Improving forecasting of adhesion conditions
An RSSB-funded project has developed guidance for monitoring and quantifying the accuracy of low adhesion forecasts in an operational environment, to give operators and other stakeholders more confidence in using forecasts to support decision-making. could have significant safety and performance benefits and represents the first step on the journey towards an industry-approved verification standard.
The research, led by the University of Sheffield in conjunction with the Met Office and with the support of Arriva Rail North, focused on demonstrating techniques for evaluating low adhesion forecasts at a range of temporal and spatial scales.
The full set of reports, now available on SPARK covers:
- The merits and limitations of existing low adhesion monitoring techniques
- An overview of how forecasts are used and verified by industry, with guidance on defining low adhesion events and potential approaches to verification
- Case study examples of verification approaches
- Guidance on how to carry out verification of low adhesion forecasts now and in the future (with anticipated technological advances)
Forecasting Adhesion competition
This project was funded through the RSSB’s Forecasting Adhesion competition, which made £300,000 available for feasibility studies to improve adhesion forecasting technology.
Two other projects from the initiative are expected to publish their findings in 2020:
- The University of Huddersfield, in collaboration with Met Office, is aiming to combine sources of data - such as signal aspects, train speed, response to acceleration or braking - with engineering models to predict braking performance and explore future improvements.
- A team from Liverpool John Moores University, working in collaboration with Merseyrail, is constructing a framework to integrate and analyse multi-score real-time data to deepen understanding of low adhesion and provide better support for mitigating its effect.
All three projects have identified the need for more data to validate their findings. Therefore, RSSB will be working with Network Rail and a number of TOCs to run a data-collection exercise during Autumn 2019, which will build a rich repository for future research purposes. If you would like to be involved, please contact email@example.com