Dimension Plate Arc Segments 2

[joshpjost]

Requesting some input from the community regarding dimensioning of plate arc segments. Design requires fabrication of rings comprised of individual plate segments. Segments will have edge preparations and will be welded together. Careful dimensioning is required to ensure arc segments during fit up meet the requirements per weld joint designs.

I have attached two different dimension conventions for the plate arc segments. Scheme A controls the overall arc angle of the part and provides a biased tolerance to ensure that all plates, when assembled, will contact each other or have a slight gap. Scheme B controls the length of the part at the inner radius and provides end cut angle. Tolerances have been calculated to ensure the plate edge gaps meet the weld joint design requirements.

I need to determine the best method to dimension these two parts. I am concerned that Scheme A’s tight angular tolerances will make inspection very difficult. I am looking for your input and comments and/or questions that require further clarification.

 

[drawoh]

joshpjost,

This is a question for forum1103.

I vote for scheme[ ]A. Use a GD&T profile tolerance all around to control dimensions.

--
JHG

 

[SnTMan]

joshpjost, what can you most easily measure? Dimension accordingly. Maybe the two chord lengths, one of which is shown in "B".

Regards,

Mike

 

[JStephen]

If this is cut out by some numerical process, I would go with option A. Can you tolerance the assembled pieces rather than each section as you cut it?

 

[desertfox]

I would go with scheme A also, however the best method of dimensioning is the one thar allows the largest tolerance band without affecting fit, form or function.

 

[joshpjost]

I am having a hard time trying to understand how a fabricator would inspect a part dimensioned like Scheme A with tight angular tolerance of .06 deg. I expect that the inspection tools would need to be an order of magnitude better than the tolerance. Can you provide an example of an inspection method for large plates like shown above in the attachment?

 

[dhengr]

Joshpjost:
Why aren’t you tolerancing the finished product (ring) and letting the fabricator do his job in the way that best fits his methods, shop and equipment. What is this ring for and how closely does the finished product have to be toleranced? How is it loaded and supported? Do the end butt joints btwn. the segments have to be full penetration welds or can they be PJP welds? Should they be single bevel or double bevel welds, and maybe made from both sides? Do you need run-out tabs at the two edges? If you don’t know how they are going to make it, you don’t know how to tolerance the individual segments to get the right finished ring. In your scheme B, I get a chord length of 85.966" not 86.68" and you wouldn’t make the 22.5̊ end cut the way you show it either. The end cuts are on radial lines, aren’t they?

It’s made out of 2" thick pl., 129.5" O.R. and 112.32" I.R., that’s basically 2" thick by 17.18" wide segments, with end cuts on the radius lines at the ends of each segment; 8 - 45̊segments, end joined to make up a 21'-7" O.D. ring. And, I don’t think most fab’ers. would tolerance the segments the way are showing or the GD&T guys might propose and then hope to end up with a finished ring. You’re putting the cart before the horse in doing this, unless you are working for the Gov’mt. They might oversize the segment width by a couple inches (say 19" wide), then nest the segments in cutting from wider plates. They would then lay out the end preps. for butt welding. They would study (run some tests) on how much the butt welding moved the pieces around or to refine the welding sequencing. They might splay the root gap a bit to account for this movement. Then they would weld two segments together, four times, and finally joint these quarters on a large layout table. Others might cut the segments out of mill width rolled plates as a trapezoids with 22.5̊ end cuts, and weld these segments together to get a slightly different rough ring. If those tolerances are fairly tight, you might want to machine the circular ring out of these rough welded rings. If the tolerances are slightly looser they might just cut the O.D. and I.D. with a cutting torch on a trammel beam. If the tolerances were slightly looser and they had considerable experience with this kind of work they might be confident enough to cut the segments to final width before welding. Remember, tight tolerances on a fab’ed. ring like this are very expensive and I would try to make my design much more tolerant of a little variation. I can always weld a shell to this ring and put a bolt circle on it, even though the edges vary a little from a perfect circle. Flatness of the finished ring may be a more important consideration for the next fit-up step. You may not want to weld the two halves together until you transport them to the job site.

 

[tbuelna]

joshpjost-

I would agree with dhengr's comments. Since your next assembly involves substantial amounts of welding and thick sections, it would be a waste of effort to require tight tolerances on the individual plate sections. If the finished ring weldment needs to be dimensionally accurate on the ID/OD, you should trim these surfaces after welding is completed.

Looking at your sketches, it appears that either of your dimensioning schemes can result in gaps along the weld joint between plate edges. The first pass during welding will cause these free edges to suck together as the weld locally solidifies and contracts. Of the two dimensioning schemes, the one using linear dimensions to define the end point locations of the joint plane seems more practical.

 

[chicopee]

I would weld the entire assembly in a fixture that defines the end product. When the pieces are assembled in a fixture, you are certain that the final product will meet your specifications.

 

[desertfox]

Looking at the two options again I notice that in option A for the 112.32 radius the chord length based on 45 degrees and then 44.64 degrees would be 85.96mm (in line with dhengr's calculation and 85.85mm using the tolerance angle.
The bit I don't understand is,if the angular tolerance is a requirement in option A then why as it been relaxed in option B.
To me option B cannot be compared to A because the values it produces are different to those using A.

 

[dhengr]

Desertfox:
If the units of measure for the chord length are really 85.96mm, you might want to consider cutting or machining this ring out of a single steel sheet, rather than as 8 segments, since the darn thing is only about 10.25" across. It must be getting late in the day over there.

 

[desertfox]

Lol I was in metric mode, anyway that should say inches but the point I am making is that both options give completely different tolerances and sadly we don't actually know which one gives the best solution.
I do think however the tolerances are to tight for a fabricationer, even .2 of a degree is hard to maintain on a fab.

Regards

Desertfox