CFAST Version History
Version 6.1.1 (February 2009)
Added a new species, labeled TS, to track flow through compartments. TS is a trace species produce during pyrolysis and is not affected by combustion. The assumption is that molecules are large enough to be filtered, but that mass is small compared to the overall pyrolysis. See NISTIR 7498 for an example of its use.
Added the ability to specify filtering for mechanical ventilation systems which applies to soot and the trace species. For the user, it is implemented with the EVENT,F command and available through the GUI. See the User’s Guide, NIST Special Publication 1041 (May 2008 Revision) for details.
Separated pyrolysis and combustion kinetics. The original purpose was to fix a bug related to pyrolysis products not being transported to the upper layer when a fire is only in the upper layer. The code was structured so that a consistent combustion model for a plume could be implemented. Species affected only by pyrolysis are hydrogen chloride, hydrogen cyanide, trace species and the concentration time product. Species affected only by kinetics are carbon-dioxide, carbon-monoxide, and soot. Species affected by both are oxygen and fuel. Nitrogen is inert and only transported.
Corrected error in mechanical ventilation that prevented flow from dropping off at high pressures like it was supposed too. An additional error caused DASSL to fail at times when the mechanical ventilation system shut off due to (the now corrected) high pressure cutoff. The solution was to not totally close the vent (the vent is left open at an extremely small value, machine epsilon for double precision) Corrected error in exehandle routine that caused model to fail with path names longer than 128 characters (windows limit is 255 characters for whole file name). Corrected error in interpolation routines for opening and closing vents that did not smoothly transition from open to closed (Since the GUI did this over 1 s, the impact should not be noticeable for previously run cases).
Incorporated Heskestad plume model as an alternative to the McCaffrey plume model (type 2). There is a small difference in the calculated results for a range of test cases (on the order of 2% to 4%). The plan is to add plume temperature to the calculation as well.
Corrected error in the GUI that incorrectly outputs the mechanical ventilation specification for the CFAST model. The fifth parameter on the line specifies a "system number" for each mechanical ventilation system. In older versions, this began with 1 for each compartment rather than being numbered sequentially throughout. The impact was that opening and closing of vents would not work when there were mechanical ventilation fans in more than one compartment. Calculation was correct for fully-opened mechanical vents.
Version 6.0.10 (January 11, 2007)
Beginning with version 6.0.10, CFAST and CEdit were compiled with Visual Studio 2005. This means that both new and existing users of earlier versions of the Windows CFAST will need to download and install Microsoft .NET version 2.0 (included in the larger download of CFAST) if it is not already installed on their computer.
Corrected subscript error in CFAST input checking for ceiling-floor conduction connections that caused an error with simulations with 5 or more compartments. Corrected an error in the definition of fire objects to allow radiative fraction to be zero, not just greater than zero. Corrected a display error for CFAST files in Smokeview that placed fires in the wrong compartment for visualization; calculations, spreadsheet, and printed output were correct.
Updated the GUI to Visual Studio 2005, including spreadsheet, auto-sizing, and preview libraries. Reversed order of CFAST log file in the message summary so that any errors would appear at the top of the list, more visible to the user. Changed limit on number of mechanical ventilation connections from 10 to 100. Corrected an error that prevented saving of an input file when compartments were removed from a simulation. Corrected an error that caused incorrect values for default fire position when engineering units were not set to m. Fixed compartment naming when a compartment is duplicated. Corrected an error when defining a new fire that did not assign a material to the fire (leading to an error 205 when running CFAST). Enabled RTI input for heat alarms (Previously, heat alarms were presumed to have a default value for the RTI). Corrected a subscript error that caused GUI error message for a compartment specified as a single zone if the compartment was the highest numbered compartment in the simulation.
Added the ability to define a general t-squared growth rate fire to the Edit Fire Objects window, allowing the user to select growth rate, peak HRR, steady burning time, and decay time, including predefined constants for slow, medium, fast, and ultra-fast t-squared fires. Added the ability to add and remove thermal properties on the thermal properties edit window. Redesigned the heat transfer connection tab to separate horizontal and vertical heat transfer specifications to allow for future enhancement of the horizontal connections to better support the HHEAT keyword in CFAST; for the moment, only a single, fractional connection is allowed on a given keyword line. Note that vertical heat transfer always assumes a fraction of one for the connection.
Version 6.0.9 (May 15, 2006)
Corrected error in specification of alternate thermal property files that would not allow reading of the alternate files nor set file changed flag when file was changed. Corrected links to GUI text boxes for HCl constants 2 through 7 that prevented editing of these constants. Corrected size of summary spreadsheets in the GUI to be consistent with CFAST design limits. This would cause the program to crash if you had more than 50 vents or targets in a single simulation.
Heat source fire on Fires tab was not actually implemented in the model; this has been removed from the GUI.
Corrected units in the GUI for HCl and HCN. They were listed as kg/s by default rather than kg/kg (dimensionless). This would lead to incorrect species yields for simulations with user selected units other than kg for mass and s for time.
Corrected error in CFAST where ignition time of a fire was off by 1 s when specified ignition time was t = 0 s.
Version 6.0.8 (February 28, 2006)
Corrected errors in the calculation of unburned fuel that caused the fuel not to burn in compartments remote from the fire room. Ignition temperature for doorway fires is now a global parameter that is entered on the Fires tab of the input program.
Beginning with version 6.0.8, default installation locations for program files and user data files follow the Windows model so that executable files are typically located in "C:\Program Files\CFAST" and user data files are located in "My Documents\CFASTData." If upgrading from an earlier version of the software, the installation program removes older executable files, but leaves existing data files in their original location. If desired, the files can be moved to the new data directory created during installation of the software.
Version 6.07 (February 1, 2006)
Corrected error in the GUI where corridor flow coefficients were incorrectly assigned a default value of 0 rather than the CFAST assumed value of -1. This caused erroneous calculation of corridor flows allowing excessive temperatures in hallways to be calculated.
Corrected an error in the horizontal flow routine that prevented printout of vent flow values for some vents
Version 6.0.5 (October 1, 2005)
Add target normal calculation to cedit, and extend object plotting in GUI to all object variables. Add error checking for objects and detectors placed outside of a compartment. Fix ceiling jet coordinates so that they agree with targets.
An inconsistent labeling of width, depth and height caused incorrect target flux calculations.
The initial/final open fraction for vertical flow and mechanical flow were not being set correctly. Both have been fixed.
The ambient flux calculations seem to be too high. The problem lies in specification of object properties. The fire has to release energy over a finite volume. As the size of this volume is reduced, the energy per unit volume goes up and thus the heat flux rises. There is a physical limit on the maximum such energy release per unit volume. As pointed out by Orloff and deRis ("Froude Modeling of Pool Fires," Factory Mutual Technical Report OHON3.BU, Norwood (1983)), there is a physical limit on the maximum such energy release per unit volume. This limitation arises because mixing and diffusion take a finite amount of time, and radiation is limited by optical depth, at least for fires of the sort we consider in CFAST, namely buoyant diffusion flames. The upper limit is 1.2 MW/m3. An object model which is the area of the object as defined in the database (.o) and the height extent which includes a maximum flame length is used to make this estimate. The input routine, INPUTOBJECT, does this calculation, with a warning at 2 MW/m3, and a stop with the error code 221 at 4 MW/m3. The minimum object volume is now 1.0x10-6 m3.
Version 6.0.1 (June 1, 2005)
Initial release of version 6 of CFAST, including a Windows version of the CEdit graphical user interface. Fires are now treated as objects, which can be placed and sized within compartments. Mechanical ventilation has been simplified to fan/opening pairs, and all vents (vertical flow, horizontal flow and forced flow) can be turned off or on (single instance).
Version 5.1.1 (May 1, 2004)
There was an error in the HCR (hydrogen/carbon ratio) calculation. It introduced about a 6% error into the pyrolysis calculation. Version 5.1.0 (March 1, 2004)
In version 5.0 and earlier, the oxygen calculation used the oxygen to fuel ratio, whereas the technical reference manual states that it uses the oxygen to carbon ratio. The model now matches the guide.
Version 5.0.1 (May 7, 2003)
Flow was not being reported correctly in some cases. The calculation was correct, but the output from CFAST (/rf) was not updated when a vent opening decreased to zero.
The vertical flow calculation has been fixed. The symptom is that it the pressure in a compartment connected by VVENT was never updated.
Version 5.0 (November 1, 2001)
The number of points in a time history has been increased. This is to allow editing of the examples, which default to 21 intervals.
Fix a printout error: the size of the fire was reported incorrectly in some cases. The calculation was done correctly, but the output from Report and ReportSS was incorrect. The problem occurs if there are two fires in a compartment, with one of them in the upper layer, and one in the lower layer. Once again, the report was incorrect, but the value used in subsequent calculations was correct.
The workspace for the solver was not dimensioned correctly, so CFAST could not do 30 compartments. This has been fixed.
(These fixes have been done in 3.1.7 as well)
Version 4.0.1 (March 1, 2000)
This release runs as an application under the Windows series of operating systems. New phenomena: horizontal heat conduction (wall/wall) (HHEAT), horizontal smoke flow in corridors (HALL), variable geometry for compartments (ROOMA, ROOMH).
Fonts must be in the "font" directory off the "bin" folder.
Version 3.1.7 (October 1, 2001)
Add initialization in dofire and fires. Fix an indexing problem in solve. Both problems appeared with the release of Windows 2000 from Windows 98.
Version 3.1.6 (November 1, 1999)
This is the last full GUI (FAST). Add "constrained/with flashover" to FAST as a fire type (note, this is not carried into CFAST!)
Version 3.1.5 (April 1, 1999)
The program Build is included, along with sample files and documentation, which can be used to generate ".pic" files for the VIEW command. Reportss has been fixed to include target temperatures in the output. New phenomena: single zone switch (single layer model).
Version 3.1.4 (July 1, 1998)
Fix compatibility between FAST and CFAST. The history files were not compatible
Version 3.1.3 (April 1, 1998)
Fix fonts for FAST (remove wddraw from an explicit include in fast and fastlite), file naming error in the spreadsheet output, incorrect entrainment when the fire is in the upper layer and eliminate heat conduction thru the ceiling/floor connection to the outside (nputp)
Version 3.1.2 (January 1, 1998)
Fix boundary condition for vertical flow to the outside. New phenomena: vertical heat flow.
Version 3.1.1 (October 1, 1997)
Fix font interface for metawindows. The fortran-c interface could sometimes cause a page fault. Fonts are now names devfntnn.fnt where nn=1 to 24 (4 is the default). Note that font changes used in labels are permanent for the session, so the bracket |Hnn .... |H04 should be used. Removed O(n**2) data structures (about 40 arrays). Added a non-rectangular room option - roomhgt, roomarea are the key words. Wind - wind now applies to vents rather than rooms (couldn't do from vs back) - effect of wind was zero for ground floor rooms. wind induced pressure rise is now calculated at the average of the floor and ceiling elevation Initialize inside and outside pressure and temperature to the same lapse rate.
Output warning for excessive number of small time steps (set in solver.ini) (DTCHEK).
Add BLACK option to have atmosphere behave as a blackbody Note: rework SHAFT option to use only 3 equations per room. Modify hybrid code to conform to new FDS heat loss algorithm
Version 3.0 (January 3, 1996)
New user interface for CEdit. Add flame height calculation to report and flame spread algorithm to the model. New module, reportss, which uses the same format as report, but puts the output into a spread sheet (ascii delimited text) output format. Added reporting option for wall and target heat flux printout. New output options are /r:winfstp. W is now wall and target heat fluxes and P is now wall temperature profiles.
New phenomena added: ceiling/floor heat transfer for inter-compartment heat transfer, CFAST keyword CFCON. and heat transfer to targets. Multiple objects in a desired position, orientation, and material. Program calculation of object temperature and impinging heat fluxes can be printed. CFAST keyword is TARGET.
Improved stability of mechanical ventilation routines so that solution converges over a wider range of input values.
Modules in the release: CFAST (main model), fast (gui data editor), cplot (plotting package), report (report utility), reportss utility to generate spread sheet format, reportg to provide graphics playback, bintoasc to convert binary history files to ascii text files, compare to compare two separate CFAST runs (using ascii text files) and compinfo to summarize output from compare .
Version 2.1.1/ H1.2 "fix on fail" update (January 1, 1995)
Fixed optional ceiling jet calculation to not take into account lower wall surfaces. Too much energy was lost through the lower wall.
Corrected EXITT and SURVIVAL to access the layer interface height correctly. Prior to this fix, layer depth was always zero, making smoke detection (by detectors or "smell") inoperable.
Fixed thermal properties in several example data files to be consistent with current THERMAL.DF file.
Version 2.0.1 H 1.2 (July 1, 1994) This release did not get a new number!
Corrected write of vertical vent information to .DAT file in CEDIT. For large vent area, old format would merge two fields together resulting in read errors with the model. Tightened differential equation solver tolerances to correct calculated species concentrations for one test case. This may make the model run slower when conditions are rapidly changing, but the answers should be more correct.
Corrected interaction between fire size and plume entrainment. Before the fix, CFAST calculated plume entrainment via a two step process:
1) Determine the plume entrainment in FIRPLM via McCaffrey's method using the fire size unconstrained by the available oxygen.
2) Once the actual fire size is calculated from the available oxygen in CHEMIE, a new estimate for the plume entrainment is determined by a simple linear correction. Since the plume entrainment is not a linear function of the fire size and the fire size depends on the oxygen entrained, this simple process can lead to an inconsistency between the fire size (calculated from the unconstrained plume entrainment) and the new estimate of the plume entrainment. For very large fires where the fire size is limited by the amount of oxygen entrained, this can lead to significant differences between the calculated fire size and the amount of oxygen actually available for the combustion. The fix is a simple one: when the fire is limited by the available oxygen entrained into the plume, the plume entrainment and fire size are both re-calculated by calling FIRPLM and CHEMIE a second time to get a better estimate of the actual oxygen available (and thus the actual fire size).
Version 2.0.1 (February 13, 1994) - H 1.2
Corrected flow through horizontal openings (VVENT). In some test cases (where significant flow occurred from upper room to lower room), model would calculate extremely low temperatures in the lower room. Corrected species mass balance. For very large fires in small rooms, species mass fractions would not sum to unity when fire became oxygen limited. Printed output file is now placed in the data subdirectory along with the input data file.
Version 2.0 (August 23, 1993)
Code restructured for future addition of a flame spread model. Consolidated calculation of hazardous conditions (and colors for plotting of them). Added a new THERMAL.DF file taken from Incorpera and DeWitt. New conduction routine. New convection routine. Use Watcom Fortran compiler. Printout routine has been totally rewritten to provide additional information with a consistent format.
Version 1.6.2 (December 1, 1992)
Implement a faster initialization routine, and better selection rules for vertical flow. We now use the Pharlap memory expander which supports more display drivers. Mechanical ventilation is now tightly coupled with vertical and horizontal flow routines.
Version 1.5 (June, 1, 1992)
Added object specification (heat loss, ...), a more robust ODE solver (DASSL), restructured code to include physical interface routines for each physical phenomenon, and CEdit support for all modeling features.
Version 1.4 (February 1, 1992)
Added multiple fires, history file compression, extended memory, vertical flow, multiwall radiation model, distributed mechanical ventilation ducts, ceiling jet and 3D positioning of fires. The number of compartments was increased to 15.
CFAST Version 1.0 (May 1, 1990)
An amalgam of FAST 18.5 and CCFM -- functionally equivalent to FAST but more modular like CCFM.
FAST version 18.5 (May 1, 1990)
HAZARD version 1.1
FAST version 18.3 (July 1988)
HAZARD I version 1.0