Research On Fault Diagnosis System For CRH5 High Speed Train Communication Network

The failure of the train communication network not only causes an increase in the bit error rate and the loss rate during information transmission,but also affects the availability of the train network control system if it cannot be detected and processed in time.The purpose of the train communication network fault diagnosis system is to realize timely diagnosis and situational awareness of the train communication network to ensure that the train is in a state of maximum safety,availability and reliability.Therefore,the research of train communication network fault diagnosis system has important application value and theoretical significance.This paper focuses on the fault diagnosis of transmission media in train communication network,and establishes a fault diagnosis system for train communication network.Firstly,the working mechanism of multi-functional vehicle bus(MVB)is studied.The relationship between common faults of MVB and electrical characteristics is analyzed.Based on this,the numerical distribution of electrical characteristics under common faults is established.Through the use of physical MVB and bus detector to fault simulation of fault electrical characteristics,the correctness of fault mechanism analysis is proved.Secondly,based on the theory of residual neural network,a fault diagnosis model of residual neural network is established.An improved residual module is proposed,which increases the model nonlinearity on the basis of reducing the weight parameter size of the residual neural network fault diagnosis model.It is proposed to establish the mapping between the fault feature data and the spectrogram by using the short-time Fourier transform,so that the fault feature data can be trained by the residual neural network.Based on this,a sample database is established for fault diagnosis model training.Thirdly,based on the theory of evidence reasoning rules,the establishment of the safety situational awareness model of train communication network is realized.Aiming at the reliability evaluation of train communication network,the reliability evaluation model of train communication network is established.Aiming at the problem of the influence of train communication network failure on the availability of Train Control and Management System(TCMS),a model for evaluating the impact of train communication network failure on TCMS availability is established.Finally,based on the train communication network reliability evaluation model and the train communication network failure to the TCMS availability impact assessment model,the train communication network security situational awareness model is established to realize the train communication network security situational awareness.Finally,the design of the train communication network fault diagnosis system was carried out.Based on the residual neural network fault diagnosis model,the fault diagnosis function module of the fault diagnosis system is designed,and a wake-up mechanism of fault diagnosis function module is proposed.A fault priority alarm level evaluation model is established for the fault sequencing problem.Based on this,the fault alarm sequencing function module of the fault diagnosis system is designed.Based on the safety situational awareness model of train communication network,the safety situation awareness function module of train communication network of fault diagnosis system is designed.Through the establishment of the data transfer process between the functional modules,the final construction of the train communication network fault diagnosis system is realized.