Effective Mitigation of Harmonics in Cascaded Multilevel Inverters: A Hybrid and Adaptive Technique

In recent years, cascaded multilevel inverters find more attention in the areas of distributed energy resources in order to connect batteries, micro turbines, fuel cells and solar cells so as to feed a load or the AC grid. In the multilevel inverters, harmonics has to be reduced effectively. Here, an efficient hybrid technique is proposed to reduce the harmonics by optimally selecting the switching angles for n-level cascaded multilevel inverters. The technique exploits neural network and genetic algorithm with adaptive mutation that determines the optimal switching angles not only with minimum total harmonic distortion but also with reduced computational time. As the genetic algorithm performance relies on the parameters such as population size, crossover rate, mutation rate and number of generations, the neural network determines the best parameters. By utilizing the obtained best parameters, the genetic algorithm determines the optimal switching angles for the cascaded multilevel inverter. As the genetic aalgorithm performs adaptive mutation, quick convergence tosolution is achieved and so the optimal switching angles are obtained in very less computational time. The implementation results show that the proposed hybrid as well as adaptive technique is effective in reducing the harmonics of the cascaded multilevel inverter.