A Dual-Filtering Complex Vector Observer for Harmonic Suppression in Sensorless PMSM Drivers

Abstract

Model-based methods stand out for their high efficiency in sensorless permanent magnet synchronous motor (PMSM) control techniques and have attracted more and more attention. However, inverter nonlinearities and flux spatial harmonics cause back electromotive force (BEMF) harmonics, which reduce the accuracy of rotor position estimation. To address this problem, a dual-filtering complex vector observer (DCVO) is proposed to accurately estimate the BEMF in this article. First, a structure of complex vector observer is constructed based on the mathematical model of PMSM and its transfer function is derived. Then, a DCVO is designed based on the frequency-domain analysis of the closed-loop transfer function. This observer accurately extracts the fundamental component of the BEMF and effectively filters out the dual harmonics, avoiding the distortion of the estimated BEMF. Compared with the existing observers, the proposed observer can effectively reduce the estimation error of rotor position by eliminating harmonics. Finally, the effectiveness of the studied observer is verified on an experimental platform.