New Sliding Mode Predictive Observer for Variable-Leakage-Flux PMSM Sensorless Drive System

The variable-leakage-flux permanent-magnet synchronous motor (VLF-PMSM) sensorless control drive system has the advantages of high efficiency, wide-speed regulation range, and high reliability, which is suitable for electric vehicles (EVs). However, the parameter variation characteristics of VLF-PMSM and the difficulty in selecting sliding-mode gain coefficient, will impact the performance of sensorless control to a certain degree. Currently, most studies target the identification of multiple parameters and the improvement of reaching laws, resulting in complex structures and limited application. In this study, a new sliding mode predictive observer (SMPO) based sensorless control strategy is proposed for the VLF-PMSM. With the proposed SMPO, the rapid response and minimal chattering in sliding-mode motion can be guaranteed without the need to select sliding-mode gain coefficients. Also, an improved forgetting factor recursive least square (FFRLS) parameter identification modular is innovatively integrated into this SMPO, which will not only reduce the computational burden but also mitigate the impact of motor parameter variations on sensorless control. Furthermore, the phase-locked loop (PLL) is redesigned by mathematical iteration method to effectively expand the bandwidth of the sensorless controller. Finally, the efficiency of this strategy is verified by experimental results.