Harmonic Detection Based on Artificial Neural Networks for Current Distortion Compensation

In this paper a method for the determination of part of the current harmonic components for the selective compensation harmonic by single-phase active power filter is presented. The non-linear load is composed by an AC controller with variable resistive load. The first six components are identified through artificial neural network. The effectiveness of the proposed method and its application in the single-phase active power filters with selective harmonic compensation are verified. Simulation results are presented to validate the proposed approach.