Computed stresses strongly depend on the mesh fineness at a notch, but it seems to be not so strict for stress intensity factors. What is the best mesh fineness, and how about the accuracy in calculating stress intensity factor?
[1] Kucewicz M, Leski A: Accuracy study of stress intensity factor solutions obtained by commercial FE software. AIP Conference Proc 2019; 2078:020100. doi: 10.1063/1.5092103.
[2] Oh G, Sasaki T, Akiniwa Y: Experimental and computational analyses on fatigue fracture and microstructure in dissimilar metal weldments with circular sharp stress raiser. Int J Fatigue 2020; 133:105422. doi: 10.1016/j.ijfatigue.2019.105422