Reducing negative influence of strong wind on normal train operations based on data mining

Abstract

Cross wind has an important influence on the safety of high-speed railways. Wind monitoring or early-warning system is widely used to provide real-time monitoring information or early warning information to train dispatching system for making appropriate measures like slowing down running speed of trains or suspending of train operations. But there is trade-off between safety and train operation efficiency for train operation in windy weather. Wind data mining method was proposed in this paper to reduce negative influence of strong wind on normal train operations and improve train operation efficiency. Three main objective factors that affects train operation efficiency were optimized, including optimizing wind alarm release time, merging wind alarm restriction sections and adding wind direction into wind alarm mechanism. Real-word case study show that wind alarm release time should be set according to the wind characteristic of different lines confronted. Wind alarm restriction sections were merged by analyzing the alarms relationship generated by adjacent wind motoring points, and the merging of wind alarm restriction sections can significantly decrease the number of wind alarm times that train dispatchers’ need to handle. Wind direction has remarkable influence to train operation and need to be added into wind alarm mechanism.