Sensorless control of permanent magnet synchronous linear motor based on improved sliding mode observer

The sliding mode observer has been gradually used to the position sensorless control of permanent magnet synchronous linear motors, because of its good robustness to system uncertainties, but in practice there are two main problems: (1) the noise problem caused by the discrete switching characteristics of the sliding mode observer; (2) the observed back EMF also generates error harmonics, which ultimately leads to errors in speed and position estimation owing to the presence of motor positioning forces. To tackle the flaws of the traditional method, this paper puts forward the following optimizations: first, the sigmoid function is employed to replace the sign function, so as to handle the chattering problem; second, a synchronous frequency-extraction filter is introduced to take the place of the low-pass filter, which helps inhibition harmonics and decrease the estimation errors of speed and position. Simulation results have confirmed that the sensorless control method for permanent magnet synchronous linear motors according to the improved sliding mode observer proposed in this paper can not only effectively restrain harmonic errors but also significantly enhance the estimation performance of speed and position.