Estimating harmonic impact of individual loads using multiple linear regression analysis

Abstract

This paper investigates the problem of how to quantify the harmonic impact of several known harmonic-producing loads on a particular location of a power network. A multiple linear regression (MLR) analysis-based technique is proposed to solve the problem. This technique overcomes a major limitation of the correlation-based methods: the difficulty of finding valid data sets when the number of suspected loads increases. As a result of addressing this limitation, all measured data can be used for harmonic impact study with the proposed method. The proposed method has been verified through simulation and field case study. Additionally, sensitivity study is carried out to address some practical considerations. The results show that the proposed technique is a reliable tool to quantify the individual harmonic impact of several large harmonic-producing loads. Copyright © 2015 John Wiley & Sons, Ltd.