Research And Implementation On Key Technologies Of Fault Diagnosis For EMU Based On Hadoop

With the continuous development of China’s high-speed railway in recent years, EMU has begun to put into mass use. At the present stage, we have accumulated vast amounts of service data of high-speed train, and these data have been increased with a speed of TB. How to analyze the massive failure date of EMU and its further maintenance service is very import for its fault diagnosis. With characteristics of diversity, body volume, high complexity and high speed shown in service data of high speed EMU, traditional data mining methods, which is time-consuming and inefficient with poor real-time capability, can’t meet the needs of EMU in handling malfunction and emergency applications. Therefore, this paper proposes the introduction of Hadoop distributed computing framework, whose map reduce programming model can solve the problem of EMU fault diagnosis during present stage according to characteristics of EMU data. Therefore the EMU fault diagnosis has certain theoretical and practical significance.This paper puts forward a solution for the fault diagnosis of EMU based on Hadoop distributed framework. Besides, through the optimization of C4.5 classification algorithm based on Hadoop, provides an effective method for improving the efficiency of the fault diagnosis of EMU and has been verified in the practical application.Several points of this thesis are as follows.Firstly, it has analyzed the core technologies of Hadoop including distributed computing framework of MapReduce, distributed file system HDFS, and data warehouse Hive. This thesis has provided solutions to the big data of analyzing EMU’s fault diagnosis based on Hadoop and has built a Hadoop cluster environment.Secondly, in the choice of algorithm it has analyzed the insufficient on EMU fault diagnosis based on Hadoop. Besides, This article propose two improved algorithms and put forward a performance improvements focus on the accuracy and scalability of C4.5 algorithm and improved the load balancing of the cluster.Thirdly, the improved C4.5 has been applied in actual scene of EMU’s malfunction in the laboratory, and produced several related experimental results. Experiments shows that the improved algorithm has a significantly improved the stability and scalability than the original C4.5 algorithm, and can meet the requirements of the fault diagnosis off EMU under the background of big data.The data mining system designed in the paper meets the specific requirements of EMU fault diagnosis, which has good concurrent mining performance and can raise the efficiency of fault diagnosis analysis of EMU.