Study on Vibration Acceleration Prediction Model of Track Inspection Vehicle Based on BP Neural Network

In order to study the mapping relationship between the track irregularity parameters, vehicle running speed and the vehicle vibration acceleration detected by the track inspection vehicle, and to provide better data support for diseases detection of track inspection vehicle and maintenance of rail line, this paper develops a BP (Back Propagation) neural network model that combines the big data analysis technology to predict the vibration acceleration of the track inspection vehicle. The sample data for this model comes from the inspection data of a passenger transport special line in East China. Based on grey relational analysis method, this research preprocesses the correction of mileage deviation for the sample data and analyzes the correlation degree between seven track irregularity parameters and two kinds of vehicle vibration accelerations, the results show that all input layer indicators of the neural network model have different degrees of influence on the output layer indicators. And the predicting results with test data shows that the average accuracy of the whole sample for vertical vibration acceleration and lateral vibration acceleration is 85.75% and 90.25%.