Learning Forum

[beyaz17]

Hello guys,

i have one more question to you.

I modeled a welding connection in Ansys and I am doing a static structural analysis. And Iam using multilinear isotropic hardening to describe my material. I solved the modell one time with the heat affected zone and the other time without heat affected zone, like you can see at pic 1 and 2... the single bodies turned into one part.  shared topology = automatic

I had expected that the result would be equial because the material of the heat affected zone are the same like the basic platte for that reason i thought that should be possible to delete the heat affected zone but you can see at pic 3 that ive gotten 2 differnet resualts..

How I can explian that?

pic1

 

pic2

 

[peteroznewman]

The difference in the result could be due to mesh sensitivity (or just a simple mistake).

It is recommended to perform a mesh refinement study to know that the result has converged and you can reliably estimate the true stress.

Regards,
Peter

[beyaz17]

Hi,

thanks for your quick reply.

Ive already done a mesh refinement study. for both modells.

here same pictures

 

R_1 belongs pic 1

R_2 belongs to pic 2 etc.

 

 

without heat affected zone. (Unfortunately I have just one picture on this computer)

[peteroznewman]

I can imagine a mistake caused by a pernicious default setting in the ANSYS software.

After attaching the new geometry, bonded contact was automatically created where you didn't want it, and that made a different model.

In the figure below, there is one edge in Pic1 where there should be no bonded contact and three edges in Pic 2 where there should be no bonded contact.

Regards,
Peter

[beyaz17]

hi Peter,

two more pictures. they are showing the deformation and you can see that there are no bond between the bodies.

thanks

 

with heat affected zone

 

without

[peteroznewman]

You used Shared Topology, so contact was not required, but I have seen ANSYS automatically add bonded contact even though it was not needed. Is that the difference?  Please set the display to show Material to verify that the correct material has been assigned to each body.  Maybe someone else has an idea?

[beyaz17]

hi,

its important to have 3edges to get the same resualt. but why? the lenght is the same. why is the number of the edges important?

 

R_3 is the new curve with the model on pic 1

[peteroznewman]

If the element size is small enough, it shouldn't matter if you have 3 edges or 1 edge.

If the element size is too large, then the effect of having 3 edges is to make smaller elements.

That assumes no mistakes were made getting the correct material assigned to each body, no mistakes were made in connecting the bodies, etc.

Since you have a symmetric model, you could take a vertical plane and slice it down the center and use symmetry. That will let you cut the element size in half and solve in the same amount of time. It will also reduce the number of bodies and the number of connections required to verify.

[beyaz17]

Hi Peter,

Ive used a very small mesh for my modell without heat affected zone like the photo belong. and now the result is fine. the question is why can I use element size 1,5mm with heat affected zone (3edge) for a good result but for the same result I need max. 0.5mm element size for the modell without heat affected zone (1edge). is ansys refinement the mesh automatically?  if yes, how can I turn it off.

 

[peteroznewman]

Hi Beyaz,

ANSYS is not automatically refining the mesh. It was just a coincidence that you got a good result with a coarse mesh for one geometry and not the other.

Please select the post that best answered your question and click Is Solution to close this discussion.