finite element analysis 2

[AAJB]

Is there any basis for allowing a stress equal to 1.5 times yield stress at a hot spot, assuming the structure will "adjust" for the load and the yield stress will increase due to strain hardening? I am unable to find any refernce for this hypothesis. How does one determine if this 1.5 times yield stress will cause failure or just a distortion? The material is steel with 50 ksi yield and 70 ksi ultimate stress levels. The material is assumed to be ductile. Von Mises indicates I am in failure territory. Any comments are very helpful.

 

[mloew]

AAJB,

The following previous threads will likely provide you with the information you are looking for:

Explanation of Very high local stresses in FEA Results thread727-52443
&
Ignoring high local stresses in FEA thread727-52469

You may also want to use a more descriptive subject line in future posts.


Best regards,

Matthew Ian Loew

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips Fora.

 

[corus]

As well as other threads you should also look at BS 5500, the pressure vessel design code, which describes stress categories and their limits which could also apply to structures in general, I believe. For secondary stress ranges the limit is twice yield, which is based on the principle that shakedown will occur. I think that ASME also describes these limits and where these categories apply. These limits are besides fatigue considerations at hot spots.

 

[brad]

I have commonly seen the 1.5 value (and also a value of 2.0) used within automotive. My understanding has been that this is a "rule of thumb" approach without any basis in any particular design codes (ASME, SAE, etc.). This seems to give reasonable answers for a variety of automotive structures. However, there is some question to applying this in a general sense, as one must realize that this assumption goes hand-in-hand with whatever loads are assumed within the analysis.

Also, this assumption can't be considered universal as it neither recognizes the distinction between various levels of ductility nor the nature of stress redistribution that occurs for different geometries.

So, is it used? Yes. Do I recommend it? No, not in the most general sense, but neither do I condemn the practice.

Brad

 

[oso]

You may also read thread727-53526 for more help.

I think first you need to review FEA model with FEA analysts to see if there are any possiblity of high stress from modeling method. Be care of high stress in
1. Between component interface in assembly
2. Boundary condition applied areas

High stresses in those areas usually could be ignored or at least be reviewed carefully.

If modelling is fine, you also have to look at the stress level around these "hot spot". If the surround area is much low than yield stress, these high stresses would be redistributed and wouldn't cause any yield.

In addition, non-linear analysis may also help... I am not very sure.

Never take a "rule of thumb" (1.5 or 2.0)for silver bullet. For example, if somebody put a sharp corner in a high stress area. The stress will continues grow to infinite with finer mesh. You will get a stress a million times yield stress