Comparative Dissolved Gas Analysis with Machine Learning and Traditional Methods

Power transformers are one of the vital equipment for power systems. Therefore, in case of outage of the service, the effects on the system are fatal. The fault diagnosis is of great importance. In this study, interpretation methods of dissolved gas analysis used in diagnosis of power transformers are examined. In the MATLAB GUI, a user interface has been created for traditional fault diagnosis. Fault diagnosis is made and the shortcomings of traditional methods have been shown. Support Vector Machine, K-Nearest Neighbors and Decision Tree algorithms are used for diagnosis of machine learning methods. Between these methods, the Python programming language is used for fault diagnosis and fault classification is made to the IEC TC 10 database. Confusion matrices and classification performance measurements of machine learning methods are obtained. The classification accuracy of the methods has been investigated and compared to each other.