Study on PMSM Integral Backstepping Controller Based on RBF Neural Network

Abstract:

In order to eliminate steady-state speed error of PMSM backstepping control system, an integral backstepping speed control algorithm is designed in this paper. By adding speed error integral factor in the speed Lyapunov function, the speed error can finally converge to zero when PMSM operates in steady-state. On this basis, an integral backstepping speed control algorithm based on RBF neural network compensation is proposed for PMSM backstepping control system used for high-altitude electric propulsion system which is vulnerable to load torque variables. The integral backstepping speed controller based on PMSM reference model can ensure global asymptotic convergence of the whole control system. In order to achieve fast robust adaptive control, the RBF neural network is adapted to online compensate dq axis current error produced by the reference speed and load torque changes. Simulink simulation results verify the feasibility of the given algorithm.

• Periodical: Applied Mechanics and Materials (Volumes 416-417)
• Pages: 599-605
• DOI: https://doi.org/10.4028/www.scientific.net/AMM.416-417.599
• Online since: September 2013
• Authors: Qian Yang, Nan Nan Zhao, Ming Hui Zhang
• Keywords: dSPACE Experiment Platform, Integral Backstepping, PMSM, Radial Basis Function (RBF) Neural Network

References

[1] Zhu Kai, Qi Naiming and Qin Changmao, adaptive sliding mode controller design for BTT missile based on backstepping control, Journal of Astronautics，vol. 31, pp.769-773, March (2010).

DOI: 10.1109/icma.2009.5244807

[2] Xue Fuqiang, Ge Lindong, and Wang Bin, RBF neural network classifier based on improved hierarchy genetic algorithm, Journal of System Simulation, vol. 22, pp.339-402, February (2010).

[3] Xie Zhenping, Liu Jihong, and Wang Shitong, RBF network learing algorithm using robust least-squares, Control and Decision, vol. 25, pp.502-504, April (2010).

[4] Ye Jian, Ge Lindong, and Wu Yuexian, a application of improved RBF neural network in modulation recognition, ACTA Automatica Sinica, vol. 33, pp.652-654, June (2007).

[5] Xu Qinhong, Dai Xianzhong, Online learning RBF neural network-based adaptive compensative control scheme for governor system, Electric Machines and Control, vol. 14, pp.15-18, February (2010).

[6] Mingjun Zhang, Jie Yang. Robot visual planar locating method based on improved RBF network. IEEE Conference on Advanced Intelligent Mechatronics, 2009: 1740 - 1744.

DOI: 10.1109/aim.2009.5229807

[7] Xia Hong, Sheng Chen. A New RBF Neural Network with Boundary Value Constraints. IEEE Transactions on Systems, Man, and Cybernetics, 2009, 39(1): 298 - 303.

DOI: 10.1109/tsmcb.2008.2005124

[8] Yu Jinpeng, Chen Bin, Adaptive fuzzy backstepping position tracking control for permanent magnet synchronous motor, Control and Decision, vol. 25, pp.1547-1551, October (2010).