A fire model is a mathematical formulation intended to predict the nature of a fire, or the effects of a fire on it’s surroundings. Fire modeling can range from a single handwritten equation, to a complex finite element computer program requiring many days to calculate on high speed computers.
Simplified fire growth calculations[1] are used to ...
A main subset of fire models is the compartment fire model, which is used to understand the effects of a fire on one or more rooms or compartments inside a building. Typical effects studied are compartment temperature, smoke production rates, gas species concentrations, visibility, fire duration, and ventilation effects. There are two main types of compartment fire models: zone and field.
Zone fire models are based on the simplifying assumption that the compartment during fire development can be split horizontally into two zones: an upper zone in which the products of combustion have stratified against the ceiling due to the buoyancy of the hot gases and smoke; and the lower zone in which the fire exists.
Field fire models use computational fluid dynamics (CFD) techniques, which are based on finite element principles.
Field fire models use computational fluid dynamics (CFD) techniques, which are based on finite element principles.
Fire process characteristics, such as overall fire size growth rate and surface flame speed, are dependant on the size, or scale, of the fire. As many equations used in fire modeling are based on empirical data, the accuracy of the predictions are related to the scale of experiments. It is common for the users of the fire model equations to consider the scale of the experiments from which the equations were derived in making decisions regarding the related limitations of the predictions.
Validation of fire models is intended to
C.4 Uncertainties in Fire Modeling[2].
Uncertainty results from the specification of the problem being addressed (fire size, location, exposures, etc.). Limitations associated with the fire models used for problem analysis can produce additional uncertainties. Specifically, limitations in the number of physical processes considered and the depth of consideration can produce uncertainties concerning the accuracy of fire modeling results. Other uncertainties can be introduced due to limitations related to the input data required to conduct a fire simulation. Other sources of uncertainty include specification of human tenability limits, damage thresholds, and critical end point identifiers (e.g., flashover).
A sensitivity analysis can be conducted to evaluate the impact of uncertainties associated with various aspects of a fire model. A sensitivity analysis should identify the dominant variables in the model, define acceptable ranges of input variables, and demonstrate the sensitivity of the output. This analysis can point out areas where extra caution is needed in selecting inputs and drawing conclusions. A complete sensitivity analysis for a complex fire model is a sizable task. Again, engineering judgment is required to select an appropriate set of case studies to use for the sensitivity analysis. The American Society for Testing and Materials also has a guide for evaluating the predictive capabilities of fire models. The recommendations in this guide should be reviewed and applied as appropriate when utilizing fire modeling.
Wildland fire models[3] are