Frazer-Nash on track with RSSB freight train project

Frazer-Nash is undertaking a project for RSSB (Rail Safety and Standards Board) to explore how data can improve the performance and short-term planning on the rail freight network.

Frazer-Nash
Freight service near Crowle, Lincolnshire (Image: Network Rail)

The project, Rapid Evaluation and Planning Analysis Infrastructure for Railways (REPAIR), is one of ten awarded funding under a grant of £4m by RSSB in collaboration with Network Rail.

Multi-million pound digital upgrade for UK freight trains

Working in collaboration with the University of Hull’s Logistics Institute, it will examine how data analytics and machine learning techniques can offer new ways of predicting and mitigating the propagation of delays on the network.

“Funded by RSSB under its Data Sandbox+1 competition, we’ll be bringing together Frazer-Nash’s skills and expertise in Machine Learning with the University of Hull’s powerful NR+ suite, to develop a predictive engine for propagation of delays on the rail network,” said Chris G Jones, REPAIR project manager for Frazer-Nash.

According to Jones, route setting and freight planning have been carried out with a combination of expert knowledge and laborious paper systems.

“The development of the University of Hull’s data tool, NR+2, has significantly improved this process – as a visual rail capability mapping system, it provides a powerful toolset for interrogating and searching the infrastructure constraints of the network,” he said. “The REPAIR project will further build upon NR+, to explore how data and machine learning can inform freight operators’ very short-term planning (VSTP) in response to incidents on their networks. It will develop a tool to enable freight operators and Network Rail to better predict the effects of the VSTP decisions they make, supplementing the functionality within NR+ to allow them to consider both current delays on the network, and future predicted delays.

“These predictions will then be embedded into NR+, and will provide an additional level of decision support for signallers and control staff when evaluating the optimum response to run a VSTP request. Once developed, the machine learning toolsets will be able to carry out ‘So What If’ testing for freight planners, signallers and controllers.”

An advisory board has been formed with ‘major players in the rail freight operating sector’ to help steer this work.