Estimation of the External Performances of PMSM Using Artificial Neural Networks Based on FE Results
Résumé
A method that combines an artificial neural network (ANN) and the finite element method (FEM) is introduced for estimating the instantaneous electromotive force (EMF) of surface mounted permanent magnet motors. Using parallel supervised multilayer neural networks, the geometrical parameters of the motors are mapped to the phase EMF. The ANN is trained using the FEM data. The ANN model is then used in an iterative designing process to achieve high torque development capability with low magnet volume.