Research project: Predicting and mitigating small fluctuations in station dwell times

The amount of time spent by a train in a station (the dwell-time) can have a big impact not just on the punctuality of the individual train, but on the performance and reliability of an entire railway network.  Delays can quickly become compounded, especially at busy and congested stations.  At one level, it is desirable to reduce dwell times, in order to improve throughput and increase the capacity of the network.  At another level, it is important to reduce fluctuations in dwell times, because these fluctuations can make it hard for operators to plan effectively to avoid disruption.  Train running and delay data are typically not recorded at a high temporal resolution (delay attribution data only provides information concerning delays of three minutes or more), making it hard to monitor smaller fluctuations in dwell times, which can still have big impacts overall on train and network performance.

This project aims to combine more detailed data sources (such as South Western Railway’s ‘Nexala’ train performance monitoring system) with those sources which are already utilised, in order to better understand variations in station dwell time.  Key aims include understanding where on the network dwell time becomes a big issue, and what factors influence it.  Using a range of data sources, dwell times will be visualised and modelled and a system developed which can predict in real time what the dwell times for given services are likely to be.  This system will be linked to a spatially-varying real time alert system and visualisation and analysis tool for delay time fluctuations.  This system will enable operating staff to identify appropriate mitigation measures and where they should be best targeted in order to reduce dwell times and improve overall network performance.