when only gas phase is present, use mole fraction times total pressure
as the partial pressure for the component. it seems like the inverse
of the partial density is not consistent with this approach. why this
is the case is a mystery to me...
Actually it can be used with any capillary pressure-saturation curve
which was obtained experimentally. Maybe ExperimentalDataLaw would be
a better name...
This significantly reduces the need for nesting typedefs because a
material law with absolute saturations as input parameters can be
defined like:
typedef RegularizedVanGenuchten<Scalar> EffectiveSatLaw;
typedef EffToAbsLaw<EffectiveSatLaw> AbsoluteSatLaw;
Basically, the material laws can now be considered complete...
I tried to catch most stuff, but occationally Dune:: prefix is
required for stuff which is provided by DUNE and sometimes the Dune::
prefix needs to be changed to Dumux:: for stuff which is part of DuMuX
- make it possible to use the 2p fluid state in conjunction with fluid
systems which require the phase composition
- make the 2pni model and tests work
- move the ifp application from appl/co2-ifp to appl/co2/ifp
- add the benchmark3 CO2 problem (i.e. kezin formation)
- make the benchmark3 problem compile. the 2pni part works, but not in parallel
- fix a most stupid bug in the h2o_n2 fluid system (temperature and pressure
parameters for the density() method were integers)
- improve the end-of-iteration output of the 2p models (now it also
prints the vertex and the primary variable which where the defect is
worst.)
- add the salinity used to calculate the CO2 properties to the
generated output. this avoids inconsistencies between the tabulation
and the brine properties used by the simulation.
- use this salinity value in the brine_co2 system
- change phaseState -> fluidState
- also use the tabulated CO2 solubility
- update the CO2 tabulation utility to generate the corrosponding tables
- in the Brine-CO2 system do not ask the CO2 for anything if the pressure is to low
- do not call updateStaticData() in initial() of the box scheme in order to avoid
the 2p2c model from breaking down if the initial condidition requires a change in
the present phases
- update the IFP problem to include a much larger CO2 table. this was made possible
by the fact that calculating the CO2 solubility in the extraction utility is now _much_
faster than it used to be
- add SimpleCO2 class (currently it is basically useless besides retrieving the triple
and critical points)
