Hybrid FUZZY-PID controller for electric shovel EKG-8И hoisting motor

Nội dung text: Hybrid FUZZY-PID controller for electric shovel EKG-8И hoisting motor

1. HỘI NGHỊ KHOA HỌC TOÀN QUỐC VỀ CƠ KHÍ – ĐIỆN – TỰ ĐỘNG HÓA (MEAE2021) Hybrid FUZZY-PID controller for electric shovel EKG-8И hoisting motor Hong Quan, Luu 1, Cao Phong, Khong 2, 1 Dong Nai Technology University, luuhongquan74@gmail.com 2 Hanoi University of Mining and Geology, khongcaophong@humg.edu.vn ARTICLE INFO ABSTRACT Article history: th Soviet electric shovel EKG-8Иs are playing an importance roll in Vietnam Received 15 Jun 2021 mining company. To improve efficiency of the shovels, a new control Accepted 16th Aug 2021 system with excitation field circuit using power electronic devices was Available online 29th Nov 2021 proposed for their Ward Leonard drive systems. In this paper, Hybrid Keywords: Fuz zy-PID controllers for current and speed control loops were suggested. Electric shovel, Ward Simulations with MATLAB Simulink for the electric shovel hoisting controllers were implemented and compared with conventional PID Leonard, PID, Fuzzy, Hybrid controller. Simulation results show that control quality of the hoisting Fuzzy drive is improved with the Hybrid Fuzzy-PID controllers. Copyright © 2021 Hanoi University of Mining and Geology. All rights reserved. 1. Introduction many other electric shovel system all over the world. For that reason, there are many practical Soviet Union shovel excavators (EKGs) have research to revamp and renovate the old electric been put into use in Vietnam mining industrial EKG5(RATHMANN, 2009), (Rathmann, 2015), since 60s of the last century. Since then, they have (Systems, 2009). In the research, there are several played a very importance role in mining solutions to revamp the antique Ward Leonard companies. There are many types of EKG shovel control systems in old electric shovels. The excavators have been used for extraction and systems are totally replaced by ABB AC drives. loading in open-pit mine. They were imported This solution is apply successfully in BE295 BII, from Soviet Union in early and Russia in recently. BE295 BIII, EKG-5, BE395 BI shovels, which The EKG excavators in Vietnam concentrate in belong to Mexico, Brazil, USA, Isarel, Germany, some medium types which have buckets from 4- Chile companies(Rathmann, 2015). To reduce 10m3: EKG 4,6; EKG 5A; EKG-8I; EKG 10. renovating cost and keep mechanical balance for The number of the EKG shovel excavators in electric shovel, only excitation circuits of DC Vietnam is about several hundreds. They are generator and DC motor are replaced. The power located mostly in open-pit coal mining companies semiconductor converters are proposed to take like Cao Son; Coc Sau; Deo Nai; Ha Tu; Nui Beo over from magnetic amplifers in excitation and in several quarry-mining companies, (Dac circuits. These proposals are analysis and Tao, 2011). compared in (Bus & Mg, 2007), (Systems, 2009). Although the electrical systems of those With the help of proposed control drives, electric shovels are old with many working electric shovels can respond faster with reference problems, the mechanical systems of EKG-8Is are signals from the drivers. However, it requires still working well. This situation is the same with research to optimize control characteristic. In 91
2. HỘI NGHỊ KHOA HỌC TOÀN QUỐC VỀ CƠ KHÍ – ĐIỆN – TỰ ĐỘNG HÓA (MEAE2021) conventional control, the proportional-integral- controller also has slow respond because of derivative (PID) is used to regulate electric motor magnetic amplifier characteristic. Magnetic because of its simplicity, applicability, and amplifier alo drifts its working point when the reliability. In complex and non-linear systems, system temperature changing after a period of Fuzzy Logic Control (FLC) is a better selection work. than PID control. The main advantage of this method is, it does not require precise mathematical models of the plant (Shaker & Al- Khashab, 2010). FLC is also good at broad range of operation system (Hu et al., 2019) and userful in power optimization system(Kholid et al., 2019). In this paper, we propose a solution to revamp the excitation system of Vietnam common electric shovel EKG-8I. With suggested system, a FLC control algorithm is implemented and simulated to compare with conventional PID control. Figure 9 Magnetic-Amplifier rectifier 2. Proposed control drive system From those disadvantages, proposals to In EKG-8I electric shovel, there are four main replace MAR with sermiconductor converters control drive system. They are Hoist drive, Crowd based on Thyristors and IGBTs are mentioned in drive, Swing drive and Propel drive. In Vietnam, (Bus & Mg, 2007). IGBTs’converter has some all EKG-8Is are working with DC drive systems. advantages over the Thyristors one. Hence, it is Each DC motor is controlled by a DC generator. DC proposed to renovate the exciter of DC motor and generators supply regulated DC voltages, which generator. The exciter is an IGBT converter with are adjusted by exictation currents with constant DC voltage in and Chopped DC voltage out (DC-DC speed rotors, to control the motors (Ward- converter Figure 10). Using chopped frequency of Leonard structure Figure 8). 20-25kHz, the suggested exciter can give out very RaG Ia RaM fast response in compare with MAR and Thyristors converter. EG Ua EM Synchronous Load Motor i i F_G uF_G F_M uF_M G-Exciter M-Exciter Control G Control M Figure 8. Ward-Leonard structure In EKG-8I, the excitation currents are still fed by magnetic amplifier rectifiers (MAR). The MAR circuit diagram is shown in Figure 9. The Figure 10. IGBT DC-DC converter disadvantages of this old MAR system are size, complicated circuit, slow response time, output 3. Mathematic model of drive system offset. Because of magnetic circuit with iron core The Ward-Leonard drive system uses two DC and copper coils, the MAR has a big size and heavy machines, one DC genrator and one DC motor, in weight. The complicated connection makes it is it’s structrure. The generator output voltage is fed difficult to repair, maintain and adjust. The to the motor to control load output speed Figure 92
3. HỘI NGHỊ KHOA HỌC TOÀN QUỐC VỀ CƠ KHÍ – ĐIỆN – TỰ ĐỘNG HÓA (MEAE2021) 8. That structure can be described by an inductances and resistances of generator field equivalent circuit shown in Figure 11. The circuit and machines armature circuit; Kf, Kg are mathematical model of the Ward-Leonard drive gain factors of generator field circuit and system can be described as follows. generator electromotive force (e.m.f); Ce is gain factor of DC motor; uref is reference control Lag Rag Ram Lam voltage;  is motor rotating speed; J is inertial La = Lag + Lam Ra = Rag + Ram moment of DC machines; M is subtraction of load moment from motor moment. Using the system of equations (1), block eg=KFff =Kgif em=Ce diagram of Ward-Leonard is presented in Figure 12. 4. PID control Figure 11. Ward-Leonard equivalent circuit diRK Result from the mathematic model above, a fff = −+ iu control structrure with three loops was dtLL f ref implemented for the IGBT exciter Ward-Leonard ff system. Each loop is regulated by a P or PI diRC Kg controller. The innermost control loop is the aae=−− ii (1) f a generator excitation current loop with PI dtLLL aaa controller. The middle control loop is the d 1 machines armature current loop with PI = M controller. The outer loop uses a P controller to dtJ  regulate speed of the DC motor. Where if, ia are generator field current and machines armature current; Lf, Rf, La, Ra are Exciter Generator Amature Figure 12. Ward-Leonard block diagram -K- Controllers Ward Leonard System 1 c_e 8 s Integrator K_i 10 4 0.257s+2.408 k_kt 1/r_kt 1/r_u 1 min Limited -K- -K- -K- s w_ref Wpc s t_kt.s+1 t_mf.s+1 t_u.s+1 MinMax Dien ap KT Tu thong BH k_mf c_m 1/j_1+j_2 Integrator Wpu -K- Kich tu TF May phat TF Dong co TF K_p -K- -C- Toc do kotb momencan -K- 1-D T(u) kot Dong i_u -K- kot3 du/dt -K- Electrical, Mechanical Limitation Derivative1 kot2 -K- koc Figure 13. PID control diagram 93
4. HỘI NGHỊ KHOA HỌC TOÀN QUỐC VỀ CƠ KHÍ – ĐIỆN – TỰ ĐỘNG HÓA (MEAE2021) In the armature control loop, a structure to electrical and mechanical limitation. After a peak make the drive system working within electrical of about 1600A, motor current is cut down within and mechanical limits is implemented. It helps the shovel characteristic curve. The simulation the system to generate electric shovel results present control ability of proposal system characteristic. When the shovel meets a heavy with P, PIs controller. load (for instance hard rock), the drive system will rapidly collapse to stall current (2-2.5 normal current). The control structure simulated in Matlab Simulink is depicted in Figure 13. To get the control ability with PID control, a set of EKG-8I parameters is used to calculate parameter values of blocks in the Simulation diagram. P and PI controllers are designed by using diagram parameters and conventional magnitude optimum method. Table 2. EKG-8I hoist DC machines Parameters ПƎМ- ДПƎ- Figure 14. PID Motor output speed 155- 82A 8K Power (Kw) 500 190 Norminal voltage (V) 560 270 Norminal curent (A) 895 760 Norminal speed (rpm) 1000 740 Pole pair 3 2 Armature resistance (Ω) 7.5E-3 3.1E-3 Field resistance (Ω) 0.96 3.5 The calculated parameters are set to the Matlab simulink diagram and simulations are implemented. In the first simulation, motor speed Figure 15. PID Armature output current is set to 100% of normal speed. Load torque in this case is set to 100% normal load. Simulation 5. Fuzzy-PID control results are motor output speed and armature As mentioned in many previous research, output current presented in Figure 15 and Figure conventional PID controllers normally produce 15. large overshoot, undershoot, and settling time. In Figure 15, simulation output speed of the This is because the conventional controllers adopt motor is stable after a response time of 1.5 second. fixed and static parameters such as scaling factor. The output current of DC motor is also limited by 94
5. HỘI NGHỊ KHOA HỌC TOÀN QUỐC VỀ CƠ KHÍ – ĐIỆN – TỰ ĐỘNG HÓA (MEAE2021) Rshunt ua ia  G M Speed controller Current controller Voltage controller |e|>e Fuzzy Fuzzy Kp Ki Kd P PI PI − − −  ref Figure 16. Proposed Fuzzy-PID control structure |u| Controllers Ward Leonard system Spee |e|1 d 'Kpid_FC.fis' F_ref Toc do P I_a du/dt 1 Dong Iu s M_loa Dev_1 'Iu_FC.fis' 10 4 I d E_g min Int Chon tai w_ref (0-10V) Wpc > du/dt MinMax FC+PI Dien ap KT Ward-Leonard D |e|>|e_0|1 Dev_2 0.257s+2.408 1*mc Mc 4*mc Mc1 s -K- Wpu1 Adaptive PID-Fuzzy Hybrid PID-Fuzzy kotb -K- 1-D T(u) kot -K- kot3 du/dt -K- 1-D Lookup Table Dev_3 kot2 -K- koc Figure 17 PID-Fuzzy control simulink diagram To improve control quality, PID-Fuzzy output range of these parameters are determined controllers are suggested in generator e.m.f and following (Zhao et al., 1993). Fuzzy algorithm is armature current control loops Figure 16. In established by setting rule and signal range in generator e.m.f control loop, Fuzzy algorithm is Matlab tool. The setting range and rule are used to tuning Kp, Ki, Kd of the PID controller. In armature current control loop, the controller is adaptively selected between PID and Fuzzy controller (Hybrid Fuzzy-PID - HFP). The outer loop is not chaned with P controller. Based on the PID control results above, parameters of PID-Fuzzy controllers are calculated. The generator e.m.f control loops is regulated by a The Fuzzy algorithm is used to derive Kp, Ki, Kd from e.m.f voltage error and derivative e.m.f error. The Fuzzy membership functions and 95
6. HỘI NGHỊ KHOA HỌC TOÀN QUỐC VỀ CƠ KHÍ – ĐIỆN – TỰ ĐỘNG HÓA (MEAE2021) presented in Figure 18 and Figure 19 Fuzzy controller is selected when the current error is large. In contrast, the PID controller is selected when the current error is small. The PID controller use the calculated Kp, Ki calculated in section 4. The Fuzzy controller is used a structrure and membership functions present in Figure 20 and Figure 21. The Fuzzy rule uses current error in range [-20, 20], derivative current error in range [-150, 150] as two inputs and output in range [-40, 40]. These input parameters are established based on the range of error value, derivative of error value which are collected from simulations with PID controllers. The output range is setting following the real excitation Figure 18. voltage of EKG-8I electric shovel. Figure 20. Fuzzy current controller structure Figure 18. Graphic user interface for Fuzzy rule Figure 21. Fuzzy current controller rule Figure 19. Self-tuning fuzzy rule (Kd output) The armature current controller is a combination of PID controller and Fuzzy controller. The output of each controller will play the role of current control signal depend on the current error and its derivative. In which, the 96
7. HỘI NGHỊ KHOA HỌC TOÀN QUỐC VỀ CƠ KHÍ – ĐIỆN – TỰ ĐỘNG HÓA (MEAE2021) control algorithms are suggested to regulate the generator e.m.f control loop and armature current control loop of the EKG-8I’s Ward-Leonard control drives. The control algorithms were implemented and simulated in Matlab Simulink. The proposed controllers were compared with the ones of conventional PID controllers. The simulation results presented a higher control quality of the proposed controllers. That proves the proposed the IGBTs conver exicter and the combination of Fuzzy and PID controllers. To renovate the old EKG-8I with the proposal Figure 22. HFP output speed exciter and control algorithm, research with practical experiments has to be implemented. References Bus, S., & Mg, I. (2007). New IGBT Exciter- Based MG Set Excavator Drive. Dac Tao, D. (2011). HỆ TRUYỀN ĐỘNG MỚI CỦA MÁY XÚC EKG•5. KHCN Mỏ. Hu, H., Wang, T., Zhao, S., & Wang, C. (2019). Speed control of brushless direct current motor using a genetic algorithm– Figure 23. HFP Armature output current optimized fuzzy proportional integral A simulation is implemented with the diagram in differential controller. Advances in Figure 17 and control algorithm above. The Mechanical Engineering, 11(11), 1–13. simulation is set with: Full load; 100% normal speed (uref=10V). Results are shown in Figure 0199 22 and Figure 23. In Figure 22, output speed of the Kholid, A., Fauzi, R. A., Yunazwin Nazaruddin, Y., motor with Hybrid Fuzzy-PID (HFP) controller & Joelianto, E. (2019). Power Optimization has a faster response (1.0 second). The output of Electric Motor using PID-Fuzzy Logic armature current in Controller. 2019 6th International Figure 23 presents a better control quality. Conference on Electric Vehicular The current is within the electrical and Technology (ICEVT), 189–195. mechanical limitation of the shovel. There is also no high peak current as in PID controler result. 19.8993984 6. Conclusions Rathmann, B. (2015). ABB Drives Retrofit The paper presented a proposal that replaces Electric Shovels. the old EKG-8I magnetic amplifier field exciter with the new IGBTs DC-DC converter field exciter. RATHMANN, B. (2009). Refurbishment of the With very fast responding abiltity of the new electrical installation of an EKG 5A shovel exciter, Hybrid Fuzzy-PID, Self-tunning Fuzzy PID excavator. ZKG International, 62(4), 80– 97
8. HỘI NGHỊ KHOA HỌC TOÀN QUỐC VỀ CƠ KHÍ – ĐIỆN – TỰ ĐỘNG HÓA (MEAE2021) 87. Systems, T. A. (2009). DC-EXX IGBT Exciter- Based MG Set Excavator Drive Overview Shaker, M. M., & Al-Khashab, Y. M. B. I. (2010). and Field Experience Presented to. Annual Design and implementation of fuzzy logic Meeting Clearwater Florida, 1–67. system for DC motor speed control. EPC- IQ01 2010 - 2010 1st International Zhao, Z. Y., Tomizuka, M., & Isaka, S. (1993). Conference on Energy, Power and Control, Fuzzy Gain Scheduling of PID Controllers. May 2011, 123–130. IEEE Transactions on Systems, Man and Cybernetics, 23(5), 1392–1398. 98