Neural network based torque control of switched reluctance motor for hybrid electric vehicle propulsion at low speeds

This paper presents a neural network (NN) based solution to reduce torque ripple of a switched reluctance motor (SRM) for hybrid electric vehicle (HEV) propulsion. Based on the high learning ability of NN, the NN controller learns off-line the non-linear torque-current-angle characteristic under twophase excitation, and finds an appropriate phase current profile for torque ripple reduction in real-time. Simulation results are presented to demonstrate that the proposed controller provides good dynamic performance with respect to changes in torque commands. The controller also satisfies the HEV propulsion requirements during starting.