TH error messaging
Hammond, Glenn E
Devs,
Issue 1:
The attached example is a TH run where THERMAL_CONDUCTIVITY_XXX are defined in MATERIAL_PROPERTY “rock” without the use of thermal characteristic curves in the input file. This should be possible. Those cards (and HEAT_CAPACITY) are intentionally left out of the other material properties to test error messaging. However, the printed error message is incorrect:
ERROR: Duplicate thermal characteristic curve <> has been detected.
Note that “<>” was added by me as this should be the name of the thermal characteristic curve (which is blank in the code).
“thermal characteristic curves” should not be discussed in the error message as in the eyes of the user, they are not employed. Their use is within PFLOTRAN and the user should not worry about them when using THERMAL_CONDUCTIVITY_XXX. This needs to be fixed.
Issue2:
If you comment out the HEAT_CAPACITY and THERMAL_CONDUCTIVITY_XXX cards in material “rock”, the code crashes because the array of thermal characteristic curves (underneath the hood in PFLOTRAN) is null.
Both of these are user errors, and code needs to be added to catch the errors and guide the user in fixing the issues.
Can someone at Sandia look into this as it all revolves around the refactor to incorporate “thermal characteristic curves”?
Thanks,
Glenn
Leone, Rosie
Sent: Friday, September 6, 2024 11:15:55 AM
To: pflotran-dev (pflotr...@googlegroups.com) <pflotr...@googlegroups.com>
Subject: [EXTERNAL] [pflotran-dev: 6282] TH error messaging
You received this message because you are subscribed to the Google Groups "pflotran-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pflotran-dev...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/pflotran-dev/PH0PR09MB7836CCD9C5375A8EA7AD33619A9E2%40PH0PR09MB7836.namprd09.prod.outlook.com.
Hammond, Glenn E
Rosie,
Note that TH also lacks a check on whether ROCK_DENSITY has been defined. The simulation will run if not defined, but gives erroneous results. It actually grinds with the linear solver frequently diverging when run in parallel.
Glenn
From: 'Leone, Rosie' via pflotran-dev <pflotr...@googlegroups.com>
Sent: Friday, September 6, 2024 10:57 AM
To: pflotr...@googlegroups.com
Subject: Re: [EXTERNAL] [pflotran-dev: 6283] TH error messaging
Check twice before you click! This email originated from outside PNNL.
To view this discussion on the web visit https://groups.google.com/d/msgid/pflotran-dev/BY5PR09MB56333B477C72F8667F3EE7F1A59E2%40BY5PR09MB5633.namprd09.prod.outlook.com.
Salazar, Alex
-
This issue is likely caused by including backwards-compatibility when TCCs were fully integrated. I do remember using an automated labeling scheme in these cases.
-
I think some associated guards can probably do the trick.
To: pflotr...@googlegroups.com <pflotr...@googlegroups.com>
Subject: RE: [EXTERNAL] [pflotran-dev: 6284] TH error messaging