Commit Graph

5878 Commits

Author SHA1 Message Date
Kai Bao
b079c0a1e5 adding flow_injecitivity executable target 2018-12-11 21:40:47 +01:00
Kai Bao
e1ede3c9d5 adding polymer molecular weight to BlackoilPhases and phaseUsageFromDeck 2018-12-11 21:31:47 +01:00
Kai Bao
93fb5a2c52 add a few fields related to injectivity to WellStateFullyImplicitBlackoil
to account for the water throughput, skin pressure and water velocity
for each perforation.
2018-12-11 21:31:47 +01:00
Atgeirr Flø Rasmussen
56cc37f255 Move throw-if-nan to after storing convergence report. 2018-12-11 13:10:22 +01:00
Tor Harald Sandve
3d866812a6
Merge pull request #1671 from atgeirr/fix-infinite-loop
Fix potential infinite loop
2018-12-07 12:41:37 +01:00
Atgeirr Flø Rasmussen
69cd2f9b84
Merge pull request #1668 from totto82/remove_logging
Remove repetitive logging info
2018-12-07 10:53:31 +01:00
Atgeirr Flø Rasmussen
cb06770b54
Merge pull request #1638 from trinemykk/addtracer
add method for getting the surfacevolume well connection rate
2018-12-07 10:38:21 +01:00
Atgeirr Flø Rasmussen
cf687c19fb
Merge pull request #1571 from kel85uk/fix_#14_aquifer_making_connections
Fixes for carter tracy aquifer model
2018-12-07 09:56:36 +01:00
Tor Harald Sandve
77869606f3 Remove repetitive logging info 2018-12-07 09:28:27 +01:00
Atgeirr Flø Rasmussen
23edd39c63 Revise logic of shutting wells vs. chopping step.
An infinite loop was possible if no wells were actually shut,
yet the step was restarted with no chopping.
2018-12-03 13:25:19 +01:00
Atgeirr Flø Rasmussen
6b896d178e Add bool return value to forceShutWellByNameIfPredictionMode().
Returns true if the well requested was actually shut.
2018-12-03 13:24:34 +01:00
kel85uk
4ccbfefe0b Factor out face area calculation to a separate function. 2018-11-29 09:43:55 +01:00
kel85uk
875fec79fc Makes the necessary interface changes to check for nullptr instead of values for the influx_coeff and aquifer initial pressure. 2018-11-29 07:56:53 +01:00
Kai Bao
575907941d removing some repeated code in phaseUsageFromDeck 2018-11-27 14:55:26 +01:00
Arne Morten Kvarving
51da91c1ad remove files moved to opm-simulators-legacy 2018-11-27 11:00:21 +01:00
Atgeirr Flø Rasmussen
b6f8128260
Merge pull request #1661 from akva2/removeBlackoilStateCorr
Remove usage of BlackoilState class
2018-11-27 10:14:06 +01:00
Tor Harald Sandve
b661791a63 Remove usage of BlackoilState class
This allows (re)moving of the following files
opm/autodiff/RateConverter.hpp
opm/autodiff/Compat.cpp
opm/autodiff/Compat.hpp
opm/core/props/BlackoilPropertiesInterface.hpp
opm/core/simulator/BlackoilState.cpp
opm/core/simulator/BlackoilState.hpp
opm/core/simulator/BlackoilStateToFluidState.hpp
opm/core/utility/initHydroCarbonState.hpp
opm/polymer/PolymerBlackoilState.cpp
opm/polymer/PolymerBlackoilState.hpp
tests/test_blackoilstate.cpp
2018-11-27 09:38:38 +01:00
Atgeirr Flø Rasmussen
7e91381cd8 Fix another parallel bug.
A WellSwitchingLogger was created in a local context (meaning not
all processes might be there), but since its destructor does
communication it must be called in a global context (guaranteeing
that all processes create it).
2018-11-26 10:21:25 +01:00
Atgeirr Flø Rasmussen
09405ce6f5 Fix bug in parallel case.
Since computeAverageFormationFactor() involves communication to reduce
values across the whole reservoir, it must be called by all processes,
also those with no wells to reopen.
2018-11-26 09:39:54 +01:00
Kai Bao
76271e3960 not shut a well because of cannot-produce-inject reason
which is a numerical reason. Removing it for now for better result.

It minght come back with more complete strategy and understanding.
2018-11-25 20:31:56 +01:00
Kai Bao
3dac7266b0 typo correction and message output adjustment for WTEST
recovering the const for updateWellStateWithTarget

there is no functional change.
2018-11-25 20:31:56 +01:00
Kai Bao
680c45b8f8 not updateWellControl for non-operable wells 2018-11-25 20:31:56 +01:00
Kai Bao
eaa3ad19f5 first version of the well testing for physical reason
the key difficulty is that we do not have reliable explicit information
to do the testing.

In this version, we try to obtain the explicit information by finishing
one converged solving.
2018-11-25 20:31:56 +01:00
Atgeirr Flø Rasmussen
48db4062b5 Count lower/upper entries simpler and faster. 2018-11-23 21:38:22 +01:00
Atgeirr Flø Rasmussen
8a038401ce Avoid in-loop allocation in convertToCRS().
This is done by counting entries first.
2018-11-23 21:38:22 +01:00
Atgeirr Flø Rasmussen
8bd3cbc1c1 Only shut down wells if under prediction-type control. 2018-11-23 12:51:13 +01:00
Atgeirr Flø Rasmussen
1979fc8f5d Avoid deep chopping by shutting down misbehaving wells. 2018-11-22 16:24:52 +01:00
Atgeirr Flø Rasmussen
422d477945 Add method BlackoilWellModel::forceShutWellByName(). 2018-11-22 16:24:17 +01:00
Atgeirr Flø Rasmussen
3bbc13a9ce Add step numbers to convergence reports, add query method. 2018-11-22 11:14:39 +01:00
Atgeirr Flø Rasmussen
9167d08839 Limit timestep when we have THP constraints present.
Using 16 days as the current limit. Could be made a user parameter.
2018-11-22 11:03:18 +01:00
Kai Bao
7d721408f2 adding canProduceInjectWithCurrentBhp to StandardWell
and use it the operability_status determination.
2018-11-22 07:05:46 +01:00
Kai Bao
f9988057af fixing comments from review of PR#1648
there should be no functional change
2018-11-20 13:56:14 +01:00
Kai Bao
8330c3a6a7 adjusting OperabilityStatus of WellInterface
hopefully, it improves the readability
2018-11-20 13:56:14 +01:00
Kai Bao
bf9b5367d9 some cleaning up, no functional changes 2018-11-20 13:56:14 +01:00
Kai Bao
90f805bf3e we need to open the cross-flow to solve the sigularity
if a well is banned from cross-flow. When it is under RATE control, its
BHP might be initialized in way causing all the drawdown in the wrong
direction. It will cause singular well equations.

here, we open the croff-flow to fix the singularity and rely on Newton
iteraton to get desired result.

possible alternative approach is to adust the BHP to avoid the situation
that all the drawdown are in the wrong direction.
2018-11-20 13:56:14 +01:00
Kai Bao
bba373c0b5 THP control tolerance changed to 1.e4
to fix some convergence problem.
2018-11-20 13:55:34 +01:00
Kai Bao
c994a99583 WCON keyword can open a well closed due to physical reason
more tests will be required to figure out more accurate specific way to
open a well closed due to physcal reason with WCON input.
2018-11-20 13:55:34 +01:00
Kai Bao
30742112c6 switching to BHP limit when THP limit is not operable
a well is specified with THP target, while it can not work under its THP
target. For this case, we will switch to its BHP limit.
2018-11-20 13:55:34 +01:00
Kai Bao
8e17d4aeba shut/close well based on physical limits 2018-11-20 13:55:34 +01:00
Kai Bao
ea42d1de9d adding function isOperable() to WellInterface
to indicate if the well is operable.
2018-11-20 13:55:34 +01:00
Kai Bao
ae3b514e0b not switching to THP/BHP limit when not operable under them 2018-11-20 13:54:45 +01:00
Kai Bao
5204bbe4ca adding function checkWellOperatability() to StandardWell
it includes BHPLimit check, THPlimit check, and also to see whether if
the well can produce/inject in the correct direction.
2018-11-20 13:54:45 +01:00
Kai Bao
dbe047a8a5 adding function operabilityCheckingUnderTHP to VFPProdProperties
To test the operability under THP limit.
2018-11-20 13:54:21 +01:00
Kai Bao
b27dc3ec2e stopping use function crossflowAllowed()
it turns on the crossflow when all the drawdown in the wrong direction,
then the well get rates in the wrong direction.

flow is not ready to handle this type of situation, then the only result
will be chopped time step.
2018-11-20 13:54:21 +01:00
Kai Bao
2e43b440e0 adding OperabilityStatus to WellInterface
to check the operability status of a well.
2018-11-20 13:54:21 +01:00
Trine S Mykkeltvedt
9fa2c09783 add method for getting the surfacevolume well connection rate
changed wellpointer from unique to shared to make it accecible from outside the wellmodel
add method for surfacevolume well connection rate
2018-11-20 08:59:50 +01:00
Atgeirr Flø Rasmussen
83ce1eb919 Ensure all residuals are checked for convergence.
Looping over all components instead of phases, to handle polymer etc.
correctly. Also slight refactoring of how component names for output
are handled.
2018-11-19 14:46:31 +01:00
Atgeirr Flø Rasmussen
b42165b560 Store convergence reports from all steps and iterations. 2018-11-19 11:48:07 +01:00
Atgeirr Flø Rasmussen
fe79a9fc07 Refactor getConvergence() to use ConvergenceReport.
Note: the communication and reduction for computing reservoir
convergence is not done by gathering ConvergenceReports, but
as before, using the convergenceReduction() method.
2018-11-19 11:48:07 +01:00
Atgeirr Flø Rasmussen
c006ea23f2
Merge pull request #1651 from totto82/speed_up_add_cell_rates
Speed up add cell rates
2018-11-19 10:49:19 +01:00
Tor Harald Sandve
9903738690 Avoid copying of SummaryConfig 2018-11-19 09:41:00 +01:00
Tor Harald Sandve
54e33a00d2 Only update intensiveQuantities for perforated cells 2018-11-19 09:41:00 +01:00
Tor Harald Sandve
3a38b9fe53 Move addCellRates to interface and add is_cell_penetrated 2018-11-19 09:41:00 +01:00
Atgeirr Flø Rasmussen
b5cb7bb829
Merge pull request #1644 from GitPaean/fixing_control_initialization
using control in DECK to initialize the control in WellState
2018-11-16 15:38:22 +01:00
Arne Morten Kvarving
f027262ec4 remove files 2018-11-16 14:53:37 +01:00
Kai Bao
7ded0907fe using control in DECK to initialize the control in WellState
otherwise, it might not be initialized if the well does not exisit in
previous well state, which will result in undefined behavoir.

it causes failure in running some realizations.
2018-11-16 13:47:43 +01:00
Arne Morten Kvarving
78158bf44a
Merge pull request #1643 from atgeirr/legacy-split-work
Avoid using legacy stuff for ebos linear solvers.
2018-11-16 13:29:24 +01:00
Atgeirr Flø Rasmussen
92a9b5fa9b Avoid using legacy stuff for ebos linear solvers. 2018-11-16 12:06:12 +01:00
Arne Morten Kvarving
fb7bdffbc7 fixed: signed/unsigned warnings 2018-11-16 08:48:53 +01:00
Kai Bao
b2be13ced4 changing std::cout to OpmLog with VFPProdProperties
for better output of information.
2018-11-15 11:46:22 +01:00
Kai Bao
76a3f2a1f5 adding function updateWellStateWithTHPTargetIPR()
this function can update the well state related based on the inflow
perfomance relationship and THP target.
2018-11-15 11:46:22 +01:00
Kai Bao
eeae6aa4fc adding two THP control related fucntions to WellInterface
getTHPConstraint() and getTHPControlIndex();
2018-11-15 11:46:22 +01:00
Kai Bao
01efbc639c adding function underPredictionMode() to WellInterface
to indicate whether a well is under prediction mode.
2018-11-15 11:46:22 +01:00
Kai Bao
9bfa39224f re-organize the interface of function findIntersectionForBhp()
hopefully, make it easier to use.

and also, there is no mistake introduced.
2018-11-15 11:46:22 +01:00
Kai Bao
8445c802c0 adding function calculateBhpWithTHPTarget to VFPProdProperties
it calculate bhp value based on THP target/limit, VFP curves and
inflow-performance relationship
2018-11-15 11:46:22 +01:00
Kai Bao
9b5e25ae0f function to calculate bhp from thp through intersection
of VFP curves and IPR relationship.

Some small adjustment of the interface will be done later.
2018-11-15 11:46:22 +01:00
Kai Bao
32b8e79eae adding function updateIPR() to StandardWell 2018-11-15 11:46:22 +01:00
Atgeirr Flø Rasmussen
0761850e7a
Merge pull request #1637 from OPM/fix-gather-convergemcereport-for-openmpi
Make MPI calls in gatherConvergenceReport with void*.
2018-11-15 11:04:35 +01:00
Markus Blatt
8cafb0e92a Make MPI calls in gatherConvergenceReport with void* instead of const void*
Fixes compilation of PR #1612 with open-mpi 1.6.5.
Compiler complained about not being able to transform const void* to void*
as needed by MPI.
2018-11-15 09:30:51 +00:00
Atgeirr Flø Rasmussen
5e51d6982b Add missing includes. 2018-11-14 21:09:40 +01:00
Atgeirr Flø Rasmussen
258c019989
Merge pull request #1617 from andlaus/sparse_matrix_abstraction
Sparse matrix abstraction
2018-11-14 16:14:00 +01:00
Andreas Lauser
e110fac52a
Merge pull request #1633 from akva2/rogue_and_unwanted
Remove some rogue includes and unwanted opm/autodiff/GridHelpers.hpp usage
2018-11-14 16:00:30 +01:00
Atgeirr Flø Rasmussen
db8c00db00
Merge pull request #1632 from akva2/use_istl_ebos
Use ebos ISTL class
2018-11-14 15:44:26 +01:00
Arne Morten Kvarving
fe8adb51d8 remove rogue includes 2018-11-14 15:37:44 +01:00
Arne Morten Kvarving
023924d057 changed: use UgGridHelpers directly
GridHelpers.hpp will be moved
2018-11-14 15:37:44 +01:00
Andreas Lauser
f58f5d3c7a MPIUtilities.hpp: include required header file 2018-11-14 14:36:11 +01:00
Andreas Lauser
d5089076d9 ISTLSolver: add now-necessary include 2018-11-14 14:35:52 +01:00
Andreas Lauser
f9104ca3d7 adapt to the eWoms interface for abstracting sparse matrices 2018-11-14 14:35:52 +01:00
Andreas Lauser
46641d5ace move the MatrixBlock class and assorted code to its own header
this avoids recompiles if something changes in ewoms because
ISTLSolver.hpp included ISTLSolverEbos.hpp . also, it is better style.
2018-11-14 14:35:52 +01:00
Kai Bao
bf967e50a7 adding hasTable() method to the VFP properties
and give a clear message about non-accessible table ID in the function
isVFPActive() in WellInterface.
2018-11-14 14:29:28 +01:00
Kai Bao
cf2eb1c336 remove a useless #if HAVE_OPENMP in BlackoilWellModel 2018-11-14 12:47:49 +01:00
Kai Bao
5a5c1bfcd3 addressing review comments from PR 1622 2018-11-14 12:45:41 +01:00
Arne Morten Kvarving
911a89fc4c fixed: use the ebos ISTL class
to avoid legacy code in tests that will sit in this repo
2018-11-14 12:27:39 +01:00
Kai Bao
afcdbcf202 detecting the NAN and INF values during VFP calculation
and give warning messages when they happen, since they can be the
culprits of the termination of simulation later.

For ADB related, I did not find a good way to do the detection, so there
is no warning message given.
2018-11-14 11:35:44 +01:00
Kai Bao
32b00b61f8 updating the thp value if VFP table is valid
through function isVFPActive().
2018-11-14 11:27:00 +01:00
Kai Bao
0d1a4b2d13 adding function isVFPActive() for WellInterface
Even the well does not have a THP target/constraint, but if it is
specified with a valid VFP table, we are supposed to update the thp
value for output purposes.
2018-11-14 11:27:00 +01:00
Atgeirr Flø Rasmussen
49a7773b30
Merge pull request #1630 from akva2/split_vfp
Split VFP classes in ebos and legacy
2018-11-14 11:04:05 +01:00
Arne Morten Kvarving
c5ae3adbbf changed: split VFP classes in base (ebos) and legacy
to get rid of eigen usage in ebos based classes
2018-11-14 09:54:33 +01:00
Atgeirr Flø Rasmussen
e1e39b4bd9
Merge pull request #1626 from GitPaean/fixing_different_number_of_wells_due_to_wtest
fixing different well number between Wells and Well_containers
2018-11-14 09:37:09 +01:00
Atgeirr Flø Rasmussen
54cc52dcaf
Merge pull request #1629 from akva2/remove_rogue_includes
remove two rogue includes
2018-11-14 09:06:13 +01:00
Arne Morten Kvarving
e8c2029166 remove two rogue includes
motivation: getting rid of BlackoilPropsAdFromDeck in ebos code path
2018-11-14 08:35:21 +01:00
Atgeirr Flø Rasmussen
8d2bc0e449 Add missing include for std::vector. 2018-11-13 20:26:44 +01:00
Kai Bao
8f56634b69 fixing different well number between Wells and Well_containers
when the solveWellEq did not get converged.
2018-11-13 20:01:38 +01:00
Atgeirr Flø Rasmussen
58642254e0
Merge pull request #1624 from akva2/split_mpi_func
split mpi function from newtoniterationutilities
2018-11-13 19:21:56 +01:00
Arne Morten Kvarving
e0a9e13a97 split mpi function from newtoniterationutilities 2018-11-13 15:26:44 +01:00
Kai Bao
2ed2c20f12
Merge pull request #1623 from atgeirr/use-convergencereport-gather
Use gatherConvergenceReport() in BlackoilWellModel.
2018-11-13 15:16:34 +01:00
Atgeirr Flø Rasmussen
d858f2901a
Merge pull request #1614 from totto82/well_ppp
Output of well potential when asked for
2018-11-13 14:29:58 +01:00
Atgeirr Flø Rasmussen
446dbdf86d Use gatherConvergenceReport().
This avoids multiple all-reduce blocks, simplifying the code and
enabling us to return the report from getWellConvergence().
2018-11-13 14:02:55 +01:00
Atgeirr Flø Rasmussen
6e7cc756de Move implementation of gatherConvergenceReport() to cpp file.
No templates involved, no reason to keep it in header. This also makes
building more robust by only invoking HAVE_MPI in the cpp file, after
including config.h.
2018-11-13 14:01:01 +01:00
Atgeirr Flø Rasmussen
566f6813dd Created gatherConvergenceReport() function and test. 2018-11-13 10:30:32 +01:00