I am trying to do a literature review on the correlation between average grain size and fatigue performance of AMS 5662 (solution annealed condition) Inconel 718. The Special Materials Corporation datasheet (SMC-045, 2007, page 16, PDF), shows a significant correlation between average grain size and endurance limit of AMS 5596. Basically the finer the grain size the higher the endurance limit, according to this reference. Some of our internal documents also refer to the "aerospace structural handbook 1995" which is nowhere to be found!
However, I am not able to find any scientific literature supporting this claim. It appears as if the correlation between average grain size and fatigue performance is non-obvious. Considering that fatigue is a chaotic phenomenon, finer grain sizes only delay the crack initiation thanks to their elevated hardness, I believe. However, as soon as a microcrack is initiated it propagates much faster in the bulk, leading to a total failure, I hypothesize. For example, the paper by Almaraz et al. reports that double-aged material lives ~30 times longer at higher loads, but the endurance limit is actually slightly worsened! My explanation is that the tougher and more ductile the material the higher its tolerance to absorb hysteresis energy.
So now my question is whether there is any conclusive correlation between average grain size and alternating fatigue endurance limit for this specific material, AMS 5662, or any other advanced alloy of a similar nature?
P.S. I also posted the question here.