We only have one gridcreator test using gmsh files. This adds
another one. The grid ordering changes and the fuzzy compare
could not sort it correctly. Might need a fix there. The
solution is identical. Update reference.

Kilian Weishaupt authored 9 years ago and Timo Koch committed 9 years ago

* Use input file for structured grids instead of dgf files for
  zeroeq2cni tests

Kilian Weishaupt authored 9 years ago and Timo Koch committed 9 years ago

* Use input file for structured grids instead of dgf files for
  2p2c implicit tests

The analytical problem cluttered the directory with hundreds of vtus.
We just write out one now and no restart file.

Kilian Weishaupt authored 9 years ago and Timo Koch committed 9 years ago

* Use input file for structured grids instead of dgf files for
  2p implicit tests

Kilian Weishaupt authored 9 years ago and Timo Koch committed 9 years ago

*The dgf files are no longer needed, therefore the folder "grids" is removed

Kilian Weishaupt authored 9 years ago and Timo Koch committed 9 years ago

* Use input file for structured grids instead of dgf files for 1p2c tests

Larissa de Vries authored 9 years ago and Timo Koch committed 9 years ago

This includes several tasks (in this commit tutorial / stokes / 1p / 2p-sequential)
* Remove unnecessary header includes
* Use vectors / FieldVectors when reading from input file more often
* Use input file for structured grids instead of dgf files for most tests
* Adapt usage message

Several severe fixes to the original commit were provided by Dennis.

Our box models only support non-overlapping partitioning.
Our cc models only support overlapping partitioning (overlap > 0).
This only concerns YaspGrid as it is the only grid manager using
overlapping partitioning.

[2cstokes2p2c] Use new fluxVars facility



See merge request !119

[fix][1p][test][sequential] Use a simple unit cube grid creator

The MPFA interaction volumes store ElementSeeds. In other to
restore entities from seed they need the grid which they obtain
from the GridCreator. Thus having a GridCreator singleton class
is mandatory for the MPFA models.

See merge request !116

The MPFA interaction volumes store ElementSeeds. In other to
restore entities from seed they need the grid which they obtain
from the GridCreator. Thus having a GridCreator singleton class
is mandatory for the MPFA models.

[Richards] remove sources from output

reference solutions had to be updated. Source terms were removed from
the reference.

See merge request !115

reference solutions had to be updated. Source terms were removed from
the reference.

[patches] adapt patches for the 2.9 release

Dumux 2.9 depends on Dune >= 2.4. Delete obsolete patch for
dune-grid, update the one for dune-istl. Adapt README.

See merge request !113

Dumux 2.9 depends on Dune >= 2.4. Delete obsolete patch for
dune-grid, update the one for dune-istl. Adapt README.

[multidomain] Construct FluxVariables in new manner



See merge request !106

Feature/fs291 solutes



See merge request !110

[test][3p][fix] Correct temperature for H2O tables

Results in immense speedup of the test of about 80%.

See merge request !112

Results in immense speedup of the test of about 80%.

Fix/2pdfm bug



See merge request !86

Dennis Gläser authored 9 years ago and Alexander Kissinger committed 9 years ago

The fracture variables and matrix saturations have to be initialized
by default in case the return functions are used for vertices that are
not connected to a fracture

Dennis Gläser authored 9 years ago and Alexander Kissinger committed 9 years ago

Many of the private variables were not necessary for the performed
calculations.

Dennis Gläser authored 9 years ago and Alexander Kissinger committed 9 years ago

the porosity is already set in the base class

Dennis Gläser authored 9 years ago and Alexander Kissinger committed 9 years ago

for the update of the new fluid state the updated primary variables have to be used

Dennis Gläser authored 9 years ago and Alexander Kissinger committed 9 years ago

the fracture permeability in the volvars was not set. It is also not called yet which is why there was no segfault. However, one might want to have the possibility to call this function.

Dennis Gläser authored 9 years ago and Alexander Kissinger committed 9 years ago

After updating the matrix saturations in the interface condition the mobilities in the matrix have to be updated. This is essential in the dfm approach and was not implemented.

Dennis Gläser authored 9 years ago and Alexander Kissinger committed 9 years ago

in the local residual the derivative dSw_m/dSw_F was multiplied to the storage term even though in an implicit scheme this is not necessary

* instead the h2o2n2 fluidsystem can be called and the phase index can be set in the proeprty system

* reference solution for box and cc changed after correcting liquidDensity function in commit 9757ac83
* reference solution for cc used wrong globalPos