search

firemodels / fds

Fire Dynamics Simulator
https://pages.nist.gov/fds-smv/
Other
642 stars 614 forks source link

Gas phase conductivity output off #13306

Closed johodges closed 2 weeks ago

johodges commented 1 month ago

Describe the bug I was using the gas phase property outputs to do some manual calculations and noticed that the conductivity value was giving me odd results. In the simplified case (one of Validation/Convection/impinging_jet with the impinging jet correlation removed and a coarser grid) I see the behavior below. The far left image shows what I get with master. In master, DUMP uses a different call for DNS versus other SIM_MODE. In non-DNS cases the call is:

    GAS_PHASE_OUTPUT_RES = MU(II,JJ,KK)*CPOPR

In the middle and right image I changed all the SIM_MODE to use the same call as DNS (see below). The middle includes the &SPEC line for air where CONDUCTIVITY is manually specified by the user. The right does not include that line.

     ZZ_GET(1:N_TRACKED_SPECIES) = ZZ(II,JJ,KK,1:N_TRACKED_SPECIES)
     CALL GET_CONDUCTIVITY(ZZ_GET,GAS_PHASE_OUTPUT_RES,TMP(II,JJ,KK))

Is there a reason we are calculating conductivity differently between the sim modes? Is there a big performance impact in ZZ_GET and GET_CONDUCTIVITY for LES calculations?

Screenshots

image

Desktop (please complete the following information):

Additional context

impinging_jet_Re_1e5_coarse.txt

drjfloyd commented 1 month ago

Subgrid turbulence makes the conductivity effectively larger than the molecular level value.

johodges commented 1 month ago

That makes sense. The reason I was using the output was to get the molecular level value used in the Nu calculations so I could do my own approximate calculations to a fictitious surface throughout the domain. I should be able to do the same thing calculating a MOLECULAR CONDUCTIVITY using MOLECULAR VISCOSITY. Would it be beneficial to add an output for MOLECULAR CONDUCTIVITY or is this too much of an edge case?

drjfloyd commented 1 month ago

I thnk having that output makes sense. Would just be a call to GET_CONDUCTIVITY.

mcgratta commented 3 weeks ago

I added MOLECULAR CONDUCTIVITY. Test it and make sure it works.