Fast Prediction Model for the Susceptibility of Balise Transmission Module System Based on Neural Network

Abstract

In this paper, the Electromagnetic Susceptibility (EMS) of BTM system is predicted by Neural Network (NN) modeling method. First, the EMS data of BTM system was obtained through actual measurement. That is, the training set and validation set needed for NN modeling are constructed. Then, the Back-Propagation (BP) NN modeling method is selected and initialization operation is carried out. Finally, a fast prediction model for EMS of BTM system based on NN algorithm is established. Validation shows that the prediction results of the NN model are accurate. This provides an effective way to study the EMS of the BTM system.