> On 5. Nov 2018, at 02:55, Ranjeet kumar <ranjeet...@gmail.com> wrote:
>
> Dear All,
> Would you please explain the followings:
>
> 1. For Neumann Boundary and Source, unit are [kg/(m^2.s)] and [kg/(m^3.s)].
> In the input file, I want to give rate [kg/s] for both Neumann bc and source.
> Do I need to divide the rate by scv's area/scv's volume in the neumann() and
> source() method respectively? Similarly, for method
> neumannAtPos()/sourceAtPos(), rate/(domain's area/volume ).
>
Hi Ranjeet,
For both functions neumann/neumannAtPos you need to divide the total mass flux
(in kg/s) by the total injection/extraction area. You need to make sure that
your mesh resolves the injection area exactly to get the exact total boundary
flux.
If this is not the case, the value will be approximated better and better with
grid refinement.
For both functions source/sourceAtPos you need to divide by the total volume of
injection. Again, you need to resolve you volume of injection exactly by the
grid.
If you divide by the area/volume of the sub control volume instead, it would
mean that you inject at your given rate for _every_ sub control volume, i.e. if
you are refining your grid you would inject more and more.
There is also point sources (see e.g. the 1p point sources test for how to use
the interface), in case you want to inject at a singular point. There you
specify the value in kg/s and give a position. Dumux is going to find the right
sub control volume for you.
> 2. In 2p cornerpoint, FLuidSystem::phase1dx is used as indices for source()
> whereas in , 1p implicit, contiW/NEqIdx used. Are both the indices same?
> Could you please explain how indices works?
The source function returns an array with the size number of equations. You
have a mass balance equation for every component. For the 2p model, phases and
components are the same, since the components are considered immiscible.
The index is determined by the fluid system. The mass balance for the first
component will have index 0, the mass balance for the second component will
have index 1. To make that clear, you can use
FluidSystem::comp0Idx/FluidSystem::comp1Idx or
(in the 2p immiscible case only) FluidSystem::phase0Idx/FluidSystem::
phase1Idx. Usually, you would rename these indices in the problem header for
better readability. If you’re dealing with water and nitrogen for example, you
might define something like
enum
{
H2OEqIdx = FluidSystem:: comp0Idx,
N2EqIdx = FluidSystem:: comp1Idx
};
or whatever you find readable, _and_ set water as your first phase and nitrogen
as your second phase in the fluid system. If you change the order in the fluid
system, the order of the equations will change as well.
Best wishes
Timo
>
>
> Thank you,
>
> Regards,
> Ranjeet
> _______________________________________________
> Dumux mailing list
> Dumux@listserv.uni-stuttgart.de
> https://listserv.uni-stuttgart.de/mailman/listinfo/dumux
_______________________________________________
Dumux mailing list
Dumux@listserv.uni-stuttgart.de
https://listserv.uni-stuttgart.de/mailman/listinfo/dumux
