Commit Graph

3309 Commits

Author SHA1 Message Date
Markus Blatt
f913baa66e Make use of template argument deduction for createAMGPreconditionerPointer 2017-06-27 10:00:31 +00:00
Markus Blatt
59ca0b4424 Fix the PR that said it activated AMG.
Actually, it did not as it did set the define eith wrong
or the wrong locations. This commit fixes this and finally
makes AMG available.
2017-06-27 10:00:31 +00:00
Atgeirr Flø Rasmussen
bf0937fae1 Merge pull request #1220 from totto82/impl_polymer_simpleWellModel
Add flow_ebos_polymer and flow_ebos_solvent
2017-06-27 10:51:42 +02:00
Atgeirr Flø Rasmussen
730ab5526c Merge pull request #1233 from nairr/term_output_mod
Multiple problem logging for linear solver failure
2017-06-26 13:58:21 +02:00
Rohith Nair
7500d3eb0c edit 2017-06-26 12:46:09 +02:00
Rohith Nair
84ca3cd9b9 Removes multiple problem logging for linear solver convergence failure 2017-06-26 12:21:44 +02:00
Tor Harald Sandve
1c05402209 Fix 2p case after rebase 2017-06-26 08:40:30 +02:00
Tor Harald Sandve
1bc2550541 Fix polymer in well model for producers. 2017-06-26 08:03:54 +02:00
Tor Harald Sandve
e9a1aa2a83 Clean up after rebase 2017-06-26 08:03:54 +02:00
Tor Harald Sandve
8088347c96 Move adjustment of water mobility caused by polymer to getMobility() 2017-06-26 08:03:54 +02:00
Tor Harald Sandve
17ada607eb Assume polymer and water is fully mixed in the well 2017-06-26 08:03:54 +02:00
Tor Harald Sandve
0068c175a7 Add polymer option to flow_ebos
No extra equation is added for polymer in the well equation.

Seperate executables are added for polymer: flow_ebos_polymer
and solvent: flow_ebos_solvent

Tested and verified on the test cases in polymer_test_suite

This PR should not effect the performance and results of the blackoil
simulator
2017-06-26 08:03:54 +02:00
Atgeirr Flø Rasmussen
2ceca48f04 Merge pull request #1206 from GitPaean/switching_well_primary_variables
Not switching the order of well primary variables and well equations
2017-06-22 19:37:40 +02:00
Robert Kloefkorn
2f13de8592 [cleanup][Ebos::getConvergence] removed unnecessary creation of vectors
for temporary quantities.
2017-06-22 14:37:43 +02:00
Kai Bao
e0e1fc65e2 removing ebosCompToFlowPhaseIdx from StandardWellsDense 2017-06-21 11:44:43 +02:00
Kai Bao
aaa66e0982 not switching the order of well equations. 2017-06-21 11:44:43 +02:00
Kai Bao
70e193696d removing the order switiching of the well primary variables.
Order swtiching of the well primary variables make the code more
complicated, and harder to understand and extend.
2017-06-21 11:44:43 +02:00
Atgeirr Flø Rasmussen
757320c57d Merge pull request #1224 from blattms/only-created-step_timings.txt-if-output
Only create step_timing.txt file if it gets written to.
2017-06-21 09:01:33 +02:00
Atgeirr Flø Rasmussen
f48acb8807 Merge pull request #1226 from blattms/use-output-to-skip-debug-or-all-files
Use output to skip debug or all files
2017-06-21 09:01:17 +02:00
Markus Blatt
a6175942fd Make lambda C++11 compliant 2017-06-20 15:49:13 +02:00
Markus Blatt
e6c2ab47b3 Equalize if statements governing step_timing.txt output 2017-06-20 15:46:04 +02:00
Markus Blatt
75561a0da7 Some cleanup concerning spaces, const, simplified logic. 2017-06-20 15:27:05 +02:00
Markus Blatt
e5b8cddfa8 Reactivate no_debug_log to indicate whether .CASE.DEBUG should be written
This switch only has an effect if output is not equal to none.
2017-06-20 13:36:22 +02:00
Markus Blatt
d7c2d58853 write init stuff only if output=all or output=true. 2017-06-20 13:21:26 +02:00
Markus Blatt
58948521bb Do not activate output writer for when output=log 2017-06-20 13:16:50 +02:00
Markus Blatt
1f8b747299 Removed leftover getDefault("output", ...) where bool was expected. 2017-06-20 13:11:16 +02:00
Markus Blatt
9c6e9b3715 Use option output to activate or (partly) deactivate output to files.
Possible values are none, log, and all. The first does not do any logging
to files. The second does log to files but does not create and log to
the DEBUG file. The latter uses all possible files.
2017-06-20 11:57:47 +02:00
Joakim Hove
eafadaa555 Merge pull request #1222 from joakim-hove/time-map-api
Changed to TimeMap api - using std::time_t
2017-06-20 10:29:29 +02:00
Markus Blatt
588111d135 Allow a flow_ebos run without creating a debug or prt file.
Adds two switches no_prt_log, and no_debug_log that deactivate
writing to PRT and DEBUG file.
One can now run flow_ebos without creating any output by
passing "output=false no_prt_log=true no_debug_log=true"
on the command line.
2017-06-20 08:33:47 +02:00
Markus Blatt
e0a7ced256 Only create step_timing.txt file if it gets written to.
If output=false is requested we created an empty file.
Now that file will only be created if output was requested
and it will only be opened on one rank in a parallel run.
2017-06-19 17:51:26 +02:00
Joakim Hove
66931cb030 Changed to TimeMap api - using std::time_t 2017-06-19 14:03:14 +02:00
Andreas Lauser
dc9ad10f87 flow_ebos: do not use BlackoilPropsAdFromDeck anymore
the only thing that was used of this class was the phase usage object,
but the phase usage object can be accessed via much leaner interfaces.

The old BlackoilPropsFromDeck (without "Ad") is still required to
compute the initial condition, but the init code should be refactored
soon anyway.
2017-06-16 15:13:47 +02:00
Atgeirr Flø Rasmussen
56cef57c8d Merge pull request #1218 from blattms/faster-ilu-old-order
This is the faster ilu pull request but with the traversal of the columns of upper in the original ordering
2017-06-15 14:23:17 +02:00
Atgeirr Flø Rasmussen
dd453d3e50 Merge pull request #1211 from andlaus/improve_cnv
Improve stability by changing the convergence check
2017-06-15 10:39:13 +02:00
Andreas Lauser
e3d0bbc144 specify the number of iterations where we use a "strict" Newton convergence criterion using a separate parameter
that parameter is called "max_strict_iter". This increases the
flexibility of this slightly and it avoids screwing up the default
value for the "max_iter" parameter in the future. The credits for this
patch go to [at]atgeirr for proposing it.
2017-06-14 14:22:22 +02:00
Andreas Lauser
0304183361 Newton: fail earlier
i.e. give up after 10 instead of 15 Newton iterations. since this now
also causes the CNV convergence criterion to be ignored after 8
instead of after 13 Newton iterations, and because CNV stagnation
seems to be the cause for the vast majority of the failures, it does
not lead to worse results or more time step chopping on Norne but
causes fewer wasted iterations.
2017-06-14 14:22:22 +02:00
Andreas Lauser
e58fed9956 ignore the CNV convergence criterion for the last few newton iterations
the cell based residual can cause problems for cells with
almost-singular shapes and/or small pore volume.
2017-06-14 14:22:22 +02:00
Atgeirr Flø Rasmussen
699b0678a0 Remove unused simulator program.
This stand-alone simulator is no longer needed since
flow_polymer supports two-phase runs.
2017-06-14 14:15:13 +02:00
Markus Blatt
c2129185ae Store columns of ParallelOverlappingILU0::upper in descending order
This is the order used by ILU in dune-istl
2017-06-14 10:06:33 +00:00
Atgeirr Flø Rasmussen
d284a657cf Merge pull request #1213 from nairr/terminal_output
remove debug info from convergence failure terminal output
2017-06-14 07:43:00 +02:00
Robert Kloefkorn
e3c898030c [cleanup] ilu_iteration --> ilu_fillin_level. 2017-06-13 15:48:01 +02:00
Robert Kloefkorn
0b90b7863b [cleanup[ use Dune::Amg::SequentialInformation instead of int. 2017-06-13 15:48:01 +02:00
Robert Kloefkorn
38300a4c9b [cleanup] avoid over aggressive memory allocation in ILU. 2017-06-13 15:48:01 +02:00
Robert Kloefkorn
1fa04096c3 [bugfix] circumvent bug in dune-istl when matrix was assembled in
row_wise mode. Also cleanup of output.
2017-06-13 15:48:01 +02:00
Robert Kloefkorn
e4f1bf9955 [feature] make ILU relaxation and iteration a dynamic parameter. 2017-06-13 15:48:01 +02:00
Robert Kloefkorn
a774128fb7 [feature] Make ILU use CRS for storing lower and upper triangular matrices.
Then the backwards in memory iteration for the upper triangular can be
avoided by storing the matrix blocks in the correct order.
2017-06-13 15:48:01 +02:00
Atgeirr Flø Rasmussen
a064f1bae2 Merge pull request #1215 from atgeirr/transport-solver-bugfixes
Transport solver bugfixes
2017-06-13 14:26:35 +02:00
Atgeirr Flø Rasmussen
f37a72da71 Merge pull request #1209 from dr-robertk/PR/remove-uninitialized-vars
[cleanup] initialized Eval variables with 0.
2017-06-13 08:56:23 +02:00
Atgeirr Flø Rasmussen
e0570fc980 Fix issues for iterated sequential implicit. 2017-06-12 15:36:56 +02:00
Atgeirr Flø Rasmussen
e01eca28ca Bugfix: avoid modification of flux data every iteration. 2017-06-12 15:36:48 +02:00