Abstract

This paper proposes a predictive torque control strategy based on look-up tables with improved delay compensation to enhance prediction capabilities. The tables or static maps contain the flux linkage and torque characteristics of the machine, which provides higher prediction accuracy compared to approximated analytical models. The delay compensation is enhanced by including a Kalman filter stage. The algorithm is validated through simulations and experiments using a 5.5 kW, 12/8 SRM. The results evidenced improved torque sharing capabilities with respect to conventional methods and other predictive control strategies by offering a trade-off between torque ripple production, average torque and average torque per ampere.