Finite-state model predictive control of NPC inverter using multi-criteria fuzzy decision-making

Abstract

Summary In this paper, the multiple objectives optimization problem in a finite-state model predictive control (FS-MPC) is formulated using fuzzy multi-criteria decision-making. Conventionally, to optimize the multiple objectives in a FS-MPC, the aggregate objective function or a single cost function is constructed, which requires the weighting factor tuning to select the appropriate switching state. Determination of these weighting factors for a particular objective function is a complex and time-consuming task as no systematic procedure is available in literature. The main aim of this paper is to replace the time-consuming task of weighting factor tuning by a simple and systematic procedure, which relies on the relative importance of the individual objective functions. The relative importance of objective functions derived from the expert's knowledge and the desired control objectives is used for appropriate switching state selection. The method is validated with the help of simulation results of a neutral point clamped inverter for the multiple objectives viz. reference current tracking, capacitor voltage balance and switching frequency minimization. The result outcomes of the proposed methodology are compared with the conventional FS-MPC and space vector pulse width modulation based current control schemes. Copyright © 2014 John Wiley & Sons, Ltd.