Recommendations on type of weld to use

[MikeG7]

Hi

I have attached a sketch to show the situation. Basically we need to split a shell due to manufacturing reasons at the location of an internal plate. Can anyone give a recommendation on the type of weld to use to at the joint?

This is an ASME VIII Div.1 vessel. The shell OD is 22 in so it is possible to get inside to add an internal fillet weld on the open end.

Has anyone encountered this situation before?

What is of concern is that if shell sections are attached by full pen "V" groove weld full penetration, the shell sections would be reliably joined, but the flat plate will have welding only around its outer edge. Not ideal to me!

Thanks

 

[SnTMan]

MikeG7, can you make your plate OD roughly equal to the vessel OD and groove weld thru both sides?

Regards,

Mike

 

[racookpe1978]

Your concern seems unwarranted:

you have a very thin 1/4 round plate of diameter 20 inches being joined to a 20 in dia flat plate 1/2 inch thick with a large (but unspecified) hole in the middle of the flat plate.

So: there is no pressure difference across the flat plate. A 1/4 fillet weld (one sided or even two sided) could hold the flat plate in place and be far stronger than the wall of the pressure vessel.You report there is access inside for welding - REMEMBER TO VENTILATE AND PROTECT THE WELDER.
There is little pressure inside the 1/4 inch thick pressure vessel. (If there were, you'd need a thicker wall PV.)

There is very little wall thickness for a weld prep at all - best would be a simple V-bevel with a 1/16 - 3/32 flat land and a 45 deg bevel on the remaining 3/16 thick wall.

The 1/2 thick plate could easily and very conveniently serve as the "backing ring" for the outside weld for simplicity and added strength of the 1/4 wall weld. If so, then add a single-sided fillet weld (staggered fillet weld even) to finish supporting the 1/2 plate on the inside.

 

[dhengr]

MikeG7:
I’m not exactly sure what the “ASME VIII Div.1 vessel” means in terms of the allowable welding details, you guys know that better than I do. But, I would want the outer shell to have the best weld integrity. Is there any pressure difference across that internal pl., what does it do, how is it loaded? I’d make that internal pl. o.d. the same as the shell i.d. and use the pl. as a sort of backer bar. Then the issue becomes edge prep. on the shell halves. If you use your .5" pl., and 2-.125" lands and leave a .25" gap btwn. the shell edges, you probably want to put a 10-15̊ bevel on each edge and fill the groove on the shell. Maybe you could get by with a .125" chamfer on the outer corners of the shells. If you used a .625" internal pl. and left a .375" gap btwn. the shell edges (that means 2-.125" lands), you could probably get by without any shell edge prep. and still have good welding access. The one potential problem with SnTMan’s suggested detail is that welding to opposite faces of a pl., as I think he is suggesting at his joint, can sometimes cause lamellar tearing in the plane of the pl., due to weld shrinkage.

 

[MikeG7]

Chaps thanks both for the input.
SnTman, I like your suggestion, and it is a viable option to consider. Seems to solve a lot of issues for me. Due to the method of assembly, we want to do the internal coating as complete as possible on the lower section and then attach it finally to the upper part (which per your idea can have the 1/2" plate pre-welded to the top shell section) and finish it off with a single full pen groove weld. We will then touch up the paint in the heat affected zone.
racookpe1978:
You are correct as there is less pressure differential across the 1/2" thick plate (a filter element fits in the hole) than as contained in the vessel. We would normally take about percentage of the shell pressure as a DP across the plate as a design basis. The shell contains 150 psi.
Your recommendations for weld prep and treating the 1/2" plate as a "backing plate" would surely suffice for the low(ish) delta P, with added strength by fillet welding.

 

[SnTMan]

Yes, lamellar tearing is possible, so UW-13 requires NDE of the plate edge. Using the plate as a backer for a single sided is workable, but you need to consider effect of joint efficiency on your shell components, see UW-12 (d).

Regards,

Mike

 

[MikeG7]

...and dhengr, thanks also I only saw your post after my reply. With the plate thicknesses as is, the .125" chamfer is no issue.
now I have options, earlier I had issues...
Cheers

 

[tbuelna]

As noted by SnTMan, you should consider the NDE requirements when deciding on which type of construction to use. Putting the bulkhead directly under the perimeter weld location for the two shells is convenient, but will you be able to perform a proper NDE on the finished weld? Locating the bulkhead offset from the shell weld will allow you to weld the bulkhead in place and inspect that weld, and then butt weld and NDE the shells using conventional practice. Butt welds are often preferred for pressure vessel joints since full penetration welds, with no structure behind the weld, and possible access to both sides of the weld, makes most NDE processes easier and more reliable.

 

[GenB]

The design looks perfect by using the ring as backup bar. But if the t. Is not In the shell to support non NDE, then the option is to weld the ring upstreem and close the lower part by itself. Genblr

 

[MikeG7]

Chaps thanks for all the input. The point about access to the main shell circ weld for NDE purposes is something I didn't consider and very valid indeed. In this case, we are designing for no RT. If it is feasible, we will shift the bulkhead plate further into the vessel top or bottom section and perform just a conventional circ. weld, that makes a lot of sense.
I think the point is that the circ. weld between the shell sections is the critical weld.

 

[racookpe1978]

I disagree. Internal NDE on the fillet weld holding the 1/2 plate in position is simple with dye penet test or mag particle exam. No continuous fillet needed either. No Xray.

The outside circumferential weld is the pressure boundary, and the 1/2 plate serves as a good available backing ring for that closure weld with a V-prep and 1/16 (nominal) flat landing.

 

[tbuelna]

If the 1/2" plate bulkhead were placed directly behind the circumferential V-groove butt weld between the outer shells, it would seem likely that there would be an unfused interface between the bulkhead OD and the shell ID right at the outer edges of the shell butt weld. Wouldn't this gap cause problems with NDE of the butt weld using mag particle, ultrasound or eddy current?

 

[MikeG7]

Are these processes, i.e. Mag Particle, Ultrasound or die-pen testing mandatory ? I am designing the main welds for zero RT and using the appropriate factors for "E" in the calculations.