CFD Online Discussion Forums

CFD Online Discussion Forums (
-   OpenFOAM (
-   -   Closing valve implementation (

simpomann October 29, 2012 13:43

Closing valve implementation
1 Attachment(s)
Hey guys,

Some weeks ago I came up with a problem that was solved in this thread (basically by Mr Gschaider): How to change a boundary condition depending on a probe set

and a final(very simple) test case

In the meantime my work changed a bit:
What I need to do is closing a valve depending on the filling level of a tank.
I think I have to implement a solution-based dynamic mesh.
Is there any tutorial for this (only found tutorials with a mesh motion that didnt depend on the solution)?

If you see the attached picture you can see a slice through my valve: It is a self floating device that will close the ventilation pipe above it.

Is there any way that I can move the valve (single stl) depending on the fluid level?

In my dreams I would like to define a translation in z-direction depending on the alpha value of a probe set that is aligned in the centre axis of the valve in z-direction.

Any suggestions? Any help? I never used a dynamic mesh and couldnt find a "master-dynamic-mesh-dictionary" with all the options (saw something like this once for sampled surfaces).

Every help is highly appreciated!
Big thanks and greetings,


simpomann October 29, 2012 16:07

Ok I guess I have to specify a dynamicMeshDict with a linearBodyMotion, but I can't find out what I coefficients I have to specify. Is there any documentation? And how does it recognize which part of the mesh to move?

I think I must specify it somewhere before, but I am absolutely clueless.

A look at the source code didnt enlighten me.

Edit: A workaround for me could be to install a small flat surface in the pipe above the valve and vary the boundary condition (this at least I know how to do with groovyBC). But are there boundary conditions that would make this surface "invisible" at first? So that there will be a flow through the surface in the original setup?
Then I could change them similiar to a wall later (e.g. zeroGradient for p,alpha and fixedValue 0 0 0 for U).

deepsterblue October 29, 2012 17:14

I remember a post by Philippose a little while ago that had a spool-valve simulation using the dynamicTopoFvMesh class. Here it is:

The setup is a little involved, but you can definitely give it a shot.

All times are GMT -4. The time now is 08:05.