ASME SECTION VIII, DIV.1 flanged pipewell calculation question

[timsch]

I am running an ASME SECTION VIII, DIV.1 calculation for a flanged pipewell (partial drawing attached). The flange is 3" 1500# F22, and the pipe is 2" 160S 347SS (customer required combination). The flange will be machined for a full penetration weld and welded from both ends. You can see from the drawing that the material removed from the flange There will be an ER309 butter layer laid on the flange, followed by PWHT of flange, and then the final attachment to the pipe with the appropriate SS filler.

The flange shown is a hubbed flange, but the welding time required will be substantial, so I'm looking at what thickness would be required for a non-hubbed blind flange. With a full penetration attachment as shown, my understanding is that the effective ID of the flange is then the ID of the pipe. If that is the case, then the diameter of the central opening is less than the short span of the flange, and I could use UG-39(b)(1) to determine reinforcement requirements. However, if the pipe ID is not the effective ID, then I have to go with UG-39(c), and ultimately Appendix 14. However, the welds referenced there in UW-13 and UG-34 aren't very representative of the type of weld we're performing in my estimation.

Is the effective ID of the flange given a weld of this nature covered in the BPVC code?

 

[SnTMan]

Suppose you run it as a loose ring w/ ID equal the pipe OD. What do you get?

Regards,

Mike

The problem with sloppy work is that the supply FAR EXCEEDS the demand

 

[timsch]

Perhaps I'm a bit dense today, but I don't see what that would show me regarding my situation. What's the similarity? Using pipe OD would make the bore greater than 1/2 the short span, and I would have to go to UG-39(c).

 

[SnTMan]

timsch. possibly I am the dense one. I gotta admit I don't know what a flanged pipewell is. But:

My assumption is you are designing a flange to a B16.5 pattern. I further assume you are trying to minimize the thickness. And further that you are casting about for a flange I.D. on which to perform calculations.

My suggestion was to run this as a loose flange under Appendix 2, with ID equal the pipe OD and see if the thickness is acceptable. This is sometimes a simpler method than treating as a flat head and having to consider reinforcement.

This whole thing, IMO, I going to hang on your weld metal being of at least equal strength to the F22 material.

Regards,

Mike

The problem with sloppy work is that the supply FAR EXCEEDS the demand

 

[marty007]

Why are you going full-pen on the welds? Customer requirement?

 

[timsch]

A pipewell is a thermowell but rather than the stem being made from bored barstock, it is made from pipe. We make multipoint thermocouple assemblies using pipewells, with the TC's running inside of the pipe. You are correct in your assumptions. A loose flange with a hole of significant size in it would probably be a hubbed flange per B16.5. I typically do recommend hubbed flanges just to keep things simple, but in this case the hubbed flange has a significant thickness greater than the standard blind, and many more hours of welding will be required. That is why I'm looking into the required thickness of the blind flange. Appendix 2 requirements generally result in a greater thickness that B16.5, which was proven again when I ran APP2 calcs for this flange, and that was without the hole accounted for yet. So, now I'm at the point where I made the original post.


Full-pen welds are a feature that my company hangs their hat on as resulting in a superior product. No need debating that can of worms, because they are not going to change their mind on that topic.

 

[BJI]

You can achieve the same results with a set-on nozzle, assuming the pipe does not need to be continuous. Much less welding and heat input = superior product. It is still a full pen weld through the nozzle wall thickness, you could also machine a small weld prep into your blind assuming you are making it custom, then butter it, gives you a butt joint instead of a corner joint. You may be able to butter up the weld prep and not need the custom flange.

If it all seems too hard then change to a partial pen weld both sides and get a superior product to the proposed welding detail.

 

[GenB]

This is a non conventional design and may require a destructive test or further analisis
Acceptable to the AI.
'cant use a modified flange as ANSI Std.
Even if camouflage the calculations.

General Blr. CA,USA

 

[BJI]

I don't think a destructive test will be necessary, not really practicable for a flange anyway. The use of a standard flange is for convenience, the design is still custom and requires calculation. There probably should be some additional testing of the HAZ and heat treatment, if the proposed welding detail is used. Could be distortion as well, so machining after welding may be needed.

I would be interested to see what the OP ended up going with.