Neural network based harmonic elimination in single-phase inverters

Abstract

Purpose This study aims to present the experimental results for neural network (NN) based harmonic elimination technique for single-phase inverters. Design/methodology/approach Switching angles applied to power switches are determined using the NN technique based on the harmonics to be suppressed. Thus, besides controlling the output voltage, NN controller provides elimination of predetermined harmonics from output signal of single-phase inverter. Simulation and experimental results for the elimination of 15 and 20 low-order harmonics are presented. The switching angle values calculated by a NN , fuzzy logic and Newton–Raphson are compared for elimination of first 10 harmonics. Findings This paper provides the harmonic spectra showing that first 15 and 20 harmonics are suppressed from output signal. The NN is proved to give closest results to angle values calculated by Newton–Raphson’s numerical solution method. Originality/value The value of this paper is to verify the simulation results with the experimental result for the elimination of 15 and 20 low-order harmonics. Both the simulation and the experimental results demonstrate the success of the NN based selected harmonic elimination.