RSSB will use AI to predict the condition of low-adhesion rails

The Rail Safety and Standards Board is collaborating with the University of Sheffield to develop an AI-powered tool to help predict low-adhesion track conditions. Railway Supply magazine writes about this with reference to the Global Railway Review.

Low adhesion track conditions are a serious safety and operational issue for the rail industry, costing around £350 million each year. It not only causes delays affecting train performance but can also result in station overruns and signals being passed at danger.

Tighina-Basarabeasca-Giurgiulesti railway section is in the process of major renovation

Temperature, humidity, and the presence of layers of leaves or other contaminants affect the level of adhesion between a train wheel and rail. The project will use artificial intelligence to analyze data and record high-resolution video to more accurately predict friction at the wheel-rail interface. One outcome of the project will be an online tool that allows users to enter data that will generate friction predictions for any location on the network by fall 2023.

“While people may think of leaves on the line as a joke, or just an excuse used when a train is delayed, the reality is that it’s a very serious issue for the rail industry,” said Paul Grey, lead technical officer at RSSB.

The new research project will use artificial intelligence and data analytics to predict and determine where and when low adhesion may occur on the rail network. This will allow targeted action in those specific locations to help manage security risks and reduce delays. Fundamental analysis of the causes of low friction as well as extensive track data collection are now coming together to enable the development of predictive AI that will help the rail industry address performance and safety challenges in the fall.

Railway news you might have missed:

The current volume of exports of corn and flour exceeds last year figures