Journal of Applied and Industrial Sciences, 2014, 2 (4): , ISSN: (PRINT),ISSN: (ONLINE)

Transcription

1 Journal of Applied and Industrial Sciences, 214, 2 (4): 188-2, ISSN: (PRINT),ISSN: (ONLINE) Research Article 188 Advanced Control of the Delayed Coking Unit in Khartoum Refinery Tomadir A. I. Hamed*1, Gurashi A. Gasmelseed2, Ibrahim H. Elamin3. 1Department of Chemical Engineering, University of Science and Technology, Khartoum- Sudan Department of Chemical Engineering, University of Science and Technology, Khartoum- Sudan gurashigar@hotmail.com Department of Chemical Engineering, University of Science and Technology, Khartoum- Sudan Ibrahimelamin@hotmail.com (Received: April 4, 214; Accepted: July 31, 214) Abstract :Delayed coking units convert heavy crude oil or topping refinery residue to more light valuable products including diesel, gasoline and naphtha by thermal catalytic treatment. The residue from the fractionation column is to 5 o C in the furnace. The thermally treated residue enters the delayed coking tower where cracking and condensation reactions occur to produce oil gas and coke. A cascade control strategy was developed to control the pressure of the coking drum using the flow of the heater fuel as manipulated variable. The block diagrams of the systems were constructed and the process transfer functions were identified using MATLAB Black Box model. Then the overall transfer functions, the open and closed-loops, and the characteristic equations were determined, and the control systems were tuned to obtain the adjustable parameters using Routh-Hurwitz, Direct Substitution, Root locus, Nyquist, and Bode methods. The adjustable parameters were appropriately inserted into the characteristic equation for the offset investigation, stability analysis and response simulation. It is found that using of PID controller for the Primary loop provides the highest gain than P and PI controllers and also it eliminates the Offset. Indexterms: Delayed coking, Thermal cracking, Advance control. 1. INTRODUCTION Delayed coking is a thermal cracking process used in petroleum refineries to upgrade and convert petroleum residuum (bottoms from atmospheric and vacuum distillation of crude oil) into liquid and gas product streams leaving behind a solid concentrated carbon material, petroleum coke. The delayed coker is the only main process in a modern petroleum refinery that is a batchcontinuous process. The flow through the tube furnace is continuous. The feed stream is switched between two drums. One drum is online filling with coke while the other drum is being steam-stripped, cooled, decoked, pressure checked, and warmed up. The overhead vapors from the coke drums flow to a fractionator, usually called a combination tower. This fractionator tower has a reservoir in the bottom where the fresh feed is combined with condensed product vapors (recycle) to make up the feed to the coker heater [1]. A fired heater with horizontal tubes is used in the process to reach thermal cracking temperatures of 485 to 55 o C (95 to 941 o F). With short residence time in the furnace tubes, coking of the feed material is thereby delayed until it reaches large coking drums downstream of the heater. Three physical structures of petroleum coke: shot, sponge, or needle coke can be produced by delayed coking. These physical structures and chemical properties of the petroleum coke determine the end use of the material which can be burned as fuel, calcined for use in the aluminum, chemical, or steel industries, or gasified to produce steam, electricity, or gas feedstocks for the petrochemicals industry [2]. In Khartoum refinery a delayed coking unit with a capacity of 1 Mt/year was constructed in Phase I and a second delayed coking unit with a capacity of 1 Mt/year was constructed in Phase II where the overall delayed coking capacity reached 2 Mt/year. A set of delayed coking unit with annual capacity of 1Mt was constructed in Phase I, which is composed of 8 sections, these are electrostatic desalting; coking tower; heating furnace; fractionation column; steaming out and venting; rich gas compression; rich gas desulfurization, coke storage, basin-settling, base coking, and cold coking water circulation. The second set of delayed coking unit consists of 6 sections, these are coking tower, heating furnace, fractionation column, rich gas compression, absorption and stabilization, LPG desulfurization and dethioalcoholization. The 3 sections coking tower, heating furnace and fractionation column are matched with the annual capacity of 1Mt; while the electrostatic desalting, rich gas desulfurization, steaming out venting, coke storage basin- settling, basin cutting coking, water circulation and cold coking water circulation match with the annual capacity of 2Mt.

2 Journal of Applied and Industrial Sciences, 214, 2 (4): 188-2, ISSN: (PRINT),ISSN: (ONLINE) Fig. 1: Delayed Coking Unit flow sheet The farction heavier than coker gas oil (CGO) in the feedstock oil flows to the column bottom together with the condensated fraction (called circulation oil) of overhead oil gas from the coking tower. At a temperature of 366ºC, the mixture is pumped to the radiation section of the furnace by heating furnace radiation section feeding pump to be promptly heated up to 5ºC. Then it enters the coking tower via four way valve, where cracking and condensation reactions occurred to produce oil gas and coke. The mixture at high temperature oil gas transferred from coking tower top to the duckbilled type tray in the lower section at the bottom of the fractionation column. Circulation oil fraction is condensated, the rest large amount of oil gas passes through washing plate to enter the vaporization stage where it rises to the distillation stage above the CGO collection pot to separate the cuts such as coking rich gas, gasoline, diesel oil and CGO [3]. One of the most useful concepts in advanced control is cascade control. It is one of the most successful schemes for enhancing the performance of singleloop control [4], A cascade control structure has two feedback controllers with the output of the primary (or master) controller changing the set point of the secondary (or slave) controller. The output of the secondary goes to the valve, as shown in fig. 2. There are two purposes for cascade control: 1. To eliminate the effects of some disturbances 2. To improve the dynamic performance of the control loop Fig. 2: Cascade Control Loop There are many reasons for installing advanced process control (APC) in a delayed coking unit, in order to maximize the throughput in compliance with all given product qualities and security limits. Unit security as well as product yields would be improved. APC monitors all the process variables of the unit and reacts on changes instantly inside the pre-specified limits [5]. The objectives of this study are to: develop a cascade control strategy for the Coking Drum of the Delay Coking Unit, Identify the Transfer Functions of control loops, Stability analysis of the control loops, Controllers tuning, Offset investigation, and response simulation. II. METHODLOGY Standards tests for raw material and products were carried out by ASTM D, the raw material standard tests are shown in table 1 as example.

3 Journal of Applied and Industrial Sciences, 214, 2 (4): 188-2, ISSN: (PRINT),ISSN: (ONLINE) 19 Item Analyzed Item Method Table 1: Raw material, parameters and methods of determination [3] 1 API degree N/A 2 Density (2ºC), g/cm 3 ASTM D Kinematic vis. 5/8ºC, mm 2 /s ASTM D Freezing point ºC ASTM D Flash point (open) ºC N/A 6 Conradson carbon residue, m% ASTM D Water content,m% ASTM D-46 8 Salt content, mgnacl/l ASTM D Acid value, mgkoh/g ASTM D Sulfur content, m% ASTM D Nitrogen content, m% ASTM D Gum, m% ASTM D Asphaltene, m% ASTM D Wax content, m% ASTM D Fe/Ni/Cu/V/Pb/Ca/Mg/Na, ppm ASTM D Characteritic factor N/A 17 Crude type N/A 18 BKP-distillate at 24ºC, m% ASTM D- 116 System Control, Tuning, and Stability analysis A control strategy was developed as shown in figure 3, the block diagrams were constructed, the transfer functions were identified using System Identification Toolbox in MATLAB and the characteristic equations were obtained. These characteristic equations are used for tuning, stability analysis and simulation responses. A system is considered unstable if, after been disturbed by an input change, its output takes off and does not return to the initial state of rest. The stability analysis of a system can be treated in a unified way independently if it is closed or open. The location of the poles of a transfer function gives the first criterion for checking the stability of a system: If the transfer function of a dynamic system has even one pole with positive real part, the system is unstable The criterion of stability for closed loop systems does not require calculation of the actual value of the roots of the characteristic polynomial. It only requires that if any root is to the right of the imaginary axis [6].The Routh-Hurwitz test is a numerical procedure to determine how many roots of a polynomial are in the Right Hand Plane and how many are on the imaginary axis. It doesn't give specific root locations but performing the test is generally far easier than factoring [7]. As we change gain, we notice that the system poles and zeros actually move around in the S-plane. This fact can make life particularly difficult, when we need to solve higher-order equations repeatedly, for each new gain value. The solution to this problem is a technique known as Root-Locus graphs. The root locus analysis is another criterion of stability. The root loci are merely the plots, in the complex plane, of the roots of the characteristic equation as the controller gain is varied from zero to infinity. Bode and Nyquist diagrams of the openloop transfer functions are used to study the stability characteristics of a closed-loop system. A feedback control system is unstable if the amplitude ratio of the corresponding open-loop transfer function is larger than 1 at the crossover frequency, this is known as the Bode stability criterion. The Nyquist stability criterion states that: if the open-loop Nyquist plot of a feedback system encircles the point (-1, ) as the frequency takes any value from - to +, the closed-loop response is unstable [6]. Tuning cascade control systems is more complex than tuning simple feedback systems, if only because there is more than one controller to tune. However, this does not mean it is difficult. Because the inner loop by itself is a simple feedback loop, this controller should be tuned as fast as possibleavoiding instability, of course. The objective is to make the inner loop fast and responsive in order to minimize the effect of upsets on the primary controlled variable. Tuning this system then comes down to tuning the primary controller [7].

4 Journal of Applied and Industrial Sciences, 214, 2 (4): 188-2, ISSN: (PRINT),ISSN: (ONLINE) Fig. 3: Control strategy physical diagram Transfer function Identification: The primary or Master loop: The secondary or slave loop: III. RESULTS AND DISCUSSION The properties of mixed crude oil {light crude (Fula-North-AG)/viscous crude (Fula-North-B) =1:3} in the six districts of Sudan are shown in Table 2. The basic characteristics of the crude oil are high salt and calcium contents, high acidic value, water content, high density, high viscosity and high light component, and low sulfur content. Tables 3.1 through 3.5 show the properties of the oils. Table 2: Properties of Mixed Crude Oil [3] Item Analyzed Item Fula-North-B (viscous crude) Fula-North-AG (light crude) Mixed Crude * 1 API degree Density (2ºC), g/cm Kinematic vis. 5/8ºC, mm 2 /s / /4.716 (5/1ºC) 267/117 (8/1ºC) 4 Freezing point ºC Flash point (open) ºC Conradson carbon residue, m% Water content,m% Salt content, mgnacl/l Acid value, mgkoh/g Sulfur content, m% ppm Nitrogen content, m% ppm 12 Gum, m% Asphaltene, m% Wax content, m% Fe/Ni/Cu/V/Pb/Ca/Mg/Na, 97.9/18.3/1.2/.9/.1/1652. ppm /8.5/ Characteritic factor Crude type Naphthenic- middle base Low sulfur middle base 18 BKP-distillate at 24ºC, m% 8.3 * = Calculated data The main products of the unit are purified dry gas, LPG, stabilized gasoline, coking diesel oil, coker gas oil and petroleum coke.

5 Journal of Applied and Industrial Sciences, 214, 2 (4): 188-2, ISSN: (PRINT),ISSN: (ONLINE) Table 3: The components of dry gas. [3] 192 Component Content (wt %) Component Content (wt %) Hydrogen 1.22 Propene 1.86 Methane Butane.6 Ethane Butene.4 Ethene 5.98 Water.5 Propane 1.95 Sulfur 2mg/Nm 3 Total 1. Table 4: Composition of LPG. [3] Component Content (wt %) Component Content (wt %) Hydrogen. Propene Methane. Butane 26.5 Ethane.27 Butene Ethene. C Propane Thio-alcohol 1ppm Total 1. Table 5: Main properties of stabilized gasoline, coker diesel oil and CGO (calculated data) [3] Stabilized gasoline Coker diesel oil Coker CGO Density (2ºC) kg/m Initial boiling point % % % Distillate 5% ºC 7% % % Dry point Table 6: Properties of coke [3] (lab. data in Dec. 21) property value real density, g/cm Volatile, m% 8.8 sulfur content, m%.41 Ash, m% 3.33 metal content Ni/V/Na/Al/Fe/Cu/Ca real density, g/cm Volatile, m% 8.8 Table 7: Main properties of coker gasoline (stabilized gasoline), diesel oil and coker gas oil. [3] Gasoline Diesel oil Coker gas oil Density (2ºC) kg/m Initial boiling point % Distillate 3% ºC 5% % % Dry point Acid value, mgkoh/1ml Actual gum, mg/1ml (m%) Induction period, min 669 Copper corrosion, (5ºC, 3h) Fail Fail Sulfur content, mg/kg (m%) Alkaline nitrogen content, mg/kg

6 Journal of Applied and Industrial Sciences, 214, 2 (4): 188-2, ISSN: (PRINT),ISSN: (ONLINE) Bromine value, g Br/1ml Conradson residual carbon, m%.27 Ni/V/Na/Al/Fe/Cu/Ca,μg/g.1/.1/1.4/1.1/3./.1/.6 Cetane number 51.6 Freezing point, ºC Flash point (close), ºC (open) Aniline point, ºC 61.8 Kinematic viscosity, 5ºC, mm 2 /s (1) Aromatic hydrocarbon, m% Asphaltene, m%.1 Saturated component, m% 71.8 Ash, m%.4 Table 8: Composition of acidic gas [3] Component Content, wt % H 2 O 4.9 H 2 S Hydrocarbon 1.49 Stability, tuning, offset and simulation Using MATLAB the transfer function for the process of the secondary loop is: The open loop transfer function using P controller is: i. Routh test: For this test the Characteristic equation of the loop is: ii. Direct substituting: Putting in characteristic equation to get:

7 Imaginary Axis Journal of Applied and Industrial Sciences, 214, 2 (4): 188-2, ISSN: (PRINT),ISSN: (ONLINE) Using Ziegler-Nichols table to calculate the controller parameter 194 Type of controller P PI.45 = 8 - For the Primary loop: When using P controller for the secondary loop: Characteristic equation is: iii. Routh test: iv. Direct substituting: s. Using Ziegler-Nichols table to calculate the controller parameter Type of controller P PI.45 =74 - PID.6 = =11.11 v. Root locus criterion: Root locus plots for the Secondary loop and primary loop is shown in figure 4, and 5, respectively root locus plot for the secondary loop Gain: 867 Pole: i Damping:.55 Overshoot (%): 98.4 Frequency (rad/sec): Gain: 867 Pole: i Damping:.55 Overshoot (%): 98.4 Frequency (rad/sec): Real Axis

8 Phase (deg) Magnitude (db) Imaginary Axis Journal of Applied and Industrial Sciences, 214, 2 (4): 188-2, ISSN: (PRINT),ISSN: (ONLINE) Fig. 4: Root locus for secondary loop root locus plot for the primary loop w ith(p secondary) Gain: 19.3 Pole: i Damping:.169 Overshoot (%): 94.8 Frequency (rad/sec): Real Axis Fig. 5: Root locus for primary loop with P controller for secondary loop vi. Bode and Nyquist plots: Bode and Nyquist plots for the Secondary loop and primary loop are shown in figure 6 through 9. bode plot for the secondary loop -5 Frequency (rad/sec): 8.75 Magnitude (db): Frequency (rad/sec): 9.19 Phase (deg): Frequency (rad/sec) Fig. 6: Bode plot for secondary loop

9 Phase (deg) Magnitude (db) Imaginary Axis Journal of Applied and Industrial Sciences, 214, 2 (4): 188-2, ISSN: (PRINT),ISSN: (ONLINE).4.3 Nyquist plot for the secondary loop db-2 db-4 db-6 db -1 db 1 db 6 db 4 db2 db db -2 db Real Axis Fig. 7: Nyquist plot for secondary loop bode plot for the primary loop w ith(p secondary) -5-1 Frequency (rad/sec): 2.72 Magnitude (db): Frequency (rad/sec): 2.72 Phase (deg): Frequency (rad/sec) Fig. 8: Bode plot for primary loop with P controller for secondary loop

10 Imaginary Axis Journal of Applied and Industrial Sciences, 214, 2 (4): 188-2, ISSN: (PRINT),ISSN: (ONLINE) db Nyquist plot for the primary loop w ith(p secondary) 6 db 4 db 2 db db -2 db -4 db -6 db -1 db db -2 db Real Axis Fig. 9: Nyquist plot for primary loop with P controller for secondary loop vii. Offset investigation: Secondary loop: For P controller For step change in input For PI controller For step change in input Primary loop: For P controller for primary and secondary loops: For step change in input

11 Amplitude Journal of Applied and Industrial Sciences, 214, 2 (4): 188-2, ISSN: (PRINT),ISSN: (ONLINE) For PI controller: 198 For step change in input For PID controller: For step change in input viii. System response: System response for secondary loop after introducing unit step change in the input is shown in figure 1, and for primary loop, system responses are shown in figure 11 through step change response for the primary loop w ith(p secondary) 1 Peak amplitude: 1.44 Overshoot (%): 45 At time (sec): 1.5 Final Value:.995 Rise Time (sec):.544 Settling Time (sec): Time (sec) Fig. 1: system response for secondary closed loop

12 Amplitude Amplitude Amplitude Journal of Applied and Industrial Sciences, 214, 2 (4): 188-2, ISSN: (PRINT),ISSN: (ONLINE).5 step change response for P primary loop.4 System: v Final Value:.34.3 System: v Settling Time (sec): 171 System: v Rise Time (sec): Time (sec) Fig. 11: system response for primary closed loop with P controller 1.8 step change response for PI primary loop w ith(p secondary) System: k System: k Settling Time (sec): 767 Rise Time (sec): 419 System: k Final Value: Time (sec) Fig. 12: system response for primary closed loop with PI controller step change response for PID primary loop w ith(p secondary) System: n Rise Time (sec): 12 System: n Final Value: 1 System: n Settling Time (sec): Time (sec) Fig. 13: system response for primary closed loop with PID controller

13 Journal of Applied and Industrial Sciences, 214, 2 (4): 188-2, ISSN: (PRINT),ISSN: (ONLINE) Delayed Coking Unit provides good processes to upgrade the heavy crude, or to convert it to more light valuable products. Coke can be more utilized if it is gasified to Syngas which is can be used to produce energy and also as a feed for petrochemical industries. Secondary loop gain using Routh test and Root locus respectively, is 435, and for P- controller and 395.5, and for PI-controller, results of Routh test are identical to the results of Direct Substitution method. Using Bode criterion 2 the value of the Gain Margin is 876, and the Phase Margin is infinity, these results are identical to Nyquist criterion. The offset is -.38 for P- controller and. for PI controller. The primary loop with P- controller for the secondary loop, the gain is 1.625, 1.463, and 1.95for P, PI, and PID respectively using Routh test, and 9.6, 8.69, and 11.6 P, PI, and PID respectively using Root locus criterion. The offset is -.64 for P- controller and. for PI, and PID controller. A summary for these results is shown in table 9. Table 9: Summary for the Comparison between Stability and tuning method for Primary loop Method K u P u Offset P : Routh_Huwrtiz PI : PID : P : Direct Substitution PI : PID : P : Root Locus PI : PID : IV. CONCLUSIONS Delay Coking Unit is one of the most valuable units in the refinery because it increases the economic benefits by converting low price residue to valuable products with higher price. Advance control has a great effect on the response of the chemical plants. It increases stability and eliminate offset. There are many methods that are used to get the adjustable parameters such as Routh-Hurwitz, Direct substitution and there are another three graphical method, Bode, Nyquist, and root locus. Ziegler-Nicholas criterion is used to tune the adjustable parameters. PID controller for the Primary loop provides the highest gain than P, and PI controllers and also it eliminates the Offset. ACKNOWLEDGEMENTS The authors wish to thank the Faculty of Graduation Studies and Scientistic Research of Karary University, and Khartoum Refinery Company for their help and support, this Paper is generated from a Thesis in partial fulfillment for Ph.D. in Chemical Engineering. REFERENCES [1] Ellis, P. J., Paul, C. A., Delayed Coking Fundamentals, Topical Conference on Refinery Processing, Paper 29, March 8-12, [2] Meyers, R. A., Hand Book of Petroleum Refining, McGraw Hill, third edition, [3] Manual of Delayed Coking Unit at Khartoum Refinery, 27. [4] Jeng, J.-C., Simultaneous closed-loop tuning of cascade controllers based directly on set-point step-response data, Journal of Process Control, Elsevier, 214. [5] Haseloff, V., Friedman, Y. Z., et al., Implementing coker advanced process control, Hydrocarbon Processing, Gulf Publishing Company, p , June 27. [6] Stephanopolous, G., Chemical Process Control, Prentice-hall of India, New Delhi, 25. [7] Tyner, M., May, F. P., Process Engineering Control, The Ronald press company, New-York, [8] P. C. Chau, Chemical Process Control: A First Course with MATLAB, 21.