Add support for using (v,e) interpolation tables in THM #31860

GiudGiud · 2025-11-06T22:15:05Z

Bunch of missing routines, enough to be using the 1 phase pipe flow model at least
EDIT: and the volume junction now. So it should be nearly all we need for regular use, and we are within striking distance if a new object comes up!

Also added:

additional checks on input variables (consistent across the entire object now)
more AD support. The "capping" of input variables does not play well with our use of "const ADReal &" but I think the capping (for robustness) is more important than performance right now, so I made copies of the ADReal inputs
more unit testing, covering previously uncovered routines
deprecated 'fp' and added a workaround as THM simulations love having 'fp' in GlobalParams and 'fp = my_name' does not work for TFP
widely buffed the ability to do partial passthrough, e.g. interpolating some quantities and using an 'input_fp' for others. It's well tested now, and you can pass-through a tabulated property (previously would not let you)

@licharlot in case you want to keep following TFPs

GiudGiud · 2025-11-11T21:04:09Z

Worked on adding support for a volume junction, so this should be enough for general use in THM now.
Bunch of (p, s), (p, h) etc calls work now

refs idaholab#20101

Add debugging output

… newton refs idaholab#20101

With this we can initialze a THM flow channel 1 phase refs idaholab#20101

- add the p/T from v e with derivatives and AD derivatives to be able to do a Newton solve with AD refs idaholab#20101

- requires T_from_p_rho and e_from_p_T with AD for pT tabulations refs idaholab#20101

- use parameter tolerance - use parameter number of iterations

…he input file - also the current iterate and targets

refs idaholab#20101

- mention two options: interpolated properties with specific name fp object

to enable using volume junctions - use newton on getting rho=1/v, then newton again to get e by inverting T(v,e)

to use TFP with volume junction When using (p,T) previously, s_from_v_e ended up computed through a p_T_from_v_e call

to satisfy the p_from_h_s call in volume junction

to enable use of v e tabuluations with volume junctions

…erties (consistency) Use the FP uo preferentially for computing grids

…rated from a FP object

Add more range check

moosebuild · 2025-11-12T02:52:05Z

Job Documentation, step Docs: sync website on 470fc75 wanted to post the following:

View the site here

This comment will be updated on new commits.

- enable when not using (p,T)->(v,e) conversion - avoid the header to be able to use for gold files - avoid a lone comma when not outputting any extra properties - avoid conversion errors with h(p,T) call

Add a dedicated test for (p,T) output

- try to guess when newton is used for initial guess checks

initialize FP earlier & use the tabulation rather than the FP

…y sepecifying the interpolated properties in the input - address porous flow failure that was relying on forwarding a property

moosebuild · 2025-11-13T05:00:25Z

Job Coverage, step Generate coverage on 470fc75 wanted to post the following:

Framework coverage

Coverage did not change

Modules coverage

Fluid properties

	b79a48	#31860 470fc7
	Total	Total	+/-	New
Rate	85.95%	86.08%	+0.13%	100.00%
Hits	8195	8466	+271	455
Misses	1340	1369	+29	0

Diff coverage report

Full coverage report

Full coverage reports

Reports

This comment will be updated on new commits.

Make sure to initialize (v,e) grid for passthrough case Add testing for (v,e) passthrough

- create a dedicated passthrough tabulation - use idg fps that have the most AD definitions created

…e FP - otherwise we would compute min/max from FP calls OOB from the tabulation - then flush to min/max that could be OOB

- T/p from v e with derivatives in the passthrough _fp case - AD e and v from p T - e from p rho using a pT tabulation - AD s from v e - T from p rho using a ve tabulation

moosebuild · 2025-11-13T23:11:41Z

Job Test, step Results summary on 470fc75 wanted to post the following:

Framework test summary

Compared against b79a48d in job civet.inl.gov/job/3380566.

No change

Modules test summary

Compared against b79a48d in job civet.inl.gov/job/3380566.

Added tests

Test	Time (s)
`fluid_properties/test:tabulated.file_writing/pT`	1.01

Run time changes

Test	Base (s)	Head (s)	+/-
`porous_flow/test:fluidstate.theis_tabulated`	7.50	3.32	-55.79%

Avoid empty columns in tabulation file

283df28

GiudGiud self-assigned this Nov 6, 2025

GiudGiud changed the title ~~Add some support for using (v,e) interpolation tables in THM~~ Add support for using (v,e) interpolation tables in THM Nov 11, 2025

GiudGiud marked this pull request as ready for review November 11, 2025 21:04

GiudGiud requested a review from joshuahansel as a code owner November 11, 2025 21:04

GiudGiud force-pushed the PR_fp_thm_ve branch from a03a926 to da808c2 Compare November 11, 2025 21:11

GiudGiud added 23 commits November 11, 2025 16:12

Check the initial guess when loading a (v,e) tabulation

a42e6c9

refs idaholab#20101

Make 2D newton inversion able to do 1D solves in degenerate case

05317d5

Add debugging output

Add ability to compute rho from p T when using v,e tabulations: using…

29ae5da

… newton refs idaholab#20101

Better error on missing routines

d21121b

Add e_from_p_rho computed from a ve tabulation

6d4ae89

With this we can initialze a THM flow channel 1 phase refs idaholab#20101

Add rho from p T with AD for ve tabulations:

1c8b4de

- add the p/T from v e with derivatives and AD derivatives to be able to do a Newton solve with AD refs idaholab#20101

Add e_from_p_rho with AD for ve tabulations

b0cfb39

- requires T_from_p_rho and e_from_p_T with AD for pT tabulations refs idaholab#20101

Adjust newton2D parameters

3259107

- use parameter tolerance - use parameter number of iterations

Add the caller name for the 2D newton solves to help debugging from t…

930f4b6

…he input file - also the current iterate and targets

Add test for degenerate Newton solve

300ace2

Add test for new fluid properties calls

2ca7275

Add more log to all Newton utils : use parameters in SPFP too

6baf412

Simplify default params for newton inversion

d219f00

Add h_from_p_T using a (v,e) tabulation for volume junctions

4dd1fd1

refs idaholab#20101

Rework again error messages:

b8a498b

- mention two options: interpolated properties with specific name fp object

Add another error for missing input

8af3db5

Add e_from_p_T using a (v,e) tabulation

f4dbb9c

to enable using volume junctions - use newton on getting rho=1/v, then newton again to get e by inverting T(v,e)

Add missing s_from_v_e with derivatives

42292fd

to use TFP with volume junction When using (p,T) previously, s_from_v_e ended up computed through a p_T_from_v_e call

Implement h_from_p_T with derivatives using (v,e) tabulation

17b05cf

to satisfy the p_from_h_s call in volume junction

Add rho from p T with derivatives

033f45e

to enable use of v e tabuluations with volume junctions

Add s from p T with derivatives with v e tabulation

33bf3a0

Use the tabulations in presence of a FP source uo when computing prop…

32716a6

…erties (consistency) Use the FP uo preferentially for computing grids

Add unit tests for properties from ve computed from a tabulation gene…

e790a17

…rated from a FP object

GiudGiud added 4 commits November 11, 2025 16:12

Label the unit tests better, highlighting the mssing fromVEFromFile case

2643621

Add some missing tests for p,T tabulation and p,T->(v,e) tabulation

930fa6b

More clarifying what fluid properties object is built how in unit tests

540404c

Simplify logic for having ve interpolations

85466cc

Add more range check

GiudGiud force-pushed the PR_fp_thm_ve branch 3 times, most recently from fd862a6 to 76fa2d6 Compare November 12, 2025 00:25

GiudGiud requested a review from cpgr as a code owner November 12, 2025 23:10

GiudGiud added 8 commits November 12, 2025 17:32

Deprecate 'fp' parameter to be able to use 'fp' in GlobalParams

562577a

Fix FP output from (v,e):

476dcd0

- enable when not using (p,T)->(v,e) conversion - avoid the header to be able to use for gold files - avoid a lone comma when not outputting any extra properties - avoid conversion errors with h(p,T) call

Add a test for (v,e) output of tabulation

6f0332d

Add a dedicated test for (p,T) output

Allow fp in GlobalParams with TFP

0059cfb

std::cout -> Moose::out

7c41a40

Address test failure:

c390cac

- try to guess when newton is used for initial guess checks

Fix test failures:

56ec22e

initialize FP earlier & use the tabulation rather than the FP

Add the ability to ignore a fluid property from the tabulation file b…

502dc25

…y sepecifying the interpolated properties in the input - address porous flow failure that was relying on forwarding a property

GiudGiud force-pushed the PR_fp_thm_ve branch from fac1771 to 502dc25 Compare November 13, 2025 00:32

GiudGiud added the PR: Ready for review/merge label Nov 13, 2025

GiudGiud added 3 commits November 13, 2025 07:24

Avoid seg fault on OOB access

109ab08

Make sure to initialize (v,e) grid for passthrough case Add testing for (v,e) passthrough

Add AD (p,T) passthrough testing:

be047cb

- create a dedicated passthrough tabulation - use idg fps that have the most AD definitions created

Use the interpolation bounds from the files in (v,e) over querying th…

ed4be65

…e FP - otherwise we would compute min/max from FP calls OOB from the tabulation - then flush to min/max that could be OOB

GiudGiud force-pushed the PR_fp_thm_ve branch from c326676 to ed4be65 Compare November 13, 2025 15:31

Examine coverage report and catch most misses:

470fc75

- T/p from v e with derivatives in the passthrough _fp case - AD e and v from p T - e from p rho using a pT tabulation - AD s from v e - T from p rho using a ve tabulation

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Add support for using (v,e) interpolation tables in THM #31860

Add support for using (v,e) interpolation tables in THM #31860

GiudGiud commented Nov 6, 2025 •

edited

Loading

Uh oh!

GiudGiud commented Nov 11, 2025

Uh oh!

moosebuild commented Nov 12, 2025 •

edited

Loading

Uh oh!

moosebuild commented Nov 13, 2025 •

edited

Loading

Uh oh!

moosebuild commented Nov 13, 2025

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Add support for using (v,e) interpolation tables in THM #31860

Are you sure you want to change the base?

Add support for using (v,e) interpolation tables in THM #31860

Conversation

GiudGiud commented Nov 6, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

GiudGiud commented Nov 11, 2025

Uh oh!

moosebuild commented Nov 12, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

moosebuild commented Nov 13, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Framework coverage

Modules coverage

Fluid properties

Full coverage reports

Uh oh!

moosebuild commented Nov 13, 2025

Framework test summary

Modules test summary

Added tests

Run time changes

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

GiudGiud commented Nov 6, 2025 •

edited

Loading

moosebuild commented Nov 12, 2025 •

edited

Loading

moosebuild commented Nov 13, 2025 •

edited

Loading