Model Predictive Control with Multi-Step Prediction for Single-Phase Inverters

Abstract

Voltage source inverters are widely used in photovoltaic power generation, uninterruptible power supply and other fields. Its control algorithm has a key impact on the inverter design. Model predictive control(MPC) is considered as an ideal control algorithm for single-phase inverters. Aiming at the time-delay problem of model predictive control, MPC with two-step prediction is proposed in this paper. However, two-step prediction only considers the optimal switching state of only one controlled cycle, so that it is conservative. To solve the conservative problem, MPC with multi-step prediction considering the optimal and suboptimal switching states of two cycles is proposed. Finally, simulation results of the proposed algorithms prove the effectiveness of two-step prediction and multi-step prediction.