A new approach to commutation in predictive control of switch reluctance motor

This paper provides a model predictive approach to control switched reluctance motors (SRM's). A local linear neuro-fuzzy model is used to model SRM. Then a predictive control schema is devised considering an appropriate energy term in the optimization phase. Commutation occurs naturally as an outcome of the predictive control design process, not as an extra step added to the control policy. From a computational view point, we use locally linear model predictive control that with a quadratic cost and linear constraints reduces to a simple quadratic program, which can be solved very fast in a closed form. Simulation studies justify applicability of our proposed method to SRM applications.