Position Sensorless Control Method of Synchronous Reluctance Motor based on Spherical Unscented Kalman Filter

The position estimation method based on extended Kalman filtering has the advantages of high estimation accuracy and high disturbance immunity. However, the magnetic saturation and cross-coupling effects of the motor make a highly nonlinear mathematical model, which leads to the non-negligible linearization error when applying it to a digital position sensorless control system of the synchronous reluctance motor (SynRM). In this paper, the SynRM position estimation method based on spherical unscented Kalman filtering (SUKF) is proposed, which avoids the need to solve Jacobian matrices and achieves higher estimation accuracy for the high-dimensional SynRM mathematical model. The computation volume of the proposed observer is minimized while maintaining the estimation performance by introducing the spherical unscented transformation algorithm. The feasibility is verified by simulation and experiment.