Artificial Intelligence based Torque Ripple Minimization of Switched Reluctance Motor Drives

The Switched Reluctance Motors (SRMs) are designed for many applications. They have wound field coils for the stator windings. Their rotor however has no magnets or coils attached. Their main drawback is the torque ripple. The torque ripple minimization is reduced in this paper using Torque Sharing Function (TSF) technique for four quadrants operation of the motor. The Artificial Neural Network (ANN) is used to translate the torque reference to suitable current. With the last advance of microcontrollers and DSPs, the implementation of the ANN becomes less complex and inexpensive. This simplifies the controller and cuts down its cost. For accurate modelling and high effective results, the authors use the Finite Element Analysis (FEA) for obtaining the magnetization characteristics of the motor. The SRM is modelled and simulated based on Matlab/Simulink. The torque ripples obtained by the proposed model are kept minor for the various operation modes of the motor.