Overhaul Helium and Hydrogen Validation Cases

[GoogleCodeExporter]

Currently the INERIS_H2 and NIST_He_2009 cases have the following problems:

1. Too much CPU time
2. Grid stretching beyond recommended limit
3. Inconsistent CSMAG and Pr numbers
4. Unnecessary and wasteful specification of thermal properties in an 
isothermal calculation
5. Unneeded parameters like SYNCHRONIZE in a serial calculation
6. Too many slice files cluttering up disks
7. Commented out input lines with no explanation
8. Inadequate references to experiments
9. Awkward specification of leakage

These cases need to be re-run before they can be documented in the 
Validation Guide.

Original issue reported on code.google.com by mcgra...@gmail.com on 9 Jul 2009 at 5:22

[GoogleCodeExporter]

Should BAROCLINIC bet set for these cases?

ISOTHERMAL hides some of the errors of the transport.  Should these be 
non-ISOTHERMAL
for V&V (we will get some temperature changes due to the density gradient)?

Original comment by drjfloyd on 10 Jul 2009 at 3:36

[GoogleCodeExporter]

Are the NIST_He_2009 cases the same documented in NHA Annual Hydrogen 
Conference, 2009
A NUMERICAL STUDY OF HYDROGEN OR HELIUM RELEASE AND MIXING IN PARTIALLY 
CONFINED SPACES
K. Prasad1, W. M. Pitts and J. C. Yang ?  

If so the paper says the source was a Fisher burner.  This is a burner designed 
to
give a pre-mixed flame.  What was the co-flow opening during these tests, and 
was any
assessment made of the amount of air co-flowing with the helium out the top of 
the
burner?  Any significant co-flow should be accounted for in the burner inlet 
condition.

Original comment by drjfloyd on 14 Jul 2009 at 5:13

[GoogleCodeExporter]

Jason, Yes the cases are documented in the NHA annual hydrogen conference 2009 
paper. 

In the experiments the helium was released through a Fisher burner. The air 
inlet 
holes of the Fisher burner were sealed off. As a result there was no air 
co-flow 
along with the helium. Helium mass flow was measured by a mass flow controller 
and 
this mass flow rate per unit area was used in the FDS simulations.

Original comment by prasad.k...@gmail.com on 14 Jul 2009 at 5:26

[GoogleCodeExporter]

We did perform simulations with the baroclinic terms set. The changes in the 
results 
was small, but the cost of the simulations increases. 

I am currently attempting different ways to mesh the geometry so as to reduce 
the 
computational costs without affecting the accuracy of the results.

Kuldeep

Original comment by prasad.k...@gmail.com on 14 Jul 2009 at 5:29

[GoogleCodeExporter]

What is the status of this thread?

Original comment by mcgra...@gmail.com on 5 Aug 2009 at 12:41

[GoogleCodeExporter]

The revised input files for the NIST_He_CaseA, B, C and D have been uploaded 
and are 
attached. Two simulations have not finished as yet and I will upload all the 
output 
files once the calculations have finished. 

Original comment by prasad.k...@gmail.com on 5 Aug 2009 at 1:09

Attachments:

• NIST_He_CaseA.fds
• NIST_He_CaseB.fds
• NIST_He_CaseC.fds
• NIST_He_CaseD.fds

[GoogleCodeExporter]

The cases still exhibit problems. Re-run with the latest version of FDS and 
review 
and/or update the Validation Guide.

Original comment by mcgra...@gmail.com on 17 Nov 2009 at 2:19

• Changed state: Started

[GoogleCodeExporter]

Uploaded the revised cases for NIST_He_CaseC.fds and NIST_He_CaseD.fds
Case D is an 8 hour long simulation, and as a result the CPU cost of that 
simulation 
scales accordingly.

Original comment by prasad.k...@gmail.com on 3 Dec 2009 at 7:14

• Changed state: Fixed

[GoogleCodeExporter]

Verified, thanks.

Original comment by mcgra...@gmail.com on 4 Dec 2009 at 10:54

• Changed state: Verified