Exit strategy

Computer simulation of disasters such as September 11 could not only radically improve survival rates but also help designers create safer structures. Andrew Lee reports.

If fire breaks out in the building where you work or the aircraft on which you are a passenger, pray that somebody thought long and hard about that eventuality while designing it.

A team of UK fire engineering experts has just been commissioned to carry out the biggest study yet into the most deadly building fire in history, the aftermath of the attack on the World Trade Centre in 2001.

The project will be led by Prof Ed Galea of the Fire Safety Engineering Group (FSEG) at the University of Greenwich, which has already raised disturbing early pointers about the design of the WTC and its implications for getting people out alive. The staircases were too close together and insufficiently reinforced. Thousands more could have died if the buildings had been full.

FSEG has equally uncomfortable questions for the engineers designing the planes, trains and ships on which we all travel. Several are particularly pressing. Do railway operators have any real understanding of how to evacuate a carriage in an emergency? Can future blended-wing passenger aircraft pass international safety tests?

Galea and his current 30-strong Greenwich team have spent almost 20 years researching the implications of fire and emergency evacuation for the engineering of buildings and other structures. They have developed a sophisticated computer modelling system called Smartfire to simulate the spread of a blaze, and a second, Exodus, to predict the behaviour of human beings inside a structure (see sidebar).

FSEG has worked with some of the biggest names in industry and with the UK and other governments, advising on fire and emergency procedures on land air and sea. This has given Galea and his colleagues a unique insight into how much, or little, thought goes into preparing for the awful eventuality of a fire or other emergency.

Galea’s harshest assessment is reserved for the UK’s railways. While acknowledging the beleaguered rail industry’s attempts to get its act together in some safety-critical areas such as signalling, he claimed it is in something approaching collective denial over how to cope with catastrophes such as the 1999 Ladbroke Grove crash.

Fire added immeasurably to the horror of Ladbroke Grove, which left passengers desperately trying to escape from derailed carriages.This and other high-profile disasters have prompted rail operators to place notices on trains advising passengers on carriage evacuation. But according to Galea, they are barely worth the paper they are printed on and represent little more than lip service by an industry with its head in the sand.

‘The UK rail industry just doesn’t care about evacuation,’ claimed Galea. ‘They think that on UK trains evacuation is a non-issue.’ Even after Ladbroke Grove, Potters Bar and the other familiar litany of rail disasters, Galea believes the rail safety orthodoxy is obsessed with keeping people on board and away from live rails, trains coming the other way and other hazards outside the carriage. ‘I’ve had numerous meetings with safety people in the rail industry, but all they can say is stay on the train. Tell that to the people at Ladbroke Grove.

‘The rationale is that it’s always safer to stay on than get off, and certainly in may cases that is true,’ said Galea. ‘But in other circumstances you need to get off quickly, which means you need to have procedures in place to evacuate rapidly and safely, from carriage to carriage if necessary.’

The aircraft-style evacuation notices that have appeared in trains cannot, Galea contends, be based on anything more than guesswork. ‘Can they tell me what the flow rate out of a carriage window is? I doubt it. I have seen no data that suggests they can. But if they are advising the public that these are viable escape routes then they need to do the work to back it up. I find it worrying that this sort of research is not being done.’

Galea contrasted the UK approach with the US rail system, which has done significant work on the evacuation of stricken carriages.If the rail industry in the UK scores poorly on Galea’s safety audit, the global aviation sector receives a generally better report – although with significant reservations.

Aviation safety is where FSEG started. In 1985 a Boeing 737 caught fire on the runway at Manchester Airport killing 55 people. Galea was at that stage a specialist in computational fluid dynamic (CFD) techniques, which he had used to model everything from industrial steel-making processes to the behaviour of the sun. He was horrified and fascinated by the disaster. ‘I couldn’t understand why so many people died in a plane that hadn’t even taken off.’

Galea decided that the principles of CFD could be applied to the spread of the fire, smoke and toxins on board the aircraft. ‘The details are different, but the basic equations that you have to solve are the same,’ said Galea.

He was able to model the progress of the fire, which began in one of the plane’s engines and spread to the fuselage and the interior of the aircraft, filling the cabin with black toxic smoke. Galea soon realised he had only half of the story. ‘I now understood how the fire had spread but to find out why 55 people had died we had to understand the human dynamics.’

If flames, heat and smoke are complex factors to model, human beings provide even greater challenges. At least the former obey certain fundamental physical laws. Where people are concerned, the reactions are less predictable.

The importance of behavioural factors in how people move during an emergency may seem obvious, but according to Galea evacuation modellers all but ignored them until relatively recently.

‘Until we developed our model the assumption was that things proceeded in a nice orderly fashion; emergency, alarm, response and then evacuation.’

People were presented in simulations as ‘water flowing through a pipe or as ball bearings, rolling neatly towards the closest exit,’ he said.

The Manchester study demonstrated with horrifying clarity how flawed such assumptions were. ‘People are not ball bearings, and if engineers are using models that suggest they are, and if they don’t do what they are supposed to do, then you have problems,’ said Galea.

‘It became clear that many people at Manchester did not react straight away. There were instances of people frozen in their seats. Others did react straight away, but didn’t know what to do. Older people reacted differently to the young. Some people were jumping over seats rather than waiting in a queue.’

The work done on the Manchester crash formed the basis of the Smartfire and Exodus simulation software. Aircraft safety remains a preoccupation for the Greenwich team. It has worked on simulations for Airbus of the A380, the 550-seat, two deck ‘superjumbo’ due to enter service soon as the world’s largest passenger plane. Like all new aircraft, the A380 will have to pass an emergency passenger evacuation test. The test is far from straightforward, and is a nail-biting and expensive experience (costing around $2m/£1m to stage each time) for the aircraft manufacturer and safety regulator alike.

During the procedure the aircraft is filled with a representative sample of passengers – old, young, male, female, able bodied and infirm. Half the available exits are rendered useless, although the crew taking part are not told which ones in advance. Everyone has to get off the plane within 90 seconds, in darkness except for the plane’s emergency lighting systems.

The evacuation test is a cause of some concern to Galea and the Greenwich team. ‘In principle it’s a good idea, in practice it is flawed. It is not representative of a likely scenario.’

There are doubts, for example, about whether the test can recreate the atmosphere of a real emergency, with all the intricacies of human reactions that this brings. This does not mean that the certification procedure is not a genuinely alarming experience. On average six per cent of the volunteers taking part will suffer injuries ranging from cuts and bruises to broken limbs. The testing programme’s darkest hour came in 1990 when a participant broke her neck and was left quadriplegic, causing much soul-searching in the aviation safety community.

Galea’s main concern is the one-off nature of the test, and its one-size-fits-all approach to aircraft that will vary widely in size and design. Big planes with many people on-board can be safely evacuated. ‘I have no doubt that the A380 is capable of passing the 90-second test,’ said Galea. ‘We’ve simulated it and I don’t think it’s got a problem.’

But as well as the A380, Galea has been working on simulations of possible emergencies on-board future blended-wing body designs for passenger aircraft, and these are proving far more problematic. Blended-wing airliners with capacities of up to 1,000 people would represent a radical departure from the ‘tube with wings’ designs of current planes. They are attractive to the aviation industry on a number of grounds, not least the potentially massive fuel efficiency gains brought about by reductions in drag.

But would they meet the evacuation criteria? In a ‘flying wing’, passengers would sit in conditions more akin to a cinema than a railway carriage, with those in the centre aisles a considerable distance from its side exits with many seats and fellow passengers in between. Galea claimed significant design challenges could lie ahead if blended wings are expected to pass the 90-second test.

While aerospace engineers are making good progress developing the concept, Galea believes evacuation could prove to be ‘one of the big showstoppers’ looming for the revolutionary design. ‘We’re looking at the concepts right now. The real issue is whether you can design the structure and the layout so you can satisfy the 90-second evacuation.’

According to Galea, the most simplistic approach – to swamp the aircraft with emergency exits and to make those exits bigger – is a red herring. ‘More and bigger exits are not the issue,’ said Galea. ‘The issue is whether you can feed those exits quickly enough, and get people to where they need to be. You need a very clever design or lots of aisles. If you have lots of aisles you are using a lot of space.’

Simulations continue in a bid to come up with a design that could pass the 90-second test. But Galea claimed that if the industry wants to see innovative aircraft such as the flying wing enter service, it may be forced to think again about whether the test itself is relevant.

The 90-second criteria has been around for decades and is based on research into the spread of fire inside standard tubular aircraft cabins. The point of 90 seconds, rather than, say, 80 or 100, is that this is the threshold at which early experiments suggested that a phenomenon called flashover generally occurs. This is a catastrophic escalation in the scale of the fire at which the chances of surviving it are reduced to virtually nil.

Galea has general doubts over the applicability of a blanket 90-second rule to today’s huge variety of aircraft.

‘It applies to a 90-seater or a 747, which doesn’t really make engineering sense,’ said Galea. ‘But let’s assume for the sake of argument that it holds true in a tube. Would it be the same in a large cavernous space such as the blended wing?’

The Greenwich team’s simulations suggest not. The types of fire that would flash over in a typical aircraft are not doing so. ‘On the other hand there are other things going on with smoke and toxicity.’

Galea’s point is that the 90-second rule could scupper an aircraft design that might actually be safe, and a design that passes might not be safe in real life. Perhaps it is time to reconsider the procedure. Instead of 90 seconds, he prefers to talk about a measure called the Available Safe Egress time – ‘how much time have you got and how much time do you need?’

This, he said, would be the logical engineering approach, and he claimed that such calculations are possible thanks to the work done at Greenwich and elsewhere.

FSEG has picked up a hatful of accolades for its work, including a Queen’s Anniversary Prize, the EU’s top award for IT innovation and recognition from the Royal Institution for Naval Architects.

However, fire and evacuation, like safety in general, still sometimes face a struggle to place themselves centre stage in the day-to-day realities of engineering and design. Like others in the field, Galea said Greenwich experiences peaks and troughs of interest in its work.

The peaks tend to come after a disaster – a fire, a crash and explosion that suddenly concentrate the collective mind of an industry. Then the interest tails off, as the focus returns to the more everyday commercial considerations of capacity, project deadlines and cost.

‘The safety business is a tough business to be in,’ said Galea. ‘People generally don’t want to spend money. But if you think safety is expensive, try having a disaster.

Sidebar:Simulating the unthinkable before it happens

Smartfire and Exodus, FSEG’s fire and evacuation computer simulation packages, represent an ambitious bid to give engineers a view of the unthinkable before it happens.

With specific versions for buildings, ships, aircraft and so on, they have been used to simulate emergencies in, among others, the Millennium Dome, the Sydney Olympic Stadium and the Royal Navy’s new aircraft carriers.

Smartfire is based on the principles of computational flow dynamics (CFD), but is designed to be used by safety engineers, architects and others with little or no expertise in the CFD field.

It tracks the spread of fire, smoke and toxins within a computer-simulated building, plane or other environment. One of its most ambitious deployments came with the investigation into a fire that caused the 1998 crash of a Swissair flight in Nova Scotia that killed 229 people.

The fire, which started above the cockpit of the Swissair MD-11, was recreated in precise detail, including complex properties of the materials on-board, their behaviour on combustion and how the airflow around the aircraft contributed to the unfolding disaster. The last factor turned out to be crucial. The simulation suggested that most of the smoke and toxic odours from the fire were drawn into the avionics compartment below the cockpit and expelled from the aircraft or backwards through the rest of the plane.

This explained one of the mysteries of the crash. The pilots were apparently under the impression until it was too late that – despite the fire raging in the ceiling cavity above their heads – they were dealing with a far less serious problem with the air-conditioning system.While Smartfire attempts to take painstaking account of the factors governing the spread of fire and smoke, Exodus charts the behaviour of people as they try to leave a simulated enclosed area.

It shows when and where they are likely to fall victim to heat, smoke and other hazards and how quickly a structure can safely be cleared before death or injury occurs.The Greenwich team has refined Exodus in a bid to capture even the smallest nuance of behaviour.

An example is hesitation time: the instinct that will make some people halt, even in an emergency, before using an aircraft’s emergency slide.