Sleeper runaway
A closed brake pipe isolation cock caused runaway of a Caledonian Sleeper ‘Lowlander’ through Edinburgh Waverley station.
A closed brake pipe isolation cock (BPIC) on a CAF Caledonian Sleeper Mk 5 coach was the main cause of a runaway on August 1 2019 that led to the Edinburgh portion of the ‘Lowlander’ Sleeper service running away through Edinburgh Waverley station ( RAIL 885).
No damage or injuries occurred, but there was the potential for a collision with other trains at Haymarket East Junction or Edinburgh Waverley station.
An investigation by the Rail Accident Investigation Branch found that the BPIC was in the closed position when the train left Carstairs station, meaning that braking was entirely dependent on GB Railfreight 92020 which was hauling the train.
The train departed from Carstairs at 0659, 19 minutes late. Five minutes later, the driver carried out a running brake test which brought the train’s speed down from 70mph to 59mph. Subsequent braking applications successfully controlled the train’s speed.
However, at a neutral section just after Curriehill station, the pantograph unintendedly dropped. This is believed to have prevented the Class 92’s rheostatic brakes, which had provided most of the braking power until then, from working. Friction braking brought the speed down to 50mph at Slateford, but just before Slateford station the train stopped decelerating on the 1-in-100 gradient with service brakes fully applied.
At 0725, the driver operated the emergency brakes to no immediate effect. As the train passed through Haymarket station, he tried to contact the train manager, but was unable to do so.
A GSM-R call was made to the signaller to request a clear route through Waverley station. As the train passed through Platform 11 at around 30mph, the train manager realised there was a problem and applied the emergency brake, bringing the train to a stand in Calton South Tunnel with the rear of the train between Platform 7 and the tunnel.
Detailed analysis of the sequence of events showed that a brake continuity test was carried out at Carstairs before the BPIC was closed, but that efforts to connect an ETS connection by the driver and Carstairs shunter could have dislodged it from its open position required for the locomotive to operate the train brakes.
RAIB found that the BPIC valve on the Mk 5 Sleeper coaches does not comply with the most appropriate standard for such equipment as the Dellner coupler is not defined as a manual coupler - even though on the vehicles in the incident they require activity at track level to make mechanical and electrical connections and operate isolation valves.
None of the parties involved in specifying, designing or approving the coaches identified BS:EN14601:2005+A1:2010 ‘Railway applications - Straight and angled end cocks for brake pipe and main reservoir pipe’ as being appropriate for the BPIC.
Two recommendations were made. The first is that RSSB should amend the wording of section
4.2 of Rule Book module TW1 (GERT8000-TW1 – ‘Preparation and movement of trains’) to make it clear that the brake continuity test should be carried out after all coupling-related activities have been completed.
The second is that Serco
Caledonian Sleepers Ltd, in conjunction with its design authority, should review the design of the brake pipe isolating cock on its Mk 5 Sleeper coaching stock - particularly in relation to its vulnerability to undetected, inadvertent operation by people during shunting or train preparation, or by objects such as fallen trees, flying ballast, or other debris. If applicable, it should implement mitigation measures to reduce the risk.
Caledonian Sleeper has since changed its procedures at Carstairs to use the Glasgow-bound locomotive to split the train before the Edinburgh-bound locomotive is coupled to its portion. This means the locomotive can be fully coupled, including ETS connectors, in one activity and that no coupling activity takes place after a brake continuity test has been completed.
CS has changed the orientation of the BPIC handle so that it points upwards when the valve is closed. If left unlatched, it should vibrate to the safe (open) position. The BPIC vent could also be made larger so venting is louder. A review is also considering whether the vent’s capacity could be sufficient to prevent the locomotive to release the brakes if a valve has been closed in the train.
Six learning points have also been identified, covering coupling, risk assessing the running brake test, using the ‘train in distress’ signal, application of standards to rolling stock, and access to recorded train data.