Estimation Procedure Based on Less Filtering and Robust Tracking for a Self-Sensing Control of IPMSM

Abstract

For the self-sensing control of interior permanent magnet synchronous motors (IPMSM), the rotor position and speed estimation based on high-frequency (HF) injection techniques generally employs two parts after the signal injection: signal processing/heterodyning process and tracking. This article proposes significant improvements for these two parts. First, the low-pass filters used in the classical heterodyning process of HF injection techniques are removed, which leads to the proposal of a novel heterodyning process in order to improve the estimation procedure in terms of delays, cost/complexity of implementation, and harmonics. Second, an independent machine parameters estimation based on step-by-step sliding mode observer using only the sign of the rotor position estimation error is developed. Thanks to this observer, the tuning of position, speed, and acceleration observer gains is achieved in a decoupled manner, which ensures states convergence separately in finite time. Third, the machine acceleration is estimated to improve the estimation procedure in transient modes. These contributions enhance significantly the estimation procedure in all operating ranges. A comparative study of the robustness/performance of the proposed strategy and classical tracking algorithms is carried out. Both simulations and experimental tests on an IPMSM rig are conducted in the framework of electric propulsion benchmark used in automotive applications, showing good agreement with the expected results.