Optimum number of cascaded multilevel inverters for high-voltage applications based on Pareto analysis

Abstract

High power megawatt (MW) scale drives and power supplies are becoming more prevalent in industrial applications due in part to the development of the Cascaded Multilevel Inverter (CMI) topology. This paper examines the selection of the number of levels for a particular application by using Pareto efficiency analysis to optimize the number of cells with respect to cost, quality, and reliability. This study has been accomplished by designing a general Model Predictive Control (MPC) as a feedback controller which can be employed for any number of levels. MPC is used in the evaluation of the performance metrics to ensure that each design operates optimally. A discrete-time model of the CMI along with a model of the load is used to predict the future behavior of the inverter output currents. The MPC controller uses that prediction along with a set of multi-objective control variables all in one cost function to produce the optimal switching signals. Three-phase AC output current and common mode voltage (CMV) are considered together when designing the controller. MATLAB simulations are presented to validate and implement these concepts.