Simulation of Predictive Control Based ANN for PMSM Servo System

In servo system driven by permanent magnet synchronous motor(PMSM),the control strategy of motor is vector control mostly.Parameter vibration,disturb of load torque have remarkable influence on dynamic character of motor.Because of lagging of control strategy,controller can′t produce ideal output by using characteristic of motor.Predictive control strategy based artificial neural networks,was proposed for PMSM servo system.The architecture and parameter are acquired by learning offline and modification online.It can be used to predict characteristic of PMSM.Optimal controller can produce more ideal output waveform by using input output of neural networks of PMSM.The control method is verified by simulation based matlab Simulink.The results of simulation indicate that predictive based ANN has better characteristic and robust,compared with traditional control strategy.