Journals Proceedings

Abstract

This Paper aims to extend the knowledge about the performance of different cascaded multilevel inverter through harmonic analysis. Large electric drives and utility require advanced power electronics converter to meet high power demands. As a result, multilevel power converter structure has been introduced as an alternative in high power and medium voltage situations. A multilevel converter not only achieves high power ratings but also improves the performance of the whole system in terms of harmonics, dv/dt stresses and stresses in the bearing of the motor. Despite of various advantages seen in multilevel inverter the total harmonic distortion has always been a matter of concern. In the presented work a neural network based model is proposed for reducing the total harmonics distortion from multilevel inverter output. For this neural network is trained to perform selective harmonic elimination. The training data is obtained by solving non-linear equation obtained from selective harmonics analysis using genetic algorithm technique. The trained neural network based controller is shown to decide the notches angle for given modulation index. The proposed model is simulated on MALAB Simulink platform. Experimental results are obtained for different modulation index and the performance of the multilevel inverter. The results are further analyzed in terms of total harmonic distortion (THD).