Another slant on one of this forum's favorite subjects - trunnions!
Consider a horizontal trunnion which undergoes a cyclic loading [high cycle / low stress range].
We can determine that the stresses in the header caused by the trunnion [under simple static loading] are within acceptable limits by using a method such as the Kellogg approach. We can also determine what the stress range at the trunnion / header node is from the Caesar model.
Combining the two correctly seems a bit harder to pin down what the combined alternating stress intensity is and so what would be the best approach to do a check on the fatigue life of the pipe at the trunnion without resorting to FE/Pipe or similar brick FEA approach?
Is there an argument form applying a SIF to the trunnion / header intersection node? Looking at the Kellogg method I think you can break the moment terms into a factor x a bending stress in the trunnion of the form:-

Stress due to Longitudinal moments = [(1.17 * R^0.5 * t) / (T^1.5)] * Bending stress in trunnion in Long. direction
Stress due to circumferetial moments = [(2 * 1.17 * R^0.5 * t) / (T^1.5)] * Bending stress in trunnion in Circ. direction.
R=mean radius of header
T = header thickness
t = trunnion thickness

But doing this will intensity the bending stress at the root of the trunnion, and also intensify the moments acting along the header, which would be overtly conservative perhaps, but that is not perhaps a 'bad' thing where fatigue is concerned given that the trunnion base will offer many stress raisers around it and beyond if a re-pad is used [not recommended by EN 13480].
NC-392 & EN 13480 both use the same method of establishing the stresses in the header caused by the trunnion and offer methods for combining them, but both are not directly applicable to B31.3 and are not that familiar to many people. They recommend that the calculated stress, Spt is added to the intensified bending stress range. Note they do do not offer any advice on the value of the intensification factor at a trunnion
what do the more learned contributors out there recommend as a practical approach for day to day use.

Still mulling....

I'm glad I'm not the only one!

I have been doing a bit of digging through some codes and I think that the NC-392/EN 13480 method of summing the bending stress range [SE] in the header pipe and calculated stress in the header [Spt] from the attachment can be modified to give sensible results if Spt is calculated from the range of forces acting on the trunnion under the cycle being considered. We are then adding two stress ranges together and we should aim to keep these within SA <EN13480 requires a SUS check to be done as well on the attachment>.
Does anybody know if Coade has, or has plans to, implement the rules of EN13480 section 11 [attachments to piping] for circular and rectangular attachments within Caesar II. It would also be helpfull if there was a way to write and run macros to peform calculations on results generated by Caesar II without resorting to exporting to Excel or similar.

I can add this to the list of "Requests", and study this section of EN-13480. This will not make the 5.10 release, as that is pretty much cast in stone.

Hmmmm ... that would be interesting ...

Richard - I hope you are saying "Hmmmm ... that would be interesting ..." in a positive way. I am thinking of something similar to AutoLisp in Autocad or Visual Basic in the MS packages. Once Caesar II has made it's run, the program contains some very important information in the form of local element forces and boundary reactions. From these everything else can be calculated.
I often [most days] use the text outputs to copy and paste or transcribe forces and moments into excel to do flange leakage calcs / trunnion calcs etc etc. Having some sort of programming / macro language would help automate some of these tasks. Of course having things automated does not relieve the analyst of any checks and reviewing that require doing, but there is no change there from doing it by hand / excel / mathcad using the results from Caesar II.