Extending the Convergence Region of Saliency-Based Rotor Position Estimation for Heavily Saturated PMSMs

In position-sensorless control techniques, saliencybased methods are commonly employed to determine the rotor position of synchronous machines in the low speed range, including standstill. The accuracy of these methods is affected by saturation and cross-coupling effects, leading to a load-dependent estimation error. In the case of heavily saturated motors, often encountered in automotive applications, rotor position tracking may even diverge, requiring the implementation of a torque limit during position-sensorless operation. A convergence analysis presented in this paper shows that the approach to compensate for the load-dependent angle error can influence the convergence of the estimation. Building on this insight, an angle error compensation method is proposed that significantly improves the convergence behavior. As a result, the robustness and overload capability of position-sensorless control can be substantially enhanced. Experimental results are provided to validate the effectiveness of the proposed method.