Added "dummy" water phase to compositional fluid system #4410
Conversation
|
Thanks for the work. I will look into the PRs soon. |
|
jenkins build this opm-simulators=5851 please |
|
After discussion with @totto82 we agreed to have both two-phase and three-phase versions of the compositional fluid system. So I will make this and the opm-simulators PR drafts again and to make the necessary changes. |
I agree and it is appreciated. And depending the testing example you have, you can create the flowexp_comp executable for that specific number of component only first. Still thinking how to avoid too many exectubales to compile but it can be for later optimization. |
svenn-t
force-pushed
the
dummy_water_compositional
branch
from
1316ad5 to
fb16da4
Compare
|
I have updated this PR and OPM/opm-simulators#5851 to have both two- and three-phase versions of the compositional simulator. I did make new executables for two-phase versions of the simulators, and which version that will be run is dependent on if |
svenn-t
force-pushed
the
dummy_water_compositional
branch
from
fb16da4 to
3bcaf72
Compare
|
jenkins build this opm-simulators=5851 please |
svenn-t
force-pushed
the
dummy_water_compositional
branch
from
e5dc750 to
086d0ac
Compare
|
jenkins build this opm-simulators=5851 please |
Water/oil/gas phase indices are permutated in compositional fluid systems.
svenn-t
force-pushed
the
dummy_water_compositional
branch
from
086d0ac to
25e3951
Compare
|
jenkins build this opm-simulators=5851 please |
|
Same here, well done PR. I will also invite @akva2 to have a look. |
| * \brief Set the pressure-volume-saturation (PVT) relations for the water phase. | ||
| */ | ||
| static void setWaterPvt(std::shared_ptr<WaterPvt> pvtObj) | ||
| { waterPvt_ = pvtObj; } |
There was a problem hiding this comment.
std::move to avoid reference work.
There was a problem hiding this comment.
Done, thanks for the comment!
tests/material/test_fluidsystems.cpp
Outdated
| FluidSystem::init(); | ||
| using WaterPvt = typename FluidSystem::WaterPvt; | ||
| std::shared_ptr<WaterPvt> waterPvt; | ||
| FluidSystem::setWaterPvt(std::move(waterPvt)); |
There was a problem hiding this comment.
Done, thanks for the comment!
| @@ -861,7 +863,7 @@ class PTFlash | |||
|
|
|||
| // getting the secondary Jocobian matrix | |||
| constexpr size_t num_equations = numMisciblePhases * numMiscibleComponents + 1; | |||
| constexpr size_t secondary_num_pv = numComponents + 1; // pressure, z for all the components | |||
| constexpr size_t secondary_num_pv = numComponents + (waterEnabled ? 2 : 1); // pressure, z for all the components | |||
There was a problem hiding this comment.
I can see it wrong, but water phase does not affect the flash calculation and I do not see that we are doing anything related to the derivative No. numComponents + 1 in the following code.
There was a problem hiding this comment.
@svenn-t , please fix this and we will get ready to get the PR in. The deviation can be addressed in a separate PR.
There was a problem hiding this comment.
I'm not so familiar with the what is happening in updateDerivativesTwoPhase_(), but I just added 1 to the second_num_pv variable if waterEnabled is true to include the additional primary variable we have in the three-phase case (which is Sw). Is that wrong or are you saying that there are no derivatives wrt to Sw in any case so adding 1 is unnecessary?
There was a problem hiding this comment.
It is the difficult part when we developed this, but we will understand better with time.
I believe that we do not need it since Sw is not in the flash calculation at all. And also, from the code, I do not see we are touching the derivative index No. numComponents + 1 at all. And it did not make any difference if we remove it.
There was a problem hiding this comment.
I think I understand, thanks for the explanation. I have reverted the changes back to the original code.
|
jenkins build this opm-simulators=5851 please |
|
With the approval and jenkins approves, I am getting the PR in now. Thanks for the contribution. |
The compositional fluid system is expanded to three phases where the water phase is a "dummy"/immiscible phase. Water phase properties are given through PVT functions the same way as in the black oil fluid system.