NIST logo
Bookmark and Share

Fire Modeling for Performance-Based Design Project


This research will address the shortcomings of CFAST (Consolidated Fire And Smoke Transport) and FDS (Fire Dynamics Simulator) that were identified during the FY11—FY13 verification and validation study. Specifically, the research will lead to (a) generalized obstruction capability in FDS to handle non-rectangular geometries, (b) complementary visualization techniques in Smokeview, (c) predictive capability of the burning rates of liquids and simple solids, (d) advanced treatment of soot transport, and (e) continued improvements to CFAST to make it more consistent with FDS.


Objective: By 20171, to develop the necessary numerical framework to (1) remove from FDS and Smokeview the constraint that all obstructions must conform to the rectilinear gas phase numerical grid, (2) predict the burning rate of a fully-furnished compartments using only thermophysical property data measured at bench-scale, and (3) develop a verification suite for CFAST that is similar in scope to that of FDS.

What is the new technical idea? In Phase 1 of this project (FY11 to FY13), NIST successfully developed extensive automated techniques and computational methods to demonstrate the predictive capabilities of FDS and CFAST. These automated processes are providing extensive verification and validation (V&V) for fire models developed by NIST and others, and is enabling improvements to international building standards. Based on the results from the FDS and CFAST V&V studies and end-user feedback for performance-based design (PBD) applications, enabling the software to allow for the direct use of computer-aided design (CAD) and coupling with structural models would result in more efficient and accurate computations for PBD.

What is the research plan? In addition to general support for the public release of both CFAST and FDS, the project is separated into five research tasks all of which are critical to the project’s objective.

Task 1 is to issue a public version of the Immersed Boundary Method (IBM) in FDS. IBM is a method of representing complex solid obstructions within a uniform rectangular mesh, allowing high fidelity large eddy simulation of fire phenomena and fire-structure interaction. IBM will 1) allow for direct use of computer-aided design (CAD) in the processing of FDS fire scenarios, 2) allow users to couple complex geometry with structural finite element (FEM) solvers (e.g., Abaqus2) and 3), with the increased resolution in the gas phase, lead to more accurate prediction of the heat flux to solid surfaces for real materials in buildings3. A proof-of-concept version of IBM is currently implemented in FDS. In year 1 (FY14), we will develop and verify an overset (Chimera) mesh capability needed to handle conservative heat and mass transfer. In year 2 (FY15), we will validate the methodology (compare simulations to experiments). In year 3 (FY16), we will demonstrate fully coupled fire-structure calculations.

Task 2 is to develop the functionality within Smokeview to visualize arbitrary geometrical shapes, complementing the development of the Immersed Boundary Method. Visualization techniques are needed to verify the underlying IBM computations and guide IBM development. Various rectilinear visualization techniques used with the existing version of Smokeview (e.g., displaying data contours on surfaces) need to be adapted for use with the IBM in order to obtain insight from IBM numerical computations. Preliminary methods for visualizing IBM objects in Smokeview have been critical to the achievement of preliminary results. In year 1 (FY14), we will develop visualization tools and techniques that verify IBM computations and Chimera mesh structure (being developed in Task 1). In year 2 (FY15), we will adapt surface data contouring for use with the IBM. In year 3 (FY16), we will implement visualization techniques for IBM objects that move.

Task 3 is to reliably predict the burning rate of simple solid materials and liquids in FDS.New data generated through solid-phase modeling will guide the development of a framework for sub-models on material burning and soot emission4. In year 1 (FY14), we will demonstrate through small scale validation work5 the feasibility of modeling real materials. Real materials are typically mixtures of different components. Progress thus far has been limited to homogenous bulk plastics. In year 2 (FY15), we will simulate single burning item experiments (e.g., furniture). In year 3 (FY16), we will simulate fires within single compartments using only properties obtained through bench-scale measurements6.

Task 4 is to develop in FDS an improved transport, agglomeration, and deposition algorithm for soot. The current algorithms used in FDS are predicting that the smoke alarms will activate earlier than experimentally measured. The assumption is that soot deposition on the detectors is slower than FDS is calculating. Verification and validation of smoke alarm algorithms will improve smoke alarm predictions in FDS. This task will be completed by FY15.

Task 5 is to develop sprinkler and water mist suppression algorithms in FDS in collaboration with United Technologies Research Center. The sprinkler nozzle characterization currently used in FDS was developed by a VTT7 Finland under the sponsorship of the Marioff Corporation. Marioff was acquired by United Technologies and much of the research has shifted to UTRC in Hartford, CT. We continue to collaborate with both VTT and UTRC, providing support for algorithm development and V&V. Specific milestones will be driven by UTRC and VTT. We plan to play a supporting role.

Task 6 is to add a suite of verification cases for CFAST and to make it more consistent with FDS in terms of input parameters and functionality. CFAST is a vital design tool for the engineering community because its calculations can be run in minutes as opposed to hours or days for FDS. The trade off for the faster processing is limited capabilities as compared to FDS. During Phase 1 of this project (FY11-13), CFAST was updated into compliance with modern Fortran standards, which enables us to now extended CFAST capabilities without sacrificing the quick processing time needed by the end-users. In year 1 (FY14), we will implement updated combustion chemistry that is more consistent with FDS and a new verification guide to improve confidence testing for the model. This will facilitate adoption of the model by providing a robust set of test scenarios for model users. In year 2 (FY15), we will update the numerical solver. In year 3 (FY16), we will improve estimation of gas temperatures near fire sources (which would improve calculation of heating and ignition of nearby targets.


1 The project is in the first year of phase 2. The long term vision is a high-fidelity, physics-based fire models, like FDS, used responsibly in the regulatory arena, in operational forecasting, and in forensic investigation outside the application space where experimental data already exists.
  • Phase 1 (FY11-13): To develop verification and validation for major existing and new algorithms in FDS, CFAST, and the visualization tool, Smokeview, so that the models can be used to accurately and efficiently predict fire conditions for performance-based design in buildings. 
  • Phase 2 (FY14-16): to develop the necessary numerical framework to (1) remove from FDS and Smokeview the constraint that all obstructions must conform to the rectilinear gas phase numerical grid, (2) predict the burning rate of a fully-furnished compartments using only thermophysical property data measured at bench-scale, and (3) develop a verification suite for CFAST that is similar in scope to that of FDS. 
  • Phase 3 (FY17-19): to make FDS work seamlessly with computer-aided design (CAD) packages. 
  • Phase 4 (FY20-22): to develop adaptive mesh refinement in FDS, in which the numerical grid adjusts to account for changes in required grid resolution.  

2 This work is in collaboration with Global Engineering and Materials, Inc., Princeton, NJ, under an SBIR from the Office of Naval Research.  

3 The opening of the National Fire Research Lab will offer new opportunities for fire-structure model validation.  

4 Includes new and ongoing research in EL by Pitts along with cooperative research at VTT in Finland.  

5 Bench-scale validation of pyrolysis models typically means predicting the burning rate of the material in the cone calorimeter using measured thermophysical property data. 

6 FDS has been used to reconstruct full-scale fires during the WTC, Rhode Island, and Charleston investigations. However, these simulations used a combination of data from bench-scale and full-scale experiments.  

7 VTT is a water mist manufacturer.


Major Accomplishments:

Potential Research Impacts:

  • McGrattan, K, et al., Computational fluid dynamics modeling of fire, International Journal of Computational Fluid Dynamics, 26, 2012.
  • Gann, R.G., et al., Reconstruction of the Fires and Thermal Environment in WTC Buildings 1, 2, and 7, 49 (3). pp 679-707, July 2013.

Realized Research Impact:

  • McGrattan K, et al. (2011) “Fire Dynamics Simulator Version 5 Technical Reference Guide” NIST Special Publication 1018-5, October 10, 2009.

Impact of Standards and Tools:

  • CFAST, FDS, and Smokeview models  have transformed PBD of fire protection in buildings and have been used to support codes and standards development worldwide.
  • Determining the need and the type of corrective fire protection in nuclear plants is based primarily on FDS, CFAST, and Smokeview. NIST wrote the US Nuclear Regulatory Commission (NRC) publications with stipulates how to use this software package for nuclear power plant applications.
Smokeview rendering of a fire in a cable spreading room of a nuclear power plant.  Image: NIST
Smokeview rendering of a fire in a cable spreading room of a nuclear power plant. Image: NIST

Start Date:

October 1, 2011

Lead Organizational Unit:



Project Leader: Dr. Kevin B. McGrattan

Associate Project Leader: Dr. Glenn P. Forney

More Information on Fire Modeling Research:


FDS and Smokeview


General Information:
Dr. Kevin B. McGrattan, Project Leader
301-975-2712 Telephone

100 Bureau Drive, M/S 8664
Gaithersburg, MD 20899-8664