Train dwell time distributions at short stop stations

The train dwell time at short stop stations would influence route conflict in the railway time table, thus plays an important role in railway timetabling and rescheduling. However, it is not well estimated due to the lack of adequate understanding of its dependencies. In this paper, the train dwell times are estimated by means of algorithms based on the track occupation data of the Dutch railway network. The algorithm takes the train driver behavior into account which makes the result more accurate. Totally 17306 trains running from 1 Sep. 2012 to 30 Nov. 2012 are processed and analyzed. The distribution of the dwell times fits a lognormal distribution. The cumulative distribution functions of dwell times in different stations are same statistically. The dwell times at short stop stations are different from large stations. In addition to the number of alighting and boarding passengers, it found that the dwell times at short stops are also influenced by different weekday, peak hour, train length, which indicates the potential for further real time prediction of train dwell times.