Rapid identification of total demand distortion using a neural network model and mitigation via an integrated filter design

Abstract

This paper proposed a filter for the mitigation of power harmonics based on an integration of filters namely doubletuned plus C-type filter (DTPC). The proposed DTPC filter is mainly aimed to filter the total demand distortion (TDD), as well as the total harmonic distortion (THD), based on IEEE-519 standards. The harmonic filter is presented within the framework of harmonic mitigation as a method of power quality control. Besides, a neural network estimation model to identify harmonic percentage in the power system is also proposed. Two modelling schemes are presented for the simulation of the harmonic filter, which are the load modelling and the source modelling, using the neural network technique. The load modelling is a scheme to predict the current distortion, while the source modelling is a scheme to predict the voltage distortion at the point of common coupling. These two methods may work as standalone tools at the customer's side, thus, it will not interfere with the online operation of the customer's power supply system. The load and source modelling are combined with the DTPC filter in mitigating both the THD and the TDD effects. As a result, the DTPC filter allows customers to maintain a THD and TDD percentage below 10%, and hence customers could meet the IEEE-519 standard.