[natalie]

Hello everyone!
I’m dealing with changing of wall profile due to flow properties. The profile isn’t known beforehand, so it’s part of solution. Water flows in tube, the problem is planar. View of .cas file is attached. Mesh view in general and zoomed in deforming region are also attached.
The desired wall profile is taken from the article and picture (in results file) is attached.
To my mind, wall deformation can be estimated according with Wall Shear variable.
So, I’ve got several questions.
1. Is it possible to solve such type of wall deformation, using CFX only (I mean, mesh deformation laws and UDFs)?
2. If not, then maybe FLUENT can do it?
3. If neither CFX nor FLUENT can do this by their own means, then is the two-way FSI proper decision?
4. Is it possible to estimate contribution of each component of Wall Shear variable in common deformation? (probable functions that I use to deform mesh are attached)

Thanks a lot for help! )

[singer1812]

CFX should be able to do it.

I assume your equations are correct, I dont know anything about them.

I assume the X-axis runs normal to the center diameter axis of your cutout. If it doesnt, you need to move it.

I dont like your mesh. I would consider changing it. It shouldnt be hard to get a smoother transition around that cutout.

Your Top wall and symmetry walls should have unspecified motion.

[Far]

try to improve the mesh topology, it will help you alot in simulation.

[natalie]

Thank you

i agree that mesh topology is not perfect, but quality nevertheless is higher than 0.15

the problem is that it doesn't move the wall to the necessary level.

I tried to solve this problem using square, not sphere caverne wall and still it doesn't fold.

Should i remesh? and how?

[singer1812]

Regarding mesh: What do you use to mesh? Put an Ogrid in there. That alone will help smooth that transition.

Regarding Walls Sheer: I took a closer look at your equations. For the Axial component you are taking the absolute value of Wall Shear in the X direction as the impetus for movement. Does that seem right? Seems fishy to me. In addition, you are reducing that value down by 0.00001. Is that arbitary? Why did you do that?

[natalie]

Thank you for piece of advice, concerning mesh. I'll try to improve it.

About Wall Shear. In fact, i don't know the proper law of movement, i just try to guess it according to the picture given in the article (this picture is attached in results.file).

The coefficient was chosen, because when it was bigger, the solver failed.

The magnitude of this coefficient was chosen, because the magnitude of Wall Shear for my problem is about 100-1000 Pa, the size of the cell is less than 1 mm. so maybe, solver fails because of too high deformation in first steps

[singer1812]

Get rid of the fancy equation at first. Do something simple.

Put in a simple term for axial displacement such as (I assume this is a transient run?) 0.001 [m s^-1]*t (or some displacement versus time). Zero out the other displacements. See if your mesh deforms in a linear fashion as dictated from above.

Also, based on your displayed mesh, any type of movement at all might cause mesh failure. It really is a bad mesh.

[natalie]

Thanks a lot

i tried different functions depending on time in axis direction. mesh was without boundary layer (and also some other topologies) it deforms quite normally, but then after some time (100 iterations for example) no changes happen.

That is the main problem. Why it doesn't move further?

[singer1812]

Did your job crash at that point?

Also, you switched views on your pics. It is hard to tell what the deformation is doing. What equation did you use? Try only moving in the axial direction, the radial direction is a bit more complicated.

[natalie]

Yes at that moment the job stopped, because no deformations happened any more for over than 1000 iterations.
If i use the greater value in the expression for wall deformation (because i'd like to model rather fast changes), then solver crashed, because deformations were too strong.
i guess i should use remeshing.. is it correct?

In fact, axial deformation turned out to be more sensitive to deformation. i need moving for about the size of the cell (it's about 1mm). Never mind, what function i use for moving: constant value or expression depending on Wall shear variable.

So, the problem is that it either moves very-very slow or crashes

[singer1812]

No your function is important.

The way you setup your moving mesh in your initial post, the wall will move to a set point based on wall shear. If wall shear is 1000Pa like you said earlier, your axial movement will be 0.001m (or 1 mm) and then stay there (assuming wall shear doesnt change). Perhaps this is why the wall didnt appear to move. It is hard to tell and if you dont post how you are trying to move your wall, people trying to help you can only guess at what you are doing.

In contrast, if you setup your function including a velocity and time (like 0.001[m s^-1]*t) the wall will constantly move 1mm each second, with no stopping.

Remember you are specifying wall displacement, not wall motion (or velocity).

[natalie]

Thanks a lot for your help.

I'll try..