Closed dkachuma closed 3 weeks ago
All modified and coverable lines are covered by tests :white_check_mark:
Project coverage is 53.74%. Comparing base (
010fc34
) to head (b952b80
).
:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.
The fix to the derivative in EzrokhiBrineDensity
has changed the response in class09_pb3_smoke_3d
. The change in pressure is slightly different and quite expected. I would say a rebaseline is OK.
Error: /Problem/domain/MeshBodies/mesh1/meshLevels/Level0/ElementRegions/elementRegionsGroup/reservoir/elementSubRegions/1_hexahedra/deltaPressure
Arrays of types float64 and float64 have 125 values of which 8 fail both the relative and absolute tests.
Max absolute difference is at index (107,): value = 15333.86215468496, base_value = 15333.859636742622
Max relative difference is at index (56,): value = 43.7011882327497, base_value = 43.70022329688072
This PR breaks up the
testMultiFluid
unit test. This test does a full fluid update call and tests the obtained values of fluid properties against expected values. It also does numerical derivative checks at selected fluid conditions (pressure, temperature and composition).This test was getting larger and taking longer to build especially with CUDA. Instead of having all the fluid types in one test, this is now broken up into different tests for the different multiphase fluid types. The common part is factored out into the
MultiFluidTest
test fixture.The numerical derivative tests in this case were not active because a very large absolute tolerance was used (see #2876). These are now activated for most of the fluid types. As a consequence, a fix on the temperature derivative was required in
EzrokhiBrineDensity.hpp
. The tests are not activated for theCompositionalMultiphaseFluidPVTPackage
because this already uses finite differences to calculate derivatives and for some of the test compositions, it is actually unable to give reliable values (the package itself needs updating).The inclusion of
testCompFlowUtils.hpp
inconstitutiveTestHelpers.hpp
was pulling in too many include files into the compilation of thee tests. The methods required for the multiphase fluid tests have now been moved to a smaller includetestFlowUtils.hpp
.This will likely require a rebaseline due to the change in the Ezrokhi density derivative.