The Nonlinear Modeling of SRM with Improved RBF Network

Abstract

The modeling of inductance of SRM (switched reluctance motor) using RBF (radial basis function) network can resolve the nonlinear problem of SRM effectively. But traditional RBF network easily bring problems such as local minimum value and slow convergent speed. In order to resolve these problems, two aspects of improvements with RBF network are given from structure and algorithm in this paper. Additionally, current PWM is regulated dynamically with inductance model. It is proved that the improved RBF network can achieve good convergent speed and control effect.