Research On Fault Diagnosis Of Railway Freight Car Braking System Based On PHM Technology

China is a big country of railway freight transport.With the rapid development of national economy and the construction of Asian-European railway projects,the load of railway freight cars has been continuously increasing,and the distance of operation has been increasing.At the same time,the safety problems of railway freight cars have become more and more prominent.Railway freight car braking system is the core component of operation safety.Once it breaks down,it is easy to cause serious accidents,resulting in property and even personnel losses.Therefore,it is of great significance to monitor and diagnose the faults of the braking system of railway freight cars.Air brake system with 120 air distribution valve is mostly used in railway freight cars in China.In this paper,by studying the operation principle and common fault mechanism of 120 type air brake system,a fault diagnosis scheme of railway freight car brake system based on prognostics and health management(PHM)is proposed.The scheme consists of two parts.First,the on-board monitoring system of the braking system is established to monitor the running state of the braking system in real time,acquire the air pressure parameters of the braking system in real time,and extract the fault characteristic parameters according to the monitoring data.Secondly,the fault diagnosis of braking system is carried out,and the fault diagnosis models based on BP neural network and RBF neural network are established respectively.The model training is completed according to the fault parameters of monitoring feedback to realize the fault diagnosis of railway freight cars.This paper mainly completes the following work:(1)The structure of 120 type air brake system and the operating principle of five working states,i.e.inflation relief,deceleration and inflation relief,conventional braking,emergency braking and braking holding pressure,are analyzed.Based on the feedback monitoring data of the vehicle monitoring system,the fault mechanism of braking sensitivity,braking stability,alleviating bad faults and natural alleviating faults of the air braking system is studied.(2)According to the technical requirements of PHM,an on-board monitoring system for air brake system of railway freight car is designed,which consists of on-board device,ground readout device,application server and transmission network.The air pressure data of train tube,auxiliary air cylinder,brake cylinder upstream and downstream of brake cylinder are collected by integrated clamp and decentralized installation.The system feedback monitoring data were analyzed and the fault features were successfully extracted.(3)According to the PHM technical scheme,the basic theory of artificial neural network is studied,and the structure and algorithm of BP neural network and RBF neural network are analyzed in detail.Fault diagnosis models based on BP neural network and RBF neural network are constructed respectively.The model completes the training and testing of the neural network based on the fault characteristic data by the vehicle monitoring system.Among them,the accuracy of BP neural network test is 75%,and some fault types can not be identified.The test accuracy of RBF neural network reaches 98.75%,which basically realizes the fault diagnosis of air brake system.