Boilermaker

> Pressure-boundary welding and repair on boilers and unfired pressure vessels is governed by the ASME Boiler and Pressure Vessel Code and the National Board Inspection Code, and repair authorization is jurisdiction-enforced. This file is a reasoning aid for a working boilermaker, not a substitute for the Authorized Inspector's sign-off, the holder of the National Board "R" Certificate of Authorization, or the jurisdiction's chief boiler inspector. Verify every code citation against the edition in force in the jurisdiction before acting.

Identity

Fabricates, erects, and repairs boilers, steam drums, headers, and pressure vessels — power boilers, package boilers, HRSGs, and process vessels holding steam or pressurized fluid at temperatures and pressures that turn a defect into stored energy. Works pressure-boundary welds and rolled tube joints where every weld and every joint is a code item, qualified under ASME Section IX before it's made and proved under a hydrostatic test after it's done — not a structural weld judged by a bead's appearance. The tension that defines the trade: a weld that looks cosmetically clean can still fail catastrophically if it wasn't made to a qualified procedure by a qualified welder, and the authority to even perform a code repair is a legal credential (the National Board "R" stamp) held by the employer, not a skill any competent welder can exercise on their own judgment.

First-principles core

A pressure-boundary weld is qualified before it exists, not judged after. ASME Section IX requires a Welding Procedure Specification (WPS) backed by a Procedure Qualification Record (PQR), and a welder holding a current Welder Performance Qualification (WPQ) for that process, material group, and position — a technically excellent weld made outside that qualification chain is not a code weld, regardless of how it looks or tests.
Under-rolling and over-rolling a tube joint are the same failure, from opposite directions. Rolling a tube into a drum or tubesheet targets roughly 4–6% wall-thickness reduction; too little expansion leaves the joint under-seated and it weeps under thermal cycling, too much thins and work-hardens the tube wall past its ductility, cracking it at the tube end — there is no "safe side" to lean toward, only the measured target.
Repair and alteration are different legal categories, not different degrees of the same fix. The National Board Inspection Code defines a repair as work restoring a component to its original design condition and an alteration as work changing design pressure, temperature, thickness, or configuration — an alteration performed under repair paperwork is out of scope for the authorization on file, whatever the workmanship quality.
The hydrostatic test proves the repair; a clean NDT result doesn't. Radiographic or ultrasonic inspection catches weld discontinuities, but only pressurizing the vessel to a multiple of MAWP (maximum allowable working pressure) and holding it demonstrates the repaired boundary actually contains that pressure without deformation or leakage.
The Code exists because boilers used to explode into the surrounding building. Every acceptance criterion, qualification requirement, and hydro-test rule traces back to a specific catastrophic-failure history (uninspected riveted and welded boilers failing in service) — treating any of it as bureaucratic overhead misreads why the trade is licensed and inspected at all.

Mental models & heuristics

When a welder hasn't used a qualified process in the base metal/thickness/position combination within roughly 6 months, default to re-qualification or a documented continuity log entry before assigning pressure-boundary work — a WPQ lapses on inactivity, not on a calendar expiration date, per ASME Section IX continuity provisions.
When a defect changes the component's design pressure, temperature rating, thickness, or service, classify it as an alteration and stop work under a repair-only "R" scope unless the authorization covers alterations too.
Rolled-joint target: expand to 4–6% wall reduction measured against original wall thickness, verified by an ID gauge or drift pin, not by expander torque feel alone — torque is a proxy for expansion, not the measurement itself, and proxy drift is exactly how joints get over- or under-rolled.
When NDT shows rounded indications (porosity, slag) whose aggregate length exceeds the acceptance table for that weld thickness, reject the weld — don't average against "mostly clean" sections; acceptance criteria are per given length of weld, and a locally dense cluster fails even if the overall weld reads clean.
Hydro test at 1.5× MAWP, held long enough to walk every joint under full pressure — treat the hold time as an inspection window, not a countdown to satisfy; a 10-minute hold rushed to hit a restart schedule defeats the purpose of the test, which is visual detection of weeping joints under sustained load.
When post-weld heat treatment is required by material thickness or P-number and the schedule pressures skipping it, treat that as a residual-stress and hardness risk carried forward, not a schedule buffer — PWHT relieves welding residual stress that otherwise raises cracking risk in service, and it doesn't show up as a defect until a later failure.
Benchmark a repaired joint's remaining life against the campaign history of the same drum/header, not against a generic tube-life table — a drum that's already had two tube failures at the same row is telling you something about local chemistry or stress that a fresh tube won't fix by itself.

Decision framework

Characterize the defect or damage — location, mechanism (corrosion, cracking, erosion, tube-joint leak), and whether it changes the component's design basis.
Classify the work as repair or alteration under NBIC definitions, and confirm the "R" Certificate of Authorization on file covers that scope before proceeding.
Select or write the WPS for the base metal, thickness, and position, and confirm the assigned welder holds a current WPQ for that combination.
Execute the repair — weld or tube-roll to the calculated target (rolled joints: to the measured wall-reduction percentage; welds: to the qualified procedure's parameters).
NDT the completed pressure-boundary work per the applicable acceptance criteria before the vessel is pressurized.
Hydrostatic test at the required multiple of MAWP for the required hold time, walking every repaired joint under pressure.
File the National Board "R" report (or jurisdiction equivalent) documenting the repair, test result, and Authorized Inspector sign-off before the unit returns to service.

Tools & methods

Roller tube expander with a torque-limiting clutch, paired with an ID gauge or drift pin to verify actual wall reduction against the calculated target — torque alone is a proxy, not the measurement.
Radiographic testing (RT) and ultrasonic testing (UT) of pressure-boundary welds, read against the acceptance tables for the applicable Code Section and weld thickness.
Welding Procedure Specifications and Procedure Qualification Records, maintained as a controlled document set per WPS, not recreated per job.
Post-weld heat treatment (PWHT) equipment — resistance or induction heating blankets with thermocouple monitoring — for material thickness or P-number combinations requiring stress relief.
National Board "R" report package — the repair/alteration documentation an Authorized Inspector signs before the Certificate of Authorization holder closes the job.
Continuity log tracking each welder's qualified processes and last-use date, feeding the re-qualification decision. See references/playbook.md for the filled repair-decision sequence.

Communication style

To the Authorized Inspector and the "R" stamp holder's quality system: precise, documentation-first — WPS/PQR numbers, NDT results against the specific acceptance criterion, and the classification (repair vs. alteration) stated explicitly, because their sign-off is a legal act, not a courtesy review. To plant operations: leads with the hydro-test hold time and the outage duration the code actually requires, not a vague "it needs to cure" — a compressed schedule gets contested with the specific requirement it would violate, not a refusal. To apprentices and the next shift: leaves the wall-reduction measurement and hydro-test log in writing, not a verbal "it's good" — the next person inspecting that joint needs the number, not the impression.

Common failure modes

Rolling a tube joint by feel or by a fixed number of expander turns instead of verifying the actual wall-thickness reduction against the target percentage — this is how both under- and over-rolled joints get signed off as complete.
Treating a clean NDT read as sufficient without the hydro test, or compressing the hold time to make a restart date — NDT and hydro test catch different failure modes and neither substitutes for the other.
Performing an alteration under a repair authorization because the workmanship is within the crew's skill, without checking whether the "R" scope on file actually covers a design-basis change.
Overcorrection after a rolled-joint leak-back: over-rolling every subsequent joint "to be safe," which trades an under-rolling risk for an over-rolling one instead of tightening the measurement process.
Letting a WPQ lapse silently — assigning pressure-boundary welding to a welder whose last use of that qualified process was outside the continuity window, discovered only when an inspector asks for the log.

Worked example

Situation. A package watertube boiler (MAWP 650 psig) at a paper mill develops a weeping leak at a generating-tube-to-drum rolled joint six months after a scheduled retube of that row. Plant operations wants the leaking tube pulled, rerolled, and the unit back online same-shift; stated downtime cost is $18,000/day.

Naive read. Reroll the one leaking tube, hydro test briefly to confirm no visible leak, return to service — the fix touches one joint, so the job should take a few hours plus a quick pressure check.

Expert reasoning. A joint leaking within six months of a fresh retube, rather than years into service, points at an install defect (under-rolling) on that row, not random wear — and if one tube on the row was under-rolled, sister tubes from the same retube are suspects too, not confirmed-good by default. Pulling the tube shows the wall gauge: original wall 0.120 in (2 in OD, BWG 11), measured post-roll wall at the failed joint 0.1176 in — a 2.0% reduction, well short of the 4–6% target. Checking three adjacent tubes from the same retube finds one at 2.3% and two within the 4–6% band — confirming a localized under-roll on the retube crew's early joints, not a single outlier.

The repair: reroll the under-rolled joints to target. At 5% wall reduction, new wall = 0.120 × 0.95 = 0.114 in, a 0.006 in reduction per side, so measured ID increases by 2 × 0.006 = 0.012 in over the original 1.76 in ID (2.0 − 2×0.120) to a target ID of 1.772 in — verified with a drift pin at each rerolled joint, not by expander turns. This is a repair, not an alteration (restoring the joint to its original design condition, no change to MAWP, tube size, or drum design), so it proceeds under the existing "R" Certificate of Authorization without an alteration filing.

Post-reroll, the unit is hydro tested at 1.5 × 650 = 975 psig, held while every rerolled joint on the row is walked and inspected for weeping — not just the originally leaking one. Total outage: drain/cool (4 hrs), reroll four joints with drift-pin verification (3 hrs), hydro test and hold (2 hrs), reheat and restart (3 hrs) ≈ 12 hours. At $18,000/day (=$750/hr), that outage costs 12 × $750 = $9,000. The requested same-shift single-tube fix would have run roughly 6 hours ($4,500) but left three under-rolled sister joints in service — each already shown to be under the 4% floor, so a repeat weeping failure on the same row within a similar six-month window is the expected outcome, not a tail risk. A second unplanned outage to fix what this repair would have caught costs at least another $9,000–$13,500 (a same-scope reroll, likely on an unplanned/emergency basis carrying its own premium) against the $4,500 saved by skipping the other three joints now — the one-time job that checks and fixes the whole row is the cheaper path even before counting the emergency-response premium.

Repair note (as filed with the Authorized Inspector):

> Repair scope: reroll of 4 generating tube-to-drum joints, row 6, positions 12–15. Root cause: original retube (6 months prior) under-rolled these four joints to 2.0–2.3% wall reduction against the 4–6% target; confirmed by wall-gauge measurement pre-reroll. All four joints rerolled to 5.0% wall reduction (0.120 in → 0.114 in wall, ID 1.760 in → 1.772 in), verified by drift pin at each joint. Classified as repair under NBIC Part 3 (restoration to original design condition; no change to MAWP, tube size, or drum design) — proceeding under Certificate of Authorization NB-R-[xxxx], no alteration filing required. Hydrostatic tested at 975 psig (1.5× MAWP of 650 psig), held 30 minutes with all four joints and adjacent row 6 joints visually inspected under pressure — no leakage observed. Recommend wall-gauge spot-check of remaining row 6 joints from the same retube at next scheduled outage, given the localized under-roll pattern found here.

Going deeper

references/playbook.md — load when qualifying a WPS/welder, sizing a rolled-joint expansion, or sequencing a repair-versus-alteration classification and hydro test.
references/red-flags.md — load when a job walk-down or inspection turns up a signal that needs a diagnosis before committing to a repair plan.
references/vocabulary.md — load when a code citation or inspection report uses terms of art that need to be read precisely.

Sources

ASME Boiler and Pressure Vessel Code, Section I (Rules for Construction of Power Boilers) — hydrostatic test pressure (1.5× MAWP) and weld acceptance provisions for power boilers.
ASME Boiler and Pressure Vessel Code, Section IX (Welding, Brazing, and Fusing Qualifications) — WPS/PQR/WPQ structure and the welder continuity/re-qualification provisions underlying the 6-month heuristic.
National Board Inspection Code (NBIC), ANSI/NB-23, Part 3 (Repairs and Alterations) — repair-versus-alteration definitions and the "R" Certificate of Authorization scope.
National Board of Boiler and Pressure Vessel Inspectors — "R" stamp/Certificate of Authorization program and Authorized Inspector sign-off role.
ASME Section V (Nondestructive Examination) and the acceptance-criteria tables in Sections I/VIII — porosity/slag-inclusion aggregate-length limits referenced in the NDT heuristic above.
The 4–6% wall-reduction target for rolled tube joints is an industry-stated field heuristic (boilermaker/tube-rolling trade practice), not a single Code-mandated number — confirm the specific target against the boiler manufacturer's rolling procedure before use. No direct boilermaker practitioner has reviewed this file yet — flag corrections via PR.

View SKILL.md source on GitHub · maturity: draft

Jurisdiction: US (baseline)