Research On Fault Early Warning Technology Based On Train Data Analysis

The train operation control system is the "nerve center" of the high-speed train control system and the infrastructure of railway transportation.Among them,the vehicle-mounted subsystem is the core part of the train control system,which is the key to ensure driving safety and improve operating efficiency.The train is affected by various factors in the process of operation,and the on-board equipment failure occurs from time to time.Currently,the fault handling mode of the on-board subsystem is still based on the experience of maintenance personnel,and there is a lack of research on safety trend prediction,risk assessment and warning,which affects the running safety and efficiency of the train.Therefore,how to improve the early warning capability of on-board equipment and develop appropriate auxiliary maintenance methods to improve the safety and reliability of trains is an urgent problem to be solved by railway operation departments.This thesis mainly studies the 300 T vehicle-mounted equipment from the following three aspects: fault diagnosis technology,reliability evaluation method,equipment maintenance strategy.The main research work and research results of the thesis are as follows:(1)To solve the problem of complex text structure and high redundancy of text failure data,this thesis proposes a feature extraction method for fault features based on text mining technology that combines TF-IDF algorithm with rough set theory to get an effective fault feature set.By establishing a BP neural network model,it is verified that the method can eliminate redundant and irrelevant features,simplify the diagnosis model,improve the diagnosis accuracy and the defect of the diagnosis model’s dependence on data quality.This model is different from the previous signal-based fault diagnosis model,and provides a new way of train fault diagnosis.(2)To improve the accuracy of system reliability assessment,this thesis considers the FMECA model based on BN improvement from the perspective of equipment failure hazards,and carries out objective and quantitative risk assessment of on-board equipment to solve the subjective arbitrariness caused by the original FMECA for early warning.From the perspective of system function,this thesis sorts the safety and reliability weights of vehiclemounted equipment based on the AHP model.Then,this thesis comprehensively considers the hazard degree and weight of the equipment,and uses the combination weighting method to carry out comprehensive early warning evaluation of the system,which changes the traditional way of evaluating failure from a single factor.Finally,based on the above two aspects,this thesis uses the grouping method to carry out comprehensive early warning evaluation of the system,which changes the traditional way of fault evaluation from a single factor.(3)In order to improve the maintenance efficiency and facilitate the staff to formulate the tasks of the next stage in advance,this thesis makes a polynomial fitting prediction on the replacement data of the on-board equipment,and formulates a maintenance plan that changes from "regular repair" to "sensible repair",based on the evaluation results,which can improve the safety and reliability of trains,reduce maintenance costs,and promote the development of high-speed railways towards intelligence.(4)This thesis uses MATLAB and C # mixed programming technology to design each functional module of on-board equipment safety analysis and early warning system.The software realizes the intelligent analysis of on-board equipment fault diagnosis,and achieves the visual management of failure early warning.