Globally Converging Observers for SPMSM Sensorless Control

This paper proposes a kind of globally converging observer for sensorless control of surface mount permanent magnet synchronous motors (SPMSM). The observers are based on rotor flux-oriented synchronous coordinates, and can converge to the real rotor position from any unknown initial positions. The observers can be performed by extended Kalman filter (EKF), state observer or model reference adaptive system (MRAS). By some compensation in q-axe equation, the unexpected equilibrium points can be eliminated and then observers can converge to the unique expected equilibrium point. Then the motors can start up from any unknown initial positions. Simulation and experimentation results have verified the effects of the method.