Mathematical Modeling of Flux-Linkage Characteristics of Switched Reluctance Motors Using Polynomial Neural Networks

Switched reluctance motor (SRM), built using revolutionary concept, has highly nonlinear flux-linkage characteristics depending heavily on phase current and rotor position. A good mathematical model for these characteristics would help to understand the workings of the motor; thus providing path for better control algorithms and motor designs. It is very much proven by many researchers in this area, that a straightforward simple mathematical model has never satisfied the complete overall characteristics. Moreover, there is no distinct guideline about what sort of mathematical model would be suitable. To overcome this problem, a self-organizing polynomial neural network is proposed in this paper. In this scheme, without any prior knowledge of the mathematical model, the model is evolved iteratively and progressively. The simulation test results verify the effectiveness of this approach.