Fault Detection of a Medium Voltage Cable Joint using Support Vector Machine Algorithm
M.Izadi1, M. Tolou. Askari2, M.Z.A.Ab Kadir3, M.Osman4, M.Hajikhani5

1M.Izadi, Institute of Power Engineering (IPE), Universiti Tenaga Nasional, Jalan IKRAM-UNITEN, 43000 Kajang, Selangor, Malaysia; Electrical Department, Islamic Azad University, Firoozkooh Branch, Froozkooh, Iran M.
2Tolou. Askari, Electrical Department, Islamic Azad University, Semnan Branch, Semnan, Iran
3M.Z.A.Ab Kadir, Advance lightning , power and energy research centre(ALPER), University Putra Malaysia,43100, Malaysia; Institute of Power Engineering (IPE), Universiti Tenaga Nasional, Jalan IKRAMUNITEN, Kajang, Selangor, Malaysia.
4M.Osman, Institute of Power Engineering (IPE), Universiti Tenaga Nasional, Jalan IKRAM-UNITEN, Kajang, Selangor, Malaysia.
5M.Hajikhani, Aryaphase Company,Tehran, Iran.
Manuscript received on November 22, 2019. | Revised Manuscript received on December 15, 2019. | Manuscript published on December 30, 2019. | PP: 4168-4171 | Volume-9 Issue-2, December, 2019. | Retrieval Number:  B4925129219/2019©BEIESP | DOI: 10.35940/ijeat.B4925.129219
Open Access | Ethics and Policies | Cite | Mendeley
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)

Abstract: Fault detection of the cable joints is one of significant problems in the electrical utilities and industrial companies to increase the network stability as the system interruption can make side effects for both power generation units, renewable energy generation units and other power sources beside of the costumers. In this paper, fault detection of a 20kV XLPE cable joint had been studied using the measured partial discharge (PD) signals and also support vector machine algorithm. In this study, the measured data had been classified based on proposed features as the indices of data classification and they had been used in the classifier algorithm to determine fault based on measured signals and the corresponding obtained features. The results show that the proposed features and applied algorithm could determine the faults in the cable joints with an appropriate range of accuracy. This study could develop the previous studies on a widely used cable joint. This research can be helpful for the electrical utilities to increase network stability.
Keywords: Partial Discharge, Cable Joint, Support Vector Machine.