Finite Set Model Predictive Control of PMSM drives with LC filter using dynamic weighting factor assignment

Abstract

This paper proposes a novel method to assign the weighting factor values online for a PMSM drive outfitted with an LC filter in model predictive control. This aforementioned system has six independent variables that are to be controlled. Therefore the cost function at the core of the predictive controller should optimize for all six variables, such that the resulting control law provides acceptable performance metrics. This can be achieved with weighting factors, that determine the importance of each variable of the cost function. In this paper we propose the dynamic assignment of weighting factors, where the cost of a given variable is calculated based on the set of predictions for the same variable. This technique can be utilized to determine weighting factor values for a given system or even improve the performance metrics compared to methods using constant weighting factors. Simulation results show the method in practice.