Rapid Evaluation and Planning Analysis Infrastructure for Railways (REPAIR) is a set of predictive tools being developed collaboratively by Frazer-Nash Consultancy and the Logistics Institute at the University of Hull. This award-winning research was created with the goal to revolutionise freight train planners’ and controllers’ understanding of the propagation of delays across the rail network, and to help them mitigate the impact of current and future network delays on their services.
Our award-winning research helps mitigate the impact of current and future network delays.
Inspired by rail freight, to deliver game-changing solutions to the sector.
The two organisations met at a Rail Freight Group event in November 2019. They realised that, combining their complementary expertise, they could offer a cutting-edge solution to support freight train planning during network delays.
Funded under a grant from Network Rail and RSSB, as part of its Data Sandbox+ competition, REPAIR combines the Logistics Institute’s comprehensive, detailed and integrated rail network database, NR+, with Frazer-Nash’s advanced Machine Learning experience from the defence sector.
Innovation supported by collaboration with the freight industry
The REPAIR project is highly collaborative. Built upon two key pieces of underlying technology, the development work is being undertaken by Frazer-Nash Consultancy and the University of Hull’s Logistics Institute under a grant agreement from RSSB and Network Rail.
REPAIR involves the combination and further development of intellectual property that both Frazer-Nash and the Logistics Institute have developed separately, to form a significantly innovative product. A close working relationship has been formed, underpinned by the complementary nature of each company’s strengths and skills.
In a further example of its collaborative nature, REPAIR is being shaped by the involvement of an advisory board, featuring stakeholders from across the freight industry. This group provides a sounding board for the development and usability of the tool, and offers influential insights that are acting as key drivers towards its success.
Combining world-leading expertise from other industries, first of a kind network databases, and broad industry engagement REPAIR aims to deliver significant benefits to the freight industry.
Planning and scheduling of freight trains is highly complex, and the information needed is very fragmented, including a variety of databases and paper-based documents. Controllers, Planners and Signallers have to deal with a large number of short-term requests (VSTP), within a myriad of infrastructure and operational constraints.
When live delays occur on the rail network, decision-making becomes hugely challenging, and the timing of freight service delivery can become highly uncertain. A cascade of delays may result, exacerbated by factors such as train crews’ hours running out as delays lengthen.
REPAIR uses significant historical timetable data, geospatial rail databases, and historical incident data to train machine learning algorithms to predict future delays and potential routing mitigations.
Combining the University of Hull Logistics Institute’s comprehensive and detailed rail network database, NR+, with Frazer-Nash’s innovative machine learning techniques honed in the defence sector, the REPAIR toolset revolutionises dynamic train planning.
How will the REPAIR tools help freight planners?
REPAIR’s visualisation tools allow the end user to see, and understand, how delays are likely to propagate across the rail network and affect their services, as well as displaying how constraints such as crew availability will limit effective recovery. It is able to suggest potential mitigation options, through highlighting alternative routes and their likely arrival times. Controllers are empowered to make faster, better decisions, to determine which specific route to adopt.
Through innovating train planning, by enabling freight train planners to understand and mitigate the impact of current and future network delays on their services, REPAIR’s predictive tools provide increased certainty that freight operations can be delivered, and help avoid unintended potential consequences of proposed actions.
Using new technologies to deliver solutions to a long-term problem
REPAIR draws upon deep learning research, proven artificial intelligence (AI) optimisation techniques, and NR+’s freight route-finding methodology to provide innovative, rapid, predictive analysis of delays and their impacts, and offers potential re-routing solutions to help mitigate the effect upon on the network.
It first enables rapid and accurate predictions of the knock-on impact of timetable changes to be generated, considered and understood. These delay predictions cover the whole rail network, based upon current delays and incidents, and can be easily interrogated by the user.
Once these predictions have been identified, NR+’s route-finding functionality suggests potential mitigation options. By enabling freight controllers to better understand their available options and likely arrival times they are ultimately able to deliver a better experience to the end customer.