Shell Course Fit-Up: Circumference, Offset and Datum
You are fitting up a 60 in. ID, 3/8 in. wall carbon steel shell course from two half-shells rolled on a pinch roll. The drawing calls for 60 in. ID with a ±1/8 in. tolerance, and longitudinal and circumferential seam alignment must meet ASME UW-33 (1/4 t maximum offset for Category A and B joints). A pi-tape reads 188 5/8 in. on the new course and 188 1/8 in. on the adjacent course already welded in place. Nominal circumference is 188.50 in. A 12 in. nozzle is laid out 3 in. from the new course's longitudinal seam, and the customer drawing dimensions that nozzle from the vessel's marked 0° datum line. Plate temperature in the sun is 130°F; the tape was calibrated six months ago.
1. Convert the two pi-tape readings to diameters and decide what passes and what does not.
Read the worked answer
Convert first, then judge
- Course A: 188.625 / π = 60.04 in. diameter.
- Course B: 188.125 / π = 59.88 in. diameter.
- Nominal is 188.50 / π = 60.00 in.
Against the ±1/8 in. tolerance (59.875 to 60.125 in.)
- Course A at 60.04 in. passes.
- Course B at 59.88 in. passes, but only by 0.005 in. - effectively at the limit, and the pi-tape itself has reading error of that order. Treat B as marginal, not comfortable.
Relative to each other
The two courses differ by 0.16 in. in diameter, or about 0.08 in. radially. That is a real mismatch that will show at the circumferential seam, and the 0.005 in. margin on B means any re-roll, weld shrinkage or tape error pushes B out of tolerance. Also check ovality separately: a pi-tape proves average diameter, not roundness. Measure across several diameters before accepting either course.
2. The circumferential seam shows a visible step. Using the tolerance given, decide whether it is acceptable, and if not, choose between re-rolling and a transition taper.
Read the worked answer
Apply the tolerance
UW-33 limit for Category A/B is 1/4 t = 0.25 × 0.375 = 0.094 in. The calculated radial mismatch between the courses is about 0.08 in., which is under the limit in theory - but the measured step at the seam will be larger than the calculated mismatch because of peaking at the longitudinal seams and ovality of the two courses in different planes.
Decision process
- Measure the actual offset with a bridge-type alignment gauge at four points around the circumference, not just at one seam intersection.
- If the measured offset is at or under 0.094 in., proceed: align and pull the seam with strongbacks, tack, and weld to the qualified procedure.
- If it exceeds 0.094 in., re-roll. Do not fix it with weld buildup on the plate edge - buildup is not an accepted correction for a mismatched shell course.
- A transition taper is for thickness changes between courses, not for correcting a rolled diameter. Trying to use one here hides an out-of-round course that will show up at the next seam and at nozzle fit-up.
The 0.005 in. margin on course B is the deciding fact: re-roll B now rather than fight every downstream seam.
3. Evaluate the nozzle layout as described and state how you would lay it out correctly and what temperature effect you would and would not correct for.
Read the worked answer
The layout as described is wrong in principle
The drawing locates the nozzle from the vessel's 0° datum. Measuring 3 in. from the longitudinal seam and calling that the position assumes the 0° mark coincides with the seam. It usually does not, and if the half-shells were rolled from plates whose orientation was not tracked, the nozzle will land rotated by an unknown angle - a mistake that is found at hydrotest or at pipe fit-up, not at layout.
Correct method
- Establish 0° from the marked datum, transferred to the course as a positive punch mark or soapstone line with a level or laser reference.
- Set the seam to its specified orientation (or note its actual angle and confirm the drawing permits it).
- Check minimum distance from the nozzle edge to the longitudinal and circumferential seams per the drawing and code.
- Wrap a flexible pi-tape around the actual circumference to convert the arc distance to a chord or arc dimension, then mark and verify with a second measurement from the opposite side of the shell.
Temperature
A steel tape over 12 in. at a 40°F difference expands roughly 12 × 6.5e-6 × 40 ≈ 0.003 in. - negligible; do not "correct" it. The plate at 130°F expands far more around a 188 in. circumference, but the part will be measured and fitted at that temperature, so the layout should be made on the part at its current temperature, and the datum should not be re-established from a cold reference.