When using Fatigue calculation using existing FEA data, What type of results are typically used and why? Example : Signed Von Mises stress, Maximum Principal stress etc.
If you want to perform a monoaxial fatigue life prediction, the fluctuation of the maximum principal stress is the parameter that you are looking for. Otherwise, you need to consider all the tree principal stresses and a multiaxial fatigue model, to obtain an uniaxial equivalent stress amplitude. Given that the analysis is limited to elastic behaviour.
If the analysis is carried out at the root of a notch, you might want to modify the stresses resulting from FEA, since the fatigue notch factor is in general smaller than the stress concentration factor.
Max principal is not a good choice. For example, if you have a uniaxial tension-compression normal stress (+sig/-sig), the max principal stress is always positive (0/+sig) because max pricipal is zero. Also von Mises needs great care, as it gives half of the amplitude of uniaxial sigma, double oscillation frequency, and it works only for in-phase multiaxial stresses
The fluctuation of the normal stress along a fixed direction is a better answer, in uniaxial case. It is probably wise to start considering the direction of the max principal stress obtained at the maximum load.
ASME Sec VIII- Div-3 code uses Maximum shear stress and mean stress on the maximum shear stress plane to estimate the fatigue life. If the loads are non-proportional, critical plane search method (CPS) is used to detect the plane with maximum shear stress. This is good for non-welded components.
For non welded components, ASME Sec VIII, Div-2 and DNV-RP-C203 standards uses von Mises stress to estimate the fatigue life.
In case of welded components, maximum principal stress is commonly used in oil and gas industry. Other codes (ASME VIII-2, -3) uses membrane and bending stress intensity for welds to estimate the fatigue.
Use appropriate stress/strain value to estimate the fatigue amplitude.