Speed Control of PMSM with Finite Control Set Model Predictive Control Using General-purpose Computing on GPU

Abstract

In this article, novel approach in implementing finite control set predictive control is introduced. Algorithm is implemented using general-purpose computing on graphics processing unit. Predictions are computed using parallel threads on the GPU. Optimal switching state is then selected in dependence on the cost function given by angular speed error and constraints on the current. The algorithm is tested in the PIL simulation using Simulink and Jetson Nano. The ability of the algorithm to ensure the reference tracking and keeping the current within its limits are discussed.In this article, novel approach in implementing finite control set predictive control is introduced. Algorithm is implemented using general-purpose computing on graphics processing unit. Predictions are computed using parallel threads on the GPU. Optimal switching state is then selected in dependence on the cost function given by angular speed error and constraints on the current. The algorithm is tested in the PIL simulation using Simulink and Jetson Nano. The ability of the algorithm to ensure the reference tracking and keeping the current within its limits are discussed.