Intelligent Fault Diagnosis Method of Capacitor Voltage Transformer based on Recurrent Neural Network

Capacitor voltage transformer (CVT) is an important part of electric energy metering device. If the fault can not be located in time and repaired, it will cause serious consequences. At present, it is usually judged by manual to check the cause of the fault, which consumes a lot of manpower and material resources and has low efficiency. In this paper, an automatic intelligent diagnosis method for CVT fault causes is proposed. A neural network model based on WHO-RNN architecture is designed. The model optimizes the parameters by Wild Horse Optimizer(WHO). Then the recurrent neural network(RNN) model is trained and used for prediction to realize the deep mining of historical data information. It can intelligently judge the cause of CVT fault through online data, and judge the fault location of CVT as early as possible. The simulation results show that the accuracy of this method is about 93%, which is significantly higher than that of ordinary RNN neural network, and meets the accuracy requirements of CVT fault diagnosis.