Hybrid Switching of Four-Voltage-Vector Model-Free Predictive Current Control for Four-Switch Three-Phase Inverter-Fed SynRM Drive Systems

Abstract

Conventional model-based predictive current control suffers from the common drawbacks of high reliance on system model parameters and the use of a single input voltage vector, which result in large pulsating current ripples and prediction errors. In the case of a four-switch three-phase inverter (FSTPI) topology, the implementation of the predictive controller is exacerbated due to the limited number of candidate voltage vectors. This article presents an integrated model-free predictive current control with a hybrid switching mechanism to solve the problem. The proposed method introduces the combined switching mechanism of input voltage vectors with fixed and variable modulations by increasing the number of switching voltage vectors. The four basic voltage vectors generated in the FSTPI create 24 new synthesized voltage vectors through fourfold linear expansions of the space vector plane. The switching durations of input voltage vectors are determined by calculating their optimal duty ratios. As a result, the proposed method improves the prediction accuracy by increasing the iteration calculations of current differences every sampling period. The proposed method, known as the hybrid switching of four-voltage-vector model-free predictive current control, is practically tested via simulation and experimental works to evaluate its effectiveness and performance improvement.