Line closures in the north of Scotland on Monday (5 January) were followed by more snow-related disruption in Wales, the North West of England and the West Midlands. Network Rail provided multiple updates on its response to the cold weather, including checking tunnels for low-hanging icicles, which can damage trains and overhead power cables, and checking points heating systems, to ensure they were not jammed by snow.
“Severe weather doesn’t respect railway timetables, which is why our teams are ready to respond throughout the year to conditions including strong winds, flooding, snow and ice,” a spokesperson told Professional Engineering yesterday (12 January).
“We use a wide range of tools to monitor and manage these risks, including our dedicated 24/7 weather desk with in-house meteorologists, our own helicopter for aerial inspections, and specialist kit such as snow ploughs and de-icing trains to keep the railway safe and operational.”
Mini snow ploughs, fitted to route-proving locomotives owned by freight companies, are the most common type used by Network Rail. Adjustable depending on snow depth, they are capable to about 18 inches (46cm). Small snow ploughs can also be fitted to passenger trains to clear up to eight inches (20cm) of snow, but heavier accumulations might require the operator’s most robust Beilhack V or Independent snow ploughs.
Ploughs can also be fitted to the front and rear of snow and ice treatment trains (SITT), which scrape ice from the top of conductor rails and spray them with hot liquid anti-icer to prevent ice sticking. 11 SITT operate in Kent and Sussex, while 24 multipurpose vehicles also offer similar capabilities on third-rail routes across a wider area. A third type of de-icing train, known as the winter development vehicle, blows hot air around points and crossings to remove snow and ice, with ‘steam lances’ for more stubborn deposits.
The operator’s most extreme options are its two snow blowers, based in Scotland. These use large propellers to cut through and blow away snow drifts, and use an internal hydraulic turntable to turn around on their own if needed.
Similar machines are often used in countries that face more frequent, deeper snow. In November last year, Norwegian infrastructure manager Bane NOR received the first of two new Schmidt Beilhack HB 1100S snow‑clearing machines. Capable of clearing up to 10,000 tonnes per hour, the 2,200 horsepower heavy rail vehicle can throw snow up to 40 metres and travel at up to 100km/h.
As climate change becomes more pronounced in future, the demand for snow-clearing services could vary. Snowy days are becoming more of a rarity in the UK as average seasonal temperatures rise – although some weather pattern changes could actually lead to lower temperatures and the potential for more disruptive weather, as highlighted at an IMechE event last year on the potential slowdown of the Atlantic Meridional Overturning Circulation.
Network Rail will spend roughly £400m to strengthen the Scottish network against extreme weather between 2024 and 2029, the spokesperson told Professional Engineering, including £100m for drainage and lineside infrastructure, and more than £40m for bridges and tunnels.
While the UK operator did not specify plans to adopt new technologies or techniques for snow clearing, autonomous and AI-based tools could offer a promising route, as with other aspects of inspection and maintenance.
An article published in Intelligent Transportation Infrastructure last year proposed a deep learning approach for detecting rapid accumulations of snow on railway switches in severe storms. Manual inspections and routine monitoring are “labour-intensive, time-consuming and susceptible to human error,” wrote the researchers, who are based at several Chinese universities. Periodic inspections mean potentially delayed responses during “dynamic snowfall”, they added.
Instead, the team set out to improve on previous deep learning techniques, which could address limitations of traditional methods. Using a combination of video images collected from trains and synthetic data, and a framework including advanced data augmentation and optimised preprocessing, they said they were able to “significantly improve the accuracy and robustness of snow detection under diverse real-world conditions”.
Lightweight models could eventually be installed on ‘edge’ devices running on rail routes, they wrote, offering another option for responding to fast-changing winter weather conditions.
Want the best engineering stories delivered straight to your inbox? The Professional Engineering newsletter gives you vital updates on the most cutting-edge engineering and exciting new job opportunities. To sign up, click here.
Content published by Professional Engineering does not necessarily represent the views of the Institution of Mechanical Engineers.