Proposed Appendix 28 Design - Comments 1

[SnTMan]

Hello All. A proposed box header is made of a welded rectangular "ring" with a bolted tubesheet at one end and bolted cover at the other end for which the ring is drilled and tapped on the faces. The dimension of the ring are approx. 92" lg x 14" wide X 12" tall. Material is 1.75" thick carbon steel.

The corner joints are to be made per Appendix 28 using the double bevel design, per Fig28-1(b), rather than UW-13 as we believe it can be done less expensively.

Joint qualification will be performed per 28-2(a), and while I will take this all up with our AI, I am hoping to help focus the issues in this forum, as we do not have experience with Appendix 28.

1) Is it permitted for the two plates to be the same thickness?

2) Assuming the weld otherwise meets the requirements of the last paragraph of 28-2(a) and the requirements of 28-2(f), is it permitted to backgouge and backweld the inside corner?

3) Any other considerations related to the joint design and construction to be aware of for this design?

Thanks in advance,

Mike

 

[thermmech]

1) I believe that since introduction of App 28, you don't have to have min. 0.25" thickness differential in between header box plates. Typically it used to be... say, 1" T/S, 0.75" end plates, 0.5" wrapper plates. Now all the plates can be of the same thickness.

However, you have to have a welding procedure that is qualified for that type of joint.

I remember that on most of the AC jobs we could have very thick wrappers, but only on some of them we had to have a bevel on both the wrapper and the T/S. On none of them did we have bevel on both sides of *wrapper* plates. On pass partition/stay plates thicker than 5/8" we would have a double bevel.

2) In principle, you should be allowed to do that in order to get good/complete penetration. However, whether you are going to do itdepends on the WPS-PQR that you want to use. Typically, on CS boxes there was no backweld. On 304 SS boxes, WPS required us to put a backweld in order to ensure complete penetration.

3) Welding engineer / QC person in your shop should have a WPS-PQR ready to use for the joint. Watch your heat input and speed, and do the long seams at the same time. We have found that boxes longer than 14 ft deform too much during welding, it is hard to get them straight again. This is extremely important in this double-cover plate design (if I understoood the design well), as you will not have a stiffening that usually T/S (tubesheet) provides in the case of single - cover plate design.

Your box is very wide and very deep. Try to stiffen it as much as you can during fabrication with tacked-on plates. I am assuming that the design pressure is really low (say les than 50#), but would also advise you to try to put vertical reinforcing bars to stiffen up the "frame" of the box in the service.

 

[SnTMan]

s27289, thanks.

In this case the seams would be about 12" long to match the 12 in "tall" dimension I gave above (maybe should have said "deep"). The 92 x 14 faces actually are open and form the gasket surfaces. Pressure for this unit is 100 psi. Because of some other "features" of the design additional analysis may be useful.

The reason for my question 1) is specificially the statement in 28-2(a)(2). It is not clear to me whether it PERMITS the plates to be the same thickness or not.

I agree it is going to be important to keep the sides straight, but quantities on this job are such that simple fixtures may be an option.

As for PQR's, etc we should already have something suitable, and I expect some process development to be necessary in order to sucessfully qualify the joint and the welders.

Thanks again, and any other input is welcome.

 

[thermmech]

I must add that I have never seen a design that has both tubesheet and cover plate gasketed/bolted.

You should consult your local AIA re the acceptability of the design, as it may not be covered under App. 13. Appendix 13 covers only welded vessels of rectangular cross-section.

My local authority (Alberta Boiler's Safety Association) questioned me couple of times on validity of plain cover-plate designs, as formulas for calculation of stresses in flat plates of header boxes are valid only if both sides of the box are welded, and not open on one (in my case) or both sides (like in your case).

Further discussion in between ABSA and us was concentrated on us proving that the stress levels in flat plates as calculated by using formulae from App. 13 were uncorsevative, but low enough so that the fact that one side of the box was open - didn't result in excessive membrane and bending stresses/deformation. To me that is the biggest concern. With only one side open, you have a restraing on the bending of the wrapper and end plates on one side at least (as tubesheet is welded to the "frame").

However,if you are open on both sides, you will not have any restraints on the wrapper/end plates.

Don't forget that once you weld such a thick frame - it might bend and twist. The machining of the gasketed surfaces should be done after PWHT, but overall deformation might be so great that the adequate seal with flat cover plates would be hard to achieve.

Regards,
Sean M.

 

[SnTMan]

s27289, we have used this type of design successfully for some other types of exchangers, however it is not an Appendix 13 design. The general design is not the issue, rather just the welding of the corners per Appendix 28, rather than UW-13. I had a visit from our AI earlier today, he is looking into this.

Fortunately I don't believe we are going to have to register under ABSA.

And yes, machining after PWHT is a must for this construction.

Thanks again.

Mike

 

[thermmech]

You're welcome, Mike.

I guess I somewhat went off the topic :-D

To repeat myself, I believe that welding of the plates should not be an issue providing that you have suitable WPS-PQR's registered with your local authority.

Regards,
Sean M.

 

[SnTMan]

To you last statement, agreed, and to your first, its not always easy to tell what is OT in mid-ramble.

Mike