The discovery of how to produce fire was a significant step in the development of mankind and is still a process which we all experience personally from a very early age. A fire may be defined as "uncontrolled combustion" in order to differentiate it from well controlled combustion as obtained, for example, inside a furnace or car engine.
Fires involve a large number of physical and chemical processes:
Chemistry: The reaction of the fuel with the oxidant (normally air). It may be possible to express the overall reaction as a simple chemical reaction, e.g.,
but, in fact, the burning process is usually very complex and involves a whole series of intermediate reactions.
Mass Transfer: If the fuel is solid or liquid it will be evaporated or volatilized to release a combustible gas which, after burning, is transported away as combustion products or soot.
Heat Transfer: The combustion of the fuel produces heat. Most of the heat initially goes into heating up the combustion products and, particularly, the inert gases such as nitrogen present in the air (this is why fuel burnt in pure oxygen produces much higher temperatures, e.g., in an oxy-acetylene torch). The heat is then transferred to the surrounding environment by convection and, principally, radiation.
Buoyancy driven flows: Fires usually depend upon buoyancy driven flows to remove the hot combustion products from the source of the fire, a process which at the same time draws in fresh air. The movement of smoke and combustion products generated by a fire is usually dominated by buoyancy as the driving force.
Fires can be categorized through the type and physical state of fuel which is being burnt, the location of the fire, and its duration. A few examples are:
Involving solid cellulosic fuels (e.g., wood). Most fires inside rooms fall into this category. Heat output is typically of the order of 1MW and flame temperatures about 600-800°C.
Involving liquid fuels, such as oil, in an open pool. Heat output is directly related to size and may be over 100MW. Typical flame temperatures are about 1100°C for most hydrocarbon fuels.
Produced when gaseous or liquid fuel, released as a jet from a pressurized pipe or container, is ignited. The shape of the flame is normally dominated by the initial momentum of the fuel. Heat output is directly related to the release rate of the fuel. Heat fluxes inside the flames can be significantly higher than inside pool fires due to the greater gas velocities and hence convection coefficient. Flame temperatures can also be higher than in pool fires by 100 to 200 °C.
Fire modelling covers a wide range of disciplines with differing objectives. Workers involved with fires inside buildings may wish to model the rate of growth of a fire, its heat output and rate of generation of toxic combustion products. Alternatively, they may not be interested in the fire source so much as the movement of the smoke and its effect upon the evacuation of personnel or the speed with which the fire is detected.
Workers involved with areas where flammable liquids and gases are processed, e.g., offshore platforms, are interested in the shape of the fire, the heat flux to objects engulfed by it and the heat flux from radiation to objects or people nearby. The efficiency of fire walls and extinguishing systems such as water sprays is also a focus of interest.
In the nuclear field, designers need to demonstrate that transport containers could survive being engulfed in a pool fire. In this case, the focus of attention is not so much on the fire itself as the performance of the container in a fire environment.
Because the objectives of fire modeling depend upon the application, a wide range of methods have been developed, each with advantages and disadvantages.
At the most basic level, basic physical principles and laws can sometimes be applied to obtain a reasonable measure of quantities such as heat output, heat fluxes, rate of heating and mass flow rates. For example, the heat flux to an object engulfed by a pool fire can be determined, assuming that the convective heat flux is small using the standard equation for radiation heat transfer:
Using reasonable estimates for the quantities involved:
Surface emissivity, ε 0.9
Effective emissivity of flame 1.0
Temperature of fire 1373K
Temperature of surface 373K (say)
gives a heat flux of 180kW/m2.
Where the situation is too complex to apply basic physical principles alone, engineering correlations have been developed based upon measured experimental data. Correlations have been developed covering a wide range of fire phenomena from flame size and shape to burning rates, heat fluxes and gas velocities.
Usually the correlations assume a given fuel or type of fuel, geometry and location. Where the situation of interest is within the range for which a correlation is valid, these correlations, because they are based upon experimental results, usually provide an accurate modeling tool.
In situations such as room fires, there are many variables, such as ceiling height, room area and size of openings which each have an influence upon the fire. However, the basic mechanisms and flow patterns are well understood (e.g., cold air entering through the lower part of the doorway, reaching the fuel which is burning, a plume of smoke and combustion products rising above the fire and forming a layer beneath the roof). In this situation, a mixture of basic physical equations and correlations can be used to model the heat and mass transfer between the effectively homogeneous "zones" of the room (e.g., cold air layer, fire plume and hot layer). Once a closed system of equations has been generated, these zone models can be used to model both steady state conditions and the transient development of a fire. Many of these zone models are well established and validated, particularly for applications such as room fires.
Occasionally it is necessary to model fires in conditions for which there are no available correlations or established zone models. This may occur, for example, when novel designs of building are developed (e.g., buildings with large atria). Under these conditions, the most reliable modeling is obtained by using a Computational Fluid Dynamics (CFD) computer code to represent the fire and flow of air and smoke. In a sense, this is just an extension of the zone models to many thousands of separate "zones", with the heat and mass transfer between neighboring zones given by basic physical equations. Although CFD is the most general of the modeling methods it still currently requires significant time to set up the model and analyze the output and also needs very considerable computing power to solve the many thousands of separate equations.
Drysdale, D. D. (1985) Introduction to Fire Dynamics, John Wiley and Sons, New York. DOI: 10.1016/0010-2180(86)90037-4
The SFPE Handbook of Fire Protection Engineering, NFPA, (1988). Fire Safety Journal, Elsevier Applied Science
Blackshear, P. L. (1974) Heat Transfer In Fires: Thermophysics, Social Aspects Economic Impact, John Wiley and Sons, New York. DOI: 10.1016/0010-2180(75)90117-0