A method based on Imperialist Competitive Algorithm (ICA), aiming to mitigate harmonics in multilevel Inverters

摘要

Selective Harmonic Elimination Pulse Width Modulation (SHEPWM) is a well-known Switching Strategy which is applied in multilevel Inverters. The aim of this strategy is eliminating low order harmonics. Basically, harmonic equations are nonlinear and solving them is a major problem for researchers. Evolutionary Algorithms have shown an effective ability for this aim, because they find solution for all operating points of Modulation Indices. Recently, Continuous Genetic Algorithm (CGA) has been widely used in this area. In a research, it has been shown that Particle Swarm Optimization (PSO) has a better performance in comparison with CGA. In this paper, a new method, called Imperialist Competitive Algorithm (ICA), is introduced for this goal. The results of comparison between these methods show an appropriate privilege of ICA over other methods. The comparisons are based on probability of converging to global minimum. Effect of number of runs is investigated. The comparisons are done for 13 and 17 level inverters as case studies.