University of Isfahan Computational Intelligence in Electrical Engineering 2821-0689 10 3 2019 09 23 Determining the Contribution of Harmonic Loads to Harmonic Contamination of the Power Network based on the Intelligent Classification of Measured Data Determining the Contribution of Harmonic Loads to Harmonic Contamination of the Power Network based on the Intelligent Classification of Measured Data 41 56 24102 10.22108/isee.2019.114825.1180 FA Mehdi Tadayon Faculty of Engineering, University of Isfahan, Iran Rahmatallah Hooshmand Professor, Faculty of Engineering, University of Isfahan, Iran Arash Kiyoumarsi Associate Professor, Faculty of Engineering, University of Isfahan, Iran Mehdi Torabian Isfahan Regional Electrical Company, Isfahan, Iran Journal Article 2018 12 31 Nowadays, determining the contribution of the individual consumers to the harmonic contamination is required for improving power quality. Due to changes occurred in lines, transformers, loads and generators over time, network operating conditions and impedance of different sectors are frequently varied resulting in the change in the contribution of harmonic sources. In this paper, an intelligent-based three-step method is developed to continuously determine the contribution of each harmonic source to the harmonic voltage of different buses without any access to the voltage phase data at different buses in the network. In this algorithm, firstly, the K-means clustering method is used for pre-processing of the measured data in order to reduce the background harmonic destructive effect on accuracy of the harmonic contribution determination. After calculating the harmonic contribution, the K-nearest neighbor method is used to generalize the results, and subsequently to create a continuous harmonic contribution matrix (HCM). Finally, the method is applied to a standard power network-based calculation example. The results demonstrate the capability of the proposed algorithm to evaluate the effects of harmonic sources in power networks. Nowadays, determining the contribution of the individual consumers to the harmonic contamination is required for improving power quality. Due to changes occurred in lines, transformers, loads and generators over time, network operating conditions and impedance of different sectors are frequently varied resulting in the change in the contribution of harmonic sources. In this paper, an intelligent-based three-step method is developed to continuously determine the contribution of each harmonic source to the harmonic voltage of different buses without any access to the voltage phase data at different buses in the network. In this algorithm, firstly, the K-means clustering method is used for pre-processing of the measured data in order to reduce the background harmonic destructive effect on accuracy of the harmonic contribution determination. After calculating the harmonic contribution, the K-nearest neighbor method is used to generalize the results, and subsequently to create a continuous harmonic contribution matrix (HCM). Finally, the method is applied to a standard power network-based calculation example. The results demonstrate the capability of the proposed algorithm to evaluate the effects of harmonic sources in power networks. https://isee.ui.ac.ir/article_24102_ce13fb7a21238307c980ddd57ed948d0.pdf