DEPARTMENT OF COMPUTER SCIENCE AND ELECTRICAL ENGINEERING GRADUATE SCHOOL OF SCIENCE AND TECHNOLOGY KUMAMOTO UNIVERSITY KUMAMOTO JAPAN

Transcription

1 Presented by: Ananto Mukti Wibowo / 091 d 9859 DEPARTMENT OF COMPUTER SCIENCE AND ELECTRICAL ENGINEERING GRADUATE SCHOOL OF SCIENCE AND TECHNOLOGY KUMAMOTO UNIVERSITY KUMAMOTO JAPAN 1

2 INTRODUCTION LITERATURE REVIEW METHODOLOGIES SIMULATION RESULTS CONCLUSIONS 2

3 Background Problem Definition Research Objectives 3

4 The influences of internal combustion systems such as cars with gasoline engines become a serious social problem because of the environment pollution. To alleviate the problems, automobile manufacturers forced to shift their part of productions from pure internal combustion systems to hybrid systems or electric systems. Electric car uses battery that have dc voltage, because of this, dc motor is usually implemented. [4] WADA Masayoshi, Research and development of electric vehicles for clean transportation, Journal of Environmental Sciences 21(2009)

5 Disadvatage of dc motor: Often needs regular maintenance Series motors cannot be used where a relatively constant speed is required under conditions of varying load (not suitable for the hilly environment) Solution: Induction Motor 5

6 Induction motor advantages: simple construction robust cheaper easier to maintain high torque characteristics 6

7 Induction Motor Speed Dynamics in Electric Car Drive Starting Starting & Accelerating Running Accelerating Decelerating & Breaking Running Decelerating Breaking Speed Time (s) Induction motor needs a controller so the dynamic speed conditions can be achieved. Proposed method: Direct Torque Control using ANN Sliding Mode Control 7

8 Develop an optimized speed controller for a three phase induction motor as an electric car drive based on Direct Torque Control using Artificial Neural Networks Sliding Mode Control. 8

9 Direct Torque Control (DTC) Sliding Mode Control (SMC) Artificial Neural Network (ANN) 9

10 DTC was presented by I. Takahashi in the middle of 1980 s. DTC is a control method where the torque and speed are controlled directly based on the electromagnetic state of the motor. The controlling variables are motor magnetizing flux and motor torque. With DTC there is no need for modulator which slows down communication between the incoming voltage and frequency signals and the need for the motor to respond to this changing signal. 10

11 DTC block diagram 11

12 12

13 Sliding Mode Control (SMC) is a procedure to design robust controllers for nonlinear processes. The SMC reachability condition is based on the Russian mathematician, Lyapunov, and his theory of stability of nonlinear systems to guarantee the stability of the closed loop system. The main advantage of SMC is the robustness under uncertainties caused by load torque [3]. [3] T.B. Reddy, J. Amarnath, D. Subba Rayuddu, "Direct Torque Control of Induction Motor Based on Hybrid PWM Method for Reduced Ripple : A Sliding Mode Control Approach", ACSE Journal, Volume (6) Issue (4)

14 h a, b, d are fixed parameters introduced by friction (B) and 2 w 1 wr* h-a a 1 s Integrator Sign Beta 1/b 1 Te* inertia constant (J). Tuned Parameters: h determines the sliding surface gain β guards the trajectory in the sliding du/dt d surface The output of SMC are the SMC block diagram in Simulink torque reference for the DTC. 14

15 ANN is an information-processing system that has certain performance characteristics in common with biological neural networks. X 1 w 1 v 1 Z 1 X 2 w 2 Y w 3 Z 2 v 2 X 3 Input Units Hidden Units Output Units 15

16 Simulation Model Generate Data for ANN Learning ANN Architecture Design Learning Results 16

17 Load Torque W_ref speed ref wr* Te* w SMC Torque1 Torque2 DC Source Vd Torque Vdc va vb TL va vb Tem Wmech i_abc Tem fl_s_0 Flux Q gate vc vc v _abc I_s flux ref Flux Sector Switching Table Inverter Induction Motor Torque plot Flux plot Sector f l_s_ab Sector Selection f l_s_ab f l_s_est Tem_est v _abc i_abc Flux and Torque Estimator Speed plot 17

18 Simulations are conducted using different speed references and observed from the motor start from time 0 to 0.02 seconds. The rise time and steady state speed error is analyzed The control performance is evaluated by the performance index (J) 18

19 Optimal gain value for ANN learning Speed ref h Β

20 X 0 =1 Z 0 =1 X 0 Z 0 Z 1 w 1,1 ω r * w 1,1 w j,1 Z j w 1,j 1 h w 20,1 2 β Z 20 w 2,20 20

21 The learning will stop under two conditions: Reach criteria function (SEE=0) Reach maximum epoch (1000) SSE=

22 The motor speed reference are changed in the process to match the dynamics of movement in the electric car. The speed steady state time of the system to reach the reference speed will be observed Verification with data training Compare system using ANN with system not using ANN Verification with other data 22

23 Speed (rad/s) Simulation data Time (s) 0 0,15 0,35 Speed ref (rad) Gain value for system without ANN h=239, β= reference without nn with nn Time (s) Without ANN With ANN Without ANN With ANN Without ANN With ANN speed ref (rad) time to reach steady state 0,02 0,02 0,05 0,1 0,08 0,05 at reference speed (s) 23

24 Simulation data Time (s) 0 0,15 0,35 Speed ref (rad) speed ref (rad) time to reach steady state at reference speed (s) 0,03 0,14 0,08 24

25 Induction motor speed drive using sliding mode control can be improved with the optimization of gain value h and β. At accelerating condition, by using ANN to tune the SMC gain is 0.05 s slower than without ANN but does not have oscillations in the response which is good in electric car dynamics. At decelerating condition, by using ANN can improve the performance by 0.03 s without any oscillations in the speed response 25

26 1. Soebagio, Teori Umum Mesin Listrik, Srikandi, Surabaya, Gigih Prabowo, Mauridhi Heri Purnomo, Soebagio, Metoda Direct Torque Control pada Pengaturan Motor Induksi tanpa Sensor Menggunakan Sliding Mode Control, SITIA (2008) 3. T. Brahmananda Reddy, J. Amarnath and D. Subba Rayudu, "Direct Torque Control of Induction Motor Based on Hybrid PWM Method for Reduced Ripple: A Sliding Mode Control Approach", ACSE Journal, Volume (6), Issue (4), Dec., L. Fausett, (1993),"Fundamentals of Neural Networks: Architectures, Algorithm, and Applications", Prentice Hall 5. Perruquetti, W., Barbot, Jean Pierre, Sliding Mode Control In Engineering, Copyright 2002 by Marcel Dekker 6. Ion Boldea, S. A. Naser, Electric Drives 2nd Edition, CRC Press Taylor & Francis Group, Ned Mohan, Advanced Electric Drives Analysis, Control and Modeling using Simulink, MNPERE, S.M. Gadoue, D. Giaouris, J.W. Finch, Artificial intell-based speed control of DTC induction motor drives A comparative study, Electric Power System Research 79 (2009) Wada Masayoshi, Research And Development Of Electric Vehicles For Clean Transportation, Journal of Environmental Sciences 21 (2009)

27 THANK YOU 27