Sensorless Control of a Surface-mounted PMSM by an Improved Flux Observer with RLS Parameter Identification Method

To eliminate the effect of parameter variation on the sensorless performance of surface-mounted permanent magnet motor (SPMSM) using a flux observer, the parameter sensitivity of the conventional flux observer is assessed in this paper. The simulation results showed that the position estimation error of the flux observer considerably depends on the accuracy of the stator resistance and rotor flux linkage, and they must be updated during motor operation. Therefore, this paper proposes an online parameter identification by recursive least square (RLS) method based on the discrete-time model of SPMSM in the stationary reference frame (RF). The simulation results demonstrated that estimated parameters follow the variation of the actual parameters very well. Furthermore, the performance of three discrete-time speed and position estimation methods in αβ RF, including a sliding mode observer (SMO), a conventional flux observer, and a flux observer improved with the proposed RLS parameter identification, are compared. The simulation results prove that not only is the flux observer improved with the proposed RLS parameter identification as robust as the SMO against parameter uncertainties, but also it has a lower position estimation error at the SPMSM rated speed.