Computationally Efficient Predictive Current Control With Finite Set Extension Using Derivative Projection for IM Drives

Abstract

This article proposes a computationally efficient predictive current control (PCC) combined with an extension of a finite set using current derivative projection, to solve the challenges experienced by finite control set PCC (FCS-PCC), that is, unsatisfied steady-state performance and high computational burden. Since only one switching sequence is utilized in the entire sampling period, FCS-PCC is inevitably penalized by the high torque and current deviations. More specifically, FCS-PCC suffers from the high computational burden caused by the exhaustive search in the optimization stage. To tackle the aforementioned issues, a reformulated objective function using current derivative projection with least-squares (LSs) optimization in PCC is presented in this work. First, PCC is geometrically described as a quadratic programming (QP) problem. To minimize the deviation between the selected and desired current derivative, the objective function is rearranged as the quadratic Euclidean norm of the derivative deviation. The exhaustive search in the optimization stage is avoided by a preselection principle. Based on the above, the optimal stator current derivatives in the consecutive sampling intervals combined with their duty cycles are optimized by the LS method. The effectiveness of the proposed method is verified by the experimental results based on a 2.2-kW IM drive platform.