CFAST

CFAST (Consolidated Fire and Smoke Transport) is a compartment fire model that represents each room with an upper and lower gas layer. It calculates changes in the thermal environment using conservation of mass and energy, with empirical relationships for processes such as plume flow and ventilation. Van Rossum identifies it as a dynamic zone model capable of representing connected spaces. Each layer has representative uniform properties, so its results describe an idealised compartment environment rather than every local variation.

Within fire modelling, CFAST provides a relatively economical way to examine changing layer conditions and intercompartment transport. Its applicability depends on the scenario, the assumptions behind its submodels and relevant validation evidence. This article describes the supplied version 7 technical reference, revision 7.7.7 dated 27 May 2026. That identified document defines the software behaviour discussed here.

Compartment and layer representation

CFAST divides each compartment into a relatively hot upper layer and a relatively cool lower layer. A fire plume carries combustion products and entrained gas towards the upper layer. Openings allow exchange with adjoining compartments or the outside. These processes change the mass and energy stored in the layers over time.

CFAST represents compartment conditions through two gas layers and exchanges between them. Based on Peacock et al. (2026), pp. 1 and 7–8.

The layer temperatures and densities are treated as uniform within their respective control volumes. This reduces a spatially varying fire environment to a small set of evolving quantities. It permits an efficient calculation, but it also removes local variation from the basic layer description. A layer average cannot describe every condition close to a flame, opening or ceiling jet.

The model advances compartment pressure, upper-layer volume and the temperatures of both layers. Its pressure variable is referenced at the compartment floor. Exchange processes supply the terms driving these changes. The formulation couples the state variables through mass conservation, energy conservation and the ideal-gas relationship.

Momentum is not resolved as a spatial field inside each zone. Flow between compartments instead uses opening relationships based on pressure and density differences. This is a central distinction from a detailed fluid-flow representation. The zone state and its connecting flow models together describe the compartment system.

The Fire Dynamics Simulator represents a different level of spatial detail. The CFAST technical reference discusses it among the more complex computational-fluid-dynamics tools used in comparative validation work. Choice between these approaches should follow the quantities and mechanisms needed for the assessment. A more detailed representation is not automatically necessary for every compartment question.

Fire prescription and required inputs

CFAST does not independently predict the growth of fire on an arbitrary burning object. The user specifies the heat release rate for one or more fires. The model includes a simplified treatment that can limit burning when oxygen availability is insufficient. Consequently, a prescribed fire history remains a major input to the calculation.

Geometry inputs include compartment dimensions and the positions and sizes of openings. Construction inputs include thermal properties and thicknesses of the bounding materials. Mechanical ventilation requires its own flow specifications. These inputs influence both the movement of gas and heat transfer to the building surfaces.

The fire description also includes information needed for combustion and species production. Some products are determined from the specified fuel composition, while minor-species yields are supplied by the user. Those assumptions influence calculated gas concentrations and smoke-related outputs. A temperature calculation alone does not establish that the species inputs are appropriate.

Targets, detectors and sprinklers require their positions and relevant properties. They introduce questions about local exposure or activation in addition to average layer conditions. Their outputs depend on the corresponding submodels as well as the compartment calculation. Specifying an object in an input file does not remove the limits of the model used to evaluate it.

The guide expressly places responsibility on users to assess default values. Numerical and physical parameters have been examined in different verification and validation studies, rather than one assessment covering every parameter. A supplied default is therefore part of the model configuration. Its suitability for a particular scenario still requires consideration.

Conservation equations and the state of a layer

Mass conservation states that a layer's mass changes at its net mass-source rate. The source terms account for additions and removals, including plume entrainment and ventilation. The sign of the net term matters: a positive value increases stored mass, while a negative value decreases it. The following equation expresses that accounting for layer i.

dm_i/dt = ṁ_i,net

PV_i = m_iRT_i

Symbol	Description	SI unit
mi	Mass in layer i	kg
t	Time	s
ṁi,net	Net mass-source rate for the layer	kg/s
P	Absolute thermodynamic pressure	Pa
Vi	Layer volume	m³
R	Specific gas constant used by the model	J/(kg·K)
Ti	Absolute layer temperature	K

The energy equation separately accounts for the change in internal energy and the work associated with changing layer volume. Its source terms include fire energy, gas transport and heat exchange with surfaces. Together with the gas law, this determines temperature and volume changes. A mass balance alone cannot predict all of those quantities.

The supplied formulation uses constant thermodynamic parameters for its gas-state calculation. It sets constant-pressure specific heat to 1,012 J/(kg·K) and the specific-heat ratio to 1.4. These give a specific gas constant of approximately 289.14 J/(kg·K). The assumption applies even though the model also tracks changing combustion-product concentrations.

Species transport follows the movement of gas between layers and compartments. Their masses accumulate according to production and flow, while oxygen depletion can influence the combustion calculation. Species mass divided by layer volume gives mass concentration. The model's smoke and radiation treatments consequently depend in part on the specified species yields.

The species calculation also distinguishes concentration from the total amount present. The same species mass can have a different concentration when layer volume changes. This matters when interpreting smoke-related outputs during heating, cooling or ventilation. The calculation therefore transports species mass and relates it to the evolving layer volume.

The model includes generic toxic and trace species as additional tracked quantities. The supplied guide states that these particular species are excluded from the overall mass balance. They are nevertheless produced and transported according to the corresponding source terms. This implementation detail illustrates why a named output must be understood within the documented species treatment.

Worked layer-state illustration

Consider a selected rectangular room measuring 5 m by 4 m by 3 m. Suppose its upper layer is 1 m deep, giving a layer volume of 20 m³. Select an absolute pressure of 101,325 Pa and an upper-layer temperature of 500 K. These illustrative values define a snapshot for the cited gas-state relationship.

Using the model's gas constant gives a density of approximately 0.701 kg/m³. Multiplication by the 20 m³ layer volume gives approximately 14.0 kg of gas. This mass includes the mixture occupying the layer, not merely soot or burned fuel. The example demonstrates how volume, temperature and pressure jointly determine stored gas mass.

For a second mathematical comparison, retain the same mass and pressure but reduce temperature to 400 K. The gas law gives a volume of 16 m³. Across the same 20 m² plan area, that corresponds to a layer depth of 0.80 m. Cooling at fixed mass can therefore change the volume associated with the layer.

At fixed mass and pressure, the illustrative lower temperature corresponds to a smaller layer volume. Based on Peacock et al. (2026), pp. 7–8.

These calculations are state comparisons, not the output of a CFAST simulation. They do not solve the fire, ventilation or heat-transfer processes needed to produce the change. A complete calculation must also conserve energy and account for simultaneous mass exchange. Holding mass and pressure fixed isolates the gas-law effect for explanation.

This distinction is useful when interpreting smoke filling. Layer descent depends on the volume associated with accumulated mass and its temperature. It cannot generally be determined by adding a fixed smoke volume to an otherwise empty room. The evolving layer state and the flows serving it must be considered together.

Openings, mechanical flow and heat transfer

Natural flow through vertical openings is based on pressure differences and the Bernoulli relationship. Floor and ceiling openings use an empirical treatment of pressure and density differences. Leakage is represented by small explicit openings. These choices make opening geometry and location relevant to the exchange between modelled spaces.

Mechanical ventilation is specified through volumetric flows entering or leaving compartments. The supplied model does not solve an entire HVAC system from fan curves and duct resistance. Its guide similarly describes forced-flow treatment without modelling the complete ductwork. A prescribed extract flow therefore carries assumptions about the service delivered to the compartment.

The forced-flow treatment also allows filters to remove selected transported species. The guide describes filtering of soot and trace species through changes to the flow source terms. This does not mean that all gas constituents are removed equally. The selected species and removal assumptions must correspond to the filtering function represented.

Heat transfer includes radiation, convection and conduction associated with the gas layers, fires, boundaries and objects. Changes in wall temperature influence subsequent heat exchange. The model therefore needs thermal properties rather than only compartment dimensions. Treating all boundaries as interchangeable would remove material information required by the calculation.

The fire's radiative representation uses a simplified point source. The guide warns that this limits accuracy within a few fire diameters. It also states that radiative exchange between separate compartments is not modelled. Results close to a flame or dependent on radiation through an intercompartment opening need interpretation within those limits.

Sprinkler activation is estimated using a ceiling-jet correlation and an activation model. A simplified suppression treatment then reduces the specified heat release rate. This represents an assumed suppression response within the model. It does not constitute a detailed prediction of water distribution and fuel cooling for every sprinkler arrangement.

Opening flow can also introduce gas into an upper or lower layer in a receiving compartment. The associated mass and energy become part of that layer's balance. A remote room can therefore change without containing the prescribed fire itself. Connected-compartment calculations depend on both the source-room conditions and the intervening openings. A relatively cool lower layer should therefore not automatically be described as free of combustion products.

Outputs and their interpretation

Typical outputs include gas-layer temperatures, oxygen and smoke concentrations, boundary temperatures and heat fluxes to targets. The model also provides opening-flow information and calculated detector or sprinkler activation times. These outputs answer different physical questions. Their availability does not imply identical accuracy for every quantity.

Different outputs depend on different input assumptions and submodels. Based on Peacock et al. (2026), pp. 4–6.

A uniform upper-layer temperature is a representative zone value. A plume-centreline temperature or a ceiling-jet result comes from a different submodel. Those quantities should not be substituted for one another merely because all are temperatures. Their locations, physical meaning and supporting correlations must remain clear when interpreting exposure.

Van Rossum describes Smokeview as a means of presenting CFAST results graphically. Such a display helps communicate the modelled layers and connected spaces. The underlying zone assumptions remain unchanged by the visualisation. A detailed-looking image should not be interpreted as evidence that a full local flow field was calculated.

The results can support assessment only when the requested output matches the decision. Average layer conditions may be useful for a compartment question, while a local target question may depend on radiation or plume assumptions. The technical guide directs attention to validation for the quantity of interest. Model selection and output interpretation therefore belong together.

Applicability, validation and version control

The guide states that CFAST is generally intended for strongly stratified compartments. Many underlying correlations were developed in relatively uncluttered, flat-ceilinged spaces resembling rooms. Corridors, tall shafts and unusual aspect ratios require attention to relevant validation evidence. The absence of a hard input limit is not a demonstration that every geometry is accurately represented.

Verification and validation examine implementation and comparison with suitable evidence. The guide describes a separate validation document that relates predictions to experimental conditions and measurement uncertainty. Accuracy is expressed for particular quantities through bias and spread. A general statement that CFAST has been validated is less informative than the evidence relevant to the actual prediction.

The empirical submodels were developed for particular processes and conditions. Combining them in one program does not remove their underlying assumptions. The guide explicitly places scenario suitability against the experimental evidence available. A valid conservation balance can therefore coexist with an unsuitable plume, vent or exposure approximation for a particular application.

Version identification matters because submodels and implementation change. The supplied guide describes a revision history and notes that some limitations reported for earlier releases were subsequently corrected. Evidence from a historical version must be examined before it is applied to another. The model version, inputs and relevant validation should accompany the interpreted results.

References

• Peacock, R. D., McGrattan, K. B., Forney, G. P., & Reneke, P. A. (2026). CFAST: Consolidated Fire and Smoke Transport (Version 7). Volume 1: Technical reference guide (NIST Technical Note 1889v1, revision 7.7.7, 27 May 2026). National Institute of Standards and Technology.
• Van Rossum, E. (2027). Introduction to Fire Safety Engineering.