Characterization of the operating periods of a power transformer by clustering the dissolved gas data

This paper presents an analysis of the different operating periods of an in-service oil immersed power transformer through dissolved gas concentrations. The unsupervised classification by k-means method allows regrouping the periods of operation into classes using the Euclidean distance as a criterion of similarity. The analyzed data describes the evolution of gas concentrations as a function of time. The classes obtained are characterized by the production activities of the different gases, various operating constraints and the incipient failures. These periods also highlight the maintenance actions carried out on the insulating oil.