Weld Shrinkage

[KentCaudill]

Who among you can comment on the amount of weld shrinkage to be encountered in welding large diameter segmented T316L SE pressure vessel heads? The heads are 38'Ø with a minimum thicknes of 1.125" and are fabricated from 13 segments. Do you know of any articles or studies addressing this issue? Please feel free to pass my request along to others if you wish.

Thanks for your help, Kent.

 

[bmoorthy]

Is it a crown and pettal arrangement. Whar is the shape of the Head.

Shrinkage can happen in both directions, that is along the petal joint (reduction in diameter) and along the Crown(Extent would be lesser here)

As far as i have seen, lower thickness material tend to shrink more as compared to high thickness material.

As a general thumb rule for Butt joint we tend to take 1.5mm to 2mm for Long seam as shrinkage (Provided Stoppers instead of Tacks are provided in setup/FitUp stage) If the Fit up is with Tacks on the Groove then you may expect 2.5 to 3mm per joint.

What we do is that at the time of Fitup/Setup the Circumference is kept 5 mm more than the permitted tolerance (Circumference tolerance specified by your client) so that after welding the finished dimension is well within tolerance and facilitate good fit up for the circular seam(Between the shell and Dish) In order to maintain good fitup at the time of Circularseam setup we roll the shell after the dish is formed.
ASME dose not explictly specify tolerance on Diameter of dishend, it is generally specified by the Designer/general spec of your client or licensor based on his calculation and the shape of the Head.

 

[KentCaudill]

Good Morning BMoorthy. Thank you for responding to my inquiry. The head in question is:
ASME 2:1 SE,
11.58 meters inside diameter,
28.58 mm minimum thickness,
50.8 mm straight flange,
Construction: 3.66 meter crown plate with 13 equal petal segments. The joint design is double bevel 60% outside with a root opening of 2.38 mm and a 3.18 mm landing. Tack welds are used for normal fit up.

We plan to roll the shell after the head is welded in order to match shell circumference to head circumference. ASME addresses the allowable over/under tolerances for formed heads in UG-81.d based on the nominal diameter of the head. However ASME doesn't appear to define what is meant by "nominal diameter" and what the permissable deviation from the "nominal diameter" might be.
From your response I understand that the buyer is required to specify a tolerance for the circumference measurement. Are you aware of any industry standard circumferential tolerance to be used in the absense of a buyer specified acceptable tolerance? Does ASME require the buyer to specify the circumference tolerance somewhere in the Code? I haven't been able to find any reference to buyer specified circumferential tolerances in the code.
At the time of Fitup/Setup do you normally allow 5 mm more than the permitted circumference tolerance for all sizes of heads or do you allow 5 mm per foot of head diameter or do you use some other formula.
Out of curiosity are you with a Head Manufacturer/Vessel Fabricator?

Thanks for your help, Kent.

 

[deanc]

Perhaps you should review UG-96,you may be asked to justify your final result. Suggest you also review Part UHA.

 

[KentCaudill]

Good Afternoon DeanC:
Thanks for taking the time to respond to my question.
However UG-96 refers me back to UG-80 and Part UHA discusses the selection of materials for vessel design. Neither gets into circumferential tolerences. But thanks anyway.

Thanks for your help, Kent.

 

[deanc]

I think the info. you need is there. UG-96 does send you back to UG-81 and also the orginal print tolerences. UG-81 gives guidence for your minumum tolerences. Nominal would be taken as the dimension on the print. I sent you to UHA because of the requirments,UHA-44 calcs. to be performed and maybe PWHT.

You need a draft proposal prior to fabrication and reach an agreement with your AIA on the distortion issue.

There may be no set answers. See U-2.

Use your AIA for help with this,that is what you pay for.

 

[bmoorthy]

Hello Kent

ASME does not specify the buyer to specify the tolerance. The diameter is not decided by the vessel (Mechanical)Designer, based on the ID or OD the Mechanical designer gets the thickness. Dia is specified in the process data sheet by the buyer,Hence as regards the increase in Dia or reduction in dia is the progative of the buyer as regards the functionality of the item goes.

You may refer to Chevron spec or UOP standards for further info (These are renowned processor licensors)

I think the intent of ASME is to check the local variations in shape and dents, which are verified with template and the dimensions of template is to be obtained from the figure refered in UG 81/80.

In your case even if the circumference is more by 50mm the template check would pass and as regards ovality one has to measure only the difference. Hence ideally there would be no deviation as regards head goes (unless ofcourse you specify the tolerance on Diameter/circumference)

This situation would lead to increase in diameter which would result in following 2 situations

1) offset during circularseam fitup
2) Change from original diameter (Taken by the designer in his thickness calculations)

Where as BS 5500 is very clear, there they have specified the "Departure from true radius" and if you refer to Q&A of BS5500 in Question 33, various of methods of checking the departure from true radius is specified.

 

[arto]

For info on distortion/ shrinkage, look in the AWS Welding Handbook 8th ed Vol. 1 Chapter 7.

There's also several good books by Masubuchi & some WRC Bulletins too.

 

[KentCaudill]

Good Morning:
Thanks to all of you who took the time to reply. You may find it interesting to note that I posted the same question on the ASME Pressure Vessel and Piping Forum. My post logged 11 reads and 0 replies. It was startling to note that the first 40 posts in that Forum including mine logged 3,740 reads with only 17 replies. Apparently the ASME board is not a good place to post questions.

Thanks for your help, Kent.