Demagnetization fault detection in permanent magnet synchronous motors based on sliding observer


Authors

Jing He - School of Electrical and Information Engineering, Hunan University of Technology, Zhuzhou, 412007, Hunan, China. - School of Mechatronic Engineering and Automation, National University of Defense Technology, Changsha, 410073, Hunan, China. Changfan Zhang - School of Electrical and Information Engineering, Hunan University of Technology, Zhuzhou, 412007, Hunan, China. Songan Mao - School of Electrical and Computer Engineering, Purdue University, West Lafayette, IN 47907, U. S. A.. Han Wu - School of Electrical and Information Engineering, Hunan University of Technology, Zhuzhou, 412007, Hunan, China. Kaihui Zhao - School of Electrical and Information Engineering, Hunan University of Technology, Zhuzhou, 412007, Hunan, China.


Abstract

This paper considers a robust multiple fault detection method for actuator failures in nonlinear systems. The actuator failures model is initially put forward. By employing the unique advantage that the sliding mode variable structure is invariance to uncertainties, a sliding mode state observer is designed to isolate the unknown input disturbance effect on residual generation. The parameters of the observers being designed are determined by the use of linear matrix inequalities techniques. Accordingly, the generated residual is only sensitive to the specific fault signals, and the fault detection accuracy is improved. This paper verifies the proposed method by its application in demagnetization fault detection for a permanent magnet synchronous motor (PMSM). Simulation and experiment results illustrate the high detection accuracy and robustness.


Keywords


MSC


References