Adaptive Observer Design for Sensorless IPMSM Drives With Known Regressors Variant of Extended EMF Model

This article proposes a known regressors variant of the extended electromotive force (EMF) model, enabling rigorous theoretical stability analysis about the joint estimation of position, speed, and stator resistance in the full-order observer design. According to the literature, adaptive observer designs for nonlinear systems often require the regressors to be known as well as the systems being in Brunovsky observer form. Unfortunately, none of the existing interior permanent magnet synchronous motor (IPMSM) models (i.e., the extended EMF model and the active flux model) are capable of meeting these requirements. By defining a new state variable, the extended EMF model of IPMSM can be transformed into its known regressors variant, thereby, ensuring global stability of the speed and stator resistance adaptive observer. In the new coordinate, the inclusion of stator resistance estimation leads to the overparameterization problem, thus an adaptive high-gain observer is proposed to allow certain nonlinear terms to appear in the regressors. Analyses of stability and parametric sensitivity are also presented, resulting in a robust tuning guideline for the sensorless control system. Simulations and experiments have been conducted to verify the effectiveness of the proposed adaptive observer in a sensorless IPMSM drive, and improvements have been observed in wide-speed operations and disturbance rejections.