From Spec Sheet to U-Stamp: The Complete Vessel Design Workflow

A code-stamped pressure vessel travels a defined path from start to finish: customer datasheet, material selection, design-by-rule calculations, calc package and drawings, Authorized Inspector review, fabrication with hold points, witnessed hydrotest, and finally the Manufacturer's Data Report, nameplate, and National Board registration. Every stage produces documents the next stage depends on, and nothing downstream can really start until upstream is settled.

The Authorized Inspector, the AI, from your Authorized Inspection Agency, is the gatekeeper who has to be satisfied before the stamp goes on. That single relationship shapes the whole workflow more than any other factor. A shop with a good working relationship with its AI moves through review quickly. A shop that treats the AI as an obstacle to get past, rather than a partner to get right, tends to find the same problems again and again, later and more expensively each time.

What follows is the whole workflow, stage by stage, with the Code paragraphs that govern each step and the practical habits that separate a smooth job from a stalled one.

Stage 1, the datasheet and the PO

Everything starts with the customer's process datasheet and purchase order: design pressure and temperature, capacity, service (is it lethal, cyclic, sour), corrosion allowance, and the commercial terms. Read the service description twice. It is easy to skim past a line item and miss something that changes the entire downstream calculation set.

The PO dictates whether a U-stamp and National Board registration are required, which in turn dictates that an Authorized Inspector will be involved and that your shop's ASME quality control manual governs the job from this point forward. The service description drives material selection, MDMT, and whether Division 1 or Division 2 is even the right ruleset for the job. Getting this stage wrong does not surface immediately. It surfaces weeks later, when a calculation someone already finished has to be redone under a different set of rules, which is exactly the kind of rework everyone downstream is trying to avoid.

Stage 2, materials and allowable stresses

Select code materials: SA-516-70 plate for shells and heads, SA-106 for nozzle necks, SA-105 or SA-182 forgings for flanges are the common defaults, though the actual service dictates the real choice. Pull allowable stresses from Section II-D at design temperature, since allowable stress is temperature-dependent and using a room-temperature value on a hot service job will understate required thickness.

The material choice does more than set a price. It fixes the UCS-66 impact-exemption curve that will govern your MDMT later, so a cheap plate grade chosen without thinking ahead can cost you an impact test several weeks and several stages down the line. This is one of the clearest examples in the whole workflow of an early, seemingly minor decision quietly determining a much bigger outcome much later. The full UCS-66 procedure is worth understanding before this stage, not after.

Stage 3, design-by-rule sizing

This is the core calculation set for a Division 1 vessel, and it is where most of the engineering hours on a routine job actually go:

• Shell thickness per UG-27 under internal pressure. See What Is MAWP for how these same equations end up setting the stamped rating, not just the required thickness.
• Head thickness per UG-32, covering 2:1 semi-elliptical, torispherical, or hemispherical geometries, each with its own governing equation and its own trade-offs in material use and fabrication cost.
• External pressure and vacuum per UG-28, using the Section II-D factor A and B charts, with stiffening rings added wherever the math demands them. Vacuum service is a different failure mode than internal pressure entirely, and it is easy to under-engineer if the design team is used to thinking mainly in terms of internal pressure.
• Nozzle reinforcement per UG-37 through UG-40, the area-replacement method covering areas A1 through A5. Our worked example walks the full arithmetic on a real nozzle.
• Nozzle neck thickness per UG-45, a separate check from reinforcement that catches thin-neck failures reinforcement alone would miss.
• Flanges per Appendix 2 where custom, or B16.5 and B16.47 ratings where standard, with standard flanges usually the faster and cheaper path when the rating allows it.
• MDMT per UCS-66, including the stress-ratio reduction. The step-by-step guide covers the curves and the credit in detail.
• Combined loadings per UG-22: weight, wind per ASCE 7, seismic, and liquid static head, plus Zick analysis for horizontal vessels resting on saddles and WRC 107 or WRC 537 checks wherever piping loads land on a nozzle.

Each of these checks feeds the others, which is what makes vessel design genuinely iterative rather than a simple checklist you work through once. Move one nozzle and reinforcement changes, UG-45 changes, local stresses change, and MDMT can change too, since the governing thickness at that location may no longer be what it was. This dependency web is exactly why revision control of calculations is the quiet hard part of vessel engineering. A change made in one calculation and not propagated to the others is one of the most common sources of a bounced calc package.

Stage 4, hydrotest calculation

The standard hydrostatic test per UG-99(b) is 1.3 times MAWP, multiplied by the lowest stress ratio across the vessel's materials, with stresses checked at test conditions rather than design conditions. The test case can govern thin components even when they cleared the design case comfortably, since test pressure runs well above design pressure and material behavior at test temperature is not identical to material behavior at design temperature. This is why the hydrotest pressure is calculated as its own step, not simply assumed to be a safe multiple of MAWP without checking.

Stage 5, the calc package and drawings

Out of stages 3 and 4 comes the reviewable calc package: the design data sheet, per-component calculations with Code citations attached to each result, the MAWP and MDMT summary, hydrotest pressures, and the nozzle schedule. In parallel, the general arrangement drawing, fabrication details, weld maps, and bill of materials come together.

The discipline that saves jobs at this stage is simple to state and easy to neglect: keep the calc revision synchronized with the drawing revision. A drawing at Rev C with calculations still at Rev A is the most common audit finding there is, and it is entirely avoidable with basic revision discipline. Every time a drawing changes something that touches a calculation, the calculation needs to move to the same revision level before either document goes out for review.

Stage 6, Authorized Inspector review

Before signing the Data Report, the AI has to verify that the required design calculations are on file at the manufacturer's plant and resolve every question raised during review. Experienced shops pre-review draft calc packages internally, or informally with their AI, specifically to pass on the first formal submission rather than cycling through multiple rounds of comments.

Two things make that review fast. First, every Code paragraph identified next to the result it governs, so the AI is never left guessing which clause a number is supposed to satisfy. Second, every intermediate value shown, so nothing has to be re-derived by hand just to confirm the final answer is right. That format, showing the work rather than just the conclusion, is the entire design philosophy behind our reports, and it is the single biggest lever a manufacturer has over how long AI review actually takes.

Stage 7, fabrication with hold points

Once calculations and drawings clear review, the quality system takes over. Fit-up inspections confirm dimensions before welding starts. Weld procedure compliance, WPS and PQR on the procedure side, welder qualifications per WPQ on the personnel side, gets verified before anyone strikes an arc. NDE runs per the joint efficiency claimed in the calc package, RT, UT, PT, or MT as applicable, since a joint efficiency claim is only as good as the inspection that backs it up. PWHT runs with recorded charts wherever required, and MTR traceability follows material from the original plate all the way through to the weld map.

Joint efficiency is a promise the shop floor has to keep, not just a number an engineer wrote down. Claim E = 1.0 with full radiography and that claim is now a contractual fact the fabrication team has to deliver on, inspection by inspection, weld by weld.

Stage 8, witnessed hydrotest

The AI witnesses the hydrotest at the pressure calculated back in stage 4. Gauges are calibrated beforehand, vents are open at the high points to bleed air before pressurizing, the water is kept above the minimum metal temperature to avoid a brittle-fracture risk during the test itself, and the hold time is honored in full rather than shortened under schedule pressure. Visual inspection at test pressure follows, conducted at a reduced pressure per the Code's own provisions for personnel safety during the inspection walk-down. A passed, witnessed hydrotest gets recorded as its own document in the dossier, not simply noted informally and moved past.

Stage 9, Data Report, nameplate, registration

The Manufacturer's Data Report, the U-1 or U-1A, is completed, certified by the manufacturer, and signed by the AI. The nameplate is stamped per UG-115 and UG-119 with MAWP, MDMT, and the Code marks, then physically attached to the vessel. The vessel is registered with the National Board, putting its data on permanent file, and from that point forward any future repairs and alterations fall under the National Board Inspection Code and the R-stamp system rather than the original construction rules.

The turnover dossier

What the customer, and the next inspector who looks at this vessel years from now, actually receives:

#	Document
1	Compliance matrix against the PO and spec
2	Design calculations, final revision
3	GA and fabrication drawings, as-built
4	Weld maps
5	MTRs, traceable to the weld map
6	WPS, PQR, and WPQ
7	NDE reports (RT, UT, PT, MT)
8	PWHT charts
9	Hydrotest record
10	Nameplate rubbing or photo
11	Manufacturer's Data Report (U-1 or U-1A)

If stage 3's calc package was clause-cited from the start and kept in revision sync with the drawings all the way through, assembling this dossier at the end is straightforward paperwork, mostly a matter of collecting documents that already exist in final form. If the calc package was not handled that way, assembling the dossier turns into archaeology, chasing down which revision actually matches the as-built vessel and reconstructing a paper trail that should have existed all along.

FAQ

How long does the whole cycle take? For a routine Division 1 vessel: engineering runs in days, AI review runs in days to weeks depending heavily on the working relationship between shop and inspector, and fabrication runs in weeks to months depending on vessel size and shop backlog. The real schedule killer is almost always rework loops, a bounced calc package sent back for corrections, or a failed hold point during fabrication that stops work until it is resolved.

Who signs what? The manufacturer's designee certifies the Data Report first. The Authorized Inspector signs it after independently verifying compliance with the Code. The shop's own QC manual defines who internally holds each level of signing authority, which can vary meaningfully from one manufacturer to another.

Does every vessel need a U-stamp? No. Jurisdiction and service decide that question, not vessel size alone. Many small or exempt vessels are legitimately built to the Code without ever being stamped. The PO and the installing jurisdiction give the real answer, and it is worth remembering that "built to ASME" without an actual stamp is not the same thing as a stamped, registered vessel, even when the construction quality is identical.

What software handles stages 3 through 5? That market, COMPRESS, PV Elite, and the rest, is mapped honestly in our 2026 comparison, which covers the practical differences between the major tools rather than just their marketing claims.

---

DeepMechanix takes you from stage 1 to stage 5 with every clause cited. See the product.