Abstract

With more electrical and electronic equipment applied into the railway system, much more data can be collected and then the big data era of railway is coming. By employing the maximal information coefficient (MIC), the big data analysis of rail equipment accidents is studied to investigate the effect of the updating of rail equipment. The rail equipment accident data set of 25 years (from 1990 to 2014) is separated into three subsets corresponding to the period of the occurrence time of accidents. For every subset, the contributing factors to accident damage, to accident severity, and to accident cause are analyzed, respectively. The results show that the variation trend of the number of rail equipment accidents is more consistent with the variety of railroad service miles rather than carloads. And the factor of highway-rail grade crossings is an important one which accords with the facts. However, a seemingly surprising result is found that there will be more contributing factors to accident severity and to accident causes with more equipment applied into the railway system as time goes on.

摘要

随着更多的电气和电子设备被应用到铁路系统中，可以收集更多的数据，然后铁路的大数据时代将到来。通过利用最大信息系数（MIC），对铁路设备事故进行大数据分析，以研究铁路设备更新的影响。25年（1990年至2014年）的铁路设备事故数据集分为三个子集，分别对应于事故发生时间的长短。对于每个子集，分别分析了造成事故损坏，造成事故严重程度和造成事故的因素。结果表明，铁路设备事故数量的变化趋势与铁路服务里程的变化而不是载货量的变化更为一致。公路-铁路平交道口因素是符合事实的重要因素。但是，发现一个看似令人惊讶的结果是，随着时间的流逝，将更多的设备应用到铁路系统中，将会对事故的严重程度和事故原因产生更多的影响因素。