Torque Ripple Suppression of PMSM Speed Regualtion System Using Neural Network

In this paper, a neural network-based torque compensator is combined with a third order ESO in ADRC system to suppress the torque ripple. The third order ESO can effectively reject the low frequency disturbance, and the NN can output a feedforward compensation to optimize the torque and smooth the speed. Moreover, this NN compensator has a simple structure, which significantly reduces training cost time.