Effective Position Error Compensation in Sensorless Control Based on Unified Model of SPMSM and IPMSM

Sliding-mode observer (SMO) has attracted extensive attention in the field of medium- and high-speed sensorless control of permanent magnet synchronous motor (PMSM) because of its strong robustness and stability. However, the traditional methods are vulnerable to dc bias caused by measurement errors and parameter changes. Therefore, in this article, an improved SMO algorithm by disturbance observer compensation is proposed. The algorithm unifies the mathematical models of surface PMSM and interior PMSM. Besides, a bandpass filter (BPF) is used to replace the traditional low-pass filter, so it can effectively suppress dc bias and high-frequency noise. In addition, at any BPF center frequency, the proposed disturbance observer with low-pass filter (LPF) characteristics can perfectly compensate the position error caused by the digital filter in real time. Moreover, through sensitivity analysis, the influence of model uncertainty on the observation position is studied, and an adaptive extended state observer is added to mitigate the influence of parameter mismatch on the performance, hence improving the estimation accuracy. Finally, a triple redundant permanent magnet-assisted synchronous reluctance motor is taken as an example to verify the feasibility and effectiveness of the proposed observer.

Automated summary

In this article, an improved sliding-mode observer algorithm with disturbance observer compensation is proposed for medium- and high-speed sensorless control of permanent magnet synchronous motors. The algorithm unifies the mathematical models of surface PMSM and interior PMSM, and uses a bandpass filter to effectively suppress dc bias and high-frequency noise. It also incorporates a disturbance observer with low-pass filter characteristics to compensate for real-time position errors caused by the digital filter. An adaptive extended state observer is added to mitigate the influence of parameter mismatch on the performance and improve estimation accuracy. The feasibility and effectiveness of the proposed observer are verified using a triple redundant permanent magnet-assisted synchronous reluctance motor as an example.