Torque ripple minimization in PM synchronous motors based on backstepping neural network

Abstract

The electromagnetic torque introduces ripples into the system due to the nonlinearity of the permanent magnetic field and the change of the stator winding, which generate speed oscillations and deteriorate the performance of PM synchronous motor (PMSM). In this paper, we have given the numerical description of the ripples and proposed backstepping Neural Network (NN) control scheme to minimize them. First, online weight-tuning based NNs compensate the torque nonlinearity and external disturbances with its approximation characteristic, and make them satisfy the linear parameter requirement of backstepping; Second, backstepping is applied based on the model of PMSM; At last, the comparison simulation with PID are added to demonstrate the effectiveness of the proposed method.