Running State Prediction And Evaluation Of Power Transformers

Power transformers arc the core of energy conversion and transmission in power grid. Once an accident occurs, it may cause huge losses such as equipment assets and power outages, and even have serious social impact. Therefore, effective state prediction and evaluation of power transformers can guide the operation and maintenance, which has important theoretical and practical significance. In this paper, the large data analysis method is used to predict and evaluate the running state of the transformer. Firstly, the running state index of the transformer is predicted based on long-short term memory (LSTM). Then, the current and future running state of the transformer is evaluated by decision tree classification algorithm. The experimental data comes from the running data of 630 kVA transformers in a certain area. The results show that the method used in the paper can effectively predict the future running state of the transformer and evaluate the abnormal state with high accuracy, which can provide reference for actual operation and maintenance.