Hi Shawn,
Am Dienstag, 3. April 2012, 11:00:33 schrieb Shawn Zhang:
> Dear All,
>
> I have a couple of general questions regarding pressure in 1p model.
>
> If the simulation is driven by a pressure drop, can the pressure be bigger
> than the pressure inlet?
well, the model conserves mass not pressure, so there is no thing as a
"pressure inlet". what happens is that some mass gets injected at the inlet
and the model adapts the pressure in a way that the mass is conserved. This is
a bit confusing in the current test 1p model. there, you see something like
void neumann(PrimaryVariables &values,
const Element &element,
const FVElementGeometry &fvElemGeom,
const Intersection &is,
int scvIdx,
int boundaryFaceIdx) const
{
values[pressureIdx] = 0;
}
but it should actually be
void neumann(Dune::FieldVector<Scalar, numEq> &flux,
const Element &element,
const FVElementGeometry &fvElemGeom,
const Intersection &is,
int scvIdx,
int boundaryFaceIdx) const
{
flux[massEqIdx] = 0; // (mind the name for the index)
}
because the neumann condition defines fluxes and not primary variables. I
suppose that's what derailed you.
> Moreover, can pressure go negative?
depends on who you ask: Physicists or mathematicians. Phyisically it cannot
normaly become negative, numerically that's not a problem.
> I have some interesting results that I would like to share.
>
> Please find attached a spead sheet with results from four simulations.
>
> For each simulation, I have a 50x50x50 box, with permeability
> heterogeneities on a cell-by-cell basis. For simplicity, I have some cells
> defined with one perm, and all other cells defined with another perm.
> The flow is driven by a pressure gradient (pin=130000, pout=100000) at the
> two boundaries in X direction.
> I got four simulations by varying the difference between the two perms.
>
> At the end of each simulation, I compute average pressure on the YZ planes,
> along X axis, which gives me one average pressure distribution curve.
> When the perm difference is 1 order of magnitude, the average pressure drops
> linearly.
> When the perm diference is 2 or 3 order of magnitude, we can see average
> pressure increase dramatically.
nice stuff. thanks for sharing it.
> Does this make any sense to some of you?
yes, I think that's reasonable. If you think of a homogenious medium Darcy's
law implies a linear pressure gradient, and if you increase the permeability
contrast enough, the cross section occupied by the low-permeability material
does not really contribute to the flux, i.e. the results will be basically the
same as if they where removed from the simullation. Quantifying this is quite
a bit more complicated, though. Nice job :)
cheers
Andreas
--
Andreas Lauser
Department of Hydromechanics and Modelling of Hydrosystems
University of Stuttgart
Pfaffenwaldring 61
D-70569 Stuttgart
Phone: (+49) 711 685-64719
Fax: (+49) 711 685-60430
www.hydrosys.uni-stuttgart.de
_______________________________________________
Dumux mailing list
[email protected]
https://listserv.uni-stuttgart.de/mailman/listinfo/dumux
