On Combined PSO-SVM Models in Fault Prediction of Relay Protection Equipment

Since there are no monitoring devices in relay protection equipment of substations, it takes up a lot of manpower and material resources in operation and maintenance (O&M), and the efficiency is very low. In this paper, we solve the giving fault prediction model by applying particle swarm optimization algorithm and support vector machine algorithm for relay protection equipment. The combined model was found to be able to establish the correlation between the variables more accurately and improve the prediction accuracy of the model by comparison. The simulation result shows that the prediction accuracy of the model is at least 91% for three different devices. The combined prediction model can not only provide strong technical support for the maintenance strategy of relay protection equipment but also improve the maintenance efficiency and reduce the failure rate of protection equipment.