FEA Mesh for fatigue assessments

[Pavel_1976]

Hello,
I'm going to do fatigue assessment according to EN 13445-3 part 18. For this reason am building FEA mesh now and I have some doubts. There are two methods how to extrapolate stress from nodes which used for stress extraction to the weld toe point. First method is by two nodes with distances of 0.4t and +0.6t. Where "t" is the thickness of the shell. Should I use linear elements for this approch? Becouse I thinck the extraction nodes shoud be at the ends of elements and not middle nodes and it looks not good if beetwen nodes 0.4t and +0.6t will be some other node (middle side nodes of quadratic elements). Or quadracic elements are accepteble and corner nodes should be used for stress extraction.
There is the same question for second method where three nodes are used like 0.4t, +0.5t, +0.5t so linear or quadratic elements should be used and what nodes corner or middle side (if quadratic elements are accepted)?
Kind regards Pavel.

 

[DriveMeNuts]

If you look at the underlying British standard for this method for fatigue of structures, it allows the use of finding the linearised membrane plus bending as an alternative to what you are trying.
This method you are trying is very time consuming and needs to be done manually in ANSYS after every analysis run. I gave up on it.