In the subsequent investigation, AEA Technology at Harwell in Oxfordshire used the digital modelling technique Computational Fluid Dynamics (CFD) to try to understand how a fire that had seemed so small – the London Fire Brigade and British Transport Police had attended but observed only a fire the size of a large cardboard box with few visible flames, and ordered a slow evacuation – could have escalated so severely. The models indicated that a hitherto unknown mechanism was at work, which the researchers named the 'trench effect'.
The simulation suggested that air currents had confined the fire in the narrow trench of the escalator, between the walls supporting the handrail, and that this had heated the wooden treads of the escalator steps and the greases lubricating the machinery beneath to the point where flammable gases were released, which then ignited to form the flame jet powerful enough to blow people off their feet.
“It was the first time that CFD was proven to make a prediction ahead of experiment,” explains Guillermo Rein, professor of fire science at Imperial College London’s Department of Mechanical Engineering.
“The effect had not been known to humanity before and it suggested that if you have fuel for a fire in a closed trench where the air cannot come in at the sides, it forces the flames to get closer to the fuel and that accelerates greatly the spread of the fire.”
The research had been done by people not specifically working on fire, but on general thermofluid behaviour, Rein says, and was at first met with scepticism. But it was then confirmed by an experiment by an eminent fire scientist, Professor Dougal Drysdale of Edinburgh University.
“Normally you’d start with an experiment and then explore further with digital models, but this was the other way around, and that’s when the use of CFD in fire science really started.” Rein says.
There are two ways of designing fire safety in buildings, according to Rein. Most building design around the world is prescriptive, he says: carried out in accordance with codes and standards. “The book would say you have a corridor of defined dimensions, and that would link to an atrium, also of defined dimensions, and so on. You’d submit this to the planning regulators, and they’d approve it because it complied with the code, even though it’s not really engineering: the designer doesn’t really have to think or calculate or make many decisions.”
Such prescriptive design is now being gradually superseded by performance-based design, which is a freer approach depending much more on CFD simulation. After producing an initial design with architects, building engineers conduct many simulations of different types of fire, in various locations in the building, with windows open or closed, with sprinkler systems operational or failed and so on, and use these simulations to demonstrate to the authorities that the design meets the level of safety equivalent to the regulations. Any questions the regulators might pose can also be addressed using simulation.
Performance-based design is expensive, and tends to need large teams involved from the start of a project, so is mainly confined to large, prestigious commercial buildings. Standard residential building tends to be prescriptive and not innovative. Changes in design to improve fire safety in such projects depend on updating building codes – a lengthy and controversial process, as is becoming clear from the enquiry into the fire at Grenfell Tower in 2017.
It is not simple to use simulation to model fire, because of the large number of factors involved in fire behaviour. “It involves chemical reactions between fuel and oxygen, it involves thermodynamics, and it involves fluid mechanics, mostly in the behaviour of gases but also to a lesser extent liquids, and all three of these interact with each other in complex ways,” Rein says.
There are always unknowns. In performance-based design, project teams must determine their 'design fires': the types of fire they design the building to resist. “You can’t design against a situation like the Great Fire of London, because you’d end up with bunkers or igloos: they’re super-safe but nobody would want to live or work in them. Defining the ‘design fire’ which you design against is a big question and needs discussion between the client, the architect, the engineers and the authorities.”
As it’s neither practical to design against huge, relatively unlikely explosions or tiny fires which would be simple to extinguish, much thinking goes into designing a 'worst credible case' against which a building would be safe.
Recent fire safety concerns include the wildfires influenced by climate change, fires resulting from ignition of lithium-ion batteries, and the risks associated with greater use of materials like wood, which have less environmental impact than conventional concrete but are potentially more flammable. All of these present new challenges to engineers using modelling tools to understand them and design ways of coping with them.
IMechE’s Improving Fire Safety Through Simulation and Modelling event takes place on 11 June 2024 at the International Convention Centre in Birmingham. Find out more and register.
Want the best engineering stories delivered straight to your inbox? The Professional Engineering newsletter gives you vital updates on the most cutting-edge engineering and exciting new job opportunities. To sign up, click here.
Content published by Professional Engineering does not necessarily represent the views of the Institution of Mechanical Engineers.