Nonlinear Flux Observer for Sensorless Control of IPMSM Based on a Novel Regression Model

The nonlinear flux observer is an effective position sensorless method that has been proven in practice. For the interior permanent magnet synchronous motor (IPMSM), this article establishes a novel regression model of the IPMSM, which is the product of one known vector and unknown vector. Based on the established model and gradient method, an improved nonlinear active flux observer is proposed to estimate the position of the IPMSM, which is essentially a sensorless method. The proposed observer considers the existence of dc bias through the flux compensation term, so it is robust against dc offset of the integrator input. In addition, it has a smaller computational time and is suitable for the IPMSM drives. The observer performance is verified and investigated on an IPMSM platform with a three-level neutral-point-clamped inverter under finite control set model predictive control method.