Research On The Technology Of Association Analysis For Operation And Maintenance Efficiency Of The Key Components Of EMU

Since entering the "Twelfth Five-Year Plan",China Rail Corp accelerated the construction of the fast railway network mainly constituted by "four vertical and four horizontal" high-speed railway and the rapid railway network basically completed.Along with the high-speed railway network rapid expanding and the continuous opening of new roads,train scheduling density increases gradually and the maintenance interval is getting shorter and shorter.This requires efficient operation and maintenance of EMU,so as to ensure the safe operation and service quality of EMU.Massive operation and maintenance data accumulated since EMU put into operation is of significant importance for improving the operation and maintenance efficiency of the key components of the EMU.How to use the efficient data mining algorithms to dig out useful information from these massive data,in order to assist the development of the effective operation and maintenance mode,is the focus of this paper.The work of this paper is as follows:(1)We analyze the present situation of the domestic and international high-speed railway EMU operation and maintenance information.The China high speed train Repair Class and Repair System and the main maintenance methods of the EMU are arranged.(2)The data related to the operation and maintenance efficiency of traction motor are sorted out based on the whole life cycle of traction motor.According to the characteristics of multi-source heterogeneous data,a data preprocessing scheme is presented,including data cleaning,data integration,data reduction and data transformation.Then,the preprocessing of traction motor operation and maintenance data is realized.(3)Aiming at the shortcoming of the traditional DHP algorithm,the approximate minimum perfect hash function is used to solve the conflict problem in DHP algorithm and the AMPHP algorithm is proposed.In order to break the limitation of single machine algorithm,the parallel improved AMPHP-SON algorithm is proposed based on the idea of SON algorithm.(4)The AMPHP-SON algorithm is implemented on the Hadoop platform.The performance of DHP,AMPHP and AMPHP-SON algorithm is compared under different data sets,which have verified the effectiveness of the two improvements of the algorithm.Finally,analysis and visualization are performed on the EMU operation and maintenance efficiency association rules dug out.Experiments show that the proposed AMPHP-SON algorithm,using approximate minimal perfect hash function to filter out all the non-frequent item sets,and adopting the thought of SON algorithm to assign the association rule mining tasks to multiple nodes to complete in parallel,can dig out relevant rules from the massive EMU Operation and maintenance data quickly.These rules can help to optimize the EMU Repair Class and Repair System effectively,and finally help to improve the efficiency of the EMU operation and maintenance.