Optimal torque sharing in direct instantaneous torque control of switched reluctance motors

In this paper, optimal waveforms to control constant instantaneous torque of switched reluctance motors while minimizing copper losses are computed by using the pattern search as an optimization algorithm. Optimal torque sharing using two quadratic and one linear functions is proposed to approximate the optimal torque waveforms accurately. The optimal torque sharing was verified with a 3.75 kW, 3820 rpm, three-phase 12/8 SRM compared to predictive PWM-based direct instantaneous torque control. Theoretical calculations validated that the optimal torque sharing approximates the optimal waveforms accurately and that it reduces copper losses and peak currents. Experimental results demonstrated that the optimal torque sharing additionally reduces torque ripple, acoustic noise, and on-line calculation time.