Adhesion Research Seminar 2025

The collision between two trains at Talerddig, Powys, Wales on 21 October 2024 once again highlighted the issue of poor wheel rail adhesion, particularly in the autumn with ‘leaves on the line’. This continues to be a risk that can be realised despite the industry spending significant sums on mitigation and management.

Rail Engineer has frequently reported on the work carried out by the Adhesion Research Group. This year’s seminar included four topics including original research, experience with variable rate sanders, and kicking off studies into techniques for improving adhesion for freight trains.

Conductive sand

During the 2024 Seminar, Dr Will Skipper from the University of Sheffield reported on the residual risk that applying sand to improve adhesion could cause trains not to be detected by track circuits (Issue 207, May/June 2024). Summarising, approximately 5% of wrong side track circuit failures (WSTCF) have been attributed to sand or sandite. This led to the notion of ‘conductive sand’ and previous laboratory work had identified two promising materials. Further track trials led to just one remining: Product D, produced by LB Foster. Will described work to demonstrate that the conductive material improved track circuit detection compared to sand (known as braking sand or 10/18 sand), that it did not cause insulated block joints to become conductive, and that it was an effective adhesion enhancer.

Laboratory tests were carried out with a 10mm end post set up between two pieces of rail in full-scale rig. The objective was to demonstrate that the resistance between the two rails remained above the required 0.5 MΩ when the adhesion enhancers were applied. Measurements were taken when material was applied and after each subsequent wheel pass – three passes for each test.

Some interesting results were obtained. Neither GB ‘standard’ sand nor Product D created issues with end post insulation in dry conditions with material deposition rates up to 15g/m. Product D created issues when applied at 30g/m and above, far in excess of the application rate permitted. In wet conditions, Product D did not cause end post failure at 7.5g/m, but, curiously, GB sand did. Will suggested that this is because GB sand is less dense, has more particles, and displaced more water. Finally, the end post was lowered to simulate wear and no change in results were observed.

Track tests were carried out on the Great Central Railway on a low voltage DC track circuit using a Class 20 locomotive on sand laid at various application rates.

On the first pass, there were some isolating effects with GB sand present at all application rates. On the second pass, loss of train detection was less severe. Product D did not lose train detection on either pass.

Tests on leaf layers were carried out on the Wensleydale railway using a sander-equipped Class 142 unit and a similar track circuit. Leaf layers were created in two patches, so that the first two wheelsets were isolated upon entering the track circuit. The sand or Product D was applied via a calibrated sander set to 2kg/min, the first pass with the sander on and the second pass with the sander off. The tests showed that: (i) Product D could be applied from a typical sander set-up; (ii) an insulating layer using actual leaves could be created; (iii) GB sand exacerbated the effect of leaves; and (iv) Product D reduced voltage levels similar to clean contact.

Finally, braking tests were carried out at RIDC Tuxford using a Class 158 two-car unit at 50mph on wetted paper tape using sanders delivering 2kg/min flow. This test showed that Product D delivered stopping distances similar to those obtained with GB sand.

Summarising, Product D does everything that GB sand does without the risk of Wrong-Side Track Circuit Failures (WSTCF). The next step is a service trial which is planned to be carried out on a Northern Trains unit on a line with historic WSTCF issues.

Sticky leaves

Research continues into the fundamental issue of why leaves stick so firmly to rails and what are the influencing factors? Jordan Brant from the University of Huddersfield presented, building on work he reported at ADHERE 2023 (Issue 202, May/June 2023). This project covered: investigating the effect of different leaf species on adhesion levels and understand the bonding mechanisms related to leaf types; defining upper and lower bounds of wetting for low adhesion and correlating wetting conditions with rain rates; and quantifying persistence of leaf layers on wheels and rails and assessing its duration under varying running conditions.

Representative tests were selected using: wetting rates based on four rain rates from 0.5mm/hr to 2mm/hr; sycamore, oak, beech and horse chestnut leaf species – dried and powdered for easy application; and typical diesel multiple unit and typical inter-city train loads. Water and load parameters were adjusted to be appropriate to the University’s HAROLD test rig.

Conclusions included:

• Wetting rate is a critical factor in bonding strength and is the sole significant input on the resulting adhesion value.
• The leaf type does not affect the bonding strength, nor is there a statistical significance on the resulting adhesion.
• Leaf layer thickness is most affected by the interaction between wetting and leaf type. The thicker the leaf layer, the lower the bonding strength; thicker layers have also been shown to be made of several layers.
• A standardised method for measuring leaf layer bonding strength has been further developed.
• A further understanding of leaf layer durability has been established. In most cases, the leaf layer was still present after 24 brake tests.

During the durability tests it was observed that the leaf layer primarily built on the wheel but subsequently transferred to the rail. Rail Engineer observes this suggests that, in the field, leaf layers could be spread further than might be assumed by the location of the leaves.

Variable rate sanders

Rail Engineer has reported extensively on the development and trials of variable rate sanders, most recently in Issue 184, May/June 2020. Phil Gray from South Western Railway (SWR) and Rob Cummings from Northern talked about their experience on Class 158 and159 diesel, and Class 323 electric fleets respectively.

Following the collision in Fisherton Tunnel, Salisbury in 2021, all SWR’s Class 158 and 159 units have been fitted with Single Variable Rate Sanders (SVRS). These have been set up so that when two units are coupled, all sanders operate. Phil said that this was the first retrofit of a variable rate sand system to an ex-British Rail DMU as well as being fitted in-house within existing resources.

In spring 2022, SWR had secured £4 million from Network Rail’s Performance Innovation Fund. Following competitive tendering, Siemens Mobility was awarded the contract in December 2022. The First in Class was fitted in May 2023, with fleet fit between August 2023 and March 2024. The 39 units went  back into service with sanding rate fixed at 2kg/min. In March 2024, successful commissioning tests were carried out at RIDC Tuxford based on those defined in RSSB’s T1107 Report (Issue 158, November 2017). SVRS was enabled by a software update that was completed by September 2024.

Phil explained that there were issues to resolve as the system was rolled out but was pleased with the installation time: one day to fit and half a day to test – total 15 hours. One significant issue was that sand consumption increased by approximately 40%, something that was not experienced in the T1107 test programme. This was not an issue for the diesel trains as their time away from depot facilities is dictated by the need to refuel, but Phil said that this would be something to take into account when fitting SVRS or DVRS to electric units.

And the results? There had been three SPAD or Station Overrun incidents in 2022, four in 2023, but none in 2024.

Rob Cummings talked about Northern’s experience with DVRS on Class 323 units. He said that the Class 323 has demonstrated that retrofitting DVRS is feasible and practical, and DVRS is capable of delivering consistent braking that provides a more predictable and repeatable braking performance in very low adhesion conditions. He added that a minimum step 2 brake application (or equivalent – circa 6% g) is required to gain the braking benefits of DVRS in low adhesion. All 34 Northern Class 323 units were fitted with DVRS prior to the start of Autumn 2024 including 17 Class 323 units transferred from West Midlands Trains.

Rob said that the results were encouraging. Compared with 2020 to 2023 where there had been three, nine, four, and two station overruns respectively. There were no station overruns in 2024.

The results from encouraging drivers to take advantage of the system were mixed, and Rob suggested that it might take three seasons for drivers to develop confidence in the system through experience. This is partly because Northern’s Class 323 drivers also drive other diesel and electric units with different braking performance. That said, driver feedback has been extremely positive with the overwhelming majority believing that DVRS has had a positive impact on stopping distances. One driver reported: “Great improvement; I now feel the train will stop where I want it to.”

There were issues with brightness of lamps in the cab and, similar to SWR’s experience, sand consumption has been higher than expected. Rob reported that there was a significant number of failures in autumn 2024 which have been investigated. Plans are in hand to carry out modifications to reduce failures and to fit data loggers and labels to help monitor sand levels.

Freight

Tim Shakerley, an independent freight engineering expert, introduced a programme to improve freight train performance and safety, known as the Freight Safe Programme. It is jointly defined, resourced, and funded by freight Duty Holders, supported by RSSB, and endorsed by ORR. Its mission is “a collaborative health and safety plan designed to facilitate freight growth and to protect and enhance the reputation of the freight sector by tackling the network risks where a unified response is essential”.

The work includes several sector health, safety & wellbeing risks being managed by the National Freight Safety Group, as well as four priority projects managed by a Programme Management Office: Wagon Condition Programme, Freight Safe Insights, Horizon Scanning, and Climate Change.

Tim described the risk assessment work of the Wagon Condition project. This is the first detailed industry-wide review of whole system risks arising from the condition of rail freight vehicles. It will be used to quantify the risks and then prioritise actions to improve wagon condition.

While rail freight is inherently safe and the risks posed by wagon condition are low, there is a modest but important business case to do more. Tim added that, although Entities in Charge of Maintenance are discharging their fundamental duties, there is an opportunity for some to improve by applying best practice. He added that perhaps the biggest opportunity to reduce risk is to improve understanding of brake drag and adhesion. The Wagon Condition Programme continues to pursue those opportunities.

Paul Gray from RSSB took over and presented three projects being kicked off to explore freight train braking and low adhesion. There are some fundamental differences between freight and passenger train braking, see Table 1.

Three projects

What seems at first sight to be an unusual research project is S386 – Knowledge Search on Freight Train Braking. Paul explained that the principles of freight train braking have not fundamentally changed for many years but are not documented in one place and rely on the knowledge of increasingly rare experts who are not getting any younger.

Train performance, especially accessing train paths is largely determined by train braking characteristics and, for freight growth (longer/heavier trains, product development, performance enhancement), understanding the principles and limitations of current systems is important as a baseline for improvement.

This work will document operations, braking principles, braking performance, wheel/rail adhesion issues, and explore freight train future developments. It will also explain brake system architecture, describe single and two pipe systems including arrangements for both locomotives and wagons, and data captured by locomotive on-train data recorders and use of sanding. Some of the adhesion issues affecting freight trains are listed in Table 2.

The second project is T1350 – Understanding and Preventing Wheel Flats in Freight. Two work packages have started. The first, WP1 is about understanding the factors that lead to wheel flats in freight trains. The aim is to develop a detailed understanding of the circumstances and mechanisms that lead to wheel flats, and to provide information to operators and maintainers that can be used to reduce the risk of them developing.

The expected benefits include providing a better understanding of the root causes of wheel flats on freight vehicles; determining flat length threshold values that might result in likely wheel locking for a range of speeds and adhesion conditions; and assessing the potential/ likelihood of self-sustaining wheel flats occurring in practice (noting that these are exceptional events).

The second, WP2, covers modelling the factors that lead to wheel flats. It is aiming to determine the relative importance of the many factors that can result in the creation and growth of freight vehicle wheel flats and how they influence the level of adhesion required for rotation of a wheelset. The modelling is likely to explore the most significant factors that impede wheelset rotation; the influence of adhesion required for wheelset rotation; why some wheel flats “round out” and others continue to grow; and the viability of a “self-sustaining” wheel flat (see panel), and what conditions are required for this.

T1350 Work Package 1 and Work Package 2 are due to deliver in Summer 2025. A further work package is planned entitled Understanding Real World Freight Train Braking Through Static Testing and a potential fourth work package is Enhanced Investigation of Wheel Flats.

The third project is T1351 which will investigate the factors that affect freight braking in low adhesion conditions. It is planned that the project will explore the interactions between various influences, such as freight braking characteristics, environmental conditions, and freight driver policies and how these differ from braking behaviours in passenger trains. It will also review available freight air braking models and evaluate their ability to account for the interactions assessed within the project. The research will then assess the resources required to optimise the most suitable model(s) for GB freight operations. This project is due for delivery in early 2026.

The ADHERE Seminar 2025 covered a wide range of topics which will in time contribute to a safer railway. It was good to see practical experience of SVRS and DVRS being reported and that attention is now being paid to the issue of braking adhesion for freight trains.

With thanks to the presenters and RSSB’s Ben Altman for their assistance.

Self-sustaining wheel flats

Self-sustaining wheel flats is a term used by the RAIB in its report on the derailment at Petteril Bridge Junction on 19 October 2022. In its recommendation 1, RAIB stated:

“Network Rail and the freight operating companies should work in collaboration with RSSB to review the risks faced by freight wagons during normal brake applications in foreseeably low adhesion conditions. This work should include a detailed assessment of the risk of individual wheelsets sliding sufficiently so that they generate self-sustaining wheel flats that can ultimately lead to derailment. It should also identify what mitigations may be necessary to ensure that these risks are adequately controlled.”

From a subsequent V/T SIC paper:

“The mechanism which is postulated is that low adhesion under a normal brake application caused one wheelset to stop rotating and that, following the brake release, the wheel continued to slide. The presence of the wheel flat prevented the wheel from resuming rolling, and sliding of the wheel over the rail caused the wheel flat to continue to grow until the ‘false flange’ was too large to allow the wheel to pass through S&C, whereupon derailment occurred. The wheel flat was therefore ‘self-sustaining’.”

Image credit: Will Skipper