*volVars(elementIdx, localIdx)

Feature/diffusive fluxes as vectors

See merge request !528

Feature/move phasepresence to privars

See merge request !530

* we use a copy instead of a reference in the signature to
  avoid linker errors related to passing a static variable
  to this function
* https://stackoverflow.com/a/29259331

Feature/improve global fv geometry

See merge request !532

The number of subEntities is wrong for non-conforming
grids. Therefore, replace it with numFaces of the mapper.

Merge branch 'feature/update-to-opm-2017.04' into 'next'

See merge request !535

Timo Koch authored 7 years ago and Kilian Weishaupt committed 7 years ago

[opm] Update dumux to newest opm release

See merge request !526

* instead, use negative value of corresponding flux
* re-include errors when trying to access diffCoeff() with phaseIdx == compIdx

* precompute density and ansatz function gradient

* for trivial intersection mapper

* it is not needed to pass this variable to the helper
* it can, if need increment a private variable with each update()
* localFaceIndex now comes from the geometryHelper

* Only cache one and convert the other one on-the-fly

* use correct component indices

* introduce update method in geometry helper
* use local face index
* use intersection mapper

[staggered][model] Fix elementSolution

See merge request !531

* use integral_constant instead of int in order to access the whole
  cell-center related privars vector

* Remove unnecessary lambdas

* Use these methods to account for mole/mass fraction

* Use these methods to account for mole/mass fraction

* We should never assume that the total molar fluxes cancel out by default

* consider vector of molecular fluxes in a more generic way
* when using moles, make sure that the total molar flux given by the
  respective diffusion law actually cancels out
* otherwise, consider all individual fluxes for total balance

* use alias

* use const more often
* use alias

* use alias for vector type
* use axpy to compute the mole fraction gradient (which erroneously
  contained  the molar density)

Katharina Heck authored 7 years ago and Kilian Weishaupt committed 7 years ago

[2pnc] add useMoles as property to make 2pnc work with new compositional localresidual and return 0 as diffusion coeffiecient if phaseidx = compIdx instead of warning that diffCoeff does not exist