At present, power transformer has become an important electrical equipment in existing industrial system. Therefore, the effective analysis method to detect the overheating and other discharge faults can guarantee the continuous process of system. However, existing detection methods concentrate on the analysis of sensors data and ignore the neural network can also achieve the faults detection. In this work, we propose a novel framework to identify the faults in power transformer by applying the neural network. Initially, the sensors data is transformed in the feature space with multiple vector...
