Computationally Efficient Model Predictive Torque Control of Switched Reluctance Motor Drives

This paper proposes a computationally efficient Finite Control Set-Model Predictive Torque Control (FCS-MPTC) algorithm to reduce the torque ripple in switched reluctance motor (SRM) drives. The computational burden is minimized in two ways, one by using a single look-up table based technique for estimating the torque and the other by reducing the switching vectors to be evaluated at any instant to a maximum of three. The single LUT-based torque estimation is done using k(θ, i) profile that gives a non-linear modulating factor to be multiplied with the linear torque. Furthermore, a controlled demagnetization of the outgoing phase is realized during the magnetization of the incoming SRM phase for the entire overlapping region thereby reducing the torque ripple. The proposed model is verified through MATLAB/Simulink on a four-phase 8/6 SRM. Using the proposed method, a torque ripple of ≈ 6% has been achieved.