Nonlinear modeling of switched reluctance motor based on combination of neural network and genetic algorithm

Switched reluctance motor is very difficult to create a precise mathematic model for its nonlinear electromagnetism characteristic. But the optimal control of SRM needs an accurate mathematical model. Neural network has the ability of self-learning and can approach the result with any accuracy. However neural network´s ability of convergence and searching a global optimal solution is poor. Genetic algorithm is able to search for a global optimal solution in a short time, so an accurate result can be speedily obtained through combining neural network and genetic algorithm. In this article the method of construct a nonlinear modeling of SRM with the combination neural network and genetic algorithm is presented. The results of simulation show that through combining neural network and genetic algorithm the speed of convergence and the accurate of computation are improved. So it can be used in optimal control of SRM.