Direct Torque Control of a Permanent-Magnet Synchronous Motor with Neural Networks

Permanent-Magnet Synchronous Motors (PMSM), characterized by high power density and efficiency, domain most of the applications that require high performance in control of position, speed, and torque. Direct Torque Control (DTC) is one of the most used techniques to control these motors. In this work an intelligent approach is proposed for the realization of the DTC for a PMSM by means of Artificial Neural Networks (RN). In addition, a comparative between the response obtained with conventional DTC (DTCC) and the neural network based DTC (DTCRN) is presented. The training data for the neural networks consisted in the set of values obtained with the DTCC. Both systems were implemented and simulated with Matlab / Simulink.