Harmonic Detection of Electric Current using Machine Learning: A Literature Review

Abstract

The increased use of nonlinear devices makes distortions in the form of current and voltage waves unavoidable in power grids. As a result, harmonics, sub-harmonics and inter-harmonics often occur in the current and voltage spectrum. Harmonics cause many problems such as overpower loss, overvoltage, imbalance, flicker, relay miss operation, etc. Harmonic distortion of electric current can be measured directly in harmonically contaminated systems. Various standards have already been established in different measurement procedures and many other techniques have been proposed for analysing, characterizing and classifying power quality data. Extensive power quality monitoring is often recommended for industrial and distribution systems. Harmonic measurements are usually performed as part of more general power quality measurements to gain insight into the distribution system's state. Detection of electric current harmonics using machine learning has been widely done and provides excellent results. This study is based on a literature review of research journals on machine learning studies that detect electric current harmonics in electric power system networks.