Showing results for 
Search instead for 
Did you mean: 

Notch factor in Nx still unclear without answer




we had a discussion about defining the Kf =notch factor (fatigue limit in smooth/notch specimen) input in the fatigue life analysis (not in the fatigue safety analysis, K-factor).


in Nx you have two option 

A- Eliminate the notch from the model; run elastic FE for you un-notched specimen, apply Kf directly (we use the empirical formal function of Kt).

B-run nonlinear analysis w/o Kf.


for option A if it was necessary to use the notched model. how to calculate the Kf for this case?

it was stated in another post that Kf here is actually the Kf/Kt, as the only difference in both models is the concentrator Kt, which needs to be taken it into account to reach the suggest case A.


However, applying the two case (A, and the alternative of the notched model) but the show totally different results as follow: 

the original model with notch has Kt=10, Kf(empirical, f(Kt))=5.


Three analysis compared were:


  • FE for the smooth specimen + Kf ------> fatigue life of 55k cycles for all the specimens except a node/element at support which fas life around 2k) (k=1000).
  • FE notch specimen + Kf/Kt=5/10=0.5 <1 , input in the notch factor box, yield infinity 
  • FE notch specimen +Kt/Kf=2 >1, (no meaning for this factor), yield fatigue life around 2k , but comparable.

Thanks for your comment




Re: Notch factor in Nx still unclear without answer

Siemens Genius Siemens Genius
Siemens Genius

Tricky one Smiley Happy


If you have a elastic notch factor Kt of 10, you have quite a sharp notch.

Did you check that the elastic stresses in your notched model get up to this high stresses ?

I would typically expect some smoothing by the finite element analysis, such that you get a higher life just by the FE modelling.





Dr.Michael Hack
Business Product Line Manager Durability
Simulation and Testing Solutions

Re: Notch factor in Nx still unclear without answer

Hi Dr. Michael;

The high value of the stress concentration is due to the sharp-notch and is comparable with the theoretical value.
The wonder still is how to implement theKf. Hope any of the experts of NX durability solver answer our questions, as the manual does not. Unfortunately did not find a book in this regard as well, and the trial and error way is not the fair option for NX users.