Research On Data Analysis Technology Of EMU Fault In Big Data Environment

By the beginning of 2019,China's high-speed railway operation mileage has exceeded 30,000 kilometers,accounting for more than 60% of the global high-speed railway operation mileage.On the Beijing-Shanghai high-speed railway,China's standard EMU runs at the world's highest operating speed of 350 km/h.In October 2019,Beijing-Guangzhou high-speed railway is expected to run at 350 km/h.In terms of the number of high-speed EMUs,China Railway Corporation(CRC)has 2808 EMUs / 3287 EMUs,accounting for more than half of the world's total.At present,China's high-speed rail operation mileage,operation speed,train group ownership are ranked first in the world.With the rapid development of China's high-speed railway,the accumulated fault data of EMU in operation and maintenance are increasing.Useful information and rules are extracted from these fault data,and targeted measures are taken to optimize and improve the software logic,structure design and maintenance scheme of EMU,which is of great significance to the application and management departments at all levels of EMU.Statistical analysis of key fault data collected directly from EMUs can not only truly reflect the speed,load and line condition changes of EMUs under specific circumstances in China,but also do not require huge manpower and material resources.On the other hand,in order to ensure the efficiency and effect of data analysis and mining,the collected fault data should have good accuracy,continuity and integrity.It is necessary to build a standardized and perfect data recording model to accurately and efficiently record the information,date of failure,operating environment,system classification,mileage,fault mode,cause and location of the fault.At the same time,through data mining technology,we can find out the fault rules from the fault data,and provide data support for EMU operation management departments to make operational and maintenance decisions.The specific work of this paper is as follows:(1)Studying and studying the theory of fault data analysis,introducing RAMS technology,FMEA,FMECA,FTA,three commonly used reliability analysis methods,and studying the relationship among them;introducing the theory of data mining technology and its development,briefly outlining the fault data association rules.Analysis and clustering analysis technology concept and implementation method.(2)According to the concept of fault mode analysis in reliability engineering,the key information elements related to fault events,such as phenomena,solutions and treatment of EMU faults,are hierarchically classified into feature dimensions according to certain principles,and the fault data recording model of EMU is established,which lays the foundation for fault data analysis and mining,and provides guarantee and support.(3)Establish a distributed EMU fault database platform based on B/S system,and classify the components of EMU according to the system to establish a standardized component basic database;Based on the fault data of EMU bogie system,use association rules and Co-word clustering analysis to data mining,and explore the fault rules.This paper designs and establishes a distributed EMU fault data acquisition platform,which unifies and standardizes the structural components and fault modes of various EMUs,and analyses and processes the major fault history data of EMUs,which provides data support for optimizing maintenance procedures and plays an active role in reducing maintenance costs of EMUs.Through standardized management,statistical analysis and Research on fault rules of EMU key faults,it helps management departments at all levels to grasp the real-time situation of EMU key faults,evaluate technical performance status,and provide technical support for ensuring the safety,stability and comfort of EMU operation.