TURNAROUNDQC
← Back to Turnaround QC
Last updated 12 August 2026
A turnaround runs three or four codes at once on the same rack. New spool to B31.3, a vessel nozzle repaired under NBIC, the whole lot inspected in service to API 570 and 510. This is the short version — which document governs what, and the numbers that get asked for at seven in the morning with a fitter waiting on an answer.
The governing code and the job spec decide, not this page. Every figure below is the common case. Editions move, owners write their specs tighter than the code, and a jurisdiction can override both. Where a number matters, take it from the edition your job is bought to and from the WPS in your hand. Nothing here replaces reading them.
| Document | What it covers | Where a turnaround hits it |
|---|---|---|
| ASME B31.3 Process Piping | Design, materials, fabrication, examination and testing of process piping. New construction rules. | Every new spool, tie-in and replacement run in process service. The default piping code in a refinery. |
| ASME B31.1 Power Piping | Power and utility steam piping, and boiler external piping. | The utility side — steam headers, boiler feed, condensate. Different examination and test rules to B31.3. Do not read across. |
| ASME BPVC Section I Power Boilers | Boiler pressure parts and the boiler proper. | Steam generation and waste heat boilers. Usually the jurisdiction and the boiler inspector are involved, not just the client. |
| Section VIII Div 1 Pressure Vessels | New vessel construction — design, fabrication, examination, stamping. | Replacement vessels and new exchangers arriving on the job. It is a construction code; once the vessel is in service, it stops being the operative one. |
| NBIC NB-23 National Board Inspection Code | Repair and alteration of pressure-retaining items already in service. R-stamp work. | Any weld on a vessel, boiler or piping system that has been in service. The rules for a repair are not the rules for building it new. |
| API 510 / 570 / 653 In-service inspection | Vessels, piping and aboveground tanks in service — inspection intervals, thickness and remaining life, rerating, repair rules. | The internal inspections and CML readings the turnaround exists to allow. These drive the scope; the construction codes govern what you then do about it. |
Section V (NDE methods) and Section IX (welder and procedure qualification) sit underneath all of it. They are referenced by the others rather than applied on their own, and the referencing code is what supplies the acceptance criteria.
The category is an owner decision made at design and it sets how hard the code leans on the work. It should be on the line list. If it is not, someone has to say what it is before examination extent can be agreed, because the same weld has four different examination requirements depending on the answer.
| Category | What it is | What changes |
|---|---|---|
| Category D | Non-flammable, not damaging to human tissue, modest pressure and temperature limits set by the code. | The lightest regime. Visual examination; the code does not require radiography. An initial service leak test is permitted in place of a hydrostatic test. |
| Normal | The default. Most refinery hydrocarbon service. | 100% visual, plus 5% random radiography of girth welds, spread so each welder’s work is represented. Hydrostatic test. |
| Category M | Fluid where a single exposure to a very small leak can do serious irreversible harm. | Examination at least to Normal, and the code adds a sensitive leak test. The hazard being designed against is leakage itself, not rupture. |
| High pressure | Designated by the owner; the code handles it in its own chapter. | 100% examination of welds by radiography or ultrasonics, and its own testing rules. Nothing from the Normal chapter carries across by default. |
| Severe cyclic | An added condition rather than a category, applied where the design sees heavy cycling. | 100% radiography of girth and miter groove welds, and tighter limits on details that concentrate stress. |
| Method | Finds | Misses |
|---|---|---|
| RT radiography | Volumetric flaws — porosity, slag, incomplete penetration, lack of fill, burn-through. Gives a permanent image somebody else can re-read. | Tight planar flaws lying across the beam. Lack of fusion on a bevel face can be all but invisible on film at the wrong angle. |
| UT manual, phased array, TOFD | Planar flaws, lack of fusion, cracking, and wall loss. Better than RT at the things that actually fail. | The near-surface dead zone, and anything outside the scan plan. It is only as good as the technique sheet and the coverage — and rough or coated surfaces defeat it. |
| PT penetrant | Surface-breaking flaws in any material. The usual method on austenitic stainless, which cannot be examined magnetically. | Everything below the surface, and anything the preparation smeared shut. Grinding a crack closed makes it vanish from the report. |
| MT magnetic particle | Surface and slightly subsurface flaws in ferromagnetic material. Fast, and the method of choice for weld toe cracking. | Non-magnetic material entirely, and flaws lying parallel to the field — which is why it is shot in two directions and why a single-direction report is half a result. |
| VT visual | Fit-up, alignment, undercut, arc strikes, reinforcement, the wrong consumable in the welder’s hand. | The inside of a closed joint. It is also the examination most often done and least often written down. |
Acceptance criteria differ by code, and by the category within the code. A film read to B31.3 Normal fluid service and the same film read to Section VIII are not the same verdict. So a report that says “acceptable” with no code, edition and criterion named beside it is not a result — it is somebody’s opinion, and it cannot be checked. Name the criterion or the reading is not evidence.
| Test | Pressure | Notes |
|---|---|---|
| Hydrostatic | 1.5 × design pressure, corrected for the difference in allowable stress between test and design temperature. | The default under B31.3. Minimum ten minutes at pressure before the joints are walked. The correction can raise the figure well above a flat 1.5 on a hot line, so calculate it rather than assuming. |
| Pneumatic | 1.1 × design pressure. | Only where hydro is impracticable, and the lower multiplier is not generosity. Compressed gas stores energy that water does not; the same volume at the same pressure fails as a bang instead of a spill. The code has you stop at a preliminary pressure — the lesser of half the test pressure or 25 psi — and check before going any higher, and the exclusion zone is a real control, not paperwork. |
| Initial service leak test | Operating pressure, examined during start-up. | Permitted for Category D. Not a substitute anywhere else without an owner-approved alternative. |
| Sensitive leak test | Per the method — bubble, halide, helium. | Required for Category M. It is looking for a leak rate, not a burst, so it proves something a hydro does not. |
Two gauges of different ranges, both in calibration, because one gauge that has quietly gone off reads exactly like a gauge that has not. Size them so the test pressure lands in the middle third of the range: a reading taken in the bottom tenth of a large gauge is a guess with a needle attached.
Chlorides in the test water matter on austenitic stainless. Chloride, tension and a bit of warmth is the full recipe for stress corrosion cracking, and the tension is already there in the weld. Job specs commonly cap test water at 50 ppm chloride and require prompt draining and drying — check yours, because the number is a spec number, not a code number. The failure this prevents is not on test day. It is water left in a dead leg over the summer.
| P-No. | Material | Typical preheat | PWHT |
|---|---|---|---|
| P-1 | Carbon steel | 50°F minimum; 175°F once the section is heavy or the carbon is high. | Normally required above about 3/4 in nominal thickness. Below that, generally not. |
| P-4 | 1-1/4Cr-1/2Mo | Around 300°F, lower only for thin light-carbon sections. | Normally required, with a thin-section exemption the code defines. |
| P-5A | 2-1/4Cr-1Mo | 300–400°F, and it must be held, not just reached. | Effectively always. Treat an unheat-treated 5A weld as a defect until someone shows you the exemption in writing. |
| P-8 | Austenitic stainless | None required. Interpass temperature is the control, and it is a ceiling, not a floor. | Not normally required, and often unwanted — time in the sensitisation range does harm rather than good. |
Soak time is a rule of thumb, one hour per inch of thickness with a short minimum for thin material. The rest is not a rule of thumb: heating and cooling rates, the width of the heated band, the number and placement of thermocouples, and where the chart recorder trace goes afterwards. These four rows are the common cases only. The WPS and the governing code decide, and where they disagree with this table, they are right and this table is wrong.
| Variable | What it controls |
|---|---|
| P-number | Groups base metals by weldability. For a welder’s performance qualification the base metal range is deliberately broad — the test is of the person, not the alloy. Procedure qualification is the opposite, and much tighter. |
| F-number | Groups filler metals by running characteristics. Qualifying on one F-number qualifies a range of others, and the range is not symmetrical — it does not follow that it works in reverse. Check the direction before you let someone burn a rod. |
| Position | 6G on pipe qualifies all positions for groove welds. 2G and 5G together do the same. A welder tested 1G flat has qualified for very little of a turnaround. |
| Diameter and thickness | Both carry ranges. Small-bore qualification does not automatically cover large bore, and a thin coupon limits the thickness qualified. |
Qualification is per process, so a welder running GTAW root and SMAW fill holds two of them — and each carries its own continuity. Continuity is the one that catches packages out. A welder’s qualification in a process lapses if he has not used that process for six months, and it is judged at the date of the weld — not at the date the certificate expires, and not at the date somebody photocopied it into the file. A certificate valid on paper today proves nothing about a joint welded in week two by a man whose last documented run with that process was seven months before it. Keep the continuity log dated and keep it with the weld map, because the two are read together or not at all.
Torque is not clamp load. Most of the torque goes into friction, and the friction is set by what is on the threads and under the nut face. At the same torque figure the spread between a dry as-received stud and a properly lubricated one reaches a factor of two, which means a torque record with no lubricant written on it does not describe the joint.
Dry and as received is the worst of it — highest friction, lowest clamp load, and the widest scatter bolt to bolt, worse again with rust or grit on the threads. General purpose oil gains clamp load and consistency. Nickel or copper anti-seize gains more, and is the usual choice where the joint has to come apart hot in five years. PTFE-based lubricant has the lowest friction of the common ones, which is exactly why it has to be torqued to the figure calculated for it and not to the dry one.
| Gasket | Where it belongs | Watch |
|---|---|---|
| Spiral wound | General refinery hydrocarbon and steam service on raised face flanges. The workhorse. | Inner ring and winding material must match the service. Fitting one without the centring ring into a raised face joint buckles it inward. |
| Kammprofile | Exchangers and flanges that rotate or see thermal cycling. Seals at lower load and recovers better. | Facing finish matters more than people expect. A scored face that a spiral wound tolerated will leak. |
| RTJ | High pressure and high temperature joints on ring-groove flanges. | Ring hardness must be below the groove hardness, and grooves get damaged during removal. Inspect the groove, not just the ring. |
| Sheet | Utilities — water, air, low pressure non-hazardous service. | Not a hydrocarbon gasket because it was in the store. Check the rating before it goes in. |
ASME PCC-1 is the reference for assembly. In one line: clean and inspect the faces, lubricate, align the flanges before pulling them together, then a cross pattern in numbered passes to roughly 30%, 60% and 100% of target, followed by at least one full circular pass at 100% until no nut moves. Record the pass and the pattern — a joint dragged to final torque in one pass is not the same joint as one brought up in three, whatever the final figure says.
| Type | What it is | For |
|---|---|---|
| Spade (paddle blind, skillet) | Solid plate with a handle, fitted between flanges. Flanges must be spread to fit or remove it. | Positive isolation of a line that stays blinded for the duration. |
| Spacer (paddle spacer) | Open ring of the same thickness as the spade. | Holding the gap when the line is back in service, so the spade can go straight back in next time. |
| Spectacle blind (figure-8) | Spade and spacer joined, permanently attached to the joint. | Lines blinded and unblinded routinely. Its great virtue is that the position is visible from the ground — you can read the isolation without a permit and a ladder. |
| Blind flange | Bolts onto an open flange end. | Closing a line that has been physically broken and left open. |
A blind is a pressure part. It has to be rated for the full pressure it may see with the line live on the other side, in the material of the line, and the rating table for standard flange line blanks is a published one. A thin plate cut in the shop to fit is not a blind, whatever it is written down as.
| Mechanism | Where it shows up | What finds it |
|---|---|---|
| Sulfidation | Carbon and low-alloy steel above roughly 500°F in sulphur-bearing streams — crude and vacuum heaters, transfer lines, hydrotreater feed. | Thickness monitoring. The classic failure is a low-silicon component thinning far faster than the run either side of it, on a fitting that was never a CML. |
| Naphthenic acid | High-TAN crude at temperature, worst where the flow is fast and turbulent — vacuum transfer lines, side-cut piping, downstream of control valves. | Scanning UT or profile radiography. Localised smooth grooving; spot readings between the grooves come back reassuring and wrong. |
| Amine SCC | Carbon steel in lean amine, at and beside welds that were never stress relieved. | WFMT after the coating comes off. Cracks run parallel to the weld in the heat affected zone. PWHT is the mitigation, which is why non-PWHT amine welds are worth knowing about before the outage. |
| Chloride SCC | Austenitic stainless with chlorides, moisture and warmth — under insulation, in dead legs, and anywhere hydrotest water sat. | PT. Fine branching cracks, often with no wall loss at all, so thickness readings say the line is perfect right up until it opens. |
| HTHA | Carbon and low-alloy steel in hydrogen at temperature and partial pressure — hydroprocessing. The operating limits come from the Nelson curves in API 941. | Advanced UT techniques, not conventional thickness readings. Damage is internal fissuring; there is nothing to see and nothing to measure until it is far along. |
| CUI | Insulated carbon steel in the wet temperature band, and insulated stainless as chloride SCC. Worst at supports, penetrations, damaged cladding and anywhere water gets in and cannot leave. | Insulation removal, or profile radiography and pulsed eddy current through it. The paint looks fine where you can see it because where you can see it is not the problem. |
| Under-deposit corrosion | Wherever solids or water settle — dead legs, the bottom of horizontal runs, overhead systems, exchanger shellsides. | Scanning the six o’clock position. A CML grid on the four quadrants at the standard elevation will miss it indefinitely. |
| Erosion and erosion-corrosion | Elbows, tees, downstream of control valves and orifices, catalyst and slurry lines. | UT scan of the outer radius and the impingement point. The thin spot is small and it is not in the middle of the fitting. |
The mechanism you expect determines the method you specify. A general thickness survey is not an inspection for cracking, and API 571 is the long version of this table.
It is not the code. It is the answer to the question that actually gets asked on site, which is never “what does B31.3 say” — it is “this is going on a sheet in ten minutes, what else has to be on it”. Where the code and the job spec differ from anything above, they govern and this page does not.
One habit is worth more than all of it. Every record hangs off a line number, a weld number or an equipment tag, and if the tag drifts — P-1201A here, P1201A there, 1201-A on the daily — nothing can be pulled together at the end. A heat number on a length of pipe is worthless once it is six pieces. Transfer it before the saw touches it, and write down who did.
Spec quick reference — turnaroundqc.com/spec-reference