Novel topologies for identification and control of PMSM using artificial neural network

Rotor speed and performance of permanent magnet synchronous motor (PMSM) suffers from accuracy due to variation of motor parameters such as stator resistance, stator inductance or torque constant. The conventional linear estimators are not adaptive. Neural networks (ANN) have shown better results when estimating or controlling nonlinear systems. In this paper an artificial neural network based high performance speed control system for a PMSM with different topologies and their performance comparison is presented. The main purpose is to achieve accurate trajectory control of the speed, when the motor and load parameters are unknown. The PMSM motor was identified using three different topologies (speed, voltage and current). The unknown nonlinear dynamics of the motor and the load are captured by the ANN. The performance of the identification and control algorithm are evaluated by simulating them on a typical PMSM motor model.