PRECISE MODEL OF A CLASS OF SWITCHED RELUCTANCE MOTORS BASED ON NEURAL NETWORK DESCRIPTIONS

It is well known that switched reluctance motor (SRM) drives are much more energy-efficient than the comparable vector induction motor drives. That means less energy consumption which brings a positive impact on the environment. This paper deals with precise mathematical models of a class of SRMs based on artificial neural network descriptions. This nonlinear highly adequate to reality models description approach permits, on the one hand, high precision simulation estimations of the designed electrical drive system, and on the other implementation of these models as virtual machines which work as variable observers or as reference models running simultaneously with the real machine. The final purpose of this paper is to generalise and refine this approach of models creation, validating them over the whole class of SRMs.