valves for the Hydrocarbon Processing Industry, Hydrocarbon Processing, August Lyons, J., Lyons Valve Designer s Handbook, New York: Van Nostrand Reinhold, Monsen, J. F., Valve Wars Rising Stem vs. Rotary, Plant Services, January Pyotsia, J., A Mathematical Model of a Control Valve, 1992 ISA Conference, Houston, TX, October Rahmeyer, W., The Critical Flow Limit and Pressure Recovery Factor for Flow Control, InTech, November Shinskey, F. G., Control Valves and Motors, Foxboro Publication No Zappe, R., Valve Selection Handbook, Houston, TX: Gulf Publishing, Whatever Happened To... A Guide to Industrial Valve and Actuator Companies Past and Present, Washington, D.C.: Valve Manufacturers Association, 1997.

304 6.22 Valve Types: Saunders Diaphragm Valves C. E. GAYLOR (1970, 1985) B. G. LIPTÁK (1995, 2005) Flow sheet symbol Design Types: Applications: Sizes: Maximum Operating Pressure: Vacuum Limits: Temperature Limits: Materials of Construction: Weir, full-bore, straight-through, dual-range Slurries, corrosive fluids at low pressure drops Standard units from 1 / 2 to 12 in. (12 to 300 mm); special units up to 20 in. (500 mm) In sizes up to 4 in. (100 mm), 150 PSIG (10.3 bar); 6 in. (150 mm), 125 PSIG (8.6 bar); 8 in. (200 mm), 100 PSIG (6.9 bar); and 10 or 12 in. (250 or 300 mm), 65 PSIG (4.5 bar). See Figure 6.22a for limits. Mechanical damage can occur when opening valve against process vacuum With most elastomer diaphragms from 10 to 15 F ( 12 to 65 C); with PTFE diaphragm from 30 to 350 F ( 34 to 175 C). See Figure 6.22a. Body materials: iron, ductile iron, steel, 302 to 316 stainless steel, Alloy 20, bronze, Monel, Hastelloy C, aluminum, titanium, graphite, plastic such as PTFE lining or solid plastics Diaphragm materials: Teflon, Buna-N, neoprene, hypalon Characteristics: See Figure 6.22b Capacity: C v = 20 d 2 ; see Table 6.22c Rangeability: About 10:1; see Section 6.7 Leakage: Costs: Partial List of Suppliers: ANSI Class IV or V; for definitions see Table 6.1gg Costs vary drastically with design and accessories. In general, the cost of conventional Saunders-type control valves is about half that of globe valves of the same size and materials. (For the costs of carbon steel and stainless steel globe valves, refer to Figure 6.19b.) ABB Inc. ( Emerson Process Management ( Foxboro-Invensis ( ITT Industries, Engineered Valves ( McCanna/Marpac ( Nibco Inc. ( Teledyne Engineering ( Velan Valve Corp. ( INTRODUCTION The advantages of Saunders valves include their relatively low costs, tight shut-off, and suitability for corrosive, dirty, viscous, or slurry services (Table 6.1a). Their limitations include their poor control characteristics, although the use of the dual-range design improves it. They are also limited in their temperature (high and low) and pressure (high) ratings and are not suited for cavitating or flashing services. The valve coefficients for a number of Saunders valve sizes with a variety of connection types and lining options are tabulated in Table 6.22c. SAUNDERS VALVE CONSTRUCTION The Saunders valve is also referred to as a diaphragm valve or less often as a weir valve. Conventional Saunders valves utilize both the diaphragm and the weir for controlling the 1348

305 6.22 Valve Types: Saunders Diaphragm Valves 1349 Operating pressure (PSIG)* Pure gum and white natural rubbers ( mm) ½'' 1'' ( mm) 1¼'' 2'' ( mm) 2½'' 3'' ( mm) 5'' 6'' (200 mm) 8'' ( mm) 10'' 12'' Natural rubber & hycar Neoprene 40 Temperature ( F)** *1 PSIG = 6.9 kpa * * C = F FIG. 6.22a Pressure and temperature limitations of the various diaphragm materials used in Saunders valves, as a function of valve size. Hypalon, white butyl Kel -F, black butyl Teflon with butyl backing % Flow or Cv Quickopening STD. saunders Dual-range saunders % Lift FIG. 6.22b The characteristics of conventional Saunders valves are nearly quick opening, while the characteristics of dual-range Saunders valve designs are closer to linear. flow of the process fluid (Figure 6.22d), while straightthrough and dual-range ones do not necessarily use a weir. The conventional Saunders valve is opened and closed by moving a flexible or elastic diaphragm toward or away from a weir. The elastic diaphragm is moved toward the weir by the pressure applied by a compressor element on the TABLE 6.22c Valve Coefficients (C v ) of Conventional Saunders Valves* Flanged Ends Nominal Valve Size Sanitary Threaded mm in. Ends Ends Unlined Neoprene or Rubber Lined Glass Lined Polypropelene Kynar or Tefzel Lined 15 1 / / / / / * Courtesy of The Foxboro Co.

306 1350 Control Valve Selection and Sizing Spindle Bonnet bolts & nuts Adapter bonnet Compressor pin Compressor Finger plate FIG. 6.22f Teflon-lined Saunders valve. Body Diaphragm FIG. 6.22d The main components of a weir-type Saunders valve. diaphragm. The compressor is connected to the valve stem, which is moved by the actuator. The diaphragm, which at its center is attached to the compressor, is pulled away from the weir when the compressor is lifted. For high-vacuum service it is often desirable to evacuate the bonnet of the Saunders valve in order to reduce the force that is pulling the diaphragm away from the compressor. This is especially desirable for large valves, where the vacuum might be sufficient to tear the diaphragm from the compressor. The compressor is designed to clear the finger plate, or diaphragm support plate, and to contour the diaphragm so that it matches the weir (Figure 6.22d). The purpose of the finger plate is to support the diaphragm when the compressor has been withdrawn. The finger plate is utilized for valve sizes 1 in. (25 mm) and larger. For valves larger than 2 in. (50 mm), the finger plate is built as part of the bonnet. A Saunders valve can be considered as a half pinch valve (Section 6.20). The pinch valve operates as if two diaphragms were moved toward or away from each other, whereas the Saunders valve has only one diaphragm and a fixed weir. Because of their design similarity, their flow characteristics (see Figures 6.20v and 6.20w for pinch valves) are also similar, as it was shown above in Figure 6.22b. Figure 6.22e shows the three basic positions of a conventional Saunders valve. Materials of Construction The body of a conventional Saunders valve (Figure 6.22d), because of its simple and smooth interior, lends itself well to lining with plastics, glass, titanium, zirconium, tantalum, and other corrosion-resistant materials (Figure 6.22f). Valve bodies Streamline flow in open position Flow control in throttling position Leak-tightness in closed position FIG. 6.22e The open, throttling, and closed positions of a conventional Saunders valve.

307 6.22 Valve Types: Saunders Diaphragm Valves 1351 Bonnet bolt Body Ball brush Compressor Diaphragm stud Compressor insert Diaphragm FIG. 6.22h Full-bore Saunders valve. FIG. 6.22g The design of a straight-through Saunders valve. are available in iron, stainless and cast steels, alloys, and plastics. Iron bodies are lined with plastic, glass, special metals, and ceramics. As it was shown in Table 6.22c, lining lowers the valve capacity of smaller Saunders valves (under 2 in., or 50 mm) by about 25% below that of the unlined ones (Table 6.22c). The diaphragm for the conventional Saunders valve is available in a wide range of materials. These include polyethylene, Tygon, white nail rubber, gum rubber, hycar, natural rubber, neoprene, hypalon, black butyl, KEL-F, and Teflon, with various backings, including silicone. Some of these diaphragms also contain reinforcement fibers. Maintenance requirements of Saunders valves are mainly determined by the diaphragm life, which is a function of the diaphragm s resistance to the controlled process fluid (which may be corrosive or erosive) and also of the operating pressure and temperature (Figure 6.22a). Straight-Through Design The valve seat of the straight-through diaphragm valve is not the conventional weir. Here the compressor is contoured to meet the walls of the body itself (Figure 6.22g). The longer stem stroke of the straight-through valve necessitates a very flexible diaphragm. The increased flexure requirement tends to shorten the life of the diaphragm, but the valve s smooth, self-draining, straight-through flow pattern makes it applicable for hard-to-handle materials, such as slurry. Springs Inner compressor Outer compressor 75% open 10% open Dual-range design Conventional design FIG. 6.22i The shape of the openings of a 10 and 75% open Saunders valves are compared in the dual-range (left) and single-range (right) designs.

308 1352 Control Valve Selection and Sizing The flow characteristic of the straight-through design is more nearly linear than those of the conventional Saunders valves. Full Bore Valve The body of a full-bore Saunders valve is modified to provide a special shape to the weir. As a result, the opening of the internal flow path is fully rounded at all points, permitting ball brush cleaning (Figure 6.22h). This is an important feature in the food industry, where a smooth, easy-to-clean interior surface is required. Dual-Range Design The rangeability and flow characteristics of a conventional Saunders valve are rather poor, and so it is not suitable if high precision control is required. The flow characteristics of the dual-range design is an improvement, in comparison to the characteristics of the conventional Saunders. The dual-range valve contains two compressors, which provide independent control over two areas of the diaphragm (Figure 6.22i). The first increments of stem travel raise only the inner compressor from the weir. This allows flow through a contoured opening in the center of the valve. This is superior to the operation of the single-range design, where the corresponding flow is the result of a slit across the entire weir. This improvement in the shape of the value opening helps prevent clogging and the dewatering of stock and it also keeps abrasion at a minimum. In this dual-range design, while springs hold the outer compressor firmly seated, the inner compressor may be positioned independently to provide accurate control over small amounts of flow. When the inner compressor is opened to its limit, the outer compressor begins to open. From this point on, both compressors move as a unit. When wide open, this valve provides the same flow capacity as its conventional counterpart. Bibliography Bialkowski, B., Coughran, M., and Beall, J., Control Valve Performance Update The Last 10 Years, Pulp Pap-Canada, (102):21-22, November Bishop, T., Chapeaux, M., Jaffer, L., et.al., Ease Control Valve Selection, Chem Eng Prog, 98 (11): 52 56, November Borden, G. and Friedmann, P. G. (eds.), Control Valves Practical Guides for Measurement and Control, Research Triangle Park, NC: Instrumentation, Systems, and Automation Society, Cain, F. M., Solving the Problem of Cavitation in Control Valves, Paper # , 1991 ISA Conference, Anaheim, CA, October Carey, J. A., Control Valve Update, Instruments and Control Systems, January Control Valve Handbook, 3rd edition, Marshalltown, IA: Fisher Controls International Inc., Control Valve Seat Leakage, FCI , Cleveland, OH: Fluid Controls Institute, Inc., 2003, ( Control Valves Globe, Plug, Pinch, Needle, Gate, Measurements and Control, February Cunningham, E. R., Solutions to Valve Operating Problems, Plant Engineering, September 4, Fernbaugh, A., Control Valves: A Decade of Change, Instruments and Control Systems, January Monsen, J. F., Valve Wars Rising Stem vs. Rotary, Plant Services, January Rahmeyer, W., The Critical Flow Limit and Pressure Recovery Factor for Flow Control, InTech, November Sanderson, R. C., Elastomer Coatings: Hope for Cavitation Resistance, InTech, April Whatever Happened To... A Guide to Industrial Valve and Actuator Companies Past and Present, Washington, D.C.: Valve Manufacturers Association, 1997.

309 6.23 Valve Types: Sliding Gate Valves C. S. BEARD (1970, 1985) B. G. LIPTÁK (1995, 2005) Flow sheet symbol Types: A. Knife gate B. V-insert C. Plate and disc (multi-orifice) D. Positioned disc Sizes: A. On/off; 2 to 120 in. (50 mm to 3 m) B. Throttling: 1 / 2 to 24 in. (12 to 600 mm) C. Throttling: 1 / 2 to 6 in. (12 to 150 mm) D. Throttling: 1 to 2 in. (25 and 50 mm) Design Pressures: Design Temperatures: A and B. Up to ANSI Class 150; higher with wedge within wedge design C. Up to ANSI Class 300 D. Up to 10,000 PSIG (69 MPa) A and B. Cryogenic to 500 F (260 C) C. 20 to 1125 F ( 29 to 607 C) Rangeability: A. 10:1 B. 20:1 C. Up to 50:1 is claimed; see Section 6.7 Characteristics: See Figure 6.23a Capacity: A. C v = 45 d 2 B. C v = 30 d 2 C. C v = (6 to 10) d 2 See Table 6.23b Leakage: Materials of Construction: Costs: Partial List of Suppliers: A and B. ANSI Class I or II with metal seat; better with soft seat or lining C. ANSI Class IV; see Table 6.1gg for definitions A and B. Ductile iron, cast iron, carbon steel, 304, 316, 317 stainless steel, Alloy 20, Hastelloy B or C. Seating can be metal to metal, nylon, or RTFE C. Body: Ductile iron, bronze, carbon steel, stainless steel, aluminum, Monel Trim: Stainless steel is standard, which can be chrome-plated for hardness or Tefloncoated for corrosion resistance; Monel or Hastelloy trims are also available The cost of V-insert-type slide gate valves is similar to, but generally less than that of, single-seated globe valves, which are given in Figure 6.19b. The cost of plate and disc valves is given in Table 8.23j. Anchor/Darling Valve Co. ( DeZurik/SPX Valves ( ITT Industries, Engineered Valves ( Jordan Valve ( Kurimoto Valves ( Red Valve Company Inc. ( Richards Industries ( Stockham Valves & Fittings ( Zimmermann & Jansen Inc. ( 1353

310 1354 Control Valve Selection and Sizing INTRODUCTION The knife gate-type slide gate valves are relatively inexpensive, have high capacity, and are suited for slurry and dirty services. On the other hand they have poor control characteristics, do not provide tight shut-off, and are not suited for corrosive services. The V-insert type variation of this design has similar features, but as illustrated in Figure 6.23a, has better control characteristics. The positioned sliding disc designs are ideal for highpressure (up to 10,000 PSIG), cavitating, abrasive, or erosive services, but are relatively expensive and are not suited for sludge, slurry, viscous, or fibrous services. The multiport plate and disc type valves are similar, but provide superior control characteristics. These valves are available as pump governors or as unusually high rangeability (> 200:1) control valves in sizes from 0.5 to 6 in. SLIDING GATE VALVE DESIGNS Knife Gate Valves Changing the process fluid s flow rate by sliding a plate past a stationary hole is one of the oldest and most basic approaches to throttling flows. The most common valve, the sliding gate valve, operates like this. Although occasionally used for automatic control, it is not considered to be a throttling control valve. It is a form of guillotine -type gate valve (Figure 6.23c) and is much used in the pulp and paper industry due to its shearing ability and nonplugging body design. The slab-type sliding gate is provided with a round opening (Figure 6.23d) and therefore its characteristics are determined by the two converging circles. Its characteristics 100 TABLE 6.23b Valve Capacity Coefficients of Open Gate, V-Insert, or Plateand-Disc-Type Slide Gate Valves Size in. (mm) Plate and Disc Valve Design Open Knife Gate V-Insert Type 0.5 (12.5) to (25.4) to (38) 9.0 to (51) 17.4 to (75) (100) (150) (200) (250) (300) (350) (400) (450) 11, (500) 14, (600) 20,550 11, (750) 31,900 17, (900) 45,700 26,200 approximates equal-percentage behavior up to about 70% of its flow capacity, and above that it becomes nearly linear. The flow rate of 70% is reached by opening the valve to about 30% of its stroke. On critical services, such as in catalytic cracking, reforming, isomerization, or coal gasification applications, the double gate valves are often considered (Figure 6.23e). In this % Flow or C v Positioned disc V-insert Plate & disc Inspection window % Lift or rotation FIG. 6.23a The characteristics of the various types of sliding gate-type valve designs. FIG. 6.23c The guillotine-type sliding gate valve.

311 6.23 Valve Types: Sliding Gate Valves 1355 FIG. 6.23f Sliding gate valve with V-insert. FIG. 6.23d The design of a slab-type sliding gate valve. design a wedge is provided within the wedge-shaped sliding gate. This inner wedge forces the two sliding gates on its two sides against the two seats, thereby guaranteeing tight closure. V-Insert Type The addition of a V-shaped insert (Figure 6.23f) in the valve opening creates a parabolic flow characteristic. As shown in Figure 6.23a, this characteristic is somewhat similar to that of the V-ported globe valve. The performance of these valves is much dependent on the type of actuator and positioner used, because the quality of control is dependent upon the ability to provide very accurate positioning of the sliding gate. Positioned-Disc Valves Rotation of a movable disc with two holes, which if rotated can progressively cover two holes in the stationary disc, can successfully throttle flow (Figure 6.23g). This variable choke was designed to control flow from high-pressure oil wells. The use of ceramic or tungsten carbide discs allows it to handle pressures up to 10,000 PSIG (69 MPa). Such valves are presently furnished in 1 and 2 in. (25 and 50 mm) sizes with areas from 0.05 in. 2 with 0.25 in. hole (32 mm 2 with 6.3 mm hole) to 1.56 in. 2 with two 1 in. holes (1006 mm 2 with two 25 mm holes). An angle version of this valve design is used for proportioning control, with an actuator capable of controlling the discharge flow at quarter-turn movement. Both linear and rotary type actuators can be used. The valve opening (the relationship between the discs) remains in the last position if power fails. A stepping actuator (Figure 6.23h) positions the inner valve disc in 1 increments as a function of a pneumatic controller input to a double-acting, spring-centered piston. Rotation occurs through a rack and pinion assembly. Limit Fully open Throttling Fully closed FIG. 6.23e The design of the double gate valve, which is also called the wedge in the wedge design. (Courtesy of Zimmermann & Jansen Inc.) FIG. 6.23g High-pressure process streams can be throttled by the positioneddisc-type slide gate valve.

312 1356 Control Valve Selection and Sizing FIG. 6.23h The actuator of a positioned-disc-type angle valve. (Courtesy of Willis Oil Tool Co.) switches are provided, and a stepping switch can be used for position transmission and for position feedback in automatic control systems. Plate and Disc Valves Control signal A wide variety of flow characteristics are available using a stationary plate in the valve body and a disc that is moved Plate Disc Cap Pressure ring Locknut Body Lower rod assembly by the valve stem. The plate (Figure 6.23i) is readily replaceable by removing a flanged portion of the body, which retains the plate with a pressure ring. Areas of the plate are undercut to reduce friction. A circumferential groove provides flexibility and allows the plate to remain flat in spite of differential pressures or expansion or contraction of the body. The stem contacts the disc by a pin through a slot in the plate. The disc is held in contact with the plate by upstream pressure and by retaining guides. The contacting surfaces of the disc and plate are lapped to light band flatness. The chrome-plated surface of the stainless steel plate has a hardness comparable to 740 Brinell to resist galling and corrosion and obtain smooth movement of the disc. The material, with the registered name of Jordanite, is reported to have an extremely low coefficient of friction, is applicable to high-pressure drops, and has great resistance to heat and corrosion. Flow occurs through mating slots in the disc and the plate. Positive shut-off occurs when the slots are separated (Figure 6.23i). Flow increases on an approximately linear relationship until the slots are lined up for maximum flow. Capacities are about C v = 6.5 d 2 through the 2 in. (50 mm) size and about C v = 12 d 2 through the 6 in. (150 mm) size. Stem travels to obtain full flow are very short due to the slot relationship, and low-lift diaphragm actuators can be used for positioning. Forces needed for positioning are low, requiring only sufficient power to overcome friction between the plate and disc, which is right-angle motion and not opposed to the direction of flow. Valve bodies are offered in sizes between 1 / 4 and 6 in. (6.3 and 150 mm) and with ratings through 300 PSIG (2.1 MPa), depending on the material, with a selection of trims and packings. Many styles of actuators are used, including one with a thermal unit and cam actuation. This body design has been adapted for extensive use in self-contained pressure or temperature regulators. This valve is also used to control the steam flow to steamdriven pumps, so as to maintain the pump discharge pressure constant. The sizes and costs of these pump governors are given Table 6.23j. TABLE 6.23j Multiple-Orifice, Plate- and Disc-Type Pump Governor Steam Valve Costs* Valve Size (Inches) Carbon Steel Stainless Steel Open Closed FIG. 6.23i This throttling control valve provides wide rangeability by the use of a plate-and-disc inner valve. (Courtesy of Jordan Valve, a division of Richards Industries.) 0.5 $2,000 $2, $2,300 $3, $2,700 $3, $3,100 $4,000 * All valves are provided with 150# RF connections (Courtesy of Jordan Valve.)