Torque ripple minimization with on-line parameter estimation using neural networks in permanent magnet synchronous motors

A neural network based estimator for torque constant and stator resistance in permanent magnet synchronous motors (PMSM) is presented. The neural weights are initially chosen randomly and a model reference algorithm adjusts those weights to give the optimum estimations. The q-axis inductance is modeled offline according to q-axis stator current. The neural network estimator is able to track the varying parameters quite accurately at different speeds with consistent performance. The neural network parameter estimator has been applied to slot and flux linkage torque ripple minimization of the PMSM. The drive system is insensitive to these parameter changes. Simulation results justifying the claim are presented.