Remedial neural network inverse control of a multi-phase fault-tolerant permanent-magnet motor drive for electric vehicles

Abstract

A five-phase in-wheel fault-tolerant interior permanent-magnet (FT-IPM) motor incorporates the merits of high effi ciency, high power density and high reliability, suitable for Electric Vehicles (EVs). A new remedial Neural Networks Inverse (NNI) control strategy is proposed to attain the post-fault operation. In this scheme, the NN is used to approximate the inverse model of the FT-IPM motor. With this NNI system and the original motor drive combined, a pseudo-linear compound system can be obtained. The simulation demonstrates that the proposed control strategy leads to excellent control performance at the faulty mode and offers good robustness against load disturbance.