Predicting and mitigating small fluctuations in station dwell times

On busy railway networks, exceeding station dwell times by only a few seconds can adversely affect overall network performance. However, while these performance impacts are well known, the causes of small dwell time perturbations are not widely understood (or are not widely communicated at an operational level) and exhibit a high level of spatial and temporal variation. A lack of information and understanding makes it difficult to implement effective mitigation measures to reduce the occurrence and impact of such delays. For this paper On Train Monitoring Recorder (OTMR) data were obtained for a large number of services over a 14month period, which included the wheel stop and wheel start timings associated with station stops at a 1 second resolution. These were combined with other relevant data sources in order to investigate small fluctuations in station dwell time. An interface for communicating these variations to railway operating staff was developed, along with models to predict future dwell time fluctuations, potentially enabling mitigation measures to be implemented.