Application of neural networks for compensation of torque ripple in high performance PMSM motor

In this paper an application of artificial neural controller for compensation of torque ripple in electrical drive with PMSM (Permanent Magnet Synchronous Motor) is presented. Due to cogging torque, electromagnetic torque ripple and also, current measurement error of the motor, are created a ripple of angular velocity, finally. This parasite phenomenon influence on the precision of control, especially in machine tool drives and robots. Therefore, in the drive systems the aim is to ensure a smooth electromagnetic torque in the motor. These unfavorable phenomenon can be successfully compensated by an artificial neural networks, what is presented in the article. In the work, the structure of the neural network was proposed and its learning method in order to reduce the torque ripple. The collected the results of simulation and experimental tests are confirmed by proper operation of the control system.