[Paraview] warp by vector field(displacement) with pressure field problem

[Moreland, Kenneth]

Umut,

In the 3D case you are still getting overlap, but you are not seeing it because of an oddity of the way ParaView renders 3D solids. When rendering a 3D solid, ParaView removes the inside of the volume and just renders the outer surface. Most of the time this is correct because you won’t be able to see anything through the outer surface. However, if the volume is intersecting itself, the interior may poke out of the “external” surface, but you won’t see that because ParaView is not rendering the interior of the surface.

In your 3D slice example, try moving the slice plane back to where the displacement dips into the volume and you will see something weird. Instead of seeing a cutaway of the dip, you will see a thin wall built where the clip plane is (see the attached screenshot). This is because, unlike rendering, the clip operation takes into account the entire volume and will find those interior cells that are overlapped. That thin wall becomes the exterior so you can see that part during rendering.

-Ken

From: Umut TABAK [mailto:umut.tabak at gmail.com]
Sent: Sunday, June 17, 2018 2:15 AM
To: Moreland, Kenneth <kmorel at sandia.gov>; Quammen, Cory (External Contacts) <cory.quammen at kitware.com>
Cc: ParaView <paraview at paraview.org>
Subject: [EXTERNAL] Re: [Paraview] warp by vector field(displacement) with pressure field problem

Dear Kenneth and Cory,

Yes, indeed, I was missing this detail. The reason what made me think wrong is that on a 3d model I was not observing this behaviour with a similar approach(At least, that is what I think). See the attached pictures on a 3d post-processing view which also has a lot of zeros in the input file except for the z-direction displacement field.

In the sliced figure for the 3d example, you can also see that the mesh elements at the top face are deformed too much. On one side of the mesh they are moving out of the box and on the other side they are moving into the box(one is basically compressing the air and the other side is expanding which is of course exaggerated too much).

I was thinking that the same approach should also hold for the 2d case. At the end, what I would like to do is to displace the structure with the vector field and show the pressure profile as a color field on this deformed mesh. Since pressure is a scalar, I thought that using the scalar pressure values at the deformed locations of the nodes would give me what I want. But apparently for 2d case, this is not happening(for 3d case, I still believe it is working if I look at the results).

Is not my interpretation correct?

Ps. I will send another mail excluding paraview list with the input vtk file for one set of data to your mail addresses for a further check if you have time.

Thanks for the help in advance.

Regards,
Umut



On 14 June 2018 at 18:22, Moreland, Kenneth <kmorel at sandia.gov<mailto:kmorel at sandia.gov>> wrote:
Umit,

I took a look at your data, and the warp by vector filter is behaving exactly how I would expect.

The displacement field in your data is 0 everywhere except at the top edge. Thus, when you apply the warp the top edge will move and everything else will stay fixed where it is. In the middle of the top edge the warp vectors are pushing down into the rest of the mesh. This is causing some self-intersection with the mesh, which is why you are seeing the pressure profile sticking out the top.

You can use the Threshold filter to remove all cells that are not moving. However, that is only going to leave the top layer of cells, which I’m not sure what you want.

-Ken

From: ParaView [mailto:paraview-bounces at public.kitware.com<mailto:paraview-bounces at public.kitware.com>] On Behalf Of Umut TABAK
Sent: Wednesday, June 13, 2018 11:38 PM
To: ParaView <paraview at paraview.org<mailto:paraview at paraview.org>>
Subject: [EXTERNAL] [Paraview] warp by vector field(displacement) with pressure field problem

Dear all,

On a 2d model, I am trying to plot some mode shape for a coupled fluid-structure problem where a beam is coupled to a cavity domain. I have two variables for this problem: scalar pressure variable for the cavity and the displacement vectors for the beam at the top edges of the domain.

I am writing the attached vtk input file with my post processing codes.

And the result is somewhat as expected except that another pressure profile with an undeformed structural part is also visible with what I actually want to get. Please see the attached screenshot with the settings and visible fields.

Do I have a misunderstanding to accomplish this task. At the end, I would like to show the pressure profile which is a scalar field in the file along with the displacement field on the beam at the top edge.

I was thinking if there is deformed+undeformed visibility selection somehwere or not, maybe the issue is completely something else.

Any comments are appreciated.

BR,
Umut

P.S. For a decent view for the deformed vector field, you can set the scale factor of the warpbyvector field to 0.1 to 0.15


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://public.kitware.com/pipermail/paraview/attachments/20180618/8b513078/attachment-0001.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Capture.png
Type: image/png
Size: 269957 bytes
Desc: Capture.png
URL: <https://public.kitware.com/pipermail/paraview/attachments/20180618/8b513078/attachment-0001.png>