1263

This paper proposes an improved sliding mode observer based on super-twisting algorithm (STA-SMO) for sensorless control scheme of permanent magnet synchronous motors. An α-β axis stator current observer is built according to the back electromotive force (EMF) model after the back EMF error signal is obtained. The chattering phenomenon inherent in the traditional STA-SMO is eliminated by restricting the equivalent disturbance term to a small range, the coefficient of observer can be kept constant with speed variations. Meanwhile, because the mismatch between actual and set resistance associated with stator temperature may lead to a large estimated error and even system instability, a parallel stator resistance online estimation scheme is presented based on modified SMO.The stability of the online stator resistance estimator is proved by Lyapunov function. Compared with the traditional STA-SMO, the improved STA-SMO not only improves the accuracy of rotor position observation, but also effectively avoids chattering. Finally, the effectiveness of the proposed improved sensorless control scheme is verified by computer simulation.