High cycle fatigue analysis: Performed fully in FEM or partly in FEM and partly in Excel?

[hootrpootr]

I’m trying to wrap my head around different ways of analyzing a components high cycle fatigue life, since I’m now just learning that FEM software is (supposedly) able to perform fatigue analyses, given I set up the model correctly.

At my company, the method I was taught was to run two static FEM analyses, one with a unit g load to determine vibratory stresses, and then another FEM with actual loading to determine steady state stresses. At this point, I would create a standard Goodman diagram in excel from the material properties (endurance limit, FTU, FTY, etc.) and plot the vibration stresses (adjusted for dynamic amplification and input) and steady state stresses. Lastly is the calculation for Goodman percent.

Please tell me if I’m wrong, but is it possible to do all of this in just one FEM? I came across some YouTube videos where fatigue analyses are performed in one model, without the need to take it to excel or run multiple models. Basically, material properties (with s-n curve info), BCs and actual loading, Goodman (or other) fatigue theory, and a few other parameters were added to make the model run. If it is all possible to run an HCF model fully in one FEM, why wouldn’t it be done like this every time?

Thanks

 

[FEA way]

Many FEA programs (but not all) can perform fatigue analysis based on the results from previous static simulation. These programs use various algorithms to account for complex effects (such as Rainflow cycle counting or Miner’s rule). They can even perform fatigue analyses under random loads. As a result you will get contour plots of fatigue life (number of cycles to failure), damage and factor of safety.

 

[Mustaine3]

I think more common are the specialized codes that take FE results and perform the fatigue calculation. The most often used are probably fe-safe, nCode and Femfat.

 

[hootrpootr]

Are the results from a fully FEM analysis typically realistic enough to accept with confidence, as a static analysis would?

 

[hootrpootr]

Mustaine3, what are the benefits of using specialized codes to determine the fatigue life as opposed to determining it within FEM? Or in other words, why would one prefer having to extract FEM results for an external calculation over keeping all the work within one program (FEA)?

Thanks

 

[GregLocock]

"Are the results from a fully FEM analysis typically realistic enough to accept with confidence, as a static analysis would?"

That depends on the analyst, not the FEA program so much. If your FEM is properly validated and you understand how to interpret it then I see no reason why it should be worse than a static analysis, and it ought to be better. However given that few industries seem to bother with proper validation then GIGO.

FWIW we can develop the calibration for all the crash sensors in FEA, crash being a bit harder than fatigue.

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[Mustaine3]

Mustaine3, what are the benefits of using specialized codes to determine the fatigue life as opposed to determining it within FEM? Or in other words, why would one prefer having to extract FEM results for an external calculation over keeping all the work within one program (FEA)?

With a neutral fatigue tool you can use several FEA tool or those, who have nothing integrated. Also, those special fatigue tools are usually much more powerful than the integrated.

But there is no right or wrong. A company or user should choose the method that works best for them.