In this challenging environment, cooperation between organisations can be harnessed to help the rail sector not only navigate these challenges, but to innovate and excel. Embracing innovation is essential for delivering efficiency and productivity to meet the complex demands of modern rail systems. Recent major advancements in machine learning and data analytics have resulted in digital skills and data science becoming invaluable for optimising processes and innovating. Typically, no single organisation has all the skills and connections within the rail sector to implement these novel solutions. Within Frazer-Nash Consultancy, collaboration across diverse stakeholders and supply chains is standard practice and required to solve our client’s problems.

Frazer-Nash is a leading engineering consultancy at the forefront of leveraging data science and machine learning to tackle real-world challenges across the transport, defence, and energy sectors. Their teams of data scientists and engineers have extensive experience in the optimisation of operations and enhancing decision-making processes. As increasingly large amounts of data are generated, these skills – required to make sense of data and build tools for predictive analytics and informed decision making – have been increasingly in demand by customers.

Through two case studies, we uncover the transformative impact that collaboration between organisations, each bringing their own specialist skills, can have towards innovation within the rail sector.

PosITrack – Frazer-Nash Consultancy and eviFile

Rail possession management is central to the safe and efficient delivery of Network Rail’s programme of maintenance and renewal works in 2024 and beyond. Possession management revolves around delivering essential works in and around the track, often requiring the diverting, blockading, and restricting of services across the rail network.

Handing back the track on time after engineering works is vital to allow services to make their journeys as planned. With increasing pressure and congestion in the rail timetable the opportunities for possessions are less obvious and the potential for delays to freight and passenger services heightened. Enabling Network Rail and their contractors to better manage and mitigate their risks from the planning stages right through to delivery is crucial to allowing the whole rail network to run to time and schedule.

eviFile has delivered a real time possession management solution for Network Rail covering multiple projects across the UK, for Tier 1 contractors such as VolkerRail and Alstom, and the entire Transpennine Route Upgrade. The possession management solution acts as the single source of truth, enabling auditable data to be collected and visualised live across possession activities including access times, activity completion and quantitative tracking.

Having met at the 2022 Railway Industry Association Innovation Conference, both organisations recognised that by combining their complementary expertise, they could offer a game changing advancement in how possessions are managed. The combination of Frazer Nash’s machine learning and data science principles with eviFile’s rich insight and field data engine provides a unique opportunity to collate, manage and provide actionable recommendations to possession planners.

Funded by the Department for Transport through a Small Business Research Initiative First of a Kind grant, Frazer-Nash Consultancy and eviFile worked together to develop a possession optimisation and forecasting tool to revolutionise the understanding and management of rail possession planning and delivery.

Using rich historical possessions data collated by eviFile, Frazer-Nash developed an AI toolset that leverages advanced statistical methods to create a digital twin of live possessions. Given live trackside data fed from the eviFile platform, the tool makes forward predictions on the likely completion time of all activities using learned historical experience, and sensitivity analysis to flag key programme delivery risks. Notably, through its interactive and visual dashboard, the tool allows the nominal schedule to be modified on the fly allowing mitigation plans to be tested and prepared when works are behind schedule.

The initial trials of this tool have shown that it can predict significant activity risks and challenges to plans ahead of time. As a result, the solution is now being developed so that it is suitable for more widespread use by Network Rail and its major contractors. This new tool will allow contractors to better plan possessions and operate to tight schedules and deadlines with greater confidence. The tool will allow possession managers who are not experts in AI to harness its power and identify bottlenecked activities at the earliest opportunity, inform impacted stakeholders, and test remedial plans of action when works are behind schedule.

REPAIR – Frazer-Nash Consultancy and Lampada

Understanding how delays will propagate and recover on the UK rail network is a complex challenge; the network has a complicated infrastructure and is used heavily by both passengers and freight. With information recorded by every train as it passes through the thousands of timing points on the network, there is an abundance of data that is too detailed to give operators the holistic view required to make informed decisions.

However, if aggregated and presented suitably, this data has the potential to provide valuable insight.

Frazer-Nash Consultancy with a strong track record in implementing innovative Machine Learning and AI solutions recognised that through collaboration, they could put the data available from Lampada Digital Solutions NR+ environment to work. Lampada’s NR+ environment is a rich source of digitalised data required for effective route planning and analysis on the UK rail network, creating a strong base for bespoke applications focussed on improving efficiency and effectiveness of the network’s capability.

Originally funded under a grant from Network Rail and RSSB as part of its Data Sandbox+ competition, Frazer-Nash Consultancy and Lampada developed REPAIR (Rapid Evaluation and Planning Analysis Infrastructure for Railways).

REPAIR applies a machine learning model trained on historic delay data to live data feeds from Network Rail. This allows predictions to be made on how delays might improve or worsen in the following hours, akin to ripples across a pond. A visualisation tool provides an overview of these delays on a map of the rail network, allowing users to quickly identify locations that have delays which may impact their operations.

Extensions to the predictive model have also been developed utilising NR+’s routefinding methodology to offer alternative routing solutions and allow the user to run ‘what if’ scenarios to evaluate the impact of different incidents on the system. These extensions were designed following engagement with an advisory board, made up of major players in the rail freight operating sector.

REPAIR could empower controllers to make faster and better decisions. In the UK, the required information for planning and scheduling is fragmented. REPAIR enables operators to at a glance visualise train delay across the entire rail network – or in specific areas – as well as to see how train delay is likely to develop over the next few hours. This improved understanding of the current and predicted state of the UK rail network has the potential to allow train operations staff and freight planners to better replan services when significant disruption is expected and so minimise the effects on services, the travelling public and freight customers.

Since completion of the Sandbox grant, Frazer-Nash Consultancy have pursued a partnership with a major OEM to produce a robust product for use in station and operations control rooms. This product was demonstrated last year in the Network Rail Communications Centre to showcase its potential to the industry. Frazer-Nash are now additionally talking to SMEs and other organisations about how best to bring REPAIR to market.

Conclusion

The benefit of collaboration between organisations is exemplified in these two case studies. Through combining their respective areas of digital expertise and leading AI capability, they have produced innovative products with the potential to improve the efficiency and reduce costs of the rail network. These are just two examples but there are likely many more similar opportunities across the rail industry waiting to be exploited.

 

Tel: 07892 705 449
Email: rail@fnc.co.uk
Visit: www.fnc.co.uk/what-we-do/your-industry/transport/rail/