[Benji]

Hi everyone!
I hope nobody is getting annoyed about me frequently posting recently, I'm still trying to get comfortable with OpenFoam.
I finally managed to generate more or less useful physical data for a river flow with interFoam.
Today I found the potentialFreeSurfaceFoam solver, which looks promising to me (especially because its computational costs seems to be significantly lower than interface-capturing solvers).
Since there is very few information about this in the Forum, I decided to open a new thread. Does anyone have experience with it?

• Klick: My casefile

I created a simple river mesh, took the Oscillating Box tutorial and adjusted my BC's: I have a constant inlet, walls and bottom with currently no friction, a free surface and a calculated outlet.

There are a few things I'm not sure about, so I'll just post what I think and hope that someone will correct me :

• p and p_gh: They are given in m2/s2, so no density is included, the difference is: p = static pressure/rho and p_gh = static pressure - hydraulic pressure = (p - g*h)/rho.

• zeta: This delivers the height information, but in my outputs I get also weird values for the x- and y-coordinates. I guess I'm just not familiar with topoSet and the output control...

Thanks already for reading my noobie-stuff!

Cheers,
Benji

[Benji]

Guess I found out now more or less about the pressure (see above), but what I can't manage is getting my system to reach an equilibrium state or even getting close to it. I assumed it has to be something with the output, but I couldn't find out what it is so far.

[Benji]

New day, another time

I started setting up another small case that should be easy to reproduce by just hand calculation.

• casefile barrier

I created a rectangular channel with a small barrier at the bottom (similar to the damBreak Mesh) and set a constant inflow of 0.5 m/s.

What I want to find out is the water level before and after passing the barrier. Then I want to compare the results from the potentialFreeSurfaceFoam calculations to the calculations with the Poleni-equation [Q = 2/3*nu*b*sqrt(2g)*h^(3/2)].

This produces some more or less physically useful data in the first few seconds, but: It seems that the calculated waves get reflected at the outlet, so there is never an equilibrium state, all the water stays "inside my Mesh" and pressure is getting higher and higher.

Does anyone have an idea how to set the outlet conditions so that this won't happen? Or is this not possible with this solver?
--> I found the waveTransmissive BC for the pressure at the outlet and right now im trying to use it, but I wasn't successful so far, the outlet still reflects the waves. Still my problem, waves get reflected, apart from this: good.

Cheers, Ben

[Benji]

Dear all,
I'll continue with my monologue here, still hoping that there'll be someone who can help

As I said before, I'm still struggling with the outlet reflecting the waves and getting higher and higher pressures & velocities. Also I'm not sure about my zeta values. The setting of the waveTransmissive BC has not helped so far, but I'm not sure if I have done this correctly.

I've been playing around now with the barrier case and a narrowing mesh.

• casefile narrowing river

• casefile barrier

I stopped the simulation before the reflection of the waves at the outlet started and tried to check the calculated values. I get acceptable values for the velocity and also for the height (calculated with (p-p_gh)/g=h), but I don't think this can be taken seriously as long as the above problem is not fixed.
Anyone who has met similar problems?

Have a nice day/afternoon/evening/night or whatever time it is at your place !

Benji

[shipman]

Hi Benji,

It is a bit late to ask but since I faced same problem I wanted to ask that did you have any improvement about your reflection problem from outlet using potentialFreeSurfaceFoam.

If so, I will be happy to discuss with you.

Thank you.

B

[Jost K]

Dear all,

I hope its okay for me to post into this thread on a slightly different note:

I am just starting to look into potentialFreeSurfaceFoam.
Apart from the free surface boundary condition the solver looks pretty much like pimpleFoam to me (looking at the code).
However, from the tutorial and some initial testing, it seems to need much smaller time step sizes.
can someone explain to me why this is the case?


Also:
Does anyone know any literature about this Solver (maybe some verification and validation)?

Thanks,
Jost