Vent Opening Combination Error

[RAS]

Good Morning,

We are operating CFAST version 7.6.0 as a V&V’d version of CFAST at a Department of Energy site.  In attempting to track software updates, I noticed a revision note for version 7.7.0 which states the following:

 

“Corrected an error in CFAST where certain vent opening combinations for cases with multiple vents in a single compartment would not initialize correctly.”

 

Our validation process considered ten validation cases from the CFAST github input file listing, four of which have multiple vents in a single compartment.  We compared all the outputs for the four cases using versions 7.6.0 and 7.7.1 with no differences.

 

Can you clarify where the error in version 7.6.0 occurs?  Which “certain vent opening combinations” cause the error, and how one can be identified when it does “not initialize correctly?”

 

Thanks in advance.

Ray

[Kevin McGrattan]

Follow this link and click on "History" and  you will get a detailed history of past code commits to CFAST.

https://github.com/firemodels/cfast/commits/master/Source/CFAST

[RAS]

Kevin,

The issue is listed in the v7.7.0 release notes as follows:

Corrected an error in CFAST where certain vent opening combinations for cases with multiple vents in a single compartment would not initialize correctly.

I looked at the commit history from the link you provided and found one from 3/11/21, “Fix bug in vent opening by flux or temperature.”  If this is not the commit associated with that release note issue, please tell me which commit is the right one.

 

I’m not well versed in the code language, but it appears the fix just changes a vent trigger setting of “flux” or “temperature” to “time.”

 

At this point, I only need to assess how this issue affects results in the previous version, v7.6.0. but I can’t seem to find enough information.  So, I’ve boiled it down to one question:

 

If I leave the vent “Open/Close Criterion” set to “time” for all vents of all three types (wall, ceiling/floor, and mechanical), does the simulation initialize correctly and give valid results?

[Kevin McGrattan]

If the cases run with the old version had multiple vents of different types, then I suspect you might have incurred an error in operation. However, if all the vents were triggered by time, the default, then I suspect that this bug would not have affected you. Rick Peacock can clarify.

[Richard Peacock]

What happened with the old version was if you had multiple vents of a given type connecting the same two compartments, the initial value for the vent opening, rather than being the initial value for the given vent, was the value for the first vent.  If all vents started out as all open or all closed, then the results should be OK.  Also, unless you specifically include vent opening and closing in the input file (i.e., vents are just as input at t=0 with no change over time, the results are fine.  

If your results depend on opening and closing multiple vents connecting the same compartment pair, you may want to rerun things with the fixed version to verify that the results are still as you expect.