The link does work, but we have had compatibility problems with some international downloads. The NIST ftp server requires all IP addresses to have an associated name. Some internet services providers do not provide a name to go along with the numeric IP address. NIST is working to upgrade the ftp server, but this may not be available soon. You can request a copy of the software by e-mail to cfast [at] nist.gov (subject: Request%20for%20copy%20of%20current%20version%20of%20CFAST) . If you don't already have the Microsoft .NET framework on your machine, here's a link to the Microsoft download. It is required for operation of the software.
The only difference between the smaller and larger download from the CFAST web site is the inclusion of the .NET framework.
Yes and no. CFAST has been compared to a wide range of experimental results. Details and references for these are available in the CFAST verification and validation guide available here. We're working to give better guidance as to where the model can be trusted and where it can't, but this is an on-going exercise. Our advice to users is, as always, to validate the model for the type of fire scenario you are interested in. That means simulate an actual experiment with geometry and fire similar to that you plan on using. If the difference in model and experiment is acceptable by whatever criteria you choose to use, go ahead and use the model. If not, let us know, and we'll see if we can improve that feature or suggest another approach.
CFAST has been compiled on several UNIX platforms. Scripts to build CFAST are available as part of the CFAST source code repository available here. The GUI front-end, CEdit, is strictly a Windows application. The CFAST source code is available on the Downloads Page, and you will have to compile the code yourself. You'll need a Fortran 2003/2008 compiler.
There is a very useful Discussion Group where you can post a question to the community. This is closely monitored by all of the developers and is the best place to get your question answered. Posting here instead of just sending an email to one of the developers, will not only open your question up to all the developers at once, but will allow you to tap into the collective expertise of the entire community.
For some users, part of the display is cut off and changing the size of the main program window does not fix the problem. window Typically, the program does not show the right-hand portion. For example; the result window for the example run shows half the column for the "Ambient Target Flux" and does not show the scroll bar on the far right. We see this most often with Dell machines. Typically it is caused by the display settings configured for larger than normal fonts. You can check and correct this by going to the control panel (Start, Control Panel, Display), select the Settings tab and then the Advanced usa-button. The DPI setting on the General tab should be set to "Normal Size (96 DPI)." If it is set to Large or Custom, CFAST will not display properly. This should not be a problem in version 7 of the model.
Most likely, the error is related to your selection of an international numeric format. CFAST is designed to use the U.S. English numeric format that uses a period as the separator between the integer and fractional part of a number. Unfortunately, the software does not recognize the comma separator and treats the input 2,9 as two separate numbers. To solve the problem, users have simply set up a new user on the machine and set the regional options to "English (United States)" for that user in the "Regional and Language Options" Control Panel applet.
Consider the following excerpt from a CFAST output.Time = 1200.0 seconds.
The mass balance is just the sum of the flow in minus the flow out. For any compartment, this is just ("Upper Layer Inflow" + "Lower Layer Inflow" + "Pyrol Rate") – ((Upper Layer Outflow + Lower Layer Outflow) with the inflow and outflow summed for each vent. Note that the mixing flows are ignored since they are just exchanges between layers in the same compartment. For the above example, it's just
(6.890E-03 + 1.34 + 5.409E-02) – (6.662E-03 + 1.40) = -0.005682 kg/s
or less than 1/2 percent of the magnitude of the inflow.
The easiest way to create your own fire file is to copy an existing one and change the values as appropriate for your specific fire scenario. To copy a fire file, select "From File" from the Fires tab. The "Duplicate" usa-button copies any selected existing fire object, creating a new object. Simply change the name of the new object and edit the values as appropriate.
If you look at the CFAST output log on the Environment tab in CEdit, you will typically see and error 205 when the model does not run. This is an error in some early versions of CFAST 6 and was fixed in version 6.0.10. The 205 error indicates that a material property is not defined. This could be in the fire definition, compartment surfaces, or targets. Typically, this is caused by a selection of "Off" in the material definition for a fire which is the default selection when you create a totally new fire object with the "Add" usa-button. You do have to select a material for the object. It is used to calculate heat transfer to the object prior to ignition.