Sensorless Control for SPMSM Using Modified Super-Twisting Observer Considering Resistance Mismatch and VSI Nonlinearity

Abstract

This article proposes a low-speed sensorless control strategy for permanent magnet synchronous motors (PMSMs) with resistance synchronous identification-based modified super-twisting algorithm-sliding mode observer (RSSTA-SMO) and amplitude-adaptive notch filter (AANF). The higher order terms of the RSSTA-SMO are introduced to enhance the response capability while suppressing high-frequency oscillations. The novel adaptive law is utilized to facilitate resistance synchronous identification and improve robustness. The overall stability of RSSTA-SMO is presented, avoiding the steady-state error that may arise from the separate construction of back electromotive force (EMF) and resistance identification. The time compensation strategy and AANF are both adopted to adaptively compensate the harmonics in the estimated back EMFs caused by the nonlinearity of the voltage source inverter (VSI). Ultimately, the superiority and effectiveness of the proposed strategy are confirmed through experimental results.