Influence of Causing Factors on CRH2 EMU Failure Consequences Using Decision Trees

摘要

Understanding the characteristics of the causing factors of CRH2 EMU (Electric Multiple Units) failure contributes to improving maintenance efficiency and reducethe level of high-speed rail operation risk. Using the CRH2 EMU failure data, this study explores the EMU train type, failure-related system, classification of causes, and train operation mileage's characteristics associated with failure consequences. The failure sequences are divided into five severity levels as no influence on operation, minor influence on operation without delay, delay, temporary repair, and outage. The decision tree models were developed to predict the failure sequences associated with the causing factors. The train running kilometers was found to be the most significance variable influencing failure consequences, followed by failure-related system. The results suggest that optimizing the train maintenance cycle, identifying critical maintenance components of train systems, and improving the quality of CRH2 EMU products before operation would significantly decrease the failure severity level.