Power Transformer Incipient Fault Diagnosis Using Support Vector Machine

• Published in: 2022 3rd International Conference on Artificial Intelligence and Data Sciences: Championing Innovations in Artificial Intelligence and Data Sciences for Sustainable Future, AiDAS 2022 - Proceedings
• Main Author: Rosli M.A.S.; Zakaria F.; Zin N.M.; Daud R.M.N.R.; Sidek M.N.; Zain M.Y.M.; Kassim A.H.
• Format: Conference paper
• Language: English
• Published: Institute of Electrical and Electronics Engineers Inc. 2022
• Online Access: https://www.scopus.com/inward/record.uri?eid=2-s2.0-85141762296&doi=10.1109%2fAiDAS56890.2022.9918809&partnerID=40&md5=662b16b162217690dbcb62f762b2fc31

Power transformer is known as an essential equipment for electrical power system. If the breakdown happens and it's associated to power transformer, power distribution and transmission operation might be halted. This condition will have resulted in high cost for repair and maintenance purposes. The reliability of the power system may be jeopardized. Thus, early detection of possible faults in power transformer is become vital and essential. In this study, support vector machine (SVM) method is proposed to diagnose and predict incipient faults in power transformers. Dissolved gas analysis (DGA) method is used for the analysis technique. Based on key-gas ratios, DGA is a standard approach for diagnosing incipient faults in power transformers. In this study, the incipient faults are categorized into six types which are Partial Discharge, Discharge of Low Energy, Discharge of High Energy, Thermal Fault (t< 300C, Thermal Fault, (300C<t< 700C and Thermal Fault (t> 700C. DGA data obtained from industry are used to develop the SVM models. MATLAB software is used for simulation process. The performance of the proposed method is analyzed in terms of accuracy and computational time. Results show that the Linear SVM has higher accuracy compared to Fine Gaussian SVM for the purpose of classifying incipient fault in power transformer © 2022 IEEE.