A Robust Hybrid Complex Coefficient Filter Without Frequency Coupling for Position Estimation Accuracy Improvement of PMSM

To reduce the harmonic distortion in sensorless control of permanent magnet synchronous motor (PMSM), the phase-locked loop (PLL) with frequency-adaptive filter has been widely employed. However, the issue of frequency coupling within the PLL is often overlooked. Through precise modeling of the PLL with frequency-adaptive filter, this paper reveals that such frequency coupling leads to PLL stability degradation and bandwidth limitation. To address this issue, a frequency-decoupled complex coefficient filter (FDF) is proposed, effectively eliminating frequency coupling within the PLL. Furthermore, building upon the FDF, a low computational burden robust hybrid complex coefficient filter (RHF) is developed using a cross-feedback network to overcome the contradiction between high dynamic performance and strong harmonic suppression in the filter. Consequently, the RHF enhances harmonic suppression without sacrificing dynamic performance, thereby improving the accuracy of position estimation and dynamic performance in sensorless control. Finally, the effectiveness of the proposed method is verified by experimental comparison with other methods on a 1.5 kW PMSM platform.