A novel memetic algorithm approach for selective harmonic elimination in multi-level inverters

Harmonic reduction issue is regarded as the significant challenge in multi-level converters. SHE is a harmonic reduction technique that aims to control harmonic content of multi-level inverters by eliminating of the lowest harmonics while keeping the magnitude of the fundamental component at its desired value. To achieve this goal a set of nonlinear transcendental equations with given certain boundary constraints should be solved and optimal switching angles would be computed. This paper introduces Memetic algorithm (MA) as a powerful tool applied to this complicated problem. The MA proposed in this paper, is a combination of Particle Swarm optimization (PSO) and Mesh Adaptive Direct Search (MADS) to gain well ability in both local refinement and global exploration. Convergence analysis of MA is also studied. Furthermore, this algorithm has been tested successfully on a three phase cascade multi-level inverter. Both simulation and Experimental results validate effectiveness of the proposed method.