Notch, Special case of stress distribution

[sartorbjk]

Hello everybody,
Currently, I am performing some linear orthotropic analysis near a notch (not exactly a notch but a relative one lets say).
I ended up with the stress distribution that you can see from the attached file.

During my bachelor, I remember that I saw this picture/stress distribution in one of my mechanical engineering classes. Max stress is not on the notch or surface but a little bit inside the material.T his stress was called something. I could not remember the name of this stress.

Could anybody help me to remember this?


Best Regards,
Sartor

 

[Spirit]

sartorbjk,

two remarks:
1) you are working with an orthotropic material;
2) you are analysing the MAX principal stress.

I expect that if you do the same run on an isotropic material and check the Von Mises stress you should obtain what you are indeed expecting (max stress at the 'notch' tip).

Regards,

Spirit



'Ability is 10% inspiration and 90% perspiration.'

 

[rb1957]

maybe the OP has an orthotropic material (so an isotropic materila would be incorrect).

i think it is usual (for metals at least) to have the stress peak away from the surface, and it's possibly a combination of material and loading effects. i don't think it has a special name. are you looking at element, or nodal stresses; if you want the stress at the surface, you need to use nodal stresses.

you could refine your mesh a bit in the area of interest, 20 node bricks ? are you still less than yield ? (often stress concentrations are plastic)

 

[rb1957]

sigh, thought i'd caught it ... "usual" of course should be "unusual"

 

[ESPcomposites]

Perhaps you could show a zoomed out picture and how it is loaded. There could be a combination of the notch effect and a secondary bending effect due to eccentricity. These two may offset each in a manner that may not be initially intuitive, but just guessing. If you can show it deformed and undeformed, that would help as well.

For a standard problem of a plate with a hole/ellipse, you would expect the peak to be at the cutout boundary, regardless of material system. You can also have a look at eCutout, which may give some insight to your problem. It handles isotropic, orthotropic, and anisotropic materials.

http://www.espcomposites.com/software/eCutout.html

Brian

http://www.espcomposites.com

 

[BrianE22]

Maybe you are thinking of "contact stress"? I believe the maximum stress (shear?) is below the surface for contact loading.

 

[corus]

This stress distribution seems related to your mesh size. Refine the mesh and you may find the max stress occurs closer to the surface and the sharp corner. Why it doesnt; peak at the surface is hard to say unless you state how the model was loaded.


Tata

 

[sartorbjk]

Hello,

I found the name of this stress distribution. It is called: Herzian stress distribution.
Thank you very much for your interest.


Best Regards,
Sartor

 

[ESPcomposites]

Care to elaborate? I am familiar with a Hertzian contact stress, but how is that relevant to your problem? In that case, as mentioned above, the max subsurface shear occurs below the surface. But that would not be a "notch" problem as you presented it. Can someone clarify this?

Brian

http://www.espcomposites.com

 

[sartorbjk]

That is what I am thinking. I have actually no contact. When I saw this stress distribution, I remembered that I had seen it somewhere in my bachelor. Obviously, it is not something related.

The problem is: this is a T shaped bending/torsion sample. I apply load to the lower end of T (T is not sharp edged but curved as can be seen on the picture) and I get two times bending. I am trying to figure out a complex loading case (not just bending, not just tension or compression). I will now describe my problem in another topic under FEA engineering subforum since I am having difficulty to interpret something.

Best Regards,
Sartor