Operators of locomotives (and trams) have used sand to enhance traction adhesion for as long as anyone can remember; the first recorded use was in 1886. Over the last 20 years or so, as the combination of autumn leaf fall and modern short trains initially led to an increase in station over-runs and signals passed at danger because of lack of adhesion, the rail industry started experimenting with using sand in braking. The first modern UK standard for sanders on multiple units was issued in 2001 by Railway Safety, the RSSB’s predecessor, and it set the requirements for sanding equipment.

This early standard, based on the available equipment at the time, required mandatory fitment for new vehicle designs and was non-mandatory for retrofit. It encouraged sanders to be controlled by the wheel slide protection (WSP) equipment, but originally only allowed sand to be applied in maximum service and emergency braking positions. Manual operation of sanders in traction was also allowed. So-called “single shot” sanders, which applied all their sand in one application, were also allowed.

As sanding developed

Sanding is not universally welcome. Infrastructure managers would be happier if sand was not dumped onto their ballast and signalling engineers have always worried that sand might interfere with the operation of track circuits and delicate points mechanisms.

There were also a number of practical issues. These included the arrangements for detecting low sand levels in sand hoppers and the organisation of sand hopper replenishment. That said, as the benefits of sanding became apparent, more and more sanders were retrofitted so now there is a reducing minority of passenger trains that do not have them.

Sanding in brake step 2 (or 50 per cent brake for trains with step-less brake control) is now allowed, following the realisation that, with very low adhesion, wheels could also lock at low brake rates, and that it is counter intuitive to manually release the brakes, so as to get the wheels turning and then re-apply maximum service or emergency brake to get the benefit of sanding.

It is now universally accepted that sanders are an important part of ensuring that trains stop in the correct distance, although adhesion-related incidents still happen and there is still room for improvement.

The original standard mandated a maximum discharge rate of 2kg/minute, which was fixed irrespective of train length or speed. This was partly a limitation of the equipment then available and partly due to nervousness about the risk of sand debris between the wheels and rails causing wrong-side track circuit failures.

Gradually, experience supported by risk assessments has shown that using sand to achieve the best possible stopping distance delivers both the least risk and enables high throughput and closer headway operation. More than ever, the railway requires confidence that trains will stop in the required distance, especially where increased capacity is sought.

Testing alternatives

It has been suggested that variable discharge rates (more sand at higher speeds) and/or distributed sanders (more sanders spread along the train) might further reduce the risk of adhesion-related incidents. This begs many questions. How much sand? How many more sanders? And on which axles should they be mounted?

This is a complex, multi-dimensional issue. Whilst some evaluation can be done with simulations, testing is still required to optimise the solution. The testing would be complex with so many variables. A test train would require a number of sanders spread down the train and each sander needs to be capable of having its sanding rate varied and having each sander isolated. Moreover, poor adhesion has to be simulated predictably.

A tall order? Possibly, yet it is exactly what RSSB is doing in research project T1107.

Rail Engineer was invited to Old Dalby to see the testing. The invitation read: “RSSB’s objective is to determine the optimum distribution of sand from trains to improve performance of the railway in low adhesion conditions.

“RSSB has collaborated with a cross-industry team using Class 387 rolling stock at the Rail Innovation and Development Centre Melton to test the performance of various sander configurations in simulated leaf fall/low adhesion conditions to determine the optimum arrangements for future fitments.

“The project is carrying out an intensive programme of test runs with multiple sanders in different locations on the train using both fixed rate and variable rate equipment, building on previously completed research work.

“This complex project includes using vehicle-mounted instrumentation and monitoring equipment, temporary fitment of additional sanders, and vehicle on-site logistical and planning activities.”

The project

On a mild, dry, slightly windy day in early October, Rail Engineer arrived at the Old Dalby Goods Yard to meet Paul Gray, professional engineering lead, and Justin Willett, professional operations and performance lead, from RSSB’s R&D team. They outlined the scale of the cross-industry team cooperation on this significant project.

RSSB is providing project management, technical direction and communications. GWR has loaned two new Class 387 EMUs to the project, with manufacturer Bombardier and owner Porterbrook supplying technical, logistical and commercial support.

For the Old Dalby test track, now RIDC Melton, operator Serco and owner Network Rail are supplying the test track, depot services and operations staff. The train drivers come from Freightliner, which is also maintaining the trains and hauling the transit moves.

Ricardo Rail is leading a multi-disciplinary team, which includes DB ESG Rail, Serco, Knorr-Bremse and Bridgeway. Together, they are contributing the design, approvals and installation of temporary modifications to the test train (including sander hardware and control systems, train instrumentation, low adhesion equipment), creation of the low adhesion railhead conditions (by application of paper tape throughout a 1km test zone), provision of test strategy and engineering documentation, test management and safety staff for track access.

The work is sponsored by industry’s Adhesion Research Group (ARG), with a project steering group drawn from industry representatives and experts.

Sander test programme

Following a recently completed research project (T1046), the standard was revised to allow for trains formed of two or more units coupled together to use the existing sanders on the non-leading units, where previously the practice has been to use only the leading sander on axle 3. For example, an eight-car train formed of two four-car units can now use two sanders compared to one. This practice has now been adopted by a number of TOCs.

Paul Gray told Rail Engineer that, to support further implementation across other fleets of multiple units, this project was carrying a number of test runs using two four-car Class 387 units coupled together, applying sand on axles 3 and 19 (the third axle on the second unit) which is a typical formation, especially in the South East.

Twenty days of testing have been organised, which allows for approximately 150 tests. Most of the tests are being carried out on a four-car Class 387 unit which, in standard trim is fitted with sanders in front of axle 3 on each cab car. However, for this work, the Ricardo /Knorr-Bremse/ESG consortium had modified these to deliver fixed and variable rate sanding. Additional sanders had also been fitted in front of axles 7 and 11, as well as control equipment to allow the testers to select which sanders to use and whether fixed or variable rate. These additional sanders only allow for testing in one direction.

The sanders were demonstrated to Rail Engineer to illustrate how sand is fired into the “nip” between the wheel and the rail, but it was stressed that operation when the train is stationary is not allowed in service.

To avoid the project having to interfere with the train’s braking system, the sander configuration and operation is controlled manually. The only interface with the braking system was to obtain a brake initiated signal. Brake cylinder pressures and axle speed probe signals are recorded.

The train has also been equipped, at the leading end, with water spray equipment able to spray at two rates with plain water and with detergent solution, as required, depending on test conditions.

The test objective is to show whether, and in what combinations, stopping distance can be further reduced by the application of sand to more wheelsets and, if so, to which axles. It is also intended to investigate how the application of sand at a variable rate with speed can deliver more sand at higher speeds without exceeding the 7.5g/m of sand per rail currently allowed. (Using a variable rate sander with a discharge rate of up to 4kg/minute can provide 7.5g/m at 20mph reducing to 2.5g/m at 60mph.)

However, these were not the only variables. Compressed and moistened paper tape has been shown to provide comparable adhesion properties to crushed leaves, and this was used for the testing. The simulated adhesion conditions depend on whether paper tape is applied and its condition (roughly five sanding test runs wears off the tape).

The volume of water applied by sprays at both ends of the train to moisten the tape, and the concentration of detergent (zero, one or five per cent), also has a bearing. Interestingly, small quantities of detergent can help maintain low adhesion when the weather and rail temperatures are warmer.

The final variable is the weather; hot, cold, wet, dry. If it is too cold/damp, the paper tape will not stick properly to the rail. If the weather is hot/dry, it can be difficult to keep the paper tape surface wet enough to give low adhesion. If there is a cross wind, it can affect the sand jets. However, on the day that Rail Engineer visited, the weather was almost ideal.

With all these variables, it’s important to make best use of the available testing time. Stuart Brown of ESG demonstrated two flow charts that help him and other test engineers set up and maintain low adhesion conditions.


Before each set of tests, the team from Ricardo Rail, supported by Bridgeway Consulting, apply the paper tape to the rails over a distance of one kilometre using a bespoke trolley.

Once the tape is in place, the train will undertake a number of runs over the tape to condition it ready for testing. Then, when the tape is ready, the train will run northbound at 20mph for a final conditioning run and also applying water to wet the tape.

The train travels beyond the tape to allow enough space to accelerate up to speed. The first test run is without any sand, to demonstrate that the tape is providing a low adhesion condition. For each subsequent run, the test configuration is chosen – which axles for sanding, fixed or variable rate, detergent concentration (or none), water spray volume. As an example – sanding on axles 3 and 7, fixed 2kg/min flow rate, one per cent detergent concentration, spray nozzle.

Finally, on the word from Steve Mills, RSSB’s project manager, driver Tony Orr from Freightliner sets off and accelerates to 55mph. Dan Hamm, RSSB’s project engineer, starts the water spray, and at the agreed point, Tony applies step 3 full service brake (9%g), and at the same time Dan starts sanding, whilst Serco’s Dan Ling controls and monitors all the instrumentation including video images of the sand applications.

There is a continuous video feed of two sand nozzles, illustrating how the sand jet scatters and how it varies with train speed and wind; a screen shot of the video shows this clearly. The sand is cut off just before the train stops and, on one or two of the runs, the wheel stopped rotating just before the train stopped. The instrumentation records all brake cylinder pressures, all axles’ speed, from which WSP activity is inferred, as well as location, distance travelled and the video camera feeds.

This instrumentation allows the engineers to see very quickly how each of the 16 axles performed during the test. The test equipment is fed with electricity from the train’s auxiliary supply supported by an uninterruptable power supply, which prompted Steve Mills to observe that the test team are fed with sandwiches and doughnuts!

After each test there is a brief review of the data and then Stuart Brown observes the condition of the paper tape to determine, from the flow chart, the flow rate and concentration of water/detergent for the return trip before the next run. Typically, after 5 or so tests (one move in each direction), the paper tape is worn out.

During the afternoon’s tests, braking performance differed markedly. For a 9%g brake demand, achieved performance varied between 3.5%g (no sand) and 7.5%g. As an observation, whilst one could hear the WSP working hard, most of the stops were very smooth with little longitudinal snatching. It is also a testament to the quality of the braking system/WSP that, after more than 100 full service brake applications, all under challenging adhesion conditions, there were no flats to see or hear.


Rail Engineer was shown some of the results to date. As the tests are not complete, this article cannot yet report on the detailed findings, but there is some emerging data that quantifies the benefits, in terms of reduced stopping distance in challenging adhesion conditions, of providing sanders on an additional axle and using variable rate discharge.

The validated results, due in early 2018, will be very interesting. Paul Gray said that the draft report would be reviewed by industry experts (including ARG) before being presented to the wider industry in a series of workshops.

Look out for the results of project T1107.

Testing at RIDC Melton

Finally, a word about the test site at Old Dalby. At the time of writing, there were at least three customers using RIDC Melton site, London Underground/Thales for signalling integration work on S Stock, Bombardier for mileage accumulation on Crossrail Class 345 Aventra trains and this project.

LU and RSSB share the shorter track, with each working on different days of the week, whilst everyone uses the stabling/maintenance facility at Asfordby.

Steve Mills said that all parties are working very well together, especially for sharing access to the single pit road at Asfordby depot.

This article was written by Malcolm Dobell.

Thanks to RSSB’s Claire Grewer and Paul Gray for organising the visit. A special thanks to the test team: Steve Mills and Dan Hamm (RSSB), Dan Ling and Paul Whitworth (Serco), Stuart Brown (DB ESG Rail), Tony Orr (Freightliner) and, on paper tape laying duty, Paul Richards and Matthew Marinaccio (Ricardo Rail) for making Rail Engineer so welcome.