Design And Development Of Big Data System For Real-time Fault Diagnosis Of High-speed Train Operation

In the process of high-speed train operation,if the occurrence and evolution of operational faults or hidden dangers cannot be detected in time,chain reaction may be triggered,resulting in accidents such as train slowdown and parking;if the fault cannot be diagnosed and dealt with in time,it may lead to disastrous consequences of car crash and death.At present,the on-board diagnostic system of high-speed trains is based on expert knowledge for fault diagnosis.It is difficult to predict and diagnose faults in real time for faults that may occur during the operation of high-speed trains.And it is difficult to carry out in-depth system analysis on the problems of safety hazards caused by the aging of components because of long-term operation of trains.At present,fault diagnosis based on big data technology for high-speed train operation is still a blank.Aiming at the above problems,a big data system for real-time fault diagnosis of high-speed train operation is designed and developed.The system can realize real-time fault warning and diagnosis of high-speed trains to ensure the safety of driving and passengers.This paper relies on the National Natural Science Foundation’s major project project"Fault Modeling Theory and Method of High Speed Train Information Control System Based on Large Data and Knowledge" to carry out research.In view of the current high-speed train fault diagnosis system at home and abroad,it is impossible to perform real-time fault diagnosis on high-speed trains with high-volume,high-frequency and real-time data,and it proposes a big data design and development system for real-time fault diagnosis of high-speed train operation.The main research work of this paper is as follows:(1)Fault diagnosis requirement of high speed train.High-speed train data has obvious big data characteristics,namely:large amount of data,various types,low value density,and strong real-time performance.However,at present,the vehicle fault diagnosis system of high-speed train can not meet the real-time fault diagnosis of mass,high frequency and real-time data.In this paper,from the point of view of high-speed train data flow,the functional requirements of the system are analyzed in detail,mainly from four aspects of data acquisition requirements,the demand of information transmission,the demand of data analysis and the demand of data storage.(2)Design the system according to the needs analysis.According to the modular design principle,the design of the big data system design and development architecture for real-time fault diagnosis of high-speed train operation is completed.The system uses the Lambda architecture and consists of three layers:the batch layer,the real-time processing layer,and the service layer.It mainly includes the following functional modules:data acquisition module,data transmission module,data preprocessing module,data batch processing analysis module,data real-time processing analysis module,data storage module and data visualization module.Among them,by constructing the data acquisition cluster architecture,the running data generated by the high-speed train in the running process is loaded into the data transmission module,which solves the real-time data collection problem;and realizes the efficient transmission of data by constructing a distributed message queue system.The real-time problem of high-speed train operation data transmission;the high-speed train data is cleaned by constructing data pre-processing tools,which solves the problem of high-speed train data shortage and format unconformity;it solves the high-speed train historical data by deploying batch processing engine Fault modeling problem;it solves the real-time fault analysis problem for high-speed train running data by deploying real-time data analysis engine;it realizes the storage problem of large-scale high-speed train running data by constructing distributed database;solve the problem of high speed through data visualization tool Visualization of train operation data analysis results.Therefore,real-time analysis of high-speed train operation data is carried out by the system to realize real-time fault diagnosis of high-speed trains during operation.(3)According to the design plan,based on the industrial cloud platform of the State Key Laboratory of Integrated Automation of Process Industry,adopting Java and Scala development language,distributed column database HBase,big data processing framework such as Hadoop,Zookeeper,Kafka,Spark,HBase And the visualization tool Tableau DeskTop,the design and development of the system is completed.Among them,the Hadoop cluster is used as the infrastructure,mainly using its distributed file system HDFS to store the massive data generated by high-speed trains,and solves the problem of large-scale data storage generated during the running of high-speed trains;Zookeeper clusters are used for resources.Provisioning and cluster management,storing metadata information of the cluster,ensuring strong consistency and high reliability of data at different nodes;Kafka cluster is used to efficiently transmit data collected by high-speed trains,and solves the data transmission of high-speed trains.Real-time problem;Spark cluster is used for offline analysis and real-time analysis of high-speed train running data,solving offline analysis of high-speed train running historical data and online analysis of real-time data;HBase cluster is used to store real-time data generated by high-speed trains It solves the storage problem of massive,high-frequency,real-time high-speed train operation data;MySQL is used to store the results of real-time data analysis and processing of high-speed trains,and as a data source for data visualization modules;Tableau Desktop is used to analyze and handle high-speed rail operation data failures.Visualize the solution For the visualization of high-speed train running data analysis of fault diagnosis,the conductor can better help locate the fault.This paper designs a big data system based on real-time fault diagnosis of high-speed train operation for high-speed train fault diagnosis system,and innovatively combines big data technology with high-speed train operation.Make full use of the high-speed train running historical data to build a fault diagnosis model,diagnose the real-time running data of high-speed trains,solve the problems of the diagnosis of the high-speed train vehicle bottom-level control system,and completely restore the relationship between the measured values,Improve the accuracy of troubleshooting,quickly locate the cause of the diagnosis,and it provides a new direction for the research of high-speed rail fault diagnosis...