Grit for Grip: How Sanders Master Railway Adhesion

At 09:03 on 17 October 2016, an East Midlands Trains Class 158 diesel multiple unit operating service 1C08 from Nottingham to London St Pancras entered the approach to Elford station in Staffordshire on a section of track that had been treated with leaf-fall contamination in the preceding days. The driver applied the service brake for the normal platform stop. The WSP system activated immediately — the wheels began to slide on the contaminated rail. The sanding system activated to restore adhesion. The train continued to decelerate, but more slowly than the stopping distance calculation required. The platform end passed the front of the train while it was still moving at approximately 8 km/h. The train stopped approximately 24 metres beyond the platform stopping mark. No passengers were injured, and the overrun did not place any passengers at risk of falling from the platform edge. The RAIB examined the incident as part of a wider study of low adhesion events across the East Midlands Routes and found that 23 of the 32 sanding nozzles on the six-car formation were partially or fully blocked — a maintenance issue that had accumulated over several service weeks without detection. With only 9 of 32 nozzles operational, the sand coverage of the rail head was too sparse to restore sufficient adhesion for the planned stopping distance. The sand was available in the hoppers; the pipes delivering it to the rail were blocked. The incident did not make the national news. It did not injure anyone. It was entirely representative of the kind of low-drama, high-frequency operational failure that the combination of leaf contamination and sanding system reliability issues produces on the UK network every autumn — an estimated 4,500–6,000 such stop overrun events per year in the October–December leaf-fall period, almost all unrecorded because they are within the station platform length. Understanding why the rail becomes so slippery, why sand helps, and why a blocked nozzle can undo the entire system is the subject of this article.

What Is a Railway Sander?

A railway sander is a pneumatic delivery system that meters dry sand from onboard hoppers through calibrated pipes and nozzles to the wheel-rail contact zone ahead of selected wheelsets, with the objective of increasing the friction coefficient at the wheel-rail interface during traction or braking when reduced adhesion conditions are detected or anticipated. Sanders are fitted to electric and diesel multiple units, locomotives, and powered vehicles of all types. They are designated as a safety-critical system under EN 15595 (Railway applications — Braking — Wheel Slide Protection) because their failure to deliver sand when commanded can extend stopping distances to the point of overrun at signals.

The governing European standard for sander system performance is EN 15595 (which specifies the interaction between WSP and sanding) and the broader sand specification requirements in EN 14067-5 (Railway applications — Aerodynamics — Part 5: Requirements and test procedures for aerodynamics in tunnels). Network Rail’s rolling stock interface standard RIS-3703-RST requires sand system performance demonstration as part of the rolling stock acceptance process for any new or modified fleet operating on its managed infrastructure.

The Low Adhesion Paste (LAP): Why Autumn Leaves Are Worse Than Ice

The popular description of leaf contamination as railway “black ice” is accurate in effect but misleading in mechanism. Ice on rail reduces adhesion because water between wheel and rail behaves as a hydrodynamic lubricant — the wheel hydroplanes on a thin water film at high speed or skids on a frozen surface at low speed. Leaf film (Low Adhesion Paste, LAP) is mechanically different and, from a friction perspective, worse.

Formation of LAP

When deciduous leaves fall onto the rail head and are subsequently compressed by passing wheel loads, the cell walls rupture and release a mixture of biological compounds: pectin (a structural polysaccharide), tannin (a phenolic compound responsible for autumn leaf colour), chlorophyll and its degradation products, and cell water. Under repeated compression and heating from wheel-rail contact, these compounds undergo polymerisation and cross-linking, forming a thin (0.01–0.05 mm), hard, black film bonded to the rail surface with remarkable tenacity. The molecular structure of this polymerised biological film produces a surface with extraordinarily low friction characteristics — measured μ values on LAP-contaminated rail of 0.02–0.05 have been documented in Network Rail’s Low Adhesion Field Trials (2011–2015), comparable to PTFE (polytetrafluoroethylene) on steel.

Two properties of LAP distinguish it from simple wet contamination. First, LAP does not wash off in rain — its cross-linked structure bonds to the rail surface and persists until mechanically abraded by wheel passage or treated chemically. The Monday morning after a dry weekend can have worse LAP adhesion than the Friday evening before, because the weekend’s leaf fall has compressed and cured on the rail without the continuous wheel-pass mechanical disruption that keeps the film partly loose. Second, LAP is electrically conductive — it provides a reasonable shunt path between the two rails, meaning that track circuits generally still detect trains on LAP-contaminated sections. Sand, by contrast, is electrically insulating.

How Railway Sanders Work: System Architecture

Components and Flow Path

A typical pneumatic sander system consists of: sand hoppers (typically 30–80 litres per hopper, mounted on the vehicle underframe); a compressed air supply (typically 6–8 bar from the main air reservoir via a pressure-reducing valve to 2–3 bar for the sand delivery circuit); a sand valve or rotary feeder (the metering device that controls sand flow rate); a sand pipe (flexible hose routed from hopper to the contact zone); and a nozzle (the outlet directing the sand stream onto the rail head ahead of the wheel). Modern systems include a sand level sensor in each hopper (typically a capacitive or vibrating fork sensor), a flow sensor at the nozzle to confirm sand is actually being delivered, and an electronic control unit that coordinates sanding with WSP demand signals and train management system commands.

Sand Delivery Rate and Coverage

The effective adhesion improvement from sanding depends on the mass of sand reaching the contact patch per unit area — the “sand coverage” expressed in grams per square metre. EN 15595 specifies minimum sand delivery rates and requires demonstration that the delivered sand restores μ to at least 0.10 on a defined contaminated rail test surface:

Nozzle Position and Aim

The nozzle must deposit sand on the rail head ahead of the wheel (for braking) or on the rail head under the wheel (for traction). The precise positioning is a detailed design requirement: too far ahead and the sand is scattered by the train’s bow wave or washed aside before the wheel reaches it; too close and sand cannot build up in the contact patch before the wheel arrives. Standard nozzle position is 50–150 mm ahead of the leading edge of the contact patch, angled slightly inward toward the rail head centreline to account for crosswind and vehicle pitch effects. On bogies with two wheelsets, sand is typically delivered to the leading wheelset only during braking (the wheel doing the most work) and to both axles during traction (distributed tractive effort).

Sand and Track Circuits: The Shunting Resistance Problem

The most operationally significant constraint on sanding is the track circuit shunting problem. A track circuit relies on the electrical resistance between the two running rails being reduced below a detection threshold when a train occupies the section — the wheel-axle provides the low-resistance shunt path. Sand deposited on the rail head between the wheel-axle shunt points and the track circuit connections can increase the apparent shunt resistance, potentially causing the track circuit relay to see an inadequate shunt and report the section as “clear” even though the train is present.

The magnitude of the effect depends on: sand grain size (smaller grains pack more densely, forming a more continuous insulating layer); sand moisture content (dry sand is highly resistive; damp sand is less so); axle spacing (shorter axle spacing means sand accumulates more rapidly between axles); and track circuit type (DC track circuits are most susceptible; audio-frequency track circuits at 83 Hz or higher are less susceptible because the sand layer’s AC impedance is lower than its DC resistance).

The automatic cutoff at low speed — typically below 10–15 km/h — is a universal feature of modern sand system designs. At very low speeds, the train is close to stopping and the contact between wheels and rails provides an adequate shunt path. Continuing to sand at low speed risks covering the contact patch itself with sand and degrading the shunt. Additionally, the track circuit shunting requirement for a stationary train is stricter than for a moving train (stationary trains in complex signalling areas may need better shunts for approach locking), so stopping the sander at low speed also reduces the risk of creating a problematic “shunt degradation” while the train is completing its stop.

Automatic vs. Driver-Initiated Sanding: Control Logic

Modern sanding systems operate in both automatic and manual modes, with the interaction between WSP (Wheel Slide Protection), ATO (automatic train operation on GoA2+ systems), and driver demand defining which mode applies at any moment.

WSP-Triggered Automatic Sanding

The WSP system monitors the rotational deceleration of each wheelset and triggers sand application when it detects a wheelset sliding. The trigger logic is: if a wheelset’s angular deceleration exceeds a threshold consistent with sliding (rather than rolling) — typically set at 3–5 rad/s² above the expected value for the measured train deceleration — the WSP reduces brake cylinder pressure on that wheelset (to allow the wheel to reaccelerate toward rolling contact) and simultaneously triggers the sander to deposit sand in the contact zone. This dual response — release brake, apply sand — is co-ordinated to restore rolling adhesion as quickly as possible. The WSP sand trigger is a safety-critical function rated at SIL 2 per EN 62061; its response time from slide detection to sand delivery must be within 500 ms.

Anticipatory (Pre-emptive) Sanding

On routes with known low adhesion locations — specific curve sections, lineside vegetation areas, station approach tracks with known leaf accumulation — the train’s GPS position or track balise triggers can initiate pre-emptive sanding before the low adhesion zone is reached, depositing a sand layer in advance of the affected section. This “anticipatory sanding” is more effective than reactive sanding because it places sand in the contact zone before sliding begins, rather than after the adhesion has already failed. Network Rail’s Rail Head Treatment Train (RHTT) programme — which treats high-priority routes with Sandite (a sand-gel mixture) and Water Jetting before the autumn leaf-fall season — follows the same logic at the infrastructure level: pre-treat the rail before adhesion fails rather than managing the consequences after.

Sandite and Rail Head Treatment Trains: Infrastructure-Level Adhesion Management

Sanders on trains are an onboard intervention. The parallel infrastructure strategy is the Rail Head Treatment Train (RHTT) — a purpose-built or converted maintenance train that pre-treats critical rail sections before the leaf-fall season and during its peak period. Two treatment methods are used:

Sandite Application

Sandite is a proprietary adhesion improver developed by Network Rail and its predecessor organisations — a paste-like mixture of sand, water, and a proprietary adhesive compound that binds the sand to the rail head surface rather than allowing it to be blown away by the train’s slipstream. Sandite is applied by the RHTT through a nozzle system at operational speed (typically 30–60 km/h). The adhesive compound allows the sand to remain on the rail head surface for multiple train passes, providing a sustained adhesion benefit rather than the single-pass benefit of locomotive-applied dry sand. Sandite composition: approximately 50% sharp silica sand (0.3–0.5 mm grains), 30% water, 20% adhesive gel (proprietary formulation). Shelf life: 3–6 months in sealed containers. Application rate: 0.5–1.0 litres/m² of rail head. UK RHTT fleet (Class 20/901 and converted Class 150 units) treat approximately 8,500 route-kilometres annually, concentrated on high-risk sections with heavy deciduous lineside vegetation.

Water Jetting

High-pressure water jetting (typically 200–500 bar) mechanically removes accumulated LAP film from the rail head, restoring the clean steel surface without depositing any material that could interfere with track circuits. Water jetting is highly effective at removing established LAP — measured μ after treatment is typically 0.28–0.35 on sections where pre-treatment μ was 0.04–0.08. Its limitation is that it requires the RHTT to pass over the treated section to maintain the benefit; within 24–48 hours of heavy leaf fall in wet conditions, the LAP begins to re-form. The optimal programme combines water jetting at the beginning of the leaf-fall season (to remove existing LAP and start with clean rail) with Sandite application during peak leaf-fall periods and dry sand delivery from onboard sanders for in-service situations.

Adhesion Improvement Methods: Full Technical Comparison

Sanding System Failures and Operational Consequences

Frequently Asked Questions

1. Why does sand restore adhesion on a leaf-contaminated rail — what physically happens at the contact patch when a sand grain is present?

The mechanism by which sand restores adhesion on LAP-contaminated rail involves both mechanical abrasion and surface chemistry. When a sand grain (typically a sharp-edged silica particle, 0.3–0.5 mm diameter) is trapped between the wheel tread and the rail head under the Hertz contact load of 700–1,200 MPa, it is simultaneously squeezed into the LAP film and the soft near-surface zone of the wheel and rail steel. The sharp edges of the grain act as miniature cutting tools, physically breaking through and displacing the LAP molecular film from the contact band. The high contact pressure also fragments the grain, spreading fine silica particles across the contact patch. These silica fragments, embedded in the thin layer between wheel and rail, create micro-asperities — microscopic raised points of hard material — that interlock with the opposite surface and provide shear resistance. The friction coefficient increase from sand is therefore primarily due to this mechanical interlocking (asperity ploughing) rather than chemical bonding. This explains two observed phenomena: sand is more effective at restoring adhesion at low speeds (more grain residence time under the contact patch) than at high speeds; and sand effectiveness improves with finer grain size up to a point (more particles per unit area, more asperity contact points) before decreasing at very fine grain sizes where the particles act more like powder than aggregate.

2. What is “Sandite” and why does it work better than dry sand in some conditions?

Sandite is a paste-form adhesion improver developed by British Railways in the 1970s and refined through successive formulations by Network Rail. Its composition — approximately 50% silica sand, 30% water, and 20% adhesive gel — gives it properties that dry sand lacks in specific conditions. The adhesive gel component (originally a proprietary aloe-vera derived compound, now typically a synthetic polymer gel) binds the sand particles to the rail head surface, preventing them from being blown clear by the slipstream of passing trains before they have had time to work on the LAP film. On a section treated by the RHTT at 06:00 and then traversed by 20 trains before 10:00, dry sand applied by the RHTT would be largely dispersed by the second or third train; Sandite remains on the rail head for 8–24 hours depending on rainfall and traffic density. The gel component also has some direct friction-modifying effect on the LAP: the polymer chains in the gel physically interpenetrate the LAP molecular network during compression under wheel loads, disrupting its cohesive structure and increasing surface roughness. Sandite’s limitation is its lower maximum μ compared to dry coarse sand — the gel matrix moderates the mechanical cutting action of the sand grain — and its shorter shelf life (3–6 months vs effectively unlimited for dry sand). For onboard sander systems, Sandite is not practical because it cannot be gravity-fed through the small-diameter sand pipes and nozzles — it requires the dedicated spray nozzle system of the RHTT. The combination of Sandite pre-treatment (deposited by RHTT) and onboard dry sand (for in-service correction) provides better adhesion management than either alone.

3. Why is morning the worst time for leaf adhesion — does the leaf film change overnight?

This counterintuitive timing is well-documented in Network Rail’s operational data and reflects two distinct mechanisms. The first is physical: overnight, with no train traffic to mechanically disturb the rail surface, the LAP film has several hours to cure and harden. Freshly compressed leaf film is a soft, loose paste with moderate lubrication properties; a film that has been compressed and then left undisturbed overnight polymerises further under ambient temperature and has a more tenacious, harder surface. The Monday morning effect — where adhesion is often worst after a weekend of minimal traffic — is an extreme version of this curing effect. The second mechanism is thermal: overnight condensation deposits a thin film of water on the rail surface. On a clean rail, this water film is readily displaced by wheel-rail contact; on a LAP-contaminated rail, the water film lies on top of the LAP rather than penetrating through it, creating a combined lubricant stack of water on LAP on steel. This combined contamination produces extremely low μ values — sometimes as low as 0.02 — in the first hour of morning traffic before wheel-pass mechanical action begins to break down the film. The first trains of the morning, operating with cold brakes, cold wheels, and cured overnight LAP, face the worst adhesion conditions of the daily cycle. This is why many UK train operators specify “autumn mode” software settings that activate more aggressive sanding and braking parameters for the first one to two service hours of the morning.

4. How does the sand hopper on a train get refilled — and how does a driver know the hopper is empty during a journey?

Sand hopper replenishment is a depot maintenance task, performed either at scheduled maintenance visits (typically every 24–48 hours on trains operating through autumn leaf-fall periods) or triggered by the hopper level sensor warning. The hopper is accessed via a filler cap on the vehicle underframe or body side, and dry sand is poured or pumped from bulk sand bags (typically 25 kg bags on older designs) or from bulk pneumatic delivery nozzles at depot filling points. The filled hopper capacity depends on vehicle design: a typical Class 158 hopper holds approximately 40 litres (approximately 60–70 kg of dry sand), sufficient for 15–30 minutes of continuous sanding at standard delivery rates. During a journey, the driver receives a low sand level warning on the cab display when the hopper level falls below approximately 20% of capacity — typically 8–12 litres remaining. Many modern systems also provide a “sand not flowing” warning triggered by the flow sensor at the nozzle, which detects both hopper-empty conditions and nozzle blockage. The key operational problem identified by RAIB in the Elford investigation (2016) was that many older fleets (including Class 158) had hopper level sensors but no nozzle flow sensors — so the system could indicate “sand present” (hopper not empty) while delivering zero sand (nozzles blocked). The retrospective fitting of nozzle flow confirmation sensors to older fleets has been recommended by RAIB and incorporated into the Network Rail rolling stock maintenance standard NR/L2/RSE/100/00550 as a mandatory retrofit requirement on safety-critical sand systems for all future maintenance contract renewals.

5. Why are there restrictions on sanding at very low speeds and at standstill — and could a train ever lose its track circuit shunt by sanding while stationary?

The answer is yes — a stationary train sanding its wheels can, in theory, lose its track circuit shunt, and this is exactly why automatic sanding is inhibited below approximately 10–15 km/h and while stationary. When a train is stationary on a track circuit section, its electrical shunt is provided by the contact between the wheel flange and tread on the rail head — a metal-to-metal contact that provides very low resistance (typically 0.01–0.05 Ω). If the train sands while stationary, sand accumulates in the contact patch and immediately adjacent areas. Dry silica sand has a resistivity exceeding 10⁸ Ω·m — many orders of magnitude higher than steel. Even a thin layer of sand grains separating the metal surfaces at the contact patch increases the shunt resistance from 0.05 Ω to potentially 0.5–2 Ω or higher, depending on how complete the sand coverage is. Track circuits are designed with a maximum shunt resistance of 0.5 Ω (UIC 756-1) — above which the relay may drop, reporting the section as clear. A train sanding heavily while stationary on a DC track circuit section could therefore cause its own track circuit to mis-indicate — a safety anomaly that would confuse the signaller and potentially set an unsafe signal aspect for the following train. The automatic sanding cutoff at low speed is therefore not simply an operational convenience but a safety interlock preventing this specific failure mode. Some modern systems go further: they monitor the track circuit state in real time (via the train’s on-board ATP system’s track circuit awareness) and inhibit sanding whenever the train-speed is below 5 km/h regardless of WSP demand — ensuring that the very last phase of braking, where the train is moving slowly onto its final stopping position in a station, does not compromise the track circuit shunt that confirms its presence to the signalling system.