ANN control of Nine Level NPC Voltage Inverter Based on Selective Harmonics Elimination

This paper is devoted to artificial neural network (ANN) control of a nine levels neutral point clamped (NPC) voltage inverter based on the selective harmonics elimination. Initially, we determine all the adequate switching angles for the control of the nine levels inverter switches, in order to eliminate the 5th, 7th and 11th harmonics. Then, a multilayer ANN is elaborated, which is able to reproduce these angles according to modulation index variations. The aim of t his technique is to improve the real-time control of the multilevel inverter.