Design And Implementation Of MVB Network On-line Fault Diagnosis System

Multifunction Vehicle Bus(MVB)undertakes the task of data interaction between devices inside the vehicle,and its reliability is the basis for stable train operation.Due to the lack of reliable online fault diagnosis system,the fault data cannot be obtained in time,resulting in the low accuracy of MVB network fault diagnosis.Therefore,it is important to study the online fault diagnosis system of MVB network to improve the accuracy of fault diagnosis and the reliability of train operation.In this paper,we take MVB network as the research object and design the software part to realize online fault diagnosis system,including data analysis and fault diagnosis software,visual network monitoring and management software,so as to realize online monitoring and fault diagnosis of MVB network.The main research contents of this paper are as follows:(1)In-depth study of MVB network communication protocols and fault analysis were conducted to collect physical layer waveform data of the network during normal and fault conditions.After data is pre-processed,characteristic indicators are extracted and fault diagnosis is realized.At the same time,the characteristic extraction and fault diagnosis algorithms are managed in the form of tasks to realize the free deployment of algorithms.(2)A large amount of extracted characteristic data such as steady-state amplitude and overshoot are statistically analyzed with the aim of finding the relationship laws between characteristics,and a rule element relationship table is established using the generative rule representation,and expert experience-based fault diagnosis is realized by querying this table.At the same time,the SVM(Support Vector Machine)algorithm is introduced to improve the reliability of system fault diagnosis.(3)The visualized network monitoring management software based on B/S(browser/server)architecture is implemented for the inflexibility of online diagnosis devices based on C/S(client/server)architecture.The software transmits waveform and characteristic indicators data processed by the server side to the front-end interface through the HTTP protocol,thus realizing online monitoring and fault diagnosis visualization of MVB network.(4)Build an MVB network communication experiment platform and simulate common network faults such as remote termination and double termination.By logging into the visualized network monitoring management software and testing the functions of data acquisition,characteristic extraction and fault diagnosis,the effectiveness of the software system and the reliability of expert experience algorithm and SVM algorithm fault diagnosis are verified.