feat: feature/byer3/thermal well #3156

tjb-ltk · 2024-06-03T14:36:16Z

Add thermal well formulation for modeling convection in compositional flow models

…fix error message for negative mass rate

…or thermal. Enabled kernels/ computes to be thermal aware via specialization or constexpr on template parameter. Most well terms have dt ders. Code tested on 3 standard isothermal models to smoke out offset errors. Lots of asserts left in code to ensure general storage arrays are correct eg.. (dProp[PresIndex]==dprop_dp).. these along with dprop_dp, dprop_dcomp etc arrays will be removed after most testing. Next step to add energy balance to well element flows and perf inflows

…coupling 3) residualnomr calcs, 4) some isothermal , 5) refactoring 6) next focus solution updating/scaling

…al flux index bug 3) 2 isothermal models tested with refactored code 4) thermal case crashes at 2nd linear solve well internal energy derivative = 0

…inear solve well methods 3) next step verify simple test case results

…ng 3) removed debug std::couts 4) next steps a) address constant temp BC b) singlephase thermo dev c) deriv checker

…l derivative checker and treating warnings as errors

…s for shutin well 3) call well initialization code for wells that come online at t> 0

codecov · 2024-06-03T16:06:09Z

Codecov Report

Attention: Patch coverage is 72.91271% with 571 lines in your changes missing coverage. Please review.

Project coverage is 56.88%. Comparing base (b4b1a3a) to head (720901c).

Files	Patch %	Lines
...ells/ThermalCompositionalMultiphaseWellKernels.hpp	48.29%	152 Missing ⚠️
...rs/multiphysics/CoupledReservoirAndWellKernels.hpp	0.00%	124 Missing ⚠️
...Solvers/fluidFlow/wells/PerforationFluxKernels.hpp	70.63%	74 Missing ⚠️
...rc/coreComponents/common/KernelLaunchSelectors.hpp	25.00%	60 Missing ⚠️
...dFlow/wells/CompositionalMultiphaseWellKernels.hpp	85.80%	43 Missing ⚠️
...ies/ThermalCompositionalMultiphaseReservoirFVM.hpp	0.00%	34 Missing ⚠️
.../physicsSolvers/fluidFlow/wells/WellSolverBase.cpp	55.35%	25 Missing ⚠️
...dFlow/wells/CompositionalMultiphaseWellKernels.cpp	76.47%	16 Missing ⚠️
...ysics/CompositionalMultiphaseReservoirAndWells.cpp	17.64%	14 Missing ⚠️
...ers/fluidFlow/CompositionalMultiphaseHybridFVM.cpp	0.00%	9 Missing ⚠️
... and 7 more

Additional details and impacted files

@@             Coverage Diff             @@
##           develop    #3156      +/-   ##
===========================================
+ Coverage    55.99%   56.88%   +0.88%     
===========================================
  Files         1053     1063      +10     
  Lines        89157    90755    +1598     
===========================================
+ Hits         49921    51623    +1702     
+ Misses       39236    39132     -104

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

.gitmodules

src/coreComponents/common/KernelLaunchSelectors.hpp

paveltomin · 2024-09-18T18:05:44Z

src/coreComponents/constitutive/fluid/singlefluid/SingleFluidBase.hpp

+ * @detail if isThermal is true, the constitutive model compute the enthalpy and internal energy of the phase.
+ *         This can be used to check the compatibility of the constitutive model with the solver
+ */
+  virtual bool isThermal() const { return false; }


do you need to add one with true in ThermalCompressibleSinglePhaseFluid ?

@tjb-ltk Is this resolved?

src/coreComponents/physicsSolvers/fluidFlow/CompositionalMultiphaseBase.cpp

src/coreComponents/common/KernelLaunchSelectors.hpp

victorapm

The linear solver changes look good. Just need to update the header comment for the new MGR strategy

.../linearAlgebra/interfaces/hypre/mgrStrategies/ThermalCompositionalMultiphaseReservoirFVM.hpp

src/coreComponents/physicsSolvers/fluidFlow/IsothermalCompositionalMultiphaseBaseKernels.hpp

paveltomin · 2024-09-18T18:16:55Z

src/coreComponents/physicsSolvers/fluidFlow/wells/CompositionalMultiphaseWell.cpp

-    currentBHP = pres[iwelemRef] + totalMassDens[iwelemRef] * diffGravCoef;
-    dCurrentBHP_dPres = 1 + dTotalMassDens_dPres[iwelemRef] * diffGravCoef;
-    for( integer ic = 0; ic < numComp; ++ic )
+    //integer constexpr NUM_COMP = NC();


remove commented

paveltomin · 2024-09-18T18:19:58Z

src/coreComponents/physicsSolvers/fluidFlow/wells/CompositionalMultiphaseWell.cpp

+        isothermalCompositionalMultiphaseBaseKernels::
+          KernelLaunchSelectorCompTherm< PressureRelationKernel >( numFluidComponents(), isThermal,
+                                                                   //isothermalCompositionalMultiphaseBaseKernels::KernelLaunchSelector1<
+                                                                   // PressureRelationKernel >( numFluidComponents(),


remove commented

src/coreComponents/physicsSolvers/fluidFlow/wells/CompositionalMultiphaseWellFields.hpp

src/coreComponents/common/KernelLaunchSelectors.hpp

src/coreComponents/dataRepository/Group.hpp

CusiniM

Let's start with this.

src/coreComponents/unitTests/fluidFlowTests/testCompFlowUtils.hpp

src/coreComponents/unitTests/wellsTests/testThermalReservoirCompositionalMultiphaseSSWells.cpp

src/coreComponents/physicsSolvers/fluidFlow/wells/PerforationFluxKernels.hpp

src/coreComponents/physicsSolvers/fluidFlow/wells/CompositionalMultiphaseWellFields.hpp

CusiniM · 2024-09-18T19:02:47Z

src/coreComponents/physicsSolvers/fluidFlow/IsothermalCompositionalMultiphaseBaseKernels.hpp

@@ -2445,6 +2454,54 @@ void KernelLaunchSelector2( integer const numComp, integer const numPhase, ARGS
  }
 }

+template< typename KERNELWRAPPER, typename ... ARGS >
+void KernelLaunchSelector_NC_NP_THERM( integer const numComp, integer const numPhase, integer const isThermal, ARGS && ... args )


why did you have to add this? didn't we already have multiphase compostional thermal kernels?

is this ever used? I can't find it...

It seems that we have too many of these dispatches. We should only have 1 dispatch for compNumber, phaseNumber and thermal and use it everywhere.

will delete... yes we have several dispatches..In the spirit of " too many" 1) derivative offset structs... since prop derivatives have thermal slot at index 1. traits classes templated on thermal are needed in isothermal runs... where should we put these structs... . I think these and dispatches should all be in some common location so the developer knows what is available... 2) further single/multiphase should be combined, and 3) specializations should only be for different variable sets or implicitness and phase comp partitioning arrays introduced ....

why not using the ones you created?

In. KernelLaunchSelector2 there is a comment // Ideally this would be inside the dispatch, but it breaks on Summit with GCC 9.1.0 and CUDA 11.0.3. I thought it would be better to make that change in different PR.

CusiniM · 2024-09-18T19:10:35Z

src/coreComponents/common/KernelLaunchSelectors.hpp

if this file sits in common it should be a lot more generic IMO.

I could move to fluidflow...

moved to physicsSolvers

src/coreComponents/dataRepository/Group.hpp

src/coreComponents/physicsSolvers/fluidFlow/IsothermalCompositionalMultiphaseBaseKernels.hpp

src/coreComponents/physicsSolvers/fluidFlow/wells/SinglePhaseWell.cpp

src/coreComponents/physicsSolvers/fluidFlow/IsothermalCompositionalMultiphaseBaseKernels.hpp

paveltomin · 2024-10-07T17:06:50Z

@CusiniM @rrsettgast can we please merge it? bunch of useful changes for wells here

CusiniM · 2024-10-09T00:09:38Z

@CusiniM @rrsettgast can we please merge it? bunch of useful changes for wells here

I had left some comments and if they were addressed then I am fine with merging it. However, some things in this PR are done in a way that it's quite different from the rest of the code so I wanted someone else to have a close look at it too.

CusiniM

I am not sure I would have put those dispatch functionalities in common/internal. Otherwise this PR looks good to me now

rrsettgast · 2024-10-23T22:46:39Z

src/coreComponents/constitutive/fluid/singlefluid/SingleFluidBase.hpp

+ * @detail if isThermal is true, the constitutive model compute the enthalpy and internal energy of the phase.
+ *         This can be used to check the compatibility of the constitutive model with the solver
+ */
+  virtual bool isThermal() const { return false; }


@tjb-ltk Is this resolved?

src/coreComponents/mesh/WellElementSubRegion.cpp

rrsettgast · 2024-10-23T22:51:26Z

src/coreComponents/physicsSolvers/fluidFlow/ThermalCompositionalMultiphaseBaseKernels.hpp

@@ -599,7 +601,7 @@ class ScalingForSystemSolutionKernel : public isothermalCompositionalMultiphaseB
    {
      // compute the change in temperature
      real64 const temp = m_temperature[ei];
-      real64 const absTempChange = LvArray::math::abs( m_localSolution[stack.localRow + m_numComp + 1] );
+      real64 const absTempChange = LvArray::math::abs( m_localSolution[stack.localRow + m_temperatureOffset] );


this change looks strange to me...is this OK?

Agree. the change was needed so that wells could use the scaling kernel implemented for the reservoir. The location of the temperature variable is different for well and reservoir, hence the offset variable. It's ok but fragile. Another PR could see this being replaced with a traits class, but defining/changing lineups with traits classes would probably need more vetting beyond this kernel.

rrsettgast · 2024-10-23T22:54:39Z

src/coreComponents/physicsSolvers/multiphysics/CoupledReservoirAndWellKernels.hpp

@@ -0,0 +1,592 @@
+/*


The changes here don't seem well related...or thermal related. I missed it? Can you explain?

Originally the well-res coupling derivatives were not in kernel classes but assembled in CompositionalMultiphaseReservoirAndWells.cpp. With the addition of thermal ,kernels were added for iso + thermal and are in CoupledReservoirAndWellKernels.hpp

src/coreComponents/unitTests/fluidFlowTests/testCompFlowUtils.hpp

tjb-ltk and others added 24 commits December 20, 2023 14:48

Add support for mass constraits (inj wells only)

3c5801c

Mass constraint test cases - fixed value & table input

de5b3da

add proper isZero check

993296d

Merge branch 'develop' into feature/byer3/mass_inj_const

7bd1201

Throw error if massinj constraint and useSurfaceConditions set to 0, …

825d27d

…fix error message for negative mass rate

run make uncrustify_style

c94eed6

fix for UseSurfaceConditions set to 1

701bce6

Thermal well devs

6858b81

Fix rel mode compiler errors ++

8fd8f40

Load well prop ders in containers similar to reservoir

384e29e

thermal devs- 1) energy eqn for perf in/out flows, 2) reservoir well …

aee32bc

…coupling 3) residualnomr calcs, 4) some isothermal , 5) refactoring 6) next focus solution updating/scaling

1) fix isthermal well field runtime error 2) fix refactored isotherm…

6323c13

…al flux index bug 3) 2 isothermal models tested with refactored code 4) thermal case crashes at 2nd linear solve well internal energy derivative = 0

thermal well devs - 1) derivative bugfixes 2) added thermal to post l…

34a2a01

…inear solve well methods 3) next step verify simple test case results

thermal devs - 1) fixes for > 1 well seg 2) added soln updating/scali…

828b009

…ng 3) removed debug std::couts 4) next steps a) address constant temp BC b) singlephase thermo dev c) deriv checker

1) Added simple temperature constraint 2) Fixes from running numerica…

8b10201

…l derivative checker and treating warnings as errors

missing merge in previous commit

25e3503

more merge fixes

aa800e3

1) add iterative solver config for thermal res+well 2) skip well calc…

c9138c5

…s for shutin well 3) call well initialization code for wells that come online at t> 0

merge dev 20240526

5d95d0f

merge with dev

b4983b0

code cleanup

d297111

Use phase fraction not phase vol fraction ..and ders

0ec258e

fix for spe11b

a489169

tjb-ltk requested review from paveltomin and CusiniM June 3, 2024 14:36

fix pedantry

420855d

tjb-ltk added 2 commits June 5, 2024 08:48

1) add Aww numerical derivative unit tests, 2) thermal index fix

32ae6b9

more pedantics

bea94d5

tjb-ltk and others added 3 commits August 12, 2024 08:54

1) dont allocate temperature_n for isothermal run, 2) uncrustify style

720901c

remove debug std::cout

fdeea3a

remove std::cout

4cf2106

tjb-ltk requested a review from CusiniM September 17, 2024 20:27

CusiniM reviewed Sep 18, 2024

View reviewed changes

.gitmodules Outdated Show resolved Hide resolved

CusiniM reviewed Sep 18, 2024

View reviewed changes

src/coreComponents/common/KernelLaunchSelectors.hpp Outdated Show resolved Hide resolved

code cleanup

35cf15b

paveltomin reviewed Sep 18, 2024

View reviewed changes

src/coreComponents/physicsSolvers/fluidFlow/CompositionalMultiphaseBase.cpp Outdated Show resolved Hide resolved

CusiniM reviewed Sep 18, 2024

View reviewed changes

src/coreComponents/common/KernelLaunchSelectors.hpp Outdated Show resolved Hide resolved

victorapm approved these changes Sep 18, 2024

View reviewed changes

.../linearAlgebra/interfaces/hypre/mgrStrategies/ThermalCompositionalMultiphaseReservoirFVM.hpp Show resolved Hide resolved

paveltomin reviewed Sep 18, 2024

View reviewed changes

src/coreComponents/physicsSolvers/fluidFlow/IsothermalCompositionalMultiphaseBaseKernels.hpp Show resolved Hide resolved

paveltomin reviewed Sep 18, 2024

View reviewed changes

src/coreComponents/physicsSolvers/fluidFlow/wells/CompositionalMultiphaseWellFields.hpp Outdated Show resolved Hide resolved

paveltomin approved these changes Sep 18, 2024

View reviewed changes

CusiniM reviewed Sep 18, 2024

View reviewed changes

src/coreComponents/common/KernelLaunchSelectors.hpp Outdated Show resolved Hide resolved

CusiniM reviewed Sep 18, 2024

View reviewed changes

src/coreComponents/dataRepository/Group.hpp Outdated Show resolved Hide resolved

CusiniM reviewed Sep 18, 2024

View reviewed changes

src/coreComponents/physicsSolvers/fluidFlow/wells/SinglePhaseWell.cpp Show resolved Hide resolved

src/coreComponents/physicsSolvers/fluidFlow/IsothermalCompositionalMultiphaseBaseKernels.hpp Show resolved Hide resolved

Code cleanup

f1f633b

CusiniM approved these changes Oct 9, 2024

View reviewed changes

move KernalLaunchSelectors from common to physicsSolvers

e659c7c

tjb-ltk added flag: requires rebaseline Requires rebaseline branch in integratedTests and removed flag: ready for review labels Oct 10, 2024

rrsettgast approved these changes Oct 23, 2024

View reviewed changes

tjb-ltk and others added 2 commits October 24, 2024 09:11

PR requested changes

3162266

dev merge

988ae2b

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

feat: feature/byer3/thermal well #3156

feat: feature/byer3/thermal well #3156

tjb-ltk commented Jun 3, 2024 •

edited

Loading

codecov bot commented Jun 3, 2024 •

edited

Loading

paveltomin Sep 18, 2024

rrsettgast Oct 23, 2024

victorapm left a comment

paveltomin Sep 18, 2024

paveltomin Sep 18, 2024

CusiniM left a comment

CusiniM Sep 18, 2024

CusiniM Sep 18, 2024

CusiniM Sep 18, 2024

tjb-ltk Sep 18, 2024

CusiniM Oct 9, 2024

tjb-ltk Oct 9, 2024

CusiniM Sep 18, 2024

tjb-ltk Oct 9, 2024

tjb-ltk Oct 24, 2024

paveltomin commented Oct 7, 2024

CusiniM commented Oct 9, 2024

CusiniM left a comment

rrsettgast Oct 23, 2024

rrsettgast Oct 23, 2024

tjb-ltk Oct 24, 2024

rrsettgast Oct 23, 2024

tjb-ltk Oct 24, 2024

feat: feature/byer3/thermal well #3156

Are you sure you want to change the base?

feat: feature/byer3/thermal well #3156

Conversation

tjb-ltk commented Jun 3, 2024 • edited Loading

codecov bot commented Jun 3, 2024 • edited Loading

Codecov Report

Choose a reason for hiding this comment

Choose a reason for hiding this comment

victorapm left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

CusiniM left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

paveltomin commented Oct 7, 2024

CusiniM commented Oct 9, 2024

CusiniM left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

tjb-ltk commented Jun 3, 2024 •

edited

Loading

codecov bot commented Jun 3, 2024 •

edited

Loading