I suppose the easyness to weld it...
Most of the times it is easier to weld a set-in nozzle.
Its also easier to position.
For ASME calculations for reinforcement, set-on and set-in are not very different.
The biggest difference in reinforcement calculations would be is you have a shell in very strong material, and a very weak nozzle material. Then, a set-on might be more interesting.
CobusvanJ, it is pretty much a matter of shop preference. HOWEVER: Set-in nozzles pretty much can always be double welded, where this is not usually possible for set-ons.
Many of the owner specifications for refinery equipment do not allow set-ons, so this is another consideration.
As McJe says, the reinforcement calculations are slightly different. As McJe also says, set-in is maybe easier to fit to the correct projection, although more grinding might be needed, if it is to finish flush with the ID.
Another factors is that set on type is not prefered because of fear of promoting delamination of the shell at the nozzle cut out. This is why if you're doing set-on type then many specs will insist you MT the steel at the cut out to check for any surface cracks while a set-in type do not need such NDE. More strigent spec will insist on UT at this location, in addition to MT.
Set-on is normally allowed for 2" nozzles or smaller due to better welding geometry (ie., you don't want to have a weld bigger than the piece you're welding on) and welding rod accessibililty issue. Of course, these issue are not an issue when you talk about nozzles larger than NPS6, in most cases.
You have hit on one of my many pet peeves. When deciding whether to use paste or plug nozzle designs be sure to consult with your NDE people if volumetric inspection is required. Too many times they (we) are only consulted after the fact, with a geometry that is difficult or impossible to inspect. I refer you to ASME Section VIII-2 Figure AD610.1 for examples of "difficult to radiograph nozzle configurations".
A paste nozzle can only be fully radiographed if the ID of the nozzle is large enough to allow film to be placed inside or if a buildup is made to the OD of the shell prior to nozzle attachment. If neither of the above are present you have created, from an RT perspective, a Tee weld. The fusion line on the shell side cannot be made visible on the film.
Another factors is that set on type is not prefered because of fear of promoting delamination of the shell at the nozzle cut out. This is why if you're doing set-on type then many specs will insist you MT the steel at the cut out to check for any surface cracks while a set-in type do not need such NDE. More strigent spec will insist on UT at this location, in addition to MT.
this only us normally on thick vessels what are MT or UT inspected at bevels and cut openings anyway.
You have hit on one of my many pet peeves. When deciding whether to use paste or plug nozzle designs be sure to consult with your NDE people if volumetric inspection is required. Too many times they (we) are only consulted after the fact, with a geometry that is difficult or impossible to inspect. I refer you to ASME Section VIII-2 Figure AD610.1 for examples of "difficult to radiograph nozzle configurations"
vesselfab,
you are correct there is no value in doing MT on thin vessel. Of course the delamination issue is more applicable for wall thickness > 1/2" wall. The cut-out for a set-on nozzle is not beveled as you know.
This whole discussion is a general discussion since the originator of this thread gave no details on what kind of vessel he is working on and for what service.
For set-in, fabrication and NDT are easier. If roundness of the vessel body is critical (like headers, manifolds ...), set-on creates less weld shrinkage and out-of-roundness.
b) Nozzles in Cylindrical and Conical Shells: radial nozzle in a cylinderical shell subject to pressure is outloned, step by step, in paragraph 4.5.5 ASME CODE SECTIN VIII, DIVISION 2, FOR NOZZLE WALL ABUTTING THE VESSEL WALL. See Attach file The attach sketch demonstrate the terminology used.
