The limiting surface is outside the brake

A brake can be mechanically capable of a strong application while the wheel and rail cannot transmit the corresponding force. Wabtec identifies rain, ice, leaves, oil and grease among the factors that can reduce the adhesion available between steel wheels and rails. The engineering problem is therefore larger than brake hardware. A command to slow the train eventually meets the condition of a small moving contact between steel surfaces. That condition can reduce the usable braking force even when the onboard equipment functions.

Sources: adhesion-wabtec

Avoiding a locked axle

Wheel-slide protection follows axle rotation and adjusts the brake force when the wheel slows too rapidly relative to the vehicle. Wabtec describes using axle-speed measurements to estimate vehicle speed and identify sliding, then changing pressure in the affected brake cylinders. Reducing excessive braking gives a slowing wheel a chance to recover rotation. A locked wheel can develop a flat spot as it slides. The objective is controlled use of the available contact force, rather than either maximum brake pressure or unrestricted release.

Sources: adhesion-wabtec

The best slip range can move

Knorr-Bremse’s WSPA-3 example compares braking force available in different slip ranges. It estimates adhesion using cylinder pressure, other braking contributions, wheelset acceleration and vehicle parameters. When its comparison favours a lower-slip range, it changes control mode and keeps checking whether conditions have changed. This is feedback: the system infers what the contact can support from its response. The particular algorithm is one implementation, but it reveals why a fixed slip target need not give the best result on every contaminated rail.

Sources: adhesion-wsp

Sand changes the available contact

Sanding addresses a different part of the problem by changing conditions at the wheel–rail interface. RSSB describes variable-rate equipment that adjusts sand delivery with train speed, and testing of two sanders working in series. In its reported four-car comparison, this configuration could halve stopping distance relative to a single fixed-rate sander. The comparison matters as much as the headline: it describes a tested train and baseline, rather than a promise for any railway. Sand delivery, train formation and the underlying contamination all belong in interpreting such a result.

Sources: adhesion-sand

Protection has physical limits

The 2021 collision at Salisbury Tunnel Junction illustrates that braking systems cannot be assumed to overcome arbitrarily poor adhesion. RAIB found exceptionally low grip from leaf contamination worsened by drizzle, alongside braking that began too late for the prevailing conditions and systems unable to mitigate them. Its recommendations addressed multiple parts of the railway, including railhead treatment, vegetation and other braking technologies. The lesson is not that feedback has no value. It is that controlling the force available at a contact cannot guarantee the contact will supply enough force.

Sources: adhesion-raib

Every brake on the axle matters

Knorr-Bremse specifies that implementing its adaptive algorithm requires control over all braking systems acting on the wheelset. The reason follows from the force balance: reducing one contribution cannot fully regulate the contact if another continues demanding too much. Its adhesion estimate therefore includes external braking contributions as well as cylinder pressure. This requirement makes wheel-slide protection a coordination problem. The controller must know what is slowing the axle, assess the contact’s response, and revise the combined demand as conditions change.

Sources: adhesion-wsp