[vlr]

Hi,

I am confused by the results of a set of simulations.
I used 2 solvers pimpleFoam and interFoam with the following boundary conditions:

Inflow: Velocity = fixedValue (0.5 0 0) and Pressure = zeroGradient;
Outflow: Velocity = zeroGradient/fixedValue(0 0 0) and Pressure = fixedValue 0;
Cavity: noslip;
Initially the cavity is completely filled with water.

If the outflow boundary is set to zeroGradient, pimpleFoam as well as interFoam run correctly as expected.
If the velocity of the outflow boundary is set to fixedValue (0 0 0), for pimpleFoam the time step reduces until a small value and the calculation stops as I expect.
However, if I change the outlet value for interFoam the calculation just runs without a problem.

Can someone clarify to me why the water leaves the domain when a zero velocity outflow is specified, and why the time step does not reduce dramatically?

Thanks in advance,
ROB

[vlr]

Forgot to upload the image.

[hulli]

Hi



can you check that
a) in your blockMesh the inlet and outlet are patches ???


and b) the boudary conditions are set like that:

in 0/U
outlet
{
type inletOutlet;
inletValue uniform (0 0 0);
value uniform (0 0 0);

}
inlet
{
type fixedValue;
value uniform (0 0 5.0);
}


in 0/p
outlet
{
type fixedValue;
value uniform 0;
}
inlet
{
type zeroGradient;
}


I hope that helps ...

[vlr]

Thanks for your reply. Inlet and outlet are indeed patches.

I changed the settings to your suggestions, but the case keeps running without reduction of the time step.

I attached the case. Is very simple and runs in a few seconds.

[hulli]

Hi man,

I just run your case with interFoam.
I played a bit arround with and without your setFields.

The case just runs smoothly.

If I understood your question correctly with the timeset not getting any smaller I guess you run into that problem wit the adjustable timestep as mentioned in that post:
pimpleFoam adjustable timestep


Does this answer your question ???



Stay home

[hulli]

one more thing ...
I just checked your fvScemes


you have a lot of upwind schemes in there

you may want to check this page of Tobias and double check your schemes to get a better result ... you also can check the tutorials there are also some free ones to double check you case setup ))



https://holzmann-cfd.com/en/openfoam...hemes/analysis

[vlr]

Hi,

To state my problem in other words.
The case runs smoothly indeed, however it shouldn't, because the outlet has a fixed zero velocity while the inlet has not.

I fully understand that this is an incorrect case, but I expect the case to drop its timeStep (which is adjustable) to zero, because it cannot be solved.
For pimpleFoam it does drop the timeStep, but for the interFoam case it does not. On the contrary it just "pushes" the water through the zero velocity outlet!
The difference between the two cases is not clear to me. For completeness you find the pimpleFoam case in attachment.

I will read your references to see if the forward differencing could be the cause. Thanks for sending them.

[hulli]

Hi man thanks for the second case...
I got your question now...


This is what happened:
a) I ran your pimple case and the case reduced the CO number as you just mentioned. After running it for 10 minutes the timestep was super small and it wrote only two files. As you said



b) In order to understand what was going checked the interFoam.C file on github. The solver also uses the pimpleloop to get the Ueq running solved so I was wondering if I get your interfoam-case running with the settings of your pimple-case. So I copied the system folder and the transport properties into the interfoam case. And it was not running!!! But it should... First OF was complaining about settings of the p_refcell (which I did afterwards) and then it was complaing about some dimensional error.


c) I compared bouth of your U files - no difference then the p files and here we go the dimenions are set different !!! BAsically you must have copied the p_rgh and renamed it as p ...This is also evident in line 14 of your p file --> object p_rgh and it should be p



dimensions [ 0 2 -2 0 0 0 0 ]; p as it should be
dimensions [ 1 -1 -2 0 0 0 0 ]; in your p file



d) Now both cases are running smoothly ))))) which does not answer your question...



Now you case try setting the boundary conditions to e.g in the U file to 0 and it will give you an continuity error because you will pump more water into the system and you do not allow anything to leave the domain which is ok )))



I hope it helps ... enjoy your week end !!!

[vlr]

If this runs, the first question that pops into my mind is: What will happen with two fixed velocity inlets (positive at minX and negative at maxX). Will that run as well ? Then it will surely be something I don't understand numerically. Perhaps I have to go deeper into the solver-realms. I will let you know if I run into something noteworthy.

Thanks for your effort. Enjoy your weekend as well.