Harmonic optimization in single DC source multi-level inverters using RBF neural networks

This paper presents a comparison between heuristic based and exact solution methods to find optimized harmonics for single DC source multilevel inverters. The methods have the benefit of high rate of convergence and precision compared to other conventional optimization methods. The proposed technique can be applied to single DC source multilevel inverters with any number of levels. The goal of optimization is to eliminate some low order harmonics and to maintain the fundamental component at the desired value. As a case study, the method is implemented and tested on a 15 level inverter. Simulation results show the effectiveness and flexibility of the proposed method.