Power Transformer Fault Prediction using Naive Bayes and Decision tree based on Dissolved Gas Analysis
Yassine Mahamdi, A. Boubakeur, A. Mekhaldi, Youcef Benmahamed
Abstract
Power transformers are the basic elements of the power grid, which is directly related to the reliability of the electrical system. Many techniques were used to prevent power transformer failures, but the Dissolved Gas Analysis (DGA) remains the most effective one. Based on the DGA technique, this paper describes the use of two of the most effective machine learning algorithms: Naive Bayes and Decision Tree for the identification of power transformer’s faults. In our investigation, 9 different input vectors have been developed from widely known DGA techniques. 481 samples have been used and 6 types of faults have been considered. The evaluation result of the implementation of the proposed methods shows an effectiveness of 86.25% in power transformer’s fault recognition.