Cbr-based Fault Diagnosis System For Vehicle Equipment Of High-Speed Railway

With the rapid development of high-speed railway in recent years,China has become the country which possesses the most comprehensive technology, the fastest operating speed, the longest operating mileageand the highest integration of high-speed rail system.The safety and efficiency of China’s high-speed railwayclosely depend onthe high reliability ofCTCS-3.The on-board vehicle equipment (referred to as "vehicle equipment")is an important part of CTCS-3, which is responsible for generating the target distance control curve, overspeed protection. Due to the importance in the protection of traffic safety, the fault diagnosis of vehicleequipment plays a very significant role in ensuring the reliable and safe operation of high speed railway systems.Currently, vehicle equipment goes ahead with the direction of complexity, system, automation. Most faults of vehicle equipment are showing characteristics of concealment, diversity and overlaps that increases the difficulty of vehicle equipment maintenance.However, the maintenance of vehicle equipment currently just relying on the experience of maintenance personnel,which cannot fullfillthe new requirement of modern high-speed railway fault diagnosis.Based on the study ofintelligent fault diagnosis technology and current situation of vehicle equipment maintenance equipment, this thesis appliesthe Case-based Reasoning(referred to as"CBR") technology into vehicle equipment fault diagnosis field so that maintenance personnel can diagnosefaults for vehicle equipment quickly and accurately to ensure the efficient operation of the train.Firstly, based on the analysis of vehicle equipment structure, the failure characteristics and faultcase records, this thesis proposes avehicle equipment fault diagnosis systemcombining with case-based reasoning technologies, which is capable of completing thefault diagnose and case retaining. On the basis of analysis of common failure case representation method, this thesiscombines a newobject-oriented representation method with database technology in the field of vehicle equipment.This thesisprovides detailed information on the case organizations, case revision, case retaining and case maintenance.Then the method of using hierarchical organization to organizethe case base is putted forward, and the entire fault case library is divided into the following three categories:system case base, representing the case base and sub-case base.Secondly,compared with traditional neighbor algorithm, the weighted K-NN algorithm whichcalculatesthe similarity among casesin this thesis canavoiddisadvantagethat all the feature attribute of the case influence the case similarity samely.This thesis uses rough set theory to reduce attribute of the case, then calculates feature attribute importance, and at last doesthe normalization to overcome poor objectivity of traditional weight calculation algorithm.Vehicle equipment fault diagnosis system uses a phased neighbor search strategy to retrievecase, which is proposed in this thesis. This strategy combines case base structure in level3to reduce the number of case retrieval so that the efficiency of case retrieval is improved.Finally,according to the system requirements,the vehicleequipment fault diagn osis system is realized on the platform ofMicrosoft Visio Studio2010,and then d iagnosis process, function module and database designofthe whole diagnostic proc ess are introduced.some examples are undertaken to verify the effectiveness of th e fault diagnostic system.we summarize the research work and point out the exi sting shortcomings and the further research work of the diagnostic system in the future.