To support this the solveSystem methods of the LinearSolverInterface gets
an optional additional template parameter of type boost::any. It can hold any
copy constructable object. In our case it is used to pass the information about
the parallelization into the solvers of dune-istl without the compiler needing to know
their type. Inside of LinearSolverIstl::solveSystem we check whether the type stored inside of
boost::any is the new ParallelIstlInformation. If this is the case we extract the information
and use the parallel solvers if available, otherwise we solve serial/sequential.
The new ParallelIstlInformation is needed as the OwnerOverlapCopyCommunication is not copy
constructable. This is indeed a design flaw that should and will fixed upstream, but for the
time being we need ParallelIstlInformation to transfer the ParallelIndexSet and RemoteIndices
objects.
This patch refactors the calls to the dune-istl solvers.
The SeqScalarProduct, and SequentialInformation is now explicitly
constructed and later used to construct the smoothers
generically. Aditionally the linear operator (MatrixAdapter) is
constructed before calling the various solver dependent solve
routines.
While this does not change the behaviour of the code it is a
preparatory step to support parallel solvers. These parallel
solvers only differ in the type of the scalarproduct and
linear operator used from the sequential ones.
This patch refactors the calls to the dune-istl solvers.
The SeqScalarProduct, and SequentialInformation is now explicitly
constructed and later used to construct the smoothers
generically. Aditionally the linear operator (MatrixAdapter) is
constructed before calling the various solver dependent solve
routines.
While this does not change the behaviour of the code it is a
preparatory step to support parallel solvers. These parallel
solvers only differ in the type of the scalarproduct and
linear operator used from the sequential ones.
This fixes further occurences of DUNE_HAS_FASTAMG that were
missed in pull request #530.
Previously we relied on the define DUNE_HAS_FAST_AMG to detect
whether these preconditioners are available. This define is only
available in the 2.2 release with cmake support. Therfore we now
addtionally test whether we are using dune-istl 2.3 or newer.
and throw an exception if "simple" is encountered...
According to Ove, gwseg should be used, because "gwseg is the model
relevant to the norne case - i.e the model eclipse uses.
The fix for the simple model has to wait for a refac of the satfunc
complex."
Previously we relied on the define DUNE_HAS_FAST_AMG to detect
whether these preconditioners are available. This define is only
available in the 2.2 release with cmake support. Therfore we now
addtionally test whether we are using dune-isl 2.3 or newer.
Patch 31c09aed was erroneous as it was trying to assing a
SaturationPropsFromDeck<SatFuncSimpleNonuniform> to a
SaturationPropsFromDeck<SatFuncSimpleUniform> in the constructor
taking the new parser. This patch fixes this.
My last merge broke compilation of opm-autodiff, as this new function
still relied on the old behaviour (the one taking a grid instead of
the data directly). This patch moved initStateFromDeck to the new behaviour.
The mentioned commit was applied before the merge of
opm-parser-integrate and therefore the changes did not carry over
to the code that uses the new parser. This code mimics the
changed behaviour for the new parser.
Closes issue #516
Instead of making well rates zero for wells that are not controlled by
surface volume, we initialize them to a small value with the correct
sign (positive for injectors, negative for producers).
It is not quite complete yet for the following reasons:
- it does not compute state.surfacevol(),
- the InitialStateComputer class does not compute Rs or Rv,
- it has not been verified.
In summary:
- added RsFunction (base class),
- made NoMixing, RsVD, RsSatAtContact inherit RsFunction,
- RS and RV are no longer template arguments for EquilReg class,
- EquilReg constructor now takes two shared_ptr<Miscibility::RsFunction>,
- use of constructor updated, mostly using make_shared.
This includes relative permeability and capillary pressure functions.
The default has been to make a monotone spline from the given table
values and use a fine, uniform sampling of that. Now the default
is to use the tables as-is. It is still possible to use the spline
approach. For example in the class BlackoilPropertiesFromDeck one
may pass nonzero values for the 'pvt_tab_size' and 'sat_tab_size'
parameters, corresponding to how fine the spline will be sampled.
This patch refactors (hopefully) all parts of opm-core that are needed
by the fully implicite black oil solver in opm-autodiff and that inherently
relied on UnstructuredGrid.
We added a new simple grid interface consisting out of free functions
that will allow us to use CpGrid without copying it to an UnstructuredGrid
by the means of the GridAdapter. Using this interface we have add methods that
allow specifying the grid information (global_cell, cartdims, etc.) wherever
possible to prevent introducing grid parameters for the type of the grid.
Unfortunately this was not possible everywhere.
This commit implements some additional scaling keywords. This includes
the ISWL-family that provide hysteresis behaviour via alternative
scaling of the tables. The old parser has been somewhat extended for
testing purposes. The commit also includes a slight refactoring of the
SatFunc-family where a new base class has been introduced.
The initial implementation of RK4IVP<>::operator() failed to take
into account the possibility that we might need to evaluate the
function outside the vertical span for which it was initially
defined. This situation occurs, for instance, in the not uncommon
cases of the GOC being above or the WOC being below the model.
This commit installs a crude Hermitian extrapolation procedure to
handle these cases. Refinements are likely.
This commit adds support for assigning the initial phase pressure
distribution to a subset of the total grid cells. This is needed in
order to fully support equilibration regions. The existing region
support (template parameter 'Region' in function 'phasePressures()')
was only used/needed to define PVT property (specifically, the fluid
phase density) calculator pertaining to a particular equilibration
region.
This commit adds a simple facility for calculating initial phase
pressures assuming stationary conditions, a known reference pressure
in the oil zone as well as the depth and capillary pressures at the
water-oil and gas-oil contacts.
Function 'Opm::equil::phasePressures()' uses a simple ODE/IVP-based
approach, solved using the traditional RK4 method with constant step
sizes, to derive the required pressure values. Specifically, we
solve the ODE
dp/dz = rho(z,p) * g
with 'z' represening depth, 'p' being a phase pressure and 'rho' the
associate phase density. Finally, 'g' is the acceleration of
gravity. We assume that we can calculate phase densities, e.g.,
from table look-up. This assumption holds in the case of an ECLIPSE
input deck.
Using RK4 with constant step sizes is a limitation of this
implementation. This, basically, assumes that the phase densities
varies only smoothly with depth and pressure (at reservoir
conditions).
The pvt interface is extended to handle wet-gas systems:
1. rvSat is added as a function in the PVT interface
2. SinglePvtLiveGas computes the pvt values and its derivatives
3. The old rbub variable is changed to rsSat for clearity
4. The new interface is tested in test_blackoilfluid with data from
liveoil.DATA and wetgas.DATA
This is intentionally black-oil specific because we presently do no
know how to handle other cases (e.g., more phases or number of phases
different from number of components).
The second input of initBlackoilStateFromDeck that computes surface
volumes based on gas/oil ratio is changed from
BlackoilPropertiesInterface to Props. To avoid duplication with the old
initBlackoilStateFromDeck its name is changed to
initBlackoilSurfvolUsingRS.
Add new function is added that computes saturation from surface volumes
solving z = As for each cell. This function is used to compute an
intial guess to the saturations in initState_impl.hpp.
The 'comp_term' is supposed to be a total divergence term, which is supposed
to be zero for incompressible flow. It was added for improved robustness in
stagnant areas, but as implemented it would not be computed properly for
oil injection scenarios, due to the convention for two-phase transport
source terms (positive terms are inflow of first phase [water], negative
terms are total outflow).
The WellsManager class handles INJECTORS by assigning a phase
distribution (W->ctrls[i]->distr) that coincides with the injected
fluid for the corresponding well (e.g., {1,0,0} for WATER injectors in
a three-phase WATER/OIL/GAS simulation). This, however, meshes poorly
with the restriction that all phase components must be ONE in the case
of wells constrained by total reservoir volume flow targets (RESV)
that was introduced in commit b7d1634.
This change-set limits the restriction on phase distributions to
PRODUCERs only and is a tentative solution to GitHub PR #360.
Otherwise the compiler will probably give us a warning that these
pragmas are unknown. By default that warning is disabled with our
own build system, but we also want to be able to link to our library
without incorporating the entire build system too.
To avoid deprecation warnings the number of smoothing steps was passed
through the Criterion instead of directly to the constructor in commit
a7f32b934b.
However, due to an insufficient test matrix this was not tested using
the fast AMG variant of DUNE so it breaks the builds if
`-DHAS_DUNE_FAST_AMG` is defined.
This change should apply the same type of change to this branch as for
the others. The number of smoothing steps is put into a constant to
avoid the magic number 1 to appear in too many places (although I am
not sure the number for pre- and post-smoothing always should be the
same).
If a function is used by a template but this template is not
instantiated, the function will still be defined in the header
of a module but it won't be callable because it is in an anonymous
namespace and thus we get a warning.
This only happens in Clang; GCC consider functions referenced from
templates as used.
fixup! Don't warn about functions not emitted
make all non-implementation headers includable without
preconditions. Also, this removes the GravityColumnSolver.hpp file,
because it tried to include a non-existing file and it was thus unused.
for some of these files this is needed to make to keep it compiling
after the next patch because the new ErrorMacros.hpp file will no
longer implicitly includes <iostream>. for the remaining files it is
just good style.
While at it, the includes for most of these files have been ordered in
order of decreasing abstraction level.
SuiteSparse may or may not be installed in a suitesparse/ directory.
FindSuiteSparse will look in a suitesparse/ subdirectory when trying
to locate umfpack.h, but it will add the full directory to the compiler
command-line (e.g. `-I/usr/include/suitesparse`) and not that of the
parent. Since the parent is usually included too, it is not noticed
that it is advertedly using another include paths than its own.
However, if we have SuiteSparse installed in a non-system location,
using the subdirectory in the `#include` statement may now cause an
error, even though configuration actually found SuiteSparse!
AGMG is now under a closed-source license, meaning that results
obtained with this solver is not freely reproducible by others.
Its use is therefore discouraged.
As of version 2.3, the DUNE AMG parts are competitive, so there
is a free and open alternative.
our policy is that we only use boost if necessary, i.e., if the oldest
supported compiler does not support a given feature but boost
does. since we recently switched to GCC 4.4 or newer, std::shared_ptr
is available unconditionally.
most of them quite insignificant, but still annoying. The only
exception is the warning about the changed alignment for the 'work'
argument of spu_implicit_assemble(). AFAICT, the only reason why it
worked was that the pointer produced by malloc() was passed
directly. (malloc() seems to fulfill all alignment criteria.) To fix
this, I've changed that argument's type from char* to double*.
The current implementations of IncompPropertiesInterface are very
all-or-nothing. In some situations, you want to read rock and fluid
properties from an Eclipse file, but use analytical functions for
the unsaturated properties. Or you want to update properties based
on a marching filter.
This patch provides a way to mix various property objects, or to
"shadow" the properties with a raw array of data, so you don't have
to reimplement the entire interface just to make a small change.
The <have_boost_redef.hpp> header was introduced (commit 82369f9) as
a work-around for a particular interaction in the Autotools-based
setup of OPM-Core and the Dune core modules. Notably, Dune's
"Enable" trick for Boost failed on some older Autoconf systems. Now
that we're using CMake, however, that kluge is no longer needed
because we (OPM-Core) always
#define HAVE_BOOST 1
i.e., as an explict true/false value.
Therefore, we need no longer include <have_boost_redef.hpp> . The
header will be removed at a later time.
After this patch one can set the prolongation factor for all
AMG calls and the number of smoothing steps for all except
of FastAMG that currently only supports on step.
Do not use the well's comp_frac member, only rely on the control's distr
member for initialization. This forced a change to WellsManager's
initialization of the distr member.
- Changed interface.
- Read tracerheads (tracer start locations) from file in compute_tof_from_files.
- Initialize tracerheads from wells in compute_tof.
Now it will no longer trigger assertation failure when grids are not
edge-conformal (faulted cornerpoint grids processed by our code will
usually not be). Minor algorithm change to handle this.
Interface change: solver now requires a linear solver (for the multi-cell blocks only).
Implementation uses new private method assembleSingleCell(), that is a modified copy
of solveSingleCell(). Should refactor.
Deleted some unused code (or moved to opm-porsol), moved all code dealing with
time-of-flight to opm/core/tof, moved code for implicit transport solver to
opm/core/transport/implicit, spu_[im|ex]plicit.[ch] to opm/core/transport/minimal.
The "maxit_" counter is an upper limit on the number of non-linear
iterations in a single cell. Declaring this as a "double" is counter
intuitive unless one expects the number to be *really* high.
Present since
- Commit 93d4bd8 (TransportModelTwophase.hpp)
- Commit e0d38cf (TransportModelTwophaseCompressible.hpp)
This commit corrects an error that has been present since the
introduction of function wellsToSrc() in commit a50bb8f but was never
detected. Although the conversion int->double->std::size_t is likely to
be loss-less, it is better to not introduce any more steps than are
actually needed.
This is a minor bugfix to account for the presence of shut wells
(characterised by "ctrls[w]->current < 0"). The existing code would
lead to indexing outside the "ctrls" array in the context of a shut
well.
Also:
- Add use_tensorial_basis parameter allowing run-time choice of basis.
- Remove degree argument from solveTof() method, degree is instead obtained
from parameters in constructors. Modified compute_tof* programs to match.
