Harmonic Optimization of Diode-clamped Multilevel Inverter Using Teaching-learning-based Optimization Algorithm

This paper proposes a harmonic elimination technique for online reducing harmonics in voltage source inverters. In the presented method, in order to provide required fundamental voltage and eliminate specified harmonics simultaneously, the value of the switching angles are determined using a new efficient optimization algorithm named Enhanced Teaching-learning-based Optimization (ETLBO). The efficiency of the presented strategy is validated on three-phase seven-level diode-clamped inverter while other structures of multi-level inverters (MLIs) can be employed. The aim is to reduce the Total Harmonic Distortion (THD) in multilevel converter. In this paper, a new optimization algorithm named ETLBO has been successfully implemented to achieve optimal switching angles in order to harmonic elimination in MLIs. A three-phase seven-level diode-clamped MLI has been selected as a test case to validate the effectiveness of the proposed method. The simulation results demonstrated the ETLBO method not only is in preference to the Genetic Algorithm and Particle Swarm Optimization for including no user-defined parameter and simple implementation, but also for providing higher quality solutions. Moreover, in order to verify the superiority of the ETLBO algorithm over conventional methods for harmonic elimination, the Newton-Raphson (NR) method was implemented to find optimal switching angles. The results obtained by the NR method compared with those of the ETLBO and the superiority of the ETLBO in terms of the both computational time and the resulted THD was concluded. As a whole, the results show that the ETLBO algorithm is expected to become widespread in power systems where on-line updating harmonics is needed since it provide accurate and high-quality solution in extremely short time.