A Novel Rotor Position Observer for Sensorless Control of Permanent Magnet Synchronous Motor Based on Adaptive Generalized Second-Order Integrator

Abstract

In order to realize sensorless control of permanent magnet synchronous motor (PMSM) with high performance in low speed region, a novel rotor position observer scheme based on finite control set model predictive control (FCS-MPC) is presented in this paper. Firstly, the FCS-MPC is used to predict the current and drive the PMSM by selecting the optimal control quantity that minimizes the cost function. Next, an adaptive second-order generalized integrator (ASOGI) with adaptive center frequency adjustment was designed to replace the band-pass filter (BPF) in the rotor position observer. The ASOGI can calculate the high frequency value that can be used for position estimation by the controller switching frequency. The current ripple inherent in the FCS-MPC is considered as the response current obtained by the high frequency injection (HFI) method. The current ripple after ASOGI filtering is input to the phase-locked loop (PLL) for phase locking to obtain the estimated rotor position. In addition, adaptive linear (Adaline) neural networks are used to identify sensitive motor parameters online to avoid mismatch of model parameters, which causes degradation of control performance. Simulation experiments and hardware experiments show that this scheme is excellent in both static and dynamic conditions.