Demagnetization fault reconstruction for six-phase permanent magnet synchronous motor by improved super-twisting algorithm-based sliding-mode observer

Abstract

Demagnetization fault causes the performance degradation of a permanent magnet synchronous motor (PMSM) drive system. Therefore, it is of great importance to reconstruct demagnetization fault in real-time. This paper firstly presents a demagnetization fault reconstruction method for a six-phase permanent magnet synchronous motor (SP-PMSM). A super-twisting algorithm-based sliding-mode Luenberger observer (STA-SMLO) is designed. It can eliminate the chattering of the sliding mode observer (SMO) and effectively isolate the influence of motor speed on the observer. The mathematical model of the demagnetized SP-PMSM is constructed by the vector space decomposition (VSD). Then, an STA-SMLO, which combines a super-twisting algorithm-based sliding-mode observer (STA-SMO) with a Luenberger-observer (LO), is designed to reconstruct the demagnetization fault of the SP-PMSM precisely. Moreover, a strong Lyapunov function is designed to ensure the stability of STA-SMLO. Experiment results of hardware-in-the-loop simulation (HILS) demonstrate the feasibility and advantages of the method.