Predictive Current Control of Switched Reluctance Machine for Accurate Current Tracking to Enhance Torque Performance

Accurate tracking of reference current wave shapes is essential to obtain the desired average torque, as well as minimize torque ripple and losses of switched reluctance motor (SRM). This article proposes a predictive control technique for accurate tracking of the current reference in an SRM drive based on a simplified piecewise linear flux-linkage model. The proposed technique eliminates extensive lookup tables and requires only a small set of model parameters, resulting in efficient memory utilization. The performance of the proposed method is evaluated for two different current shapes at various operating speeds and load torques. Simulations and experiments show that the proposed technique offers better or comparable performance over state-of-the-art lookup table-based gain-scheduling PI, sliding mode, and two existing predictive control methods. Further, the proposed technique offers the lowest torque ripple among all the above controllers. Since predictive control is based on machine models, any error in modeling can potentially lead to deterioration in tracking and instability. It has been theoretically proved that the proposed controller ensures stability even with modeling errors as high as 200% in incremental inductance. Additionally, simulation results are presented to show the effect of varying three crucial model parameters on the current tracking performance of the proposed controller. The model parameters are varied within a range of -30% to +50% of their nominal values at three different speeds. The experimental results confirm very good current tracking performance even in the presence of large modeling errors.