Fixed switching frequency strategy for finite-control-set model predictive control based on cost function reconstruction

Abstract

Considering the spread spectrum problem of finite-control-set model predictive control (FCS-MPC), a fixed switching frequency strategy based on cost function reconstruction is proposed in this paper. It achieves a fixed switching frequency by selecting the optimal voltage vector through the redefined cost function. A frequency coefficient is added to this to change the selecting area of the optimal voltage vector. The proposed strategy reshapes the spectrum of the inductor current and focuses the harmonics near the switching frequency so that the design of the output LC filter is easier. Without adding any modulators or greatly increasing the number of computations, the proposed strategy reserves the excellent dynamic performance of the FCS-MPC and is convenient in terms of implementation. Finally, a 1 kW single-phase inverter is built and experiments are conducted. Experimental results demonstrate that the proposed control strategy realizes a fixed switching frequency. In addition, it has a comparative dynamic response speed and number of calculations when compared to dead-beat-based FCS-MPC. These features are beneficial to its application.