A Method for Remaining Useful Life Prediction of Transformer Based on the CNN-LSTM Networks

The transformer is one of the core equipment of the power distribution network, and its reliability determines the safety of the distribution network. Therefore, it is significant to study the prediction of the remaining useful life (RUL) to prevent the catastrophic failures of the distribution network. In this paper, an advanced RUL prediction model based on the CNN (convolutional neural network) -LSTM (long short-term memory network) network is proposed. The proposed model can identify the fault characteristics of the transformer from the original data, and then accurately predict the RUL of the transformer. Compared with the traditional data-driven method, the proposed method does not rely on signal processing technology and the prior knowledge of diagnostic experts. The method in this paper has better performance during complex operating conditions. The experimental results validate the effectiveness of the proposed method.