The Adhesion Research Group (ARG) is a sub-group of the Vehicle/Track System Interface Committee (V/T SIC), facilitated by RSSB. ARG sponsors the ADHEsion REsearch challenge programme (ADHERE). Both groups promote an annual seminar V/T SIC & ADHERE. The aim of ADHERE is to achieve ‘adhesion conditions that are unaffected by and independent of the weather and climate’.
Four presentations explored characterising adhesion in service, assessing applicability and business case for adhesion improving technologies and occupational health risks from sand used as an adhesion improver.
Wheel slide protection
John May, from Knorr-Bremse Rail Systems UK and Louis Schmandt, from Chiltern Railways presented the results of tests to provide adhesion calculations from the Wheel Slide Protection (WSP) system of a Class 165 two-car unit. The question “how bad is the adhesion?” has hitherto been hard to answer. This trial sought to provide an insight into not only where and when low adhesion is occurring but also to what extent and notably, not just within the thresholds of WSP activity; a stark contrast to common, high-level reports to date.
This required information from several systems.
The train sends brake demand, speed, GPS location, and an adhesion calculation to a cloud server. These data are combined with other sources (e.g., TRUST, schedules) to provide information that includes low adhesion hotspots (with alerts where necessary), precise GPS locations, infrastructure issues, braking exceptions, brake demand plotting, and speed on braking.
A year long trial has been carried out involving Knorr-Bremse, Chiltern, Angel Trains, ARG/RSSB, and the University of Huddersfield. Unit 165004 was upgraded with KB’s Electronic System for Railway Applications (ESRA) WSP system. 4G/5G was used to transmit data and GPS location to the cloud. The system was tested on the Chinnor and Princes Risborough railway in July 2025 and has been providing in service data since September 2025. From just this one unit a huge amount of information has been learned.
The information included:
- Wheelslide and traction slips by formation. About 75% of the unit’s mileage was in multiple with one or more other units. The unit was more prone to wheelslide and much more prone to traction slips when operating on its own.
- Adhesion by time of day. The worst adhesion times were between midnight and 09:00.
- Leaf fall/winter adhesion by day. By far the worst month for adhesion was in October, but there were hotspots which mostly coincided with named storms.
- WSP activity by axle. WSP activity was highest for the leading bogie (axles 1-2) which is in front of the single sander on the third axle. WSP activity was low on axles 3 and 4 and gradually increased on the other two bogies.
- Braking. Over 95% of brake applications were in step 1 which is in line with Chiltern’s professional driving policy especially in the autumn. 99% of brake events did NOT result in a slide. Comparing deceleration against brake demand showed that the brake system was working well.
- Low adhesion hotspots. A list of stations by probability of wheel slide showed that some stations might experience wheelslide regularly (up to 50% of stops) whilst others experienced them rarely.
- Traction slips. Tended to mirror wheel slide activity. Only four of eight axles are powered on a Class 165 so they need more adhesion for a given acceleration.
Based on your writer’s experience, if this much knowledge is gained from a single unit, enabling more units should enable a more complete picture to be obtained and, is likely to deliver further insights as engineers gain experience and discover they can learn even more.
Louis explained that this trial is part of a greater ARG goal to develop an “adhesion expert train”, which is able to determine exactly where and when low adhesion is likely to occur as well as offer insights on what mitigating measures may be undertaken as a result (e.g. vegetation management, railhead treatment). The aim is to achieve this by fitting additional hardware on Unit 165004 to enable further validation of the low adhesion against data from other systems monitoring trackside vegetation through Light Detection and Ranging (LiDAR), and railhead condition through Automated Intelligent Video Review (AIVR).
Making improvements
Two presentations proposed techniques to help justify improvements, both of which were reporting on RSSB projects inspired by Recommendation 9 from the Salisbury Collision in 2021. This stated:
“The Rail Delivery Group working with the train operating companies and Rail Safety and Standards Board should create a framework and mechanism for the assessment of future technologies to enable trains to better cope with low adhesion when braking.
“The framework should set out criteria and establish the process for cost benefit analysis to apply to the assessment of future technologies as they arise.”
RAIB intended the recommendation to encourage realisation “…of potential benefits of future technologies to enable trains to better cope with low wheel/rail adhesion when braking”. Two projects were set up.
Dr. Martin Evans of University of Huddersfield and Dr. Will Skipper of University of Sheffield presented T1341 ‘Framework for assessment of technologies to deliver reliable braking in low adhesion’. The aim is to develop a technical framework which will set performance targets and assessment criteria against which low adhesion mitigations can be assessed. This will support a strategic approach to deploying existing mitigations and provide for the assessment of solutions to manage low adhesion. The project covers both infrastructure and vehicle mitigations. Two key concepts were proposed:
Signal Adhesion Class: “adhesion required to stop in time for a Danger signal from its first caution signal, with consideration of the risk and consequences of low adhesion” – expressed as %g adhesion. A series of these can be aggregated to a Route Adhesion Class.
Vehicle Adhesion Class: “the ‘assured adhesion’ a vehicle is able to provide during Full-Service braking, even in severe low adhesion conditions” – expressed as %g deceleration on level track.
In practice the assessments will be agnostic of gradient, so it is convenient to use % adhesion for both Signals and Vehicles.
Based on minimum signalling braking distances defined in standard GKRT0075 (appendices A, B, and C) and gradients, the following Signal Adhesion Classes were defined:
- Class A, (8%). For passenger trains operating on lines signalled to Appendix C, or other lines where speeds exceed 110 mph.
- Class B, (6.5%). Passenger trains operating on lines signalled to Appendix B, up to 110 mph, or Appendix A between 75 mph and 110 mph.
- Class C, (5%). For passenger and freight trains operating on lines signalled to Appendix A up to 75 mph.
- Class D (Declassified – consider as “Dry”). Low risk and low consequence of low adhesion –considering Network Rail’s ‘Leaf Fall Risk Assessment’ and hazards beyond the danger signal.
Similarly, Vehicle Adhesion Class has been defined according to the level of “assured adhesion” a vehicle is able to provide during Full-Service braking, given as a percentage:
- Class A+ (Seasonally Agnostic Train). Capable of ‘dry’ braking performance (up to ≥9%), regardless of railhead adhesion conditions.
- Class A (8%). Able to restore reduced adhesion to achieve 8% adhesion, with reducing percentages for classes B, (6.5%), C (5%), D (3.5%), E (2%). Finally, Class U (Unclassified) covers trains that cannot meet the requirements of Classes A to E. There will be two methods of assessment: ‘Equipment Definition’ – a conservative classification based on knowledge of the train’s braking system(s) and ‘Simulation’ which is expected to be more accurate and less conservative.
An example application of this approach was illustrated for the area between signals SY29R and Fisherton Tunnel, the location of the low adhesion that was instrumental in the Salisbury collision. The actual adhesion level/deceleration rate achieved by the collision train was 2.9% putting it in Class E (2%). This means that the assessment indicated that there was significant risk of serious consequences in poor adhesion and further mitigations are required.
Potential use cases of this approach include:
- Assessing the capability of vehicles against their routes (i.e. “Service Group”).
- Prioritisation of routes for Railhead Treatment Trains.
- Identifying suitable braking points for signals and/or stations, based on vehicle performance.
- Supporting the technical case for acquiring or retrofitting rolling stock with additional/future adhesion mitigation equipment.
The team proposes to carry out case studies on other sections of railway with different types of rolling stock and service patterns.
Cost impact
Dr. James Jackson from SYSTRA described T1340 ‘Economic evidence to underpin investment decisions for adhesion mitigations’. The study aims to establish a contemporary understanding of the cost impacts of low adhesion with enough detail to help inform investment decisions for specific routes and fleets.
The study covers both financial and socio-economic impacts i.e., people and money, safety, train performance, and cost/savings. Cost Benefit Analysis (CBA) aims to capture all these but, inevitably, the financials are more precise than the socio-economic benefits.
Previous work on business cases for adhesion mitigations (prior to this study) has primarily been based on network level business cases which sometimes lack focus on specific adhesion issues.
Business cases have also tended to be purely financial, capturing cost savings and performance benefits estimated through Schedule 8. Socio-economic benefits tend to be excluded which means the appraisals deviate from DfT TAG good practice.
In seeking to understand the adhesion issue, James had analysed delays based on type of operation (commuter, Inter City etc.),Train Operator, and by service groups. He illustrated two case studies.
First, for the Calder Valley, the fitting of Single or Double Variable Rate Sanders (SVRS/DVRS) to classes 158 and 195 in two, three, and four-car formations demonstrated a benefit to cost ratio of 0.9 to 2.09 depending on the amount of journey time saved through use of sanders.
The second case study suggested removing Rail Head Treatment Trains from the Midland Main Line between Bedford and Kettering and using the savings to fund SVRS or DVRS on 12 units working the Matlock/Worksop/Skegness service. This produced a financially positive case, although caveated that assuming RHTTs could be removed still needed to be validated.
The output is an appraisal toolkit which provides a consistent approach to the assessment of the costs and benefits associated with adhesion mitigations. The tool can deal with local, regional, or national level interventions. As always, the output is dependent on the quality of the inputs and the results sometimes need more evidence about the efficacy of some mitigations. That said, the results are promising with the scope for some financially positive outcomes and it shows the benefits of a whole industry approach.
Health risks
The final presentation explored the occupational health risks in depots relating to silica dust for sanders delivered by Dave Flower from the Institute of Occupational Medicine (IOM).
Sand contains silicon dioxide (SiO2), also known as silica. The Health and Safety Executive (HSE) has noted that some silica dust is “…fine enough to get deep into…” the lungs. This fine dust is known as Respirable Crystalline Silica (RCS) (see Panel). Prolonged and heavy exposure to RCS can lead to lung cancer and other serious lung diseases including silicosis and chronic obstructive pulmonary disease. Non-crystalline (amorphous) silica is considered less hazardous than RCS however some reports suggest that amorphous silica may potentially cause respiratory diseases.
ARG requested research (T1393) to understand the risk associated with sander replenishment, maintenance and testing, and what sufficient risk reduction or control should look like. Some sand alternatives have been trialled by other researchers and ARG also requested that IOM assess the health risks of some of these be compared with current GB rail sand products. The presentation provided the project’s emerging findings.
Data was gathered about personal exposure at five depots between November 2025 and January 2026. This included taking samples and reviewing control measures being used, including respiratory protective equipment. Typical results showed that the level of RCS was below the UK Workplace Exposure Limits (WEL) – these being legal, maximum concentrations of hazardous substances in the air, averaged over set periods, designed to protect workers’ health – and in most cases, below the analytical limits of detection.
Respirable amorphous silica concentrations were below the analytical limit of detection and WELs, and respirable dust concentrations were well below the WELs. There was still ongoing assessment work to perform a more formal analysis of exposures and the needs for additional control measures to minimise worker health risks.
A literature review focussed on sand handling activities (including for industries other than rail) was undertaken to identify high risk scenarios, use of control measures and their effectiveness, as well as exposure levels experienced by workers in these sectors. High risk scenarios are linked to agitation of dry sand, bulk transfer and cleaning, and confined/poorly ventilated environments.
Other key findings were that: (i) material characterisation of rail sands should be improved; (ii) there is a need to strengthen documentation of tasks and job roles; (iii) quantitative research on rail-specific engineering control performance should be improved; and (iv) the findings can be used to optimise risk management strategies for rail sand handling activities.
Toxicology assessments were carried out on GB Rail Sand (fine and coarse) and three sand alternative products using respectively aluminium oxide/titanium oxide, aluminium oxide, and crushed glass. Analysis suggested that titanium oxide is possibly a carcinogen while aluminium oxide and glass risk dust-related respiratory irritation and particulate lung effects at sufficiently high airborne levels. This compares with GB Rail Sand (fine and coarse) dust which is carcinogenic in humans (due to RCS) and also risks other diseases from long term inhalation of respirable dust.
Next steps include completing exposure and risk comparisons using data collected from depots and development of a good practice guide.
Conclusion
Assessing the level of adhesion directly from how trains perform has long been an ambition so that a real time picture of conditions can be built. The assessment and cost benefit work will provide tools that will help implement improvements more quickly. For example, SVRS and DVRS were demonstrated to give a significant improvement in stopping capability in 2017, yet today only a handful of trains have been equipped. Finally, it is good to see that whilst sanding improves the safety of the railway, the health and safety of the people handling and using it must not be neglected.
Image credit: Chiltern Railways