Kelvin Chiddick (Technical Consultant to Whitmore Rail), Brad Kerchof and Kevin Conn (Norfolk Southern Corporation)

Transcription

1 Considerations in Choosing a Top-of-Rail (TOR) Material Kelvin Chiddick (Technical Consultant to Whitmore Rail), Brad Kerchof and Kevin Conn (Norfolk Southern Corporation) Introduction Controlling the friction level at the wheel/rail interface by using a top-of-rail (TOR) material has become an accepted track maintenance strategy throughout the railroad industry. With the increasing number of materials available to use, it is understood that field testing has become an important means to assess the comparative effectiveness of these materials This paper will describe the material performance characteristics selected by Norfolk Southern for testing and the results of those tests. Topics that will be discussed include; TOR Benefits 1. TOR benefits, including reduced rail wear, rolling contact fatigue and fuel consumption. 2. The capability of TOR materials to reduce lateral curving forces. 3. The impact of TOR materials on vehicle steering, locomotive adhesion, rolling contact fatigue and track circuit shunting. The use of TOR materials to reduce lateral curving forces offers many benefits as shown in table 1. Of these benefits, reduced rail wear, rolling contact fatigue and fuel consumption will be discussed. TOR Impact Reduced track gage widening Reduced fastener/tie failure Reduced rail & wheel wear Reduced rolling contact fatigue Reduced rail rollover derailments Reduced train resistance Table 1 Potential Benefit Reduced track gaging maintenance Increased fastener/tie life Increased rail and wheel life Reduced rail grinding Reduced service interruptions Reduced fuel consumption Rail Wear By changing friction, TOR materials can reduce the low rail wear in curves by reducing the stick-slip at the wheel/rail interface. This phenomenon occurs because wheels on the high rail of a curve must travel a longer distance than wheels on the low rail. To compensate for this difference, one of the wheels must slip to stay on rotation with the other wheel. This slip event occurs when the resultant of longitudinal and lateral forces exceeds the frictional force that can develop between the wheel and rail. A TOR friction material reduces the coefficient of friction which reduces the friction force and results in lower longitudinal and lateral forces. This reduces rolling contact fatigue (RCF) damage and ultimately, rail wear. Figure 2 shows the results of a 4-year study on the impact of TOR on rail wear rate. Testing was done on a line that had 20 miles of new similar metallurgy rail installed. The line was divided into a Gage Face (GF) Only zone and a Gage Face (GF) + Top-of-Rail (TOR) zone. Within each zone, curves were chosen with similar curvature, direction and super-elevation. Rail profiles were taken before and after rail grinding and the tonnage monitored for each curve. Figure 2 shows the total area loss (due to both wheel wear and grinding) of the low rails as a function of degree of curvature. Each point represents the average wear rate for a particular curve over the four year period. Regressed lines have been drawn through the points for the GF Only and GF+TOR conditions, showing the impact that TOR has on low rail wear rate. AREMA

2 Area Lost (Square Inches/MGT) Condition GF Only GF + TOR Degree of Curvature Figure 2 Rolling Contact Fatigue (RCF) Cracks in the rail surface are a type of RCF that is caused by longitudinal and lateral wheel creep forces that exceed the strength of the rail steel. They are common in curves, where creep forces are higher. Cracks weaken the running surface of the rail and can lead to spalls, shells and transverse defects. Cracks are removed by rail grinding. Because TOR materials reduce creep forces, they can reduce the rate of crack formation and thus the amount of grinding required. Figure 3 shows the difference between two running surfaces of two similar curves, with and without TOR. As can be seen, the curve with TOR applied has cracks that are smaller and fewer in number. Figure 3 AREMA

3 Fuel Consumption Total train resistance is the sum of wheel/rail rolling resistance, the resistance provided by curves, roller bearing resistance and grade. TOR can affect both rolling and curve resistance because it controls the coefficient of friction between the wheel and rail. By reducing rolling and curve resistance, TOR can impact the energy required to move a train. It is worth emphasizing that fuel savings can only be realized when the train is under power and not braking. This suggests that TOR fuel savings are route specific and that applying the TOR-generated fuel savings achieved on one route to the entire network may not be an accurate prediction of savings. As an example (see figure 4), consider a train that uses a total of 2500 gallons of fuel, 500 of which are used when the train is operating under dynamic or air braking. Of the 2000 gallons used while the train is under power, 1500 are used to overcome the grade resistance, 50 to overcome bearing resistance and 450 to overcome rolling and curve resistance. To determine the amount of fuel that might be saved by using TOR, the fuel savings rate must be applied to the gallons used to overcome rolling and curve resistance while the train is under power. In this example, a 5% TOR savings rate would be applied to 450 gallons, yielding a savings of 22.5 gallons. Figure 4 Figure 5 shows the results of a TOR fuel study. Fuel consumption was monitored for trains operating over a 140 mile route for two weeks one week without TOR and one week with TOR applied (gage face lubrication was a constant for both weeks). Locomotive throttle notch dwell times and fuel burn rates (specific to the model of locomotive) were used as a proxy for fuel consumption (gallons). Train tonnage and distance traveled were used to normalize the data (i.e. ton-miles). Results showed that TOR reduced fuel consumption by 7%. AREMA

4 Figure 5 Lateral Curving Force Reduction Consider a four-wheel drive (4WD) vehicle that locks in the front axle, making it similar to a solid axle wheelset. When making a turn in 4WD on dry pavement, the vehicle feels like it fights the turn and tries to push it further out further in the turn. Making the same turn on wet pavement, the vehicle feels easier to turn and does not fight the turn. This difference is caused by the lower coefficient of friction of the wet pavement compared to the dry, allowing the wheels to slip sooner than they would on dry pavement. A railway wheelset is similar to the locked front axle of a 4WD vehicle insofar as its performance on a curve and reacts similarly to a change in friction coefficient. A TOR friction material reduces the coefficient of friction which reduces both longitudinal and lateral forces, allowing the wheel to slip more easily. The degree of the force reduction depends on a number of factors. These are the carry distance, the application rate, the conditioning effect and train handling characteristics, which could disrupt the material on the wheel or rail (e.g. tread air brake application, sanding, etc.). Carry distance refers to how far the TOR material is effective in reducing lateral forces a desired amount (e.g. 35%). This reduction is determined by measuring the difference in lead axle lateral forces of similar cars in an instrumented force curve with and without TOR. Figure 6 shows the carry distance for a TOR material to be 2 miles while maintaining a 35% force reduction. AREMA

5 . Figure 6 Application rate is the amount of material used to get a certain force reduction at a particular distance. This rate is typically determined in volume per 1000 axles and requires an accurate measurement of the volume used and the number of axles passed. Each TOR material has an optimum application rate which if, too low, produces no force reduction effect or if too high, leads to material waste. As shown in figure 7, carry distance is increased when the application rate is increased from gallons/1000 axles to gallons/1000 axles. However, when increased to gallons/1000 axles, no further carry distance is achieved. Figure 7 Some TOR materials can have a conditioning effect in that the wheels and/or the rail have to become saturated before achieving their maximum carry distance. This is due to the third body layer, which controls friction, not being fully established between the wheel and the rail. The result of this is that the spacing between the first and second applicator may have to be reduced before going to the maximum applicator spacing. Consider the example in figure 8, which shows the carry distance for one applicator versus two, spaced 1 mile apart. Carry distance improves from 0.5 to 2.0 miles. AREMA

6 Figure 8 Because TOR materials are applied to the wheel and rail surfaces, anything that could wear the material off faster would reduce carry distance. Examples would be heavy locomotive sanding on ascending grades or sustained tread air braking on descending grades. To compensate for this, applicator spacing may have to be decreased and application rate increased in these areas to maintain the required level of force reduction. Figure 9 shows the effect of sustained air brake application on carry distance and application rate. Figure 9 With these parameters, a metric (dollars/axle/mile) for each TOR material can be calculated to determine what material makes the most economic sense. AREMA

7 Vehicle Steering How a wheelset steers in a curve depends on the force moments created at the wheel/rail interface. The direction and magnitude of these moments (M) are a function of the wheel rolling radius difference and the friction at the contact point. As shown in figure 10, the direction of the moment can reduce the angle of attack (AoA) of the wheelset (+M), not change it (M=0) or increase it (-M). Figure 10 The magnitude of these moments is shown in figure 11 as a function of friction level (µ). Lowering the friction reduces the anti-steering moment of the trail wheelset (decreases the angle of attack) and reduces friction forces/gage widening forces of the lead wheelset. This is the fundamental concept of TOR to reducing forces. Figure 11 However, if the friction level is allowed to go too low, the lead wheelset flanging force will no longer be sufficient to turn it. This results in the wheels sliding instead of rolling and increases of the angle of attack. As a result, this indicates that the friction level at the wheel/rail interface should be managed or controlled within a range to maintain proper steering. Low friction levels can also cause problems with locomotive adhesion, which will be discussed later. AREMA

8 Therefore, to manage or control the friction at a level that reduces lateral curving force and maintains sufficient wheelset steering, a TOR material should have the following properties. 1. Maintain a coefficient of friction low enough to reduce lateral curving forces 2. Maintain a coefficient of friction high enough to maintain wheelset steering and locomotive adhesion 3. Exhibit positive friction characteristics Most materials in nature exhibit negative friction. For example, consider a clean steel wheel on a clean steel rail. When rolling, the coefficient of friction is 0.50 but when sliding, it reduces to 0.26 (see figure 12). Because the friction decreases as the creep increases, the coefficient of friction/creep slope shown in the graph is negative and these materials are said to have negative friction. Compare this to a lubricant in a flooded condition which has a constant coefficient of friction and is shown as a horizontal line on the graph. Figure 12 Once slip occurs, the only way to reduce the slip between the two sliding steel surfaces is to increase the friction. This can be done by using a friction modifier at the wheel/rail interface. In contrast to a lubricant, a friction modifier is a combination of a solid lubricant and friction modifiers, which acts as a lubricant when the wheel is rolling, but causes an increase in friction when sliding occurs. A friction modifier allows the wheel to go from sliding back to rolling more rapidly, reducing the stick-slip phenomenon. Because the friction increases as the creep increases, a friction modifier is said to have positive friction. The relationship between coefficient of friction and creep is shown with a positive slope in the graph. It is possible to control the coefficient of friction with a lubricant by using a non-flooded condition. However, this requires careful control of the amount applied and cannot increase the friction when slip occurs, putting the wheel back into rolling (i.e. positive friction). AREMA

9 Locomotive Adhesion Because locomotive adhesion is a function of the friction level at the point of wheel-rail contact, the potential of a TOR material to reduce friction to an unacceptable level should be considered. Available tractive effort of a dieselelectric locomotive is a function of its weight, speed, throttle notch setting, friction at the wheel-rail interface and wheel slip control system. Loss of adhesion or wheel slip occurs when the total resistive forces of the train (e.g. grade, rolling, curving, etc.) exceed the available tractive effort that the locomotive can produce. Therefore, these factors must be taken into account when testing. Figure 13 Figure 13 shows the relationship between speed and tractive effort for a particular throttle-notch setting for a dieselelectric locomotive. In the area under the curve, wheels maintain adhesion due to the tractive effort being greater than the resistive force of the train. Above the curve, locomotive wheels slip due to the tractive effort being less than the resistive force of the train. In order to properly evaluate the impact that a material has on adhesion, testing should be performed at a point on the curve, not above or below it. If performed below the curve, adhesion loss may be masked if the available tractive effort is much greater than the resistive force of the train. Additionally, to keep the tractive effort constant, testing should be done at a constant speed. A practical way to do this is; 1. Perform the testing on a constant ascending grade. 2. Operating the locomotive at a fixed throttle notch setting. 3. Adjust the trailing tonnage of the train to the point to where it almost stalls. This operating plan results in the grade becoming the largest train resistive force, helps keep the speed constant, and allows the test to be performed on a point on the locomotive tractive effort-speed curve. AREMA

10 Once the appropriate tonnage and speed are determined, runs with and without with TOR material being applied can be made. Assessment of the impact of TOR material on adhesion can then be evaluated by monitoring tractive effort (if available from the locomotive) or drawbar horsepower (through the use of an instrumented coupler). Of particular importance is to observe the tractive effort or drawbar horsepower at the TOR applicator location, which is the point where wheel slip is most likely. Because wheel slip control systems can vary among locomotives, the model chosen to test with should be representative of the locomotive fleet. An example of a TOR material adhesion test is shown in figure 14. Rolling Contact Fatigue Development (RCF) Figure 14 Rolling contact fatigue begins with hairline cracks in the surface of the rail head and can develop to include more serious conditions including deeper cracks, spalls, shells, corrugations and plastic flow. RCF is the result of the small contact area between the rail and the wheels of heavy axle load (HAL) cars, experiencing internal shearing stresses greater than the strength of the rail. Left unaddressed, RCF can develop into rail defects and failures. While a benefit of TOR is reduction of curving forces and henceforth RCF growth, an opposing concern is that incompressible liquids in the TOR material might induce hydraulic cracking and increase RCF growth. This phenomenon is well known with water and has been observed with gage face and TOR systems that were mistakenly installed in the spiral of a curve instead of in a tangent. Hydraulic cracking or hydro-pressurization occurs when an incompressible fluid, such as water or oil, is trapped in a crack and becomes pressurized when the wheels roll over it. This cyclic pressurization can cause a crack to propagate and grow faster. The degree to which this occurs is also a function of rail metallurgy and the amount and viscosity of the incompressible fluid(s) in the TOR material. Considering the potential consequences, this effect should be investigated before any large scale implementation of a TOR material is made. There are two ways to look at this, in the laboratory and in the field. In the laboratory, the University of Sheffield U.K. has developed a SUROS (Sheffield University ROlling Sliding) twin-disc machine that can simulate the wheel/rail contact. As shown in figure 15, this machine rotates two metal discs relative to one another and offers the ability to control the pressure, slip, and speed of the discs. Additionally, discs can be made from steels that are representative of wheel and rail metallurgies and a TOR material can be applied to alter the coefficient of friction at the wheel/rail interface. AREMA

11 Figure 15 Once tested, discs are sectioned, polished and photomicrographed (see figure 16) to determine the impact of TOR materials on crack initiation and RCF. Figure 16 In the field, testing should compare two areas that have curves of similar direction, curvature, elevation and tonnage with the only variable being TOR applied or not. As shown in fugure 17, dye penetrant testing can then be used to make qualitative assessment of the RCF surface conditon. AREMA

12 Figure 17 Other testing options are availabe to make semi-quantitative measurements of crack depth, such as the eddy currentbased Draisine system, which allow crack depth characertization of an entire curve as opposed to a couple of points. Shown in figure 18 is the device along an example of its output in figure 19. Figure 18 AREMA

13 Crack Depth Relative Location Figure 19 The Draisine, with its eddy current technology, has been shown to be effective in measuring crack length in rails. To determine crack depth, the Draisine combines the mesured crack length with an assumed crack angles. If the assumed angle is close to the actual angles, the correlation between Draisine-measured crack depth and actual crack depth is good. Unfortunately, there is no way to determine the actual crack angle without destructive testing. Train Signaling Systems Railroad signaling systems detect the presence of a train by its wheels shunting a track circuit, creating an electrical short circuit between the two rails. This detection system is used to control train movements and activate warning devices at highway-rail grade crossings, as well as to detect broken rails. Loss of shunt occurs when the wheels do not complete an electrical circuit between opposite rails. Because TOR materials are applied to the running surface of the rail, the impact that this could have on the signaling system must be investigated. While shunt performance depends on many factors, the key factors are; 1. The pressure at the wheel/rail contact point. 2. The wheel/rail profile in regards to contact point location. 3. The length of the cars or number of wheels inside the track circuit that can shunt the rail. In view of these factors, the worst case testing senario would be for low axle weight cars that are long and are on tangent track. Figure 20 shows a setup used by Transportation Technology Center, Inc. (TTCI) to evaluate track circuit shunt performance that would also be applicable to the testing of TOR materials. Test components include a track shunt circuit, a data collection system to monitoring shunt circuit performance, wheel sensors to determine train presence, direction and speed and an AEI reader to identify car types. AREMA

14 Figure 20 Shown in figure 21 is data collected for a train exhibiting poor shunt performance. The increase in island and relay voltage indicate loss of shunt problems. AREMA

15 Figure 21 In regards to using this procedure to test TOR materials, the TOR applicator should be close to the test circuit and application rates should be the highest that are recommended for that material. Conclusion TOR materials reduce longitudinal and lateral curving forces which result in reduced rail wear, rail grinding and fuel consumption. When selecting TOR materials to use, not only their cost but the carry distance, application rate and factors that affect their effectivness like sanding and air brake application should be considered. Additionally, their impact on vehicle steering, locomotive adhesion, rolling contact fatigue and track circuit shunt should also be investigated. This will result in choosing the best material to use. AREMA

16 Choosing a Top-of-Rail (TOR) Material Kelvin Chiddick (Consultant to Whitmore Rail) Brad Kerchof (Norfolk Southern Corporation) Kevin Conn (Norfolk Southern Corporation) AREMA

17 Why TOR? Because of its Benefits Reduced gage spreading of track Reduced fastener/spike failure Reduced curve rail & wheel wear Reduced spalling & contact fatigue Reduced rail rollover derailments Reduced train resistance White, WV Example of Rail Wear Benefits 20 miles of new similar metallurgy rail installed between Old Joe and White, WV (July 2009). Area was divided into a GF+TOR and GF only zone, with a dry down zone in between. Curves with similar curvature, direction and elevation were chosen in each zone. Rail profiles were taken before and after grinding to determine wear. Tonnage monitored to determine wear rate (wear/mgt). GF + TOR zone Dry down zone GF only zone Old Joe, WV 4-Year (170 MGT) Low Rail Results st (inches 2 /MGT) Head Area Los R&L 5-6 L 5-6 R 6 R 12 R Degree of Curvature GF+TOR GF Only Head Area Lost 4-Year (170 MGT) High Rail Results t (inches 2 /MGT) R&L 5-6 L 5-6 R 6 R 12 R Degree of Curvature GF+TOR GF Only Curvature vs. Low Rail Wear Rate Degree of Curvature ss (inches 2 /MGT) Head Area Los GF Only GF + TOR Danville, KY Example of Fuel Savings Benefits CNO&TP District 75 MGT line, mph track. Danville, KY and Oakdale, TN (140 miles). 100% GF lubrication, 80% TOR friction control. One week with GF lubrication only (149 trains), one week with GF lubrication and TOR friction control (181 trains). Locomotive throttle notch dwell times and fuel burn rates used to determine fuel consumption in gallons. Train tonnage and distance travelled used to determine ton-miles. Metric compared was gallons per 1000 ton-miles. Oakdale, TN AREMA

18 Result of Fuel Savings Benefit 0 ton miles gallons./ GF Only GF + TOR Gallons used per 1000 ton-miles decreased by 7% with GF+TOR lion Ton Miles Gals. Saved/Mil Fuel Savings Prediction Model Non Braking Curve Density (%) Study substantiates other TOR fuel savings results. Fuel savings can only BC Rail be realized when throttle Best Fit is in power (not in NS dynamic or air brakes applied). Fuel savings are direction and route specific. Why Do These Benefits Occur? TOR changes the friction at the wheel/rail interface, reducing the force of the slip-stick phenomenon. Slip-stick occurs in curves because high-side wheels must travel a longer distance than low-side wheels. To compensate, one wheel must slip to stay on rotation with the other. This slip occurs when the longitudinal & lateral forces applied to the wheel exceed the friction force holding the wheel. Reducing the COF decreases the longitudinal and lateral forces that build up, imparting less energy into the track. Why Do These Benefits Occur? In regards to rail wear, this decreases the formation of RCF which decreases the amount of grinding required, increasing rail life. In regards to fuel consumption, this decreases the curving and rolling resistance of the train, saving fuel. What Hinders the Use of TOR? The largest operating cost is the material. To reduce this cost (dollars/axle-mile), one can use a material that costs less per gallon carries a further distance before having to be re-applied requires less material to be applied while maintaining the same level of lateral force reduction. How would this be evaluated? TOR Material cost Carry Distance Carry distance is how far a material is effective in reducing lateral forces to a desired level. It is done by Locating an applicator a certain distance from an instrumented curve. Measuring the difference in lead axles forces of similar cars in that curve, with and without TOR. It should be looked at under different scenarios which could change the carry distance, for example When air brakes are not applied When air brake are applied Under heavy sanding AREMA

19 Carry Distance Test Setup nce Number Test Seque Milepost TOR Applicator Force Site 1 Force Site 2 e Reduction % Force Carry Distance Results material A material B Miles from Applicator to Instrumented Curve e Reduction % Force Air Brake Carry Distance Results Brakes Not Applied Brakes Applied* * Higher application rate required Miles from Applicator to Instrumented Curve Conditioning Effect Some materials have a conditioning effect in that the wheels/rail have to become saturated before achieving their maximum carry distance. Analogy is similar to putting paint on a paint roller. This is due to the third body layer between the wheel and the rail, which controls the friction, not fully being established. The result is that the spacing between the first and second applicator may have to be reduce before achieving the maximum applicator spacing. 0.5x 1.0x 1.0x TOR 1 TOR 2 TOR 3 TOR 4 Conditioning Effect Results One applicator Two applicators, spaced 1 mile 5 apart Miles from Applicator to Instrumented Curve Reduction % Force Application rate is the amount of material used to obtain a certain force reduction at a particular distance. If too low, no force reduction will occur. If too high, waste and no increase in force reduction will occur. Determination requires an accurate measurement of the volume or weight of the material used and the number of axle counts (i.e. gals./1000 axles). Application Rate AREMA

20 Application Rate Results e Reduction % Force Miles from Applicator to Instrumented Curve gals gals gals. Rates (gals./1000 axles) Is It Only About Force Reduction? Not from NS s perspective. Additional material considerations should include Its friction level and relationship to creep Its impact on Locomotive adhesion RCF formation due to hydraulic cracking Signal compatibility Train signal system (loss of shunt) Rail defect testing (attenuation of signal) Friction Level TOR reduces forces by reducing friction. If level is too low Wheels will not maintain sufficient traction, causing them to slide instead of roll at an increased angle of attack (AoA). Locomotive adhesion may be compromised. Therefore, friction i levels l should be managed or controlled so that they are Low enough to reduce lateral curving forces. High enough to maintain wheel set steering and locomotive adhesion. They also should have positive frictional-creep characteristics. nt of Friction Coefficien Positive & Negative Friction Positive friction Negative friction Percent Creep Lubricant (flooded) Steel on Steel Friction Modifier Negative friction is when the friction decreases as the slip increases. Positive friction is when the friction increases as the slip increases. Once slip occurs, the only way to reduce it is to increase the friction. Effort Tractive E Locomotive Adhesion Column0 Locomotive loses adhesion and slips Adhesion loss could be masked if tested in this area Locomotive maintains adhesion and does not slip Speed Testing should be performed at a point on the curve Because adhesion is friction dependent, the potential of a material to impact it should be considered. Loss of adhesion occurs when the total resistive train forces (e.g. grade, rolling, curving, etc.) exceed the available tractive effort. Testing should be performed at a point on the curve, not above or below it. Hydraulic Cracking & RCF Liquids in a material may increase RCF due to hydraulic cracking. This has been seen with water and gage face and TOR systems installed in spirals of curves. This phenomenon occurs when an incompressible fluid is entrapped in a crack and is pressurized when the wheels roll over it. This cycle can lead to increased crack propagation and growth. AREMA

21 Quantifying RCF Signal Compatibility Dye Penetrant SUROS Test Rig University Sheffield UK Draisine Train signals, grade crossings and broken rail circuits rely on wheels to shunt the track for detection. Rail flaw systems use ultrasonics and induction to detect rail defects. Because TOR materials are applied to the running surface of the wheel/rail, their impact should be investigated. Factors to consider are Contact point pressure Wheel/rail profile No. of shunting wheels Loss of Shunt Test Setup Summary Wheel Sensor Island Circuit AEI Reader Data Collection System Wheel Sensor age Volta Time Rl Relay Island TOR friction control provides tangible benefits. These benefits can be increased if the largest operating cost, the material, can be reduced. When selecting a TOR material Not only the cost per gallon but the cost per axle-mile must be considered. Effectiveness to reduce lateral forces should not be the only thing to consider. Material impact on vehicle steering, locomotive adhesion, RCF formation and signaling/rail defect testing systems should also be considered. This will result in choosing the best material for use. TTCI/AAR Used with Permission Questions? Thank You for Your Attention! AREMA