Improved Finite-Control-Set Model Predictive Control for Cascaded H-Bridge Inverters

Roh Chan,Sangshin Kwak

doi:10.3390/en11020355

Abstract

In multilevel cascaded H-bridge (CHB) inverters, the number of voltage vectors generated by the inverter quickly increases with increasing voltage level. However, because the sampling period is short, it is difficult to consider all the vectors as the voltage level increases. This paper proposes a model predictive control algorithm with reduced computational complexity and fast dynamic response for CHB inverters. The proposed method presents a robust approach to interpret a next step as a steady or transient state by comparing an optimal voltage vector at a present step and a reference voltage vector at the next step. During steady state, only an optimal vector at a present step and its adjacent vectors are considered as a candidate-vector subset. On the other hand, this paper defines a new candidate vector subset for the transient state, which consists of more vectors than those in the subset used for the steady state for fast dynamic speed; however, the vectors are less than all the possible vectors generated by the CHB inverter, for calculation simplicity. In conclusion, the proposed method can reduce the computational complexity without significantly deteriorating the dynamic responses.

Highlights

Multilevel converters generally consist of power switch elements and DC voltage sources such as independent sources or capacitors, which enables the synthesization of output voltage waveforms with several steps
This paper proposes an finite-control-set model predictive control (FCS-MPC) algorithm with reduced computational complexity and fast dynamic response for cascaded H-bridge (CHB) inverters, in which different candidate vector subsets to search for an optimal voltage vector are developed for the respective steady and transient states
Because the proposed determinant algorithm is based on voltage vectors determined in the αβ plane, the distinction between steady state and transient state is free from switching ripple components and noise

Summary

Introduction

Multilevel converters generally consist of power switch elements and DC voltage sources such as independent sources or capacitors, which enables the synthesization of output voltage waveforms with several steps. Calculates all the resulting voltage vectors from all the possible switching states to regulate the load currents of the converters This basic principle of the FCS-MPC method, which predicts the step behaviors using all possible voltage vectors, results in a problem of computational complexity for multilevel converters with a high number of voltage levels. This paper proposes an FCS-MPC algorithm with reduced computational complexity and fast dynamic response for CHB inverters, in which different candidate vector subsets to search for an optimal voltage vector are developed for the respective steady and transient states. The number of voltage vectors is reduced by eliminating the redundant vectors, as indicated, the MPC-conv method considering all non-redundant voltage vectors generated by the multilevel CHB inverter still suffers from a rapidly increased computational load as the voltage level increases. This is due to the limited candidate voltage vectors in the search process for an optimal vector

Proposed Predictive Control Method for Single-Phase Inverter

Simulation and Experimental Results

Findings

Conclusions