Pioneers of Survival
All of the safety features you take for granted, from firefighting gear to seat
belts, from airplane chutes to life rafts, have gone through extensive design
and testing, often with real people serving as guinea pigs. Meet some of the
leaders in the field of safety research—people who, in some cases, have not
hesitated to put their own lives on the line to improve your safety.
Professor Edwin Galea, Director of the Fire Safety Engineering Group at
Greenwich University, has worked in fire safety research for almost 15 years.
His work in fire safety engineering began after the tragic Manchester B737 fire
in 1985. Since then, it has expanded to include the modeling of evacuation,
fire and smoke spread, and fire extinguishment. Professor Galea's work in fire
safety research covers the field of aviation, buildings, rail, and marine
Professor Edwin Galea
NOVA: Why do you think it is that people so often underestimate fire?
Galea: I think most people aren't faced with fire in everyday life. We no
longer have open fires in our houses, we have central heating. We no longer
have fire for cooking, we have electric cooking ranges, and so on. So most
people aren't faced with fire. The only time they would probably come across a
fire is possibly in a barbecue, or possibly when they have a bonfire. So when a
fire incident happens, they forget about how quickly fire can spread, and they
forget about how dangerous the actual fire products—the gases and the smoke—can be. So that certainly doesn't help in people's response, or their sense
of respect for how dangerous a fire can be.
NOVA: What have you learned about human behavior in a situation where there are
a lot of people in a building with a fire, and all those people need to get
When a fire broke out during a football match in England, there was little reaction from the nearby crowd who watched as it grew.
Galea: The way that people behave in structures is often misunderstood and
misreported. If there is an evacuation in a building, what generally hits the
headlines is that people panicked, there was a mass stampede, and so on. When
you actually go and analyze what happened in these evacuations you often find
that in fact people have not panicked. The response that people have to a known
fire is to run away from it. And that's not a panic reaction. That's a natural
reaction. If you have a fire that's about to burn you, you run away from it as
fast as you can. It's a rational response to threat.
What we find, though, is that in most instances people are not actually aware
that there is a life-threatening situation developing. They might hear an alarm
go off and think, "Well, that's usually a false alarm, it could be somebody's
car alarm, it could be a burglar alarm," and they tend to ignore it. So they
don't respond immediately to the call to evacuate. They mill around, they
continue doing what they are doing until they see some sign of a
life-threatening situation or until they start seeing other people actually
fleeing. Then they turn around and escape.
NOVA: Can you give an example?
Soon after it began, the fire in the Bradford Football stadium transformed into an inferno that claimed the lives of 53 people.
Galea: Well, in the Bradford Football Fire in England, quite a large fire
developed in the stands. People were still sitting around watching the game,
watching the fire, and not really trying to get away from it as quickly as
possible. And within seconds that small fire developed into a flashover event
which took the entire stand. So that's a very good example of where people
have just forgotten about the impact of fire, and how quickly fire can
NOVA: What is your advice for how crowds should be handled in a fire
Galea: I think people need to be made aware of fire events. I think we should
convey information to people on the scene, rather than just an alarm sounding.
We should convey, for example, a verbal warning—"there is a fire present,
please evacuate"—rather than just a bell going off, or a siren going off.
People will then tend to take that more seriously, and they tend to react more
quickly to those sorts of responses than just simply hearing a bell.
NOVA: Can you describe why the Kings Cross Fire in the London Underground in
November 1987, was such a dreadful fire?
In 1987, a small fire under an wooden escalator quickly spread to the rest of the Kings Cross Underground Station in London.
Galea: It started off as quite a small fire under an escalator and rapidly
spread onto the escalator. Once it got hold of the wooden escalator, the shape
of the escalator and the slope of the escalator caused the rising plume to be
entrained down onto the actual rising escalator. Because of that the fire had a
channel to spread the gases and the heat and the top products right up into the
ticket hall, very, very rapidly.
On top of all of this you had trains approaching and departing the actual
Underground station itself. Now when a train approaches a station, it's coming
in in a tight tunnel, and it's pushing in all the air in front of it. As it
enters into the actual platform area all that air is pushed up the escalators,
up the staircases. That fanning effect accelerates the combustion process and
accelerates the spread of flames on the escalator. That effect spread the fire
very, very rapidly. One of the key things that came out of the Kings Cross fire
was how quickly the fire developed—once it actually gripped the whole of the
escalator—how quickly the whole fire spread and how quickly it grew to a
very large size and actually spread in the ticket hall.
NOVA: Later on this fire was modeled. What was the significance of this
Galea: The computer model of the Kings Cross fire was quite pivotal in the
development of fire modeling in that we had a fire at Kings Cross that no one
could understand. The fire experts couldn't explain why the fire was so severe
and why it progressed so rapidly. When the Harwell Laboratory modeled the Kings
Cross fire the model suggested a mechanism that explained the entire fire
development and why it developed so rapidly. Initially the results were not
believed. Then an experiment, where a model of the escalator was built and a
fire was lit, confirmed what the computer model had predicted. So here we had
for the first time a fire model explaining the cause of a fire that had baffled
the experts. Usually it happens the other way around. You do the experiments,
you have an explanation, and then you do some modeling.
NOVA: Could you explain how computer modeling buildings can help save lives?
Galea: By performing a computer simulation of a potential fire and how smoke
spreads in a structure, we can get a feel for how the structure will interact
with the generated smoke products and the actual fire itself. We can get a feel
for how rapidly the structure is going to fill with smoke. We can get a feel
for where the smoke is going to go and how the fire is going to spread. So we
can design a structure so that—rather than causing a fire to spread more
rapidly—it could possibly contain the fire. We can use a computer model
while the structure is still on the drawing board to hopefully design a safer
NOVA: Are these models hard to make?
Galea: Yes. When you run a fire field model similar to what was used to
simulate the Kings Cross fire, you are taking the volume of the room and
dividing it into thousands of little volumes. And in each one of those little
volumes you are solving the fundamental equations of fire development. There
are only seven or eight of those equations. But they are very complicated, very
difficult equations to solve. And you are solving them in each one of these
thousands of volumes. So, in fact, you end up with millions of equations to
solve. Those millions of equations need to be solved at each point in time; you
mark your calculation through maybe every second of the actual fire event. So
you not only have millions of equations to solve at one point in time, you have
thousands of those millions of equations to solve. So you end up with millions
of equations that you have to solve to end up with your fire simulation. This
takes a lot of time even on a very powerful computer. Even on a supercomputer
it could take weeks to generate a simulation of a fire in a structure, and
because of that, the technique is quite expensive.
NOVA: What are the limitations on fire modeling aside from the time and
Computer models, such as these, are used with increasing frequency to study fires.
Galea: Fire field modeling is still a developing and evolving technology, and
there are limitations. Predicting the fundamental fire spread on a solid
material surface is quite difficult and is an area of research at the moment.
Predicting the onset of flashover and also predicting back draft are other
areas that are currently under development. We have developed our own computer
models and are beginning to be able to simulate flashover and back draft in a
fundamental way, but the models are still in the early days of development, and
they are restricted to very simple types of solid fuel, such as cellulosic
fuels and wood products.
NOVA: You did some interesting analysis on a fire in 1993 in Basingstoke,
England. Tell us what you learned from that.
Galea: The Basingstoke building was a modern structure. It had only recently
been built, and it had been designed to be more or less fireproof on the
inside. They had designed it using passive fire prevention means, so they had
good insulation on all the walls, they had pressurized stair cases, the whole
structure was fire blocked, which meant that if a fire started on a floor it
would be contained on that floor and it would not spread to the rest of the
structure. However, on the evening that they had the fire, the fire started on
one of the upper floors, rapidly developed, and engulfed the entire floor.
Eventually the exterior glass panels broke, and flames emerged from the fire
compartment. What the designers had forgotten to take into account in designing
the structure was that it was possible for fire to spread to higher floors via
the exterior portion of the building. So when the fire broke the glass, flames
emerged from the side of the building and spread fire to the higher floors.
NOVA: What does this suggest in terms of the design of tower blocks and high
building? What are you suggesting? Balconies?
Professor Galea demonstrates that computers not only can predict the behavior of fires—they can model human reactions to a fire.
Galea: Yes. You could have a balcony structure—a fixed permanent structure
that is perpendicular to the face of the building—or you could have a
deployable device so that when a fire breaks out, you have a deployable ledge
that comes out perpendicular to the side of the building and deflects the
upward moving flame. Computer simulations suggest that this would provide a
great deal of protection to the higher floors. As far as I know, no buildings
have purposely put in this device as a means to prevent fire spreading to
NOVA: In addition to developing fire simulation models, you've also developed
evacuation models. Can you explain how you predict how people behave in your
Galea: In the Exodus evacuation model, we predict the behavior of individual
people. The model looks at people-people interaction; how people interact with
each other. It looks at people-structure interaction; how people react with the
actual building itself, and people-environment; how people respond to the fire,
the smoke, the heat, and the toxic gases. We describe people as unique
individuals. They have their own sets of attributes, how quickly they can walk,
how quickly they're likely to respond to a fire, their knowledge of the
building, which exit they're likely to use, their drive, and how keen they are
to get out of the structure. So we have a collection of physical attributes
that describes individual people. We even have respiration rates—different
breathing rates—for people as part of the description of each individual
person. All these parameters then interact to describe how that person is
going to respond to the particular incident that they're faced with.
NOVA: And how do you use that information?
Galea: Well, we're using the Exodus computer model while the building is still
in the planning phase, to try and design the structure so that it is more
efficient in terms of evacuating people. For example, where do you have the
exits? What type and how wide should the exits be? What sort of staircases
should you have? How many staircases? What sort of procedures should you have?
Should you have trained staff, fire marshals in the structure? Where should
they be? What should they be doing? So we're using the computer model to try
and help design structures and also to look at things such as procedures—how
you can train people to manage a fire situation more efficiently.
NOVA: Do you think it's possible to design a fireproof building?
Galea: I don't think it's possible to design a structure that is one hundred
per cent fire proof. Any structure has the capacity to burn. And because it's
going to burn, it's going to produce smoke, heat, toxic gases and so on.
Because of that, you need to make sure that you can get people out of that
structure as quickly and efficiently as possible, and that's why you need
evacuation models to help design structures so that the evacuation of people is
going to be as efficient as possible.
NOVA: What advice do you have for young people who are interested in getting involved in this field?
Galea: Fire safety science and engineering is a fascinating subject, because it brings together many different disciplines in the pursuit of a safer environment. A good fire safety scientist/engineer, for example, needs to have an understanding of math, chemistry, physics, physiology, psychology, computer science, engineering, economics, architecture, and many more subjects. This is one of the reasons it is such an interesting subject: You rarely get bored doing the same thing over and over again, and there is always something new to learn.
Escape Through Time |
Survivor Stories |
Pioneers of Survival |
Teacher's Guide |
Site Map |
Editor's Picks |
Previous Sites |
Join Us/E-mail |
About NOVA |
Site Map |
PBS Online |
NOVA Online |
© | Updated November 2000
Support provided by
For new content
visit the redesigned