基于大数据的动车组故障诊断方法研究

Research on Fault Diagnosis Method of Electric Multiple Units Based on Big Data

近年来,我国高速铁路发展迅速,取得令人瞩目的成就,动车组因为其安全性、高速性、舒适性、准时性、环保性等特性,越来越得到人们的认可。随着动车组大规模的投入运营,目前已经获得了大量的故障诊断与状态监测数据,通过挖掘和利用蕴含在这些数据背后的价值和规律,来进一步指导动车组的运用维修工作,这对于提高动车组的行车安全具有重大意义。为此通过特征提取、数据分析和数据挖掘等方法,运用Hadoop平台,采用一种基于ADASYN-GBDT决策树算法的动车组故障诊断方法,该方法结合了ADASYN算法和GBDT算法的优点,不仅能够弥补数据分布不平衡给故障诊断带来的缺陷,而且还能够解决现阶段动车组故障诊断中的维修不足、维修过剩、设备故障率问题、维修成本高等一些问题,该方法具有一定的理论和现实意义。

In recent years,China’s high-speed railway has been developed rapidly and achieved remarkable achievements.EMU have gained more and more recognition among Chinese citizens in term of safety,high speed,coziness,punctuality and environmental contribution.With the large-scale operation of the EMU,a large number of data have been obtained on fault diagnosis and condition monitoring up to now.By exploring and utilizing the values and laws behind these data,we can do more work to guide the operation and maintenance of the EMU,which is of great significance.This paper explores and analyzes the data via the Hadoop platform,and adopt EMU fault diagnosis method based on an ADASYN-GBDT decision tree algorithm.This method combines the advantages of ADASYN algorithm and GBDT algorithm,which can not only make up for the defects brought by the imbalanced distribution to fault diagnosis,but also can solve some shortcomings such as insufficient maintenance,over-maintenance,high equipment fault rate as well as high maintenance cost in the current EMU fault diagnosis.Therefore,this method has both theoretical and practical significance.