Is there life after yield...

[ZattleBone]

Hi there

I have a particular application where the compression of my part induces stresses WAY beyond the yield stress. The small amount of elasticity remaining in the part is fundamental to the design as it has to survive approx one million cycles of +/- 20 microns. Unfortunately on the current design cracks are appearing and we need to somehow improve it. The part has a section of only a few mm, is made from stainless-steel and the compression is across the section NOT along the length.

My question is can fatigue calculation software model this kind of fatigue problem with high plasticity, low strain cycling and a million cycles? I'm looking at packages like MSC.Fatigue as we already have an account with them.

Any help or advice on my problem is appreciated.

Thanks in advance

Zatt.

 

[corus]

I'd be wary of considering only one load cycle in your assessment as the high compressive stresses will result in residual tensile stresses once the structure is unloaded. The stresses in the elastic part will also redistribute in subsequent cycles. The resultant tensile stress range between cycles is likely to contribute to the fatigue failure. If you're using a stress-strain curve, rather than an ideal elastic-plastic material, then you'll also find that the yield stress will change with subsequent cycles. Stresses in subsequent cycles may lie in the elastic range of the material.
The resultant stresses you get could be simply assessed against a simple S-N curve of the material but if you've got a load history and complex tri-axial stresses then it might be worth considering some fatigue software.

corus

 

[prex]

Your problem should be not very difficult to assess even with simple analytical calculations (however the geometry is not very clear to me).
If the initial deformation is not too big (say less than 1% strain) it should not significantly impair the residual fatigue strength of the item. If on the contrary the initial strain is large or there are local effects (e.g. discontinuities at supports or load location) that are difficult to model, then even the best software wouldn't be of much help.

prex

http://www.xcalcs.com

Online tools for structural design

 

[ZattleBone]

Mmm, thanks for the responses guys. I think I should explain in a little more detail though.

I have a wire 2mm thick and I want to compress it across it's thickness by 0.5mm (i.e. 25%). I then want to completely remove the load (allowing it to recover approx 20 microns) then reapply the same compression. Repeat this a million times and there's my problem. Most of the time the wire survives intact (albeit deformed). But sometimes the wire cracks.

I'm been researching this all day and am coming to the conclusion that Wohler and Goodman just can't be used beyond yield as they both predict almost instant failure.

I think I'll dig out my old Uni notes tonight and give myself a headache.

Zatt

 

[feajob]

ZattleBone,

There are 3 different methods for Fatigue Analysis:

1) S-N Method (Stress Life)
2) E-N Method (Strain Life, or Crack Initiation)
3) Crack Propagation Method

I think that Stress Life method is not a good approach for you, because, in this method you basically ignore plasticity. Strain life method can handle better your case. It is the most common life prediction method used in industry. It is also called the local strain approach. Defining the material properties for this method is more difficult than Stress life method. In MIL handbook you can find some S-N curves.

Regards,
Ali

http://www.geocities.com/fea_tek/asd/

 

[ning]

I'd say you will have plastic hardening effects on your material after load and unload. At first cycle you have strain 0.5 mm/ 2 mm =25 %, then the second cycle you will have less plastic strain due to yield stress now higher. Seems elastic part of this strain is 20 micron/ 1.5 mm =1.3%. I would say if geometry of the problem is not complex, the material property is more important than using FEA to find detailed stress distributions.