PWM-Based Optimal Predictive Direct Torque Control of Switched Reluctance Machine

This paper proposes a novel PWM-based predictive torque controller for Switched Reluctance Machines (SRM). The phase currents are predicted using a simplified flux linkage model, and phase torques are predicted from static-torque characteristics. Both predicted currents and phase torques are then expressed in incremental duty ratio. A cost function to minimize the primary objective of instantaneous torque error and the secondary objective of RMS currents is then solved online to obtain the optimal duty ratio at every switching period. The impact of weight given to the RMS currents in the cost function is studied through experiments and simulation. An appropriate weight to ensure acceptable torque performance and RMS phase currents is selected. The proposed controller is validated through experiments and simulations on a 4-phase, 8/6 pole SRM at different operating conditions.