Research on Evaluation Method of Equipment Fault in Electric Power Communication Network

The research on reliability of electric power communication network originates from the research on equipment fault in it, which is fully significant. In this paper, the equipment fault cases are classified by K-means clustering method, the neural network is trained for fault cases classification based on radial basis function (RBF), the network inputs dimension is reduced by rough set method before training, and the fault cases level are determined through the comprehensive indexes which could reflect it. Finally, a practical problem is resolved by this method to demonstrate its advantage that avoiding subjective interference effectively.