Power Transformer Classification through Dissolved Gas Analysis Utilizing Least-Square Support Vector Machine

• Published in: SSRG International Journal of Electrical and Electronics Engineering
• Main Author: Yasin Z.M.; Zakaria F.; Salim N.A.; Ab Aziz N.F.
• Format: Article
• Language: English
• Published: Seventh Sense Research Group 2024
• Online Access: https://www.scopus.com/inward/record.uri?eid=2-s2.0-85185768631&doi=10.14445%2f23488379%2fIJEEE-V11I2P107&partnerID=40&md5=04f05fb5da2f34c10fdb23198dd37a08

- This article proposes the utilization of the Least-Square Support Vector Machine (LS-SVM) approach to ascertain the presence of a fault in power transformers. Power transformers are essential elements of electrical power systems. The failure of a power transformer can cause a disturbance in the functioning of power distribution and transmission systems. This situation will result in an increase in operating expenses due to the need for repairs and maintenance. The reliability of the electrical grid may be compromised. Therefore, it is crucial to identify any flaws in the power transformer at an early stage. In this paper, the LS-SVM utilizes Dissolved Gas Analysis (DGA) data as its input. The DGA methodology is widely accepted as the prevailing method for identifying the early stages of defects that arise in power transformers by analyzing the ratio of essential gases. The simulation data acquired from the industry comprises a standard state and six distinct fault types of transformers, which are utilized as input for the LS-SVM models. The suggested model underwent testing in multiple scenarios, yielding a maximum accuracy of 97.37%. © 2024 Seventh Sense Research Group. All rights reserved.