Sect VIII Corner Joints

[BigTank]

Is a size-on-size corner joint allowed in a vessel of non-circular cross-section? It appears that with the full-penetration joint option in UG-34, size-on-size is precluded by the 1.2*ts requirement. Can anyone find an exception to this in either UG-34 or UW-13?

--------------------------------
Fitter, happier, more productive

 

[gr2vessels]

Please excuse my ignorance, but what do you mean size-on-size and where did you get this expression from? I have checked many codes and standards with affiliation to the world's pressure vessel standards, including ASME BPVC, but I couldn't identify/locate those words. I'll try to add my 2 cents to the forum, but I have trouble with the size-on-size expression.
Cheers,
gr2vessels

 

[BigTank]

That's okay. I think we all have our dilect with specific geometries/materials/and sometimes fittings. It means two plates of the same thickness coming together to form a corner, i.e. 1/2 welded to 1/2 at a 90° angle.

I'm not sure where I picked up the term, but it's not in the code.

--------------------------------
Fitter, happier, more productive

 

[heatexguy]

UG-34 states 1.2*ts but need not be greater than t. Acceptable joint in my opinion.

 

[BigTank]

See, the way I interpret that is that 1.2*ts is a minimum weld requirement that may also drive the head thickness to something thicker than required due to stress concentration at that corner joint. Unless of course both t and ts are grossly oversized, but the code is being conservative in this requirement.

'tw=2*trmin nor less than 1.2*ts but need not be greater than t'

By the UG-34 equations, 't' will most often be an order of magnitude greater than 'trmin', leading the designer to generally choose a plate material for the flat head thicker than the shell material.

There is a large stress concentration at that joint, and the 1.2*ts is accommodating that concentration. Otherwise, there would be no need for the 1.2*ts requirement, just force tw=t max.

--------------------------------
Fitter, happier, more productive

 

[BigTank]

Disclaimer: I'm sorry if I'm overthinking this or being thick-headed. I just think that the combination of 1.2*ts and 'not be greater than t' is hiding a motivation that I cannot find.

--------------------------------
Fitter, happier, more productive

 

[BigTank]

Let me rephrase:

With a 'size-on-size' joint, the weld requirement would not be satisfied as it is worded.

--------------------------------
Fitter, happier, more productive

 

[Duwe6]

BigTank, sometimes the answer to "What color was Geo. Washington's white horse?" is just "White".

Code [UG-34] sez: but need not be bigger than T. Thus a 3/8" x 3/8" Tee-joint need not have a fillet-weld bigger than 3/8" weld size.

Now, the reason it bothers you is probably that you have never seen a size-on-size square corner pressure-boundary weld on material thinner than 1/2" or bigger. Neither have I. A 3/8" shell 2-3 feet in daiameter will have a 'flat-cap' 1" to 1.5" thick.

 

[BigTank]

I suppose you're right, Duwe6. I understand the popular interpretation, but I still have a problem with the wording in that part of the code.

I'm only offering this in the effort to clarify my thinking:

'tw=2*trmin nor less than 1.2*ts but need not be greater than t'

COULD be:

'tw=2*trmin but not less than 1.2*ts unless greater than t, then tw=t.'

--------------------------------
Fitter, happier, more productive

 

[GenB]

hi.
1.2 the t of the weld applied to the corner joint
if the inner plate is thicher than the shell, not a problem, you just fill the
joint 1.2 the st and you're in compliance,
not greater than t means once you comply with the 1.2 you need no
more weld deposit.. usually it is filled all the way if the plate is not too thick that it will add considerable cost. there are other rules if the plate is to be left unwelded

if the inner plate is the same t as shell then you will have to insert the
inner plate far enough into the shell to fill more weld,
a fillet will suffice.

 

[BigTank]

GenB, I initially thought the same thing. The only problem I have with that is understanding the benefit of that fillet weld beyond the outside of the flat head.

--------------------------------
Fitter, happier, more productive

 

[GenB]

That is beyond Code question.
ASME compliance by putting the right amount of weld,
to hold the pressure. and that is the only way as the shell has to be covered, so you cannot just add weld w/o shell
so you put the plate inside the shell... to comply.
not only that, the fillet weld has to cover the groove weld.

 

[heatexguy]

Nowhere in the Code does weld size drive up plate thickness. Plate thickness is determined by pressure. In the case of equal thickness corner joint, tw=t.

UG-34(d) " The size of the weld tw in sketch (g) shall not be less than 2 times the required thickness of a seamless shell nor less that 1.25 times the nominal shell thickness but need not be greater than the head thicknes...".

Shell thickness = 3/8", head thickness = 3/8", weld size = 3/8"

 

[BigTank]

So we can agree that in order to satisfy the weld requirements for a size-on-size joint, the head must be set back into the 'shell' enough to get a fillet on top of the groove weld. I can't imagine what advantage this would have in strength of the joint, but I could suppose it's an unintended consequence of the joint requirements when a size-on-size situation is applied to the code here.

I feel it is my responsibility to understand the engineering inherent in the code. I can understand why the weld would not need to be greater than 't'. I question this because the wording can be interpreted at least a couple of different ways. I understand the popular interpretation, I'm just asking if anyone has explored an engineering verification.

--------------------------------
Fitter, happier, more productive

 

[heatexguy]

Lets assume we are designing a rectangular box header for an air cooled heat exchanger per Appendix 13. Calculations determine that the long side plates (tubesheet and plugsheet) need to be 1.0" thick. It is determined that the short side plates (top and bottom plates) need to be 0.625" thick. The weld joint between these plates is as UW-13(c). All is well so far. Now for the end closure (or end plate) UG-34 calculation sets the nominal thickness at 0.375". We now have plates of 3 different thicknesses. The end plate has weld joints at the top and bottom plates AND at the tubesheet/plugsheet. There is no reason for me to increase the end plate thickness to 1.25" just so I have 1.2 times the tubesheet thickness as weld. Same with top and bottom plate welds. This is where "need not be greater than t" comes into play. The end closure on a rectangular box header is almost always the smallest and thinnest plate. There is no reason to nearly triple the thickness to meet weld criteria that clearly states "need not be greater that t". We have yet to have our design questioned by the Authorized Inspector and have yet to have a failure.

 

[BigTank]

I agree that there is no reason to nearly triple the thickness, however the weld criteria clearly states that the tw=2*trmin NOR LESS THAN 1.2*ts. If it is less than 1.2*ts, the weld fails the requirement. In other words, for a size-on-size joint consideration with the words 'nor less than', the requirement 1.2*ts and 'not greater than 't'' are mutually exclusive.

Am I the ONLY one who can see this?

--------------------------------
Fitter, happier, more productive

 

[BigTank]

To rephrase: am I the only one that can see it this way...i.e. (3) requirements chained together, and in this case (size-on-size) the last (2) being mutually exclusive?

Something that has not been mentioned is the consideration of the stress concentration(s) at the head corners when considering that the 2 adjacent sides of the head are welded to different plate thicknesses.

I also think UG-34 is designed for the head to be bound on all sides/ around the circumference by the same thickness 'shell'.

--------------------------------
Fitter, happier, more productive

 

[BigTank]

Looks like I'm the only one that can interpret it this way.

--------------------------------
Fitter, happier, more productive

 

[GenB]

BigTank, the code is clear it just happen that you cannot interpret it well.
need not be grater than T,
if you comply with the 1.2st and the weld size first; then you need not be greater than T only if the ht is greater than the required. say you meet the required t 1/4", and yiour ht is 3/4', you only you do not need anymore weld pouring.
but if you decide to fill it up the groove, you need not be greater than T. it is very clear. it is a matter of interpreting the Code if you do not understand it.
I know it because I do a lot of flat heads designing.
All sections of the Code have same or similar requirements
as I have most stamps S,U,H,M,UM,R

 

[BigTank]

I agree. The code has more than one consideration for the word 'nor', however I do not.

if you comply with the 1.2st and the weld size first

A size-on-size joint as I've described above does not comply with the 1.2*ts.

--------------------------------
Fitter, happier, more productive