Axisymmetric stress analysis with rigid surface, stress distribution seems strange 1

[jonask]

Hi all,

I run an axisymmetric analysis of a part that comes into contact with a rigid surface and is then loaded with pressure from the inside. My analysis is linear elastic and I use quadratic elements.
The analysis steps are:

- Use BCs to fix the part in space and move the rigid surface down into contact.​

- Apply a small force in direction of the rigid surface to keep the contact in place.​

- Remove the BC from the part to "relax" the contact.​

- Load the piece with pressure from the inside.​

As soon as I get into contact, there are high von Mises-stresses in a position I find strange. If you look at the screenshot attached, you will see that contact occurs in element 3454, and the highest stress is interpolated at the nodes in the second row from the contact. Does this stress distribution seem normal to you?

Best,
Jonas

 

[corus]

In herzian contact the peak stress intensity is below the surface due to shear so the stress distribution isn't too unusual. In your case the mesh is too coarse to accurately portray the true stress distribution though. You'll also need to consider large deformation and/or plasticity as you basically have point contact with the two curved surfaces in contact and the stress intensity will tend to infinity with a refined mesh.

 

[jonask]

Hi corus,

thanks for the reply! If I understand correctly, mesh refinement would give me a more correct stress distribution, but the stresses will tend towards infinity unless I include plasticity or large deformations into the model. I will give it a shot with a refined mesh and an elastoplastic material model.

Best,
Jonas

 

[pnavy]

Hello, from me to.

2D mesh has it's downside, many times you have to analyse results with mes refinement if the peaks in the results are due to the nature of the 2D element.

Just as a thought if you put force on the line that has thickness of 0 and is a part of a analysed plane the stress would go rise infinity closer you get to the loaded line.

 

[izax1]

Hi jonask

I agree with corus. Your mesh is way to coarse to get the max contact stress. Espeicially in the area where you get in contact with the rigid surface you will need a much finer mesh.