A cascade of risks

Over the period before Christmas, the operations of cross-Channel rail company Eurostar came to resemble a disaster zone. Five trains failed inside the Channel tunnel more or less simultaneously, causing massive delays in highly distressing circumstances to thousands of passengers. A consequent three day shut-down of the service caused delay and distress to thousands more at one of the busiest times of the year

The board of Eurostar, encouraged by the British and French governments, commissioned an independent report into what had happened. The report appeared on 2 February and is exhaustive in its analysis of the failure. It should be essential reading for anyone in the business of drawing up contingency plans. It shows how easily a developing crisis can be rapidly worsened by risks which may have been overlooked or deemed to be of minor importance, but which come to sudden prominence as the contingency plan runs up against the unexpected.

The root cause of Eurostar’s problems was a relatively simple engineering defect in its power cars. Inadequate screening allowed fine snow (which covered much of northern France) to be drawn into the power cars by the powerful ventilation systems. Inside the tunnel, where the temperature is between 25 and 28 degrees as opposed to the sub-zero temperatures outside, the melting snow caused arcing in the electric motors, leading to burn out and failure. Between 21.00 hours on 18 December and 01.00 hours on 19, five northbound trains failed inside the tunnel, leading to a complete blockage of both lines. Two further trains were held overnight outside the tunnel.

As the entity with immediate responsibility for the smooth running of the Channel tunnel, Eurotunnel has a raft of formal procedures covering crisis management and detailed instructions on evacuation of passengers from inside the tunnel. The immediate burden of decision making rests with the rail control centre. On the ground, FLOR (first line of response) teams organise the evacuation from within the tunnel, following detailed operating procedures to ensure safety requirements are met. Evacuation is not desirable: the preferred option is to attempt to rectify the fault in the train. Failing that, towing the stranded train from the tunnel using one of a pair of dedicated rescue engines, is the second most desirable option.

This emphasis on safety is understandable; the downside being the time required to decide the strategy, revise it when it fails and ensure the laid-down procedures are followed. And, as the report makes clear, the delays had immediate effect on the stranded passengers. The safety risk was mitigated, but only at the expense of other aspects of their welfare.

Conditions on train 9057, which was returning to London from Marne-la-Valee (the stop for Paris Disneyland) were especially bad. At 22.40, it stopped in the tunnel behind a previously failed train, and waited for an hour before the decision was taken to return it towards Calais. This movement did not commence for a further hour. At that point, 9057 itself failed and lost its lighting and air conditioning. At 01.26 the driver reported major problems with lack of air, heat, and increasingly angry passengers. Before official permission was given, passengers started to open the doors themselves to allow air to circulate. Eventually evacuation to a Eurotunnel vehicle shuttle began. This was completed by 03.52, with the shuttle arriving back at the Calais terminal at 04.16. Passengers – with many young families among them – had spent 5 ½ hours in the tunnel. For a further 1 ½ hours, the train remained at Calais – passengers being allowed onto the platforms in small numbers. At 05.44, the train went back through the tunnel to Folkestone, arriving 30 minutes later. At this point the plan was to transfer the passengers from the vehicle shuttle to a Eurostar train from London to take them back to St Pancras. It was a further two hours (08.15) before the transfer began – passengers being kept on board the vehicle shuttle the whole time . Eventually the relief Eurostar train left Folkestone at 10.30, was delayed for a further 30 minutes, and arrived in London at 11.53.

The 644 passengers on this train therefore had to endure 12 hours of what was effectively confinement within one or another train – four of them on a vehicle shuttle not intended for the purpose of conveying large numbers. There were reports from this – and other - trains of the train staff locking themselves away from the angry passengers, of few, or no announcements made to passengers, of calls being made to emergency services, and of sufficient disturbance in one case for the police to have to be called. The Eurostar report mentions more than once that passengers felt they were being ‘held captive’.

It can be argued that it takes considerable provocation to turn a train load of passengers into a state of fury sufficient to cause train crew to lock themselves away. Yet this is what appears to have happened on more than one Eurostar train. What went wrong with the risk management to allow this to happen?

One answer is that Eurostar and Eurotunnel were hit by a cascade of risks which were not fully anticipated in their contingency planning. Among them were:

• Time of year: Many more people than usual were travelling in the pre-Christmas period.

• Weather: Snow caused the initial problem, but it also meant that alternative means of transport to move passengers affected by the three-day termination of services were difficult to come by

• Politics: The fact that Eurostar trains are ‘sealed zones’ for customs and immigration purposes meant passengers could not be allowed off trains unless in Paris, Brussels or London.

• Accommodation: Trains stranded in France were not allowed to return to Paris because there were insufficient rooms available in hotels

• Incompatible equipment: Platforms for Eurotunnel vehicle shuttles are at a different height from Eurostar platforms, making transfer of passengers between the two types difficult and time consuming

• Communication (1): Although the drivers on Eurostar trains can speak directly to the rail control centre, train managers (in charge of passengers) could not do so. All information had to be relayed through the driver, whose priority was not passenger welfare

• Communication (2): The report mentions instances where announcements by French train crews were incomprehensible to English passengers, or were distrusted

• Communications (3): Many passengers mentioned that they saw no Eurostar staff walking through the train from one end of the crisis to the other.

All of these contributed in one way or another to the delays or frustrations suffered by thousands of people. But there is a further consideration, which the Eurostar report emphasises – but perhaps not sufficiently.

That is, a contingency plan which looks at technical solutions first and at customer (or staff) morale and welfare as a poor second, may well be successful at preserving life and property, but may fail badly on other grounds. The hit to Eurostar’s reputation and the distress of its passengers had a great deal to do with the fact that in the welter of safety procedures and decision taking about train movements, the human beings at the end of the chain felt themselves to be overlooked, or treated as impersonal numbers to be shifted hither and thither without explanation.

Human beings react badly to being treated as objects. On the whole, they react well when their co-operation or expertise is asked for and when they are fully informed of what is happening and why. A long history of crises, from the Titanic onwards suggests the truth of this. It is a lesson which Eurostar had seemingly paid insufficient attention to.