Robust Variable-Gain Proportional-Integral Back-EMF Estimator for Measurement Noise and Position Sensorless Control of SPMSM

Abstract

This article presents a variable-gain back-electromotive force (back-EMF) estimator and position-sensorless control of the surface-mounted permanent magnet synchronous motor (SPMSM) to enhance robustness for measurement noise. From the frequency response analysis of the conventional proportional-integral (PI) back-EMF estimator system, we newly present a variable-gain proportional-integral (VG-PI) back-EMF estimator and its gain tuning process which enables the back-EMF estimation system to be robust to measurement noises. Owing to the robustness of the measurement noises of the VG-PI structure, the low-pass filtering of the estimated back-EMF signals is removed, which reduces the phase delay of the estimated position and enhances the sensorless speed control performance of the SPMSM. Comparative studies show that the proposed method significantly reduces high-frequency components in the VG-PI back-EMF estimation compared to conventional fixed-gain PI observers and adaptive gain super-twisting sliding mode observers (AGST-SMOs). Furthermore, experimental results demonstrate that the proposed VG-PI method outperforms the conventional method in terms of angular position phase delay, reducing it by more than 18°.