Research On Operating State Monitor And Fault Diagnosis Of Railway Freight Cars' Air Braking System

With the development of "heavy load and speed-up" of railway freight cars in China,higher requirements are put forward for the safety,application and maintenance of railway freight cars.Air braking system is one of the most critical systems to ensure the operation safety of railway freight cars.Its fault has certain complexity,it is not easy to judge and deal with,and the early detection and location of the fault is beneficial to the maintenance and repair work in the later stage.Therefore,it has strong engineering value and practical significance to design a system which can monitor the operating state of air braking system in real time and make fault identification and positioning according to the obtained data.According to the structure and working principle of railway freight car air braking system,the mechanism of its common faults is deeply studied,and the operating state monitoring and fault diagnosis system of railway freight car air braking system is designed.The system is mainly composed of on-board device and ground analysis system,which can complete the acquisition,transmission,feature extraction and fault discrimination of working parameters of the air braking system.The main work of this study includes the following three parts:1.According to the common fault discrimination mechanism of railway freight car air braking system,the design scheme of air braking system running state monitoring and fault diagnosis system is determined.The hardware design of on-board monitor is also completed,and the functions of four-point pressure acquisition,information storage and information transmission are realized.2.Wind pressure data discrimination and fault feature extraction are completed.Four kinds of common brake fault classification methods based on support vector machine are proposed.Wind pressure data are used to obtain training set and test set data,and the verification test of training and classification results is completed.3.A fault diagnosis model of braking system based on Bayesian network is established.Based on the expert experience and correlation analysis,the content of each node and the node dependence in the fault diagnosis model are determined,and the Bayesian network is established.Through Bayesian visual analysis software,the model structure is optimized by using the running fault statistics of the railway freight car braking system,and finally the data is classified according to the fault characteristics.Besides,the brake fault diagnosis and positioning are completed with the help of Bayesian network reasoning.