By G.S.Crisi 
Date 05-15-2011 19:02
Edited 05-15-2011 19:08
If the ring OD is 5200 mm and the duct OD is 5000 mm, the ring width will be a few millimeters less than 100 mm, because you'll have to leave some clearance (say 3 mm in the diameter, or 1,5 mm in the radius) between the duct and the ring.
The weld will heat up some millimeters of material beyond it, both of the duct and the ring. Let's suppose that about 15 millimeters of the ring width, or perhaps 20, will be heated up, will expand a little bit and, after cooling down, will return to its original dimension. The rest of the ring (80 - 85 mm wide) will remain at ambient temperature.
In my opinion, the dimensional change of the ring OD, if any, will be negligible.
Giovanni S. Crisi
Sao Paulo - Brazil
PS. Just a curious question: where is going to be used this huge diameter duct? Perhaps in a sulfuric acid plant? Or will it be the penstock of a hydro electric power station?
Shrinkage; a volumetric change due to solidification is on the order of 3% of volume of the liquid undergoing the phase change from liquid to solid. Contraction is approximately 6.5X10^6 in/degree times the change in temperature.
Total shrinkage and contraction will be affected by a number of variables including the degree of restraint offered by the joint and anything that restricts the free movement of the weld during the welding operation.
While I can’t provide chapter and verse, I believe there are some guidelines in the Lincoln's Design of Welded Structure and/or Welding Procedure Handbook.
Best regards – Al
Accurate estimation of shrinkage can be made through computer model for such a problem, number of programs are available to do the same (like VrWeld from Goldak technologies, Sysweld from ESI). - Best, Rakesh
Al,
6,5 x 10 elevated at the sixth power or minus sixth power?
Giovanni S. Crisi
Good catch Giovanni!
I’m sure I could put that material to good use if my coefficient of thermal expansion was correct! Maybe it was Unobtainium I was thinking of.
Best regards - Al
maybe it's the formula for Adamantium.
By DaveBoyer 
Date 05-17-2011 07:27
Edited 05-17-2011 07:29
In the tool & die shops We used 6.4 micro inch per inch per degree Farinhight for expansion/contraction of steel.
The simple answer is no.
Al said it best, there are guidelines.
There are too many variables involved.
Repeatability and emperical data are the best source for anticipating shrinkage.
JS55 has the correct answer, shrinkage is hard to predict as there are too many factors. You either go all in and use FE models like sysweld or generic deformation simulators like deform or abaqus. Or you use a simple approximation.
Disrtortion is even harder to predict.
Boon,
Rather than worrying about shrinkage, which in my opinion is negligible, if any, I suggest that you be concerned about warping (distortion) of the ring as a consequence of welding.
I also suggest that you envision a welding procedure suitable to minimize that ocurrence.
Giovanni S. Crisi
Yes, shrinkage should be manageable once you experienced it but distortion is worst.
Btw the large diameter ring and duct I mentioned is just an example. Actual frame is not as simple and I am referring to section between GT and HRSG.
I agree with Giovanni thet the actual shrinkage won't ammount to much, but the distortion may make assembly dificult.
