Stress analysis - different thicknesses of pipe / elbow / tee

[kurtz]

If you are doing a pipeline model in a stress analysis software and for example the pipe is 219 x 5mm, but the elbows and tees are 219 x 6mm.
Do you model/calculate with different thicknesses for each elements or model the pipeline of the same thickness 5mm for pipe and elbows/tees.

 

[balamurugan02]

kurtz, It is recommended to model the piping components at their actual sizes (As built / available market sizes) in order to obtain accurate dead loads, after all it is your primary load for support designs.

 

[kurtz]

It depends of the software, for example modeling different thickness of reducer in Autopipe it's quite complicated and time consuming. Impact on the result - negligible.

 

[XL83NL]

I use AutoPIPE as well. The modelling approach to modify this is not that complicated, imo, unless you have very big models. If the impact is negligible, then why do you ask?

 

[pennpiper]

kurtz
(A) Are you in the Engineering office and taking the material thicknesses from Catalogs or Internet searches?
Or
(B) Are you in the piping material warehouse at the jobsite and you are taking measurements from real thicknesses from real pipe and fittings for this specific application?

If you are in the "A" mode, then you must recognize that all your calculations are pure hypothetical. What you see in the charts may or may not be what you actually get six months later.

Don't try to "Nit-Pick" the little things.


Sometimes its possible to do all the right things and still get bad results

 

[kurtz]

1. XL83NL: Do you have experience in the Tee type A acc. to EN10253 in Autopipe? It should be modeled as "unreinforced"? What about type B or ASME elbows? DO you model nominal thickness eg. 219x8mm or the real one? The real thickness of tee 219x8mm is 19,4mm and 14,5mm!
2. Pennpiper: For example: pipe thickness is 3,2mm, elbow, reducer thickness is 3,6mm - I got a comment that the model should be corrected - it is negligible to me.

 

[XL83NL]

kurtz, yes I do have experience. with AutoPIPE, a lot. With EN 13480, a fair share (not too much, fortunately). I have found my way through EN 13480, and as much as I like the code, it still has too many flaws and bugs making it (too) challenging, at times, for users to use. Look up my history of posts where Ive made my concerns towards EN 1480 clear a few times.

We used type B tees on a job 1 time. Spec called for simple 316 (1.4401/1.4404 dual certified). Couldnt find them in the EU, unless we did a mill run (since we needed maybe say 50 tees in all sorts of sizes, that wasnt an option). Had to switch to 1.4571 or 316Ti, not sure which one. Then found out a lot of the tees (I believe the DN50 ones) had a wall thickness of about 13 mm, instead say 4 mm. They were machined from bar.

We didnt use type A on purpose as we then had to engineer a whole lot more. Especially for the PN100 spec we had, this would become challenging. So, I have no experience with type A in EN 13480 using AutoPIPE. In my stress analysis, I model the nominal wall thickness. Please bear in mind both B31.3 and EN 13480 are for new construction. The codes dont require you to model the measured wall thickness, just whats defined in the standards youre ordering to. If the actual wt is 1 mm more than ordered, I usually wouldnt be too concerned. However, in our case, we had a huge deviation. Thus potential stress raiser, as all of a sudden we went from 4 to 13 mm. That had to be checked on mechanical and process (flow) impact. In your case, if you ordered a tee with wt of 8 mm, and got something in the range of 15-20 mm, I would definitely check that; but more from a material receiving point of view. Accepting such fittings should always come with an NCR. The only reason to accept them is when they dont pose an issue to pipe stress anaylis model (outcome), or any other item in your total piping system. It may also cause local flow or porcess issues.

2. Pennpiper: For example: pipe thickness is 3,2mm, elbow, reducer thickness is 3,6mm - I got a comment that the model should be corrected - it is negligible to me.

Who ever says that is an id(&^%

 

[KevinNZ]

You need to remember that B16.9 butt weld fittings do not need to have any particular thickness. Only at the butt welded ends do they need to match the nominal pipe thickness. The rest of the fitting should be thicker to ensure the fitting burst pressure is higher than the pipe.

The B31 codes say use the nominal thickness for the fitting in the stress model. Only in short runs between fittings would overly thick fittings but high stress into the attached pipe.