- a solvent specific updateState is used to assure that the correct oil
saturation is used to detect phase transision
- presence of gas is compensated for in the oil phase
The solvent specify part of computeWellConnectionPressures is factored
out to computePropertiesForWellConnectionPressures in order to reuse
the computeWellConnectionPressures from the base model.
The following changes are done in order to remove the duplicated code in
assemble().
- extractWellPerfProperties takes SolutionState as input (only used in
the solvent model)
- the computation of effective parameters is moved to computeAccum()
With these changes the solvent model can use assemble() from the base
model.
The Todd-Longstaff model is extended to incorporate pressure effects
The solvent viscosity is then caculated as
mu_eff = mu_s^(1-\alpha * \omega) * mu_mix^(\alpha * \omega)
where \omega accounts for the porous media effects and \alpha =
\alpha(pressure) accounts for the miscibility of the solvent and oil
when contacted.
The \alpha values can be given using the TLPMIXPA keyword
If no entries are given to TLPMIXPA the table specified using PMISC will
be used as default.
IF TLPMIXPA does not appear in the grid \alpha = 1 and the pressure
effect is neglected.
This is tested in test_solventprops_ad.cpp
The relative permeabilty endpoints are scaled by the miscibility
function. The endpoints is not supposed to be negative therefore all
negative values are replaced by zero.
The most significant bug was in the computation of effective saturation
fraction for the gas phase. The current implementation differs from the
manual.
Add method to calculateEffectiveProperties (viscosity and density)
using the Todd and Longstaff model.
The effective properties are stored in mu_eff_ and b_eff_ in the
BlackoilSolventModel and accessed using fluidReciprocFVF and
fluidViscosity
When running parallel one cannot use Eigen::Array::mean() for this
as the it is just a local part of the complete array. With this commit
we correctly compute the number of global cells in the grid and use this
together with a parallel reduction to compute a global mean value.
