Please help me validating following queries on Large diameter Branch connection:

a) To satisfy the D/T<100 need of ASME B31.3 for SIF evaluation easy option of increasing run pipe thickness to fall with in D/T<100 and by doing this I can directly use SIF provided by code

b)Other alternative is to have Reinforced branch connection using Reinforcement pad (RF). Even though the total thickness of Pipe and RF pad will satisfy D/T<100, is it accpetable. If yes, can some one suggest how big the RF should be to take care of local stresses.

Thanks in Advance

Any views on this one. I hope I made my question clear enough

Thanks in advance for help

You are outside the scope of the piping codes. Your best course of action is a finite element analysis to determine the appropriate SIFs (such as FE/PIPE provides).

Additionally for large D/T and d/D ratios, DO NOT use pad widths equal to d/2. Use smaller widths on the order of 2.5*SQRT(rt) to d/4. This also helps avoid plug welds since the pads widths will typically not exceed 10" (250 mm).

I know the minions are saying, "But Bob, the CODE area replacement rules show and we always use d/2." I performed an FEA Pressure rating study about 4 years back on headers ranging in size from 24" NPS to 96" NPS with reinforced branch connections >= D/2. Of course I had done this for years earlier based on case by case FEA work.

What happens as far as pressure loadings is concerned is that the pressure capacity will DECREASE AS THE PAD WIDTHS increase past the above recommendations.

This is due to the header buldging/straining at the pad edge. For smaller pads there is less of a buldge since the header is not constrained (remember barrel hoops) as much as a gigantic d/2 pad.

Thanks to Tony Paulin's (PRG) FEPipe released in the early 90's, the stress engineer has a powerful tool to provide practical and safe designs instead of relying on wife's tales, rules of thumb and loud mouth "experts". But you still need to continuously educate yourself on the use of these tools and more importantly the proper interpretation of the results.

Be careful out there and show your work.

Hello ver43138

a) There is note in B31.3 Appendix "D" that cautions the user that the use of Code provided SIF's should be limited to piping systems in which the piping has a D/T<100. This implies that if the piping system at issue comprises piping with a D/t ratio greater than 100, SIF's calculated using the equations of Appendix "D" might not be valid. I see two alternatives to satisfying this D/t requirement so that I might use the SIF that I calculate using the equations of Appendix "D". One alternative is to increase the thickness of the run pipe such that the D/t ratio requirement is met. I assume that if I do this, I can use the SIF that I calculate using the equations of Appendix "D".

b) The other alternative that I see is to install a reinforcing pad that complies with the Code rules for area replacement, thereby increasing the effective run wall thickness local to the branch intersection such that the sum of the run wall thickness and the reinforcing pad thickness (in the affected area) would satisfy the D/t ratio for the run pipe. I assume that if I do this, I can validly use the SIF that I calculate using the equations of Appendix "D".

Regarding a):

First of all what sort of “fitting” are you considering? Is this a fabricated (by welding) branch connection or is it a manufactured (by forging or hot forming) “Tee”? There are advantages to the later in that there is a degree of “fiber continuity” at the formed “crotch radii that cannot be achieved in “cylinder-to cylinder” configuration of welded fabrication. Next consider that there are limits to the valid application of beam theory.

What are SIF's? They are coefficients that are applied to BEAM THEORY calculations to calculate an "effective bending stress" in a local area for the purpose of addressing fatigue (reference Markl). Now, you must remember that the SIF equations provided in Appendix "D" are based upon data that resulted from testing NPS 4, standard wall thickness, carbon steel piping and pipe fittings. The data was extrapolated to apply to piping of other sizes. It is not too much of a stretch of credibility to accept such extrapolation down to NPS 1/2 pipe but when extrapolating the data to apply to pipe sizes greater than NPS 4, one must consider what the limit of credibility must be. Testing that was done at Oak Ridge subsequent to the Markl (and team) testing shows that at about NPS 14 the Code SIF's may no longer be conservative (various D/t ratios are the determining factor here). One question to be considered is: "at what D/t ratio will the bending stresses calculated by beam theory no longer be the highest stresses affecting the pipe? Clearly, it can be seen that local membrane stresses will at some point be of primary importance and beam theory will no longer be a valid methodology for calculating the limiting stresses. The designer using the Code had to be reminded to consider these limitations and therefore, the note was added.

I would favor the a) approach because your are stiffening the run “beam” for some distance away from the affected area and thereby assuring that it still acts as a beam and also because increasing the section modulus of the run pipe reduces the degree of “ovalization” local to the branch intersection that would occur if a bending moment were applied to a beam of lesser thickness (with that, I may have broken my standing record for the length of a sentence). Even with this approach I might want to apply an SIF that is 20 percent greater that the SIF’s calculated using the equations of Appendix "D". As is the case with all fabricated branch connections, good fit-up, and welding (as assured by adequate NDE) is indicated. Testing by Rodabaugh and Woods has shown that the weld quality comes to the fore in determining the survivability of fabricated branch connections when cyclic loadings are applied in testing.

Regarding b):

I have never been enthusiastic about the assumption that when a “doubler plate” is welded to the OD of a pipe or a vessel it provides the same strength as a plate of the same continuous thickness. Yes, I know it is an assumption that is commonly used at reinforcing plates applied to vessel shells at nozzles when WRC-107 type analyses are performed. But the larger the reinforcing plate, the further away the attaching welds will be from each other. If the reinforcing plate were to be drilled prior it being rolled to the contour, and prior to attaching it to the shell, and plug welds were provided to lessen the distance between the attaching welds, I would be more inclined to agree with the “same effective thickness” assumption. But of course that is not commonly done. And the reinforcing plates at nozzles are usually small enough that the separation of the attaching welds may not be an issue.

As we all know, reinforcing pads on piping branch connections are intended to address circumferential stresses (“hoop” direction) and all the “replacement material” must be placed in the “zone of reinforcing” the edges of which are within a distance one half the OD of the branch pipe from the branch intersection. Materials outside the prescribed “zone of reinforcing” are not to be considered as reinforcing. Markl’s testing indicated that branch connections reinforced in this manner had better survivability than did unreinforced fabricated branch connections when cyclic loadings are applied in testing. We have no evidence to support the notion that reinforcing pads of diameters greater than the prescribed “zone of reinforcing” would be more survivable. Maybe yes, maybe no.

But I would contend that for D/t greater than 100 with a reinforcing pad your calculations would be closer to the truth if you used the WRC-107 (et. al.) approach to evaluating the stresses at the intersection and just forget the SIF’s entirely. And ideally, just use Paulin's Nozzle Pro for these analyses of local stresses and apply the rules Of the ASME B&PV Code, Section VIII, Division 2. The CAESAR II model would have a Code flexibility factor of 1.0 at the branch connection so the forces and moments that you take into the “local” Nozzle Pro analysis would likely be conservative. I have nothing against reasonable conservatism.

Now all we need to do is tie this into our trunnion analysis “discussion” and this thread could last forever.

Geez you guys certainly did seem to make this complex.... just kidding!!

One more thing a D/t ratio of 99.9999 is just as bad as 100 and thus the warning was never intended to be a less than 100 no worries sort of cut-off!

Proceed with caution as you move towrds the 100 mark!

Hello again,

Further to Bob Zimmerman's comments there is a really interesting Welding Research Council Bulletin (335) that was done by Ev Rodabaugh that will really open your eyes on the subject. It is WRC Bulletin 335, A Review of Area Replacement Rules for Piping Connections in Pressure Vessels and Piping, August, 1988, by E.C. Rodabaugh. It is well worth reading if you can find a copy.

Also, one of the lesser known papers by A.R.C. Markl looks at branch connections that are subjected to cycling pressure. That paper is "Why Branch Connections Fail", A.R.C. Markl, H.H. George and E.C. Rodabaugh, 1955. This is just one of the papers that are included in the Tube Turns Book "Piping Engineering". As we all know, all the more well known Markl papers are included in that book and the book IS AVAILABLE for $35. Any student of the art of piping engineering would be wise to get this book and read it three times a year.

Regards, John.