UW-12, Appendix L joint efficiencies

[TJOrlowski]

Background: We're evaluating the ASME VIII, Div.1 mechanical software from a very well known company. We have the full version of the program they currently have on the market, but for a limited time.

Yesterday I was rating some old jobs, comparing our calculations ('07 edition), to the program's (using the same edition). I plucked a TEMA type BEM job, where the shell was 24" OD, attached to tubesheets on both ends [Fig. UW-13.2 - sketch (i)]. The shell had only one(1) course - tube length was ~120. Material was SA-53 Gr. B carbon steel ERW pipe. No radiography on the job.

After running, I was looking at the UG-27 and Appendix 1-1(a) calcs for the main shell cylinder. The report showed joint efficiencies for the circ stress AND long stress of 1.0, With stress values of 14,600 for long and circ. The stress value already incorporates a .85 efficiency factor from Section II (note G3), but can 1.0 be right for the UG-27 and Appendix 1-1(a) efficiencies? I questioned them on it and they are standing behind it - their reasoning being that there are no butt welds in the shell, so E=1.

After scouring UW-12 and Appendix L, I can't reconcile it. They're also using much more aggressive stresses and efficiencies on the same material when used as a front or rear head cylinder, welded to an Appendix 2 flange, and a 2:1 elliptical head.

What am I missing?

-TJ Orlowski

 

[SnTMan]

TJOrlowski, for the shell section I would agree that E=1, given that an efficiency is contained in the allowable stresses. Welded tubing is regularly treated this way as well. Additionally it is permitted, for longitudinal stresses, to divide the allowable by 0.85.

Not sure what is meant by "more aggressive stresses", but for butt weld circ seams (in this context) E is only equal to 1 if each seam is spotted. The dreaded UW-11(a)(5)(b).

Regards,

Mike

 

[TJOrlowski]

Not sure what is meant by "more aggressive stresses", but for butt weld circ seams (in this context) E is only equal to 1 if each seam is spotted. The dreaded UW-11(a)(5)(b).

This is where I'm getting hung up. No radiography. Suppose we have a 20" OD SA-53 Gr.B ERW pipe, using it as a channel in a TEMA Type A front head, where both ends of the channel have appendix 2 ring flanges, attached with double fillet welds. So for my Cat C&D calculations, table UW-12, type(7), note(5), I can use E=1.0

But for the Appendix 1-1(a) circ (Cat A) calculation, the argument would be that because the material stress has a .85 efficiency built in, I can use E=1.0 for my t=(PRo)/(SE+.4P) calculation, even though I haven't satisfied the requirements of UW-11(a)(5)(b)?? (we don't regularly RT long seams on pipe that comes from the mill)

-TJ Orlowski

 

[SnTMan]

TJOrlowski, I am reconsidering my previous reply now. I also cannot find any explicit basis in the Code for using E=1 for a welded pipe shell, unless it perhaps exists in an interpretation. Nor can I find in SA-53 where any radiography is requred. My apologies. I will only plead that I don't normally use welded pipe for shells.

On the other hand it doesn't seem right to me to have to apply what amounts to a joint efficiency twice to the same component.

Any help out there?

Regards,

Mike

 

[jamesl]

There was a discussion by HSB on this issue. They used interpretation VIII-1-89-97 to support E=0.85 for ERW pipe when the category B&C joints have no radiography. They pointed out the code is intended to have double penalty when the ERW pipe already has 0.85 built-in its allowable stress.

It is true E=1 if the corner joints are fillet welds.

 

[TomBarsh]

This isn't 'my' software but I find the values correct.

Previous posts cover the paragraphs, etc, and this is summarized in the flowcharts of Appendix L.

See Figure L-1.4-1. Follow the "No" path from "Full RT Mandatory" to the "Seamless Section", follow the "Yes" branch (because we are to treat the welded pipe as a seamless pipe per UW-12(e)). Follow the first branch that drops down and extends to far right to "Category C Corner Joint". The flow chart indicates that the Category A 'seam' obtains E = 1.0 and the Category C joint also obtains E = 1.0.

Thus a welded pipe welded to flat covers each end with corner joints would use the tabulated allowable tensile stress, with joint efficiency of 1.0, when calculating required thickness for circumferential stress. And would use the tabulated allowable stress divided by 0.85, with joint efficiency of 1.0, when calculating required thickness for longitudinal stress.

 

[jamesl]

After reading the original post again, it seems the problem is the joint type used in construction. Fig UW-13.2(i) is a corner joint, not a butt weld. Therefore UW-11(a)(5)(b) does not apply. If the joint is one of Figure UW-13.3, then we will have RT or no RT (therefore joint efficiency) issue.

 

[GenB]

.85% per UW12 penalty is inevitable.
the program should indicate in its formula calculation.

 

[SnTMan]

Agree w/ TomBarsh, per UW-12(e) this may be treated as a seamless section. Don't know why I overlooked this yesterday, too much hurry I suppose.

Regards,

Mike

 

[TJOrlowski]

Here's my interpretation:

UW-12(e) Welded pipe or tube shall be treated in the same manner as seamless, but with allowable tensile stress taken from the welded product values of the stress tables, and the requirements of UW-12(d) applied

So now we're in UW-12(d), where:

UW-12(d) Seamless vessel sections or heads shall be considered equivalent to welded parts of the same geometry in which all Category A welds are Type No. 1. For calculations involving circumferential stress in seamless vessel sections or for thickness of seamless heads, E=1.0 when the spot radiography requirements of UW-11(a)(5)(b) are met. E=.85 when the spot radiography requirements of UW-11(a)(5)(b) are not met, or when ...

So ERW pipe is equivalent to seamless pipe, but UW-12(d), and therefore UW-11(a)(5)(b), must be satisfied. Seamless pipe has no CatA butt weld, so UW-11(a)(5)(b) doesn't apply. ERW pipe does have a CatA butt weld, so UW-11(a)(5)(b) is applicable, and we'd have to spot the long seam in order to use E=1.0.

When I read the L-1.4-1 flowchart, it's in my mind that CatA joints still need to meet the requirements of UW-12(d), so when using ERW pipe, the circ stress (long joint) would be E=0.85 when there is no RT.

Thoughts?

-TJ Orlowski

 

[jamesl]

As I pointed out in my previous post, UW-11(a)(5)(b) is about RT of category B&C butt weld joint. Since you don't have Category B&C butt weld, you don't need to satisfy UW-11(a)(5)(b) to use E=1.

 

[GenB]

I see.
.85% per UW12 penalty is inevitable.the program should indicate in its formula calculation.

not for the shell long seam but for the efficiency and degree of examination according to UW-12(d)

 

[GenB]

TJOrlowski there is no- circunf. weld here so .85 does not apply.
other, you do not have to re-test pipe as ''spot'' because the pipe is already cert. (A53E/B) longseam. 15% already taken from stress, see stress tables deductions from seamless.
test spot is for the circ. weld only when attached and buttwelded.

 

[TJOrlowski]

I understand the interpretation of the code, but to me, there's a disconnect when looking at it from purely an engineering standpoint.

Example:
Take a 10 foot piece of 16" OD SA-53 Gr. E/B pipe, weld appendix 2 loose ring flanges to both ends. In this case, I can use E=1.0 or less when examining the circ stress in the middle of the piece.

Now take a 10 foot piece of 16" OD SA-53 Gr. E/B pipe, weld a LWNF to one end, and a 16" OD 2:1 elliptical the other end. In this case, I need to use E=0.85 or less when examining the circ stress in the middle of the same piece.

As far as the middle of the piece is concerned, what has changed?

-TJ Orlowski

 

[TomBarsh]

TJ, you're absolutely right that nothing has changed in "the middle of the piece". But the rules for joint efficiency apply "at the joint". Sort of an application of St. Venant's principle. So the Code allowable stresses are based on condition at the end joints.

For example, take a piece of plate, roll it into a cylinder and weld it. No x-ray. So the joint efficiency is 0.7 (with a type 1 joint). The required thickness is based on the allowable tensile stress and the joint efficiency. But now consider the material "in the middle" at 180° back from the weld seam. This material is affected the same whether the weld seam is full RT or spot or none. Anyway, I think you have a handle on the general principles.

 

[GenB]

"In this case, I need to use E=0.85 or less when examining the circ stress in the middle of the same piece."

Not middle but at the joint,
Code says E=1 if you examine it or .85 if yo don't.
End plates has stress at the end fully so no joint is penalized there.
circ. stress at girt is different.