Finite Element for High Cycle Fatigue

[EngAP]

I am looking for a finite element package to model high cycle fatigue failure. Are there any free packages available and if so, which are recommended?

Thanks.

 

[corus]

There is an add-on piece of software from FE_SAFE that will look at fatigue. I've also seen Cosmos Solidworks assess fatigue too, but if you're looking at welds then it's better to manually assess them.

corus

 

[cbrn]

Hi,
unfortunately, all the fatigue add-ons I know are not freeware. Even the CosmosWorks' module Corus refers to is not free. The same is for Ansys Workbench Fatigue feature. Moreover, all these work inside programs which are very expensive by nature...
However, provided that the free FE code you want to use is able to perform results' combinations (i.e. linear combinations of loadcases), you can do high-cycle fatigue assessment "by yourself", IMO without paying the price for any more software.

Regards

 

[MechEng2005]

While I don't do fatigue analysis on a regular basis, I recall doing it while in college. Couldn't you just calculate the endurance stress and then used FEA to see if the stress in a static model (with worst case loads) exceeds that stress?

The book covering this material was "Mechanical Engineering Design" by Shigley, Mischke, and Budynas

-- MechEng2005

 

[ariansdad]

MechEng2005 I have to agree, fatigue analysis doesnt have to be complicated if you want to just look at the endurance limit, we regularly use this method. If you have an S-n curve you can effectivly predict the life of the component based on the peak stress in the structure, although you should also apply some statistical analysis to your S-N data.

 

[TGS4]

Depends on how you perform your FEA. If you have details such as welds, or even (gasp!) fillet welds, you could get one of two answers: either you'll have a numerical singularity location, in which case you'll have "infinite" stress or you'll not model some detail or forget to add an FSRF (fatigue strength reduction factor).

Does your "widget" have any of these features?

 

[corus]

If it's a weld then the practice is to take the nominal stress at the weld, ie. that which is not affected by the weld geometry. To me it's all a bit of a guesstimate, though with some methodology. I know of no software that could even attempt that. If it is welds then the only way is by manually looking at your results and use the nominal principal stresses that act in a particular direction to the weld in order to get the classification of the joint. If it's plain material, however, then FE codes with inbuilt fatigue assessment would be of use.

corus

 

[EngAP]

TGS4 - our widget does not have these features - it is a simple shaft with a filleted shoulder subject to rotational low and high cycle fatigue. It is believed that the shaft is in elastic bending hence is not rotating concentrically - leading to the high cycle fatigue.

MechEng2005 - Yes, I agree. This will be achievable for both the fatigue AND nominal stress calculations.

 

[corus]

For a rotating shaft you should simply consider the endurance limit but taking into account the overall size of the component and the surface finish. Compare the maximum stress (including stress concentrations) with this factored endurance limit.

corus