Human reliability assessment of door release
Human reliability assessment (HRA) is a method used to estimate the likelihood of human error based on characteristics of a human task and the task context. Two potential applications of HRA are:
- Making predictions about safety in situations where there is no incident data
- Understanding the extent to which human performance can be improved by changes to the task context. This is done by comparing real-world (eg incident) data on human performance with predictions from HRA.
In 2012 LOROL requested that RSSB provide a human factors review of train door release. In the 30-month period from October 2010 they had experienced 42 door release incidents at stations. This type of incident is not unique to LOROL. There is potential for harm to railway passengers if train doors are operated where there is no platform to disembark onto, although in practice there is no record of passengers having fallen from a train as a result of a door release incident on the LOROL network. In order to assist with risk mitigation, this study investigated train driver errors that lead to doors being released on the wrong side of a train. The study was conducted on parts of the LOROL network where the driver is responsible for door release.
This study found that on the LOROL network, door release incidents were only occurring roughly only once in every 300,000 times doors were released at stations. Nevertheless, on LOROL services train doors were released over five and a half million times per year. This meant that exposure to risk was high, resulting in over 18 incidents per year. LOROL therefore considered it desirable to make changes to reduce the likelihood of wrong side door release. When compared with the error rate predicted by HRA, the performance of LOROL’s drivers was exceptionally reliable, and this was at least in part supported by good interface design and strategies adopted by drivers to reduce the likelihood of such incidents.
HRA also informed the nature of the changes to reduce the likelihood of door release incidents. Given how reliably drivers released the doors on the correct side of the train, and the fact that most of the underlying factors were not easy to address (eg removing island platforms), a step change in human performance was unlikely to be achieved. One potential design solution was correct side door enabling equipment, which is now being introduced on LOROL routes. This equipment provides the driver with a message if he or she attempts to release the doors on the wrong side of the train and requires the driver to carry out the door release process again. As the equipment is still new it its effect on the frequency of door release incidents has not yet been evaluated.
In response to this study, LOROL commented that, “The research project and report was a thorough and professional piece of work. We hope to have the opportunity to work with RSSB’s Human Factors team in the future to provide an innovative approach to help us improve safety performance and also benefit the wider industry with the results of this work.”