Machine learning methods for detecting anomalies in a power transformer by monitoring its hot-spot temperature

This paper analyzes and compares different machine learning methods such as decision trees, SOMs, MLPs and rough sets for the classification of the operation condition of a power transformer. The purpose is to construct a classification model able to estimate the hot-spot temperature as a function of other external input variables. The classifier would then be used to detect anomalous operation conditions of the transformer by comparing the observed and estimated hot-spot temperatures.