Thomas Fetzer authored 8 years ago and Timo Koch committed 8 years ago

2p2c updated the reference solutions (changed phase presence indices)
2pnc changed: plSg --> pnsw; pgSl --> pwsn, removed plSg + pgSl from indices
2pncmin changed: plSg --> pnsw; pgSl --> pwsn as for 2pnc
2p2c, 2pnc: added the formulation changes to CHANGELOG.md

(cherry picked from commit b03738a2)

[2pnc][test] Clean-up spatialParams

* Remove deprecated, energy-related functions
* Test is isothermal anyway

(cherry picked from commit 182a63f2aae439f63dffae9f825006fed2cef4c6)

See merge request !257

[2pnc] Fix/2pnc

See merge request !258

Kilian Weishaupt authored 8 years ago and Timo Koch committed 8 years ago

* Remove deprecated, energy-related functions
* Test is isothermal anyway

(cherry picked from commit 182a63f2)

Fix/mpfalowdim

Improve/fix handling of branching points when dealing with network grids (2D in 3D)

See merge request !254

for dim < dimWorld (i.e. 2 < 3), the algorithm for the setup of the interactionvolume seed is more complex
than the one for dim = dimWorld = 2. This commit introduces a forwarding to the more simple algorithm if the
face does not touch a branching point.

we don't want to compare positions for the detection of branching points.
Note that the new implementation fails for hanging nodes on a 2d grid embedded
in a 3d domain.

Dennis Gläser authored 8 years ago and Timo Koch committed 8 years ago

for hanging nodes accessing the localscvfindexset was wrong, as indexInInside()
might be the same for two different intersections.

Dennis Gläser authored 8 years ago and Timo Koch committed 8 years ago

we introduced an element flux variables cache update in the cclocaljacobian
to make use of the fact that the transmissibilities are potentially not solution
dependent. In that case this process is skipped.

Dennis Gläser authored 8 years ago and Timo Koch committed 8 years ago

The stencils vector class is now set separately for the specific cell-centered
methods. Some methods might lead to non-symmetric global matrices, thus, the stencils
have to be set up differently.

Dennis Gläser authored 8 years ago and Timo Koch committed 8 years ago

We introduce method-specific implementations that should inherit from a
mpfa-specific base class (which inherits from the scvf base class). We define
an interface for the constructor that has to be met by all the implementations.
We do this so that the classes instantiating scv faces do not have to be
method-specific. We add a local index to the constructor which corresponds
to the local vertex index an scvf is connected to for the mpfa-o-fps scheme.

Dennis Gläser authored 8 years ago and Timo Koch committed 8 years ago

If dimWorld > dim, using a GlobalPosition vector is wrong, we have to
use a vector of size=dim here.

Dennis Gläser authored 8 years ago and Timo Koch committed 8 years ago

Up to now the types for the inner and boundary interaction volume seeds
were always identical. For the mpfa-l and other methods, it would be more efficient
if these types could differ, as e.g. the sizes of the containers are known at compile time
for the l method. This commit introduces the capability of the types to differ.

Dennis Gläser authored 8 years ago and Timo Koch committed 8 years ago

when UseTpfaBoundaries is deactivated, the scaling of the neumann flux
has to reoccur using the deflected vol vars. otherwise the flux calculation
in the fluxVariables goes wrong. We have to keep in mind, that using the global
flux var cache for the mpfa also means that the neuman boundaries are assumed to be
constant over time!

Dennis Gläser authored 8 years ago and Timo Koch committed 8 years ago

we reintroduce this property tag because for mpfa methods there can
be saved a lot of computational time.