Commit Graph

50 Commits

Author SHA1 Message Date
Markus Blatt
815480d052 Only call writeTimeStep for real sub steps.
Previously, we also called it when the full time step was done.
As the simulator writes that information anyway and we cannot call
it a sub step, we omit the final write in the adaptive time stepper.
2016-09-26 11:51:17 +02:00
Tor Harald Sandve
1bde4f4a22 Small fixes hardcodedTimestepControl
-- avoid using eof()
-- add comments
-- no longer assumes two lines of comments.
-- revert change to default value for timestep.initial_step_length
-- make contructer explicit
-- pass reference
2016-09-21 09:39:36 +02:00
Tor Harald Sandve
380fe6e2fd Timestepper based on userInput
A new timestepper that reads timesteps from a file generated using
ecl_summary "DECK" TIME
and applies it to the simulator

Also a parameter timestep.initial_step_length (default 1 day) is added
to controll the frist timestep.
2016-09-21 09:39:36 +02:00
babrodtk
63da817852 Initial version for outputting cell data 2016-09-01 14:37:41 +02:00
Atgeirr Flø Rasmussen
d31081992b Merge pull request #1052 from andlaus/pass_timer_instead_of_dt
pass the timer object instead of the time step size to the simulators
2016-08-02 11:05:33 +02:00
Liu Ming
61889dafbd fix indentation. 2016-07-14 10:30:20 +08:00
Liu Ming
62134b4a0f drop useage of std::numeric_limits 2016-07-14 10:27:13 +08:00
Andreas Lauser
0dda8d49c5 pass the timer object instead of the time step size to the simulators 2016-07-05 12:23:55 +02:00
Liu Ming
b8ef9cb630 output well iterations if it is a valid number. 2016-06-30 09:03:30 +08:00
Liu Ming
30d9c34481 Add space. 2016-06-28 15:26:06 +08:00
Liu Ming
b1b464535a output well iterations and non-linear iterations. 2016-06-28 13:40:32 +08:00
Liu Ming
63f244a36d mark time stepping messages type as OpmLog::note 2016-06-21 08:44:32 +08:00
Jørgen Kvalsvik
18fda2d7f0 WellState::report() to make opm-output Well data 2016-06-06 12:40:44 +02:00
Liu Ming
248ea780b3 add missing braces. 2016-05-16 09:04:54 +08:00
Liu Ming
a4dc135045 use OpmLog only for error messages. 2016-05-10 14:13:33 +08:00
Liu Ming
40dbcc496a output timer messages for terminal and log file. 2016-05-09 13:31:04 +08:00
Atgeirr Flø Rasmussen
0de2cff986 Follow class renaming. 2015-11-13 11:37:53 +01:00
Robert Kloefkorn
d3af817ff2 timeError --> relativeChange. 2015-11-10 09:53:40 -07:00
Robert Kloefkorn
005a0c2930 AdaptiveTimeStepping: pass object to compute time error to time step control. This
allows us to shift the computation of the error to the physical model.
2015-11-10 09:45:07 -07:00
Robert Kloefkorn
4f8906c5fc AdaptiveTimeStepping: also apply terminal output to solver_verbose. 2015-09-16 15:19:19 +02:00
Robert Kloefkorn
8c93936fd5 AdaptiveTimeStepping: pass variable terminal output to avoid multiple outputs in
parallel.
2015-09-16 14:33:35 +02:00
Tor Harald Sandve
6d28f19009 Change default values
- use time stepping algorithm pid instead of pid + iter
Adjusting the time-steps on the number of linear iterations does
currently not give any improvents on the time-stepping.
- Change the pid tolerance. The time-stepper will take longer time-steps
and thus reduce the simulation time significantly. The Norne and the SPE
results does not degrade
- Less aggressive reduction of time-steps after convergence problems
2015-09-08 10:56:58 +02:00
Tor Harald Sandve
dbce88aa2c Limit the timestep growth for all timestepping algorithm
The time step restriction is moved to AdaptiveTimeStepping_impl.hpp to
make it apply to all time-stepping algorithms.
2015-09-03 12:52:47 +02:00
Tor Harald Sandve
0227165bd5 Rename last_timestep_ to suggested_next_timestep_ 2015-08-25 11:16:36 +02:00
Tor Harald Sandve
81b8b631a6 Use estimated time-step instead of average in next report step 2015-08-25 06:32:39 +02:00
Atgeirr Flø Rasmussen
42464cbcb7 Match init order to member order.
In a constructor initialisation list, the order should be the same
as the order in which the variables actually are initialised, which
is given by the order they are declared in the class and not by the
order in the initialisation list.
2015-08-19 11:33:29 +02:00
Kai Bao
a7f177e35e adding command line option full_timestep_initially_
when the option is true, for each report step, the size of the adaptive
time step always beginning with the size of the report step.
2015-08-17 15:56:31 +02:00
Atgeirr Flø Rasmussen
2dfbb4ed82 Correct comment. 2015-08-17 13:02:37 +02:00
Atgeirr Flø Rasmussen
e41b766780 Catch MatrixBlockError in the adaptive time stepper.
This can be thrown by the ILU0 preconditioner when using the
interleaved solver approach.
2015-08-17 12:59:31 +02:00
chflo
bfa1d476e3 OPM-218: Write restart files at same interval as eclipse 2015-08-05 13:38:52 +02:00
Markus Blatt
384ef84556 Makes the time step control parallel.
The only stage where parallelism changes the adaptive time
stepping is when some inner products on the saturation and
pressure are computed.
This commit makes this part parallel by added an additonal boost::any
parameter to the time stepping and the controller. Per default this
is empty. In a parallel run it contains a ParallelIstlInformation object
encapsulating the information about the parallelisation. This then used
to compute the parallel inner product.
2015-05-27 11:07:16 +02:00
Atgeirr Flø Rasmussen
b6d5a3cad5 Adjust parameters after testing. 2015-04-22 13:03:19 +02:00
Atgeirr Flø Rasmussen
2dc02043e2 Modify default parameters to be suitable for the Norne case. 2015-04-21 11:41:45 +02:00
Tor Harald Sandve
9272f2d8bf Add dune/istl/istlexception.hh to header
Fix to make it compile with dune-istl 2.2
2015-04-10 14:03:27 +02:00
Robert Kloefkorn
7f6008a80e AdaptiveTimeStepping: also catch ISTLError casued in AMG when time step is to large. 2015-03-30 11:07:08 +02:00
Robert K
4a885989dd AdaptiveTimeStepping: remove initial_fraction (use restart factor instead) and use
average of previously used time steps to suggest next time step.
2015-02-10 13:10:39 +01:00
Robert K
26f113b649 AdaptiveSimulatorTimer: -improvement in time step adjustment near end of time interval
-max time step parameter

PIDTimeStepControl --> TimeStepControl:
- added simple iteration count time step control
- bug fix in PIDAndIterationCountTimeStepControl

AdaptiveTimeStepping: apply the above changes.
2015-02-06 13:59:50 +01:00
Robert K
ae683efcfc AdaptiveTimeStepping: make output more easy to read. 2015-01-16 16:08:59 +01:00
Robert K
15c2e91730 AdaptiveTimeStepping: Switch suggested time step to max of the previous taken substeps. 2015-01-16 16:08:59 +01:00
Robert K
04e8533cee AdaptiveSimulatorTimer: initialization of first time step size follows the same rule as
for later steps.
PIDTimeStepControl: added maxgrowth factor which indicates the maximum allow groth of
                    the time step from one to the next value.
2015-01-16 16:08:59 +01:00
Robert K
757bc67b80 cleanup: reportStepIdx --> writeStepIdx.
startDate     --> startDateTime
         removal of debug output.
2015-01-09 16:22:03 +01:00
Robert K
25af9e0033 - adjust OutputWriter to SimulatorTimerInterface
- remove WriterTimer from EclipseWriter and use SimulatorTimerInterface
- adjust to the above in AdaptiveTimeStepping.
2015-01-09 16:22:03 +01:00
Robert K
aa9fe2a631 EclipseWriter: allow for writing of substeps in addition to report steps. 2015-01-09 16:21:03 +01:00
Robert K
f16f4a2e98 fix problem with solver reastart, variable was not reset.
Also, the time step does only grow moderately after a solver restart.
2014-10-24 12:32:00 +02:00
Robert Kloefkorn
16624f6f4e address Atgeirs comments. 2014-10-20 12:32:11 +02:00
Robert Kloefkorn
40d851e89f also catch std::runtime_error, i.e. convergence of linear solver failed. 2014-10-14 15:18:36 +02:00
Robert Kloefkorn
1bda36b251 added initial fraction to initialize last_timestep. 2014-10-13 11:17:37 +02:00
Robert Kloefkorn
3b58ad9aa4 Two small adjustments that came up when running in debug mode. 2014-10-10 13:55:28 +02:00
Robert Kloefkorn
246acea765 use unit::convert::to instead of hard coded 86400 factor. 2014-10-07 09:48:57 +02:00
Robert K
294b899ee8 the adaptive time stepping utility classes. 2014-10-06 14:06:07 +02:00