API RP 574 — Inspection Practices for Piping System Components

API RP 574 is the recommended practice that tells inspectors how to examine piping components — where to place condition monitoring locations, how to measure and correct wall thickness readings, and how to establish minimum required thickness. It supports API 570 but never replaces it: 570 sets the mandatory intervals and the fitness decision, while 574 supplies the technique.

Scope covers pipe, tubing, fittings, flanges, valves, bolting and the supports that carry them, together with the inspection tools and the thickness data practice that keeps a circuit auditable. It does not set inspection intervals, does not classify piping, and does not decide whether a corroded component may stay in service — those belong to API 570 and, for anything past the code minimum, to API 579-1/ASME FFS-1. The requirements that bite in practice are the ones about data quality: readings taken at the same physical point each time, ultrasonic velocity corrected for metal temperature, mill under-tolerance recognised before a corrosion rate is calculated, and CMLs sited where the damage mechanism actually puts metal loss rather than where scaffolding is cheap. Most findings written against a 574 programme are data-integrity findings, not measurement findings. Auditors sample the trail from drawing to reading to rate to next inspection date, and a break anywhere in it invalidates the interval.

Source: Sources: API RP 574, Inspection Practices for Piping System Components; API 570, Piping Inspection Code — In-Service Inspection, Rating, Repair, and Alteration; ASME B31.3 Process Piping; ASME B36.10M Welded and Seamless Wrought Steel Pipe; ASTM A106/A53 pipe specifications (mill tolerance); API RP 571 Damage Mechanisms Affecting Fixed Equipment; API 579-1/ASME FFS-1 Fitness-For-Service; API 598 Valve Inspection and Testing.

Technically reviewed by Anoop Rayavarapu — ASNT NDT Level III (UT, RT, MT, PT, VT)
Three different "minimum thicknesses" that are routinely confused on the same piping circuit
NPSSch 40 nominal wall (in)As-delivered floor at −12.5% mill under-tolerance (in)Default minimum structural thickness, carbon and low-alloy steel (in)
10.1330.1160.07
20.1540.1350.07
30.2160.1890.08
40.2370.2070.09
60.2800.2450.11
80.3220.2820.12
120.3750.3280.12
200.5940.5200.13
Structural minima are the default values commonly tabulated in API RP 574 for carbon and low-alloy steel pipe on conventional support spans; they are not pressure-design thicknesses and they do not apply to alloy pipe, to piping carrying unusual external or vibratory loads, or to any circuit where the engineer has calculated a different value. The governing minimum required thickness is the greater of the pressure design thickness computed to the construction code and the applicable structural minimum. Consult the current edition of each standard for the values that apply to your circuit.

What API 574 covers, and the three things it deliberately leaves alone

API RP 574 describes how to inspect the physical components of a piping system: pipe and tubing, fittings, flanges and gaskets, bolting, valves, and the supports, shoes, hangers and spring cans that carry the line. Alongside that it covers the inspection tools — ultrasonic thickness instruments, radiographic profile techniques, boroscopes, hardness testers — and, most usefully, the practice for recording and treating thickness data so that a corrosion rate calculated from it means something. It is written as the technical companion to API 570.

Three things it deliberately does not do. It does not set inspection intervals or classify piping by consequence — that is API 570. It does not tell you whether a component measured below its minimum required thickness may remain in service — that is API 579-1/ASME FFS-1, and it needs an engineering assessment, not an inspector's judgement. And it does not identify or explain damage mechanisms; API RP 571 does that, and without it a CML plan is guesswork with a spreadsheet attached.

The distinction is not academic. In audit after audit, an owner justifies a gap by citing 574's permissive language against a 570 requirement that is mandatory. Where the interpretation is genuinely contested — mixed-code circuits, alloy piping, retirement thickness disputes — it is worth putting the position in writing before the audit, which is one of the routine outputs of ASNT Level III consulting: a documented technical basis signed by someone competent to hold it.

Minimum required thickness: two numbers that are not the same number

Every piping component has a pressure design thickness, calculated to the construction code — for process piping, ASME B31.3 — from design pressure, outside diameter, allowable stress, joint efficiency and the appropriate coefficients. It also has a structural minimum: the wall required to carry the pipe's own weight, the contents, insulation and normal external loads on the spans the line is actually supported at. API 574's tabulated structural minima exist because the pressure design thickness for a large-diameter low-pressure line can come out absurdly thin — thin enough that the pipe would sag or buckle long before it burst.

The governing minimum required thickness is the greater of those two. The recurring error runs both ways. Programmes that use the structural table as the retirement thickness on a high-pressure line will run a component past the point where it can hold pressure. Programmes that use pressure design thickness alone on a large low-pressure header will retire the line on a number that ignores dead load entirely, and will be unable to defend it if the line deflects or a support fails.

Two further traps sit inside the same table. The default structural values are for carbon and low-alloy steel — they do not transfer to stainless or nickel alloys with different moduli and different support behaviour. And they assume conventional support spans with no unusual external, thermal or vibratory loading; a line on long spans, or one carrying a heavy valve mid-span, needs the number calculated, not looked up.

Where CMLs go, and why the damage mechanism chooses the location

A condition monitoring location is only as good as the reasoning that put it there. The damage mechanisms credible for the circuit dictate the placement: erosion-corrosion puts metal loss on the outer radius of elbows, at the branch of tees, in reducers and immediately downstream of control valves, orifices and any point where the fluid flashes. Under-deposit and dead-end corrosion put it at low points and at the stagnant ends of deadlegs. Corrosion under insulation puts it at penetrations, at insulation terminations, at low points and under damaged cladding — nowhere near the points chosen for straight-run thickness readings.

This is why a CML plan is a technical document with a rationale, not a list of accessible points. The most common structural weakness in an otherwise well-run programme is that the CML set was established at commissioning, never revised after a damage mechanism review, and never extended when process conditions changed — new feedstock, higher sulphur, a revamp that moved a mixing point. The readings are good; they are simply being taken where nothing is happening.

Volume is the second problem. A moderately complex unit carries tens of thousands of CMLs across hundreds of circuits, each with a history, a rate and a next-due date, and spreadsheets stop being defensible somewhere around the first turnaround. That data belongs in an inspection data management system that holds the isometric, the CML, the reading history and the calculated rate as one linked record rather than four files that agree only by luck.

Thickness readings that survive a corrosion rate calculation

A corrosion rate is a difference between two measurements, which means the error in the rate is the sum of the errors in both. Three sources dominate. First, repeatability: a reading taken 40 mm from last time's reading on a generally corroded surface is not the same reading, and on a pitted or grooved surface it may differ by more than a decade of real loss. Permanent marking, a photograph in the record, and a grid rather than a point are what make the second survey comparable to the first.

Second, temperature. Sound velocity in steel falls as the metal heats, so an instrument calibrated at ambient reports a hot component as thicker than it is by roughly one percent per 100°F above calibration temperature. Applied consistently and never corrected, that bias produces a circuit that appears to be corroding slowly while the wall quietly disappears — and it produces the reverse artefact, an apparent negative corrosion rate, whenever a shutdown survey is compared against an on-stream one.

Third, the starting point. Corrosion rates computed against nominal wall are fiction. Seamless pipe to ASME B36.10M is permitted a 12.5 percent under-tolerance, so a nominal 0.280 in Sch 40 six-inch pipe may have left the mill at 0.245 in with nothing wrong with it. Where the installed baseline was never measured, the honest treatment is to say so and to calculate the rate from measured data only. Where the numbers already carry a decision — a retirement, a turnaround scope, a fitness-for-service submission — independent review of inspection reports before the decision is cheaper than defending the arithmetic afterwards.

Injection points, deadlegs and small bore: the three circuits that actually fail

Injection points concentrate damage over a short length. Where a chemical, a wash water or a corrosion inhibitor enters the process, mixing, flashing and impingement can remove wall at rates an order of magnitude above the parent circuit. The inspection circuit is therefore drawn separately, beginning at least 12 inches or three pipe diameters upstream of the injection point, whichever is greater, and extending downstream to the second change in flow direction. Inside that boundary, single-point readings are not adequate — the loss is local, and a grid or a profile radiograph is what finds it.

Deadlegs behave differently. Stagnant fluid, water drop-out, settled solids and the absence of inhibitor make the far end of a deadleg the worst part of the circuit, and it is invariably the least accessible. Long-standing deadlegs from decommissioned equipment are the classic case: nobody owns them, they are not on the current isometric, and the first sign is a leak. The reliable remedy is a documented deadleg register with each one either monitored at its stagnant end or physically removed.

Small-bore piping fails by a mechanism a thickness programme does not measure. Vibration-induced fatigue cracks socket welds and unsupported branch connections; the wall is full thickness right up to the moment it separates. Threaded connections add a second mode. Managing small bore therefore means a plan built around bracing, support and visual or surface examination — not a thicker grid of UT points.

Valves, flanges, bolting and supports — the components the thickness grid skips

API 574 covers more than pipe wall, and the parts it covers beyond pipe wall are the parts most programmes under-inspect. Valve bodies and bonnets thin like any other pressure boundary, particularly in erosive or wet service, and body-wall readings are rarely on the CML list. Seat tightness and shell testing follow API 598 when a valve returns from overhaul, and the shop record for that test is the evidence an auditor asks for when a valve is credited as an isolation point.

Flanged joints fail for reasons that have nothing to do with general corrosion: face damage, radial scoring across the seating surface, warped or thermally distorted flanges, incorrect gasket material, and bolting that was neither torqued to a procedure nor recorded. Bolting itself corrodes preferentially in wet insulated service and at the nut-to-flange contact, and stud condition is a legitimate inspection item, not a maintenance afterthought.

Supports are the most consistently missed. The pipe-to-shoe and pipe-to-support contact point traps moisture and debris and produces external corrosion that a UT grid on the top and sides of the line will never see; the metal loss is on the underside, at exactly the location carrying load. Spring can travel indicators that are bottomed out, hangers taking no load, and guides seized by corrosion all change the load distribution the structural minimum thickness assumed in the first place.

Turning 574 data into an interval you can defend

The output of the whole exercise is one number: the date of the next inspection. Remaining life is the measured thickness less the required thickness, divided by the corrosion rate. API 570 requires both a long-term rate, computed from the earliest reliable reading, and a short-term rate from the most recent readings, with the more conservative of the two used unless an engineer documents why not. The short-term rate exists precisely to catch a process change that a fifteen-year average will smother.

The interval is then the lesser of half the remaining life or the maximum permitted for the piping class, and where the remaining life is short, the interval collapses accordingly rather than defaulting to the class maximum. Auditors sample this arithmetic more often than they sample the readings, because it is where the programme either converges on reality or quietly drifts away from it — a mis-set required thickness, a stale rate, an interval carried over from the previous cycle without recalculation.

At programme scale that calculation has to run automatically against current data, with the underlying readings, the rate basis and the required thickness all visible on the same record. Where the circuits fall inside an OSHA process safety management programme, the same records are what demonstrate mechanical integrity for covered equipment, which is why owners increasingly run them inside mechanical integrity software rather than reconstructing the trail from archived reports each time an auditor asks.

The findings that recur against a 574-based programme

Six findings appear over and over. Required thickness recorded as the structural minimum on a circuit where pressure design governs, or vice versa. Corrosion rates calculated against nominal wall, ignoring mill under-tolerance, on lines whose installed baseline was never measured. On-stream readings never corrected for metal temperature, and compared directly against shutdown readings. CML sets unchanged since commissioning despite a documented change in service. Injection point circuits monitored as ordinary straight-run piping. Deadlegs on the isometric with no reading at the stagnant end.

What links all six is that none of them is a measurement error. The technician did the job correctly; the framework around the reading was wrong. That is why the corrective action for a 574 finding is almost never retraining the UT technician — it is revising the inspection plan, the required thickness basis, or the data handling rules, and then re-examining the population of circuits the same error touched.

Extent of condition is the part owners underestimate. A single mis-set required thickness is one finding; the same error copied across every circuit built from the same template is a programme finding, and the auditor will ask how many circuits share the template. If you would rather find that out before an audit than during one, a documented gap review against your own written programme is the cheapest version of the exercise — start a conversation with the scope of the circuits in question and the standards your programme cites.

Procurement traps: buying pipe against a mill certificate that no longer describes the pipe

API 574's CML and TML methodology assumes the starting point — the mill certificate thickness — is still true when the survey begins. It frequently is not. A spool ordered as Schedule 40 arrives with a mill certificate showing 0.365 in nominal, but the mill's own under-tolerance of −12.5% means the as-delivered wall could legitimately be as thin as 0.319 in on the day it was installed, before a single day of corrosion. A baseline TML reading taken and compared against the nominal drawing thickness rather than the certified as-built thickness produces a corrosion rate that is partly real and partly an artefact of mill tolerance, and the error compounds every survey afterward.

The fix is procedural rather than technical: retain the mill certificate against the circuit isometric, and use the certified thickness, not the nominal schedule thickness, as the CML baseline. Where a certificate cannot be located for legacy piping, treat the first survey as the baseline rather than backdating a corrosion rate to an assumed construction thickness — a rate calculated from an assumed starting point is not a measured rate, whatever the report format implies.

This is a recurring finding in third-party audits of piping inspection programmes precisely because it looks correct on paper. The report shows a baseline, a current reading, and a calculated rate, and none of those three numbers is wrong in isolation — the error is that the baseline was never verified against what was actually installed.

Is API 574 mandatory, or is it only a recommended practice?

API RP 574 is a recommended practice. Its language is "should", not "shall". API 570 is the inspection code and carries the mandatory requirements — classification, maximum intervals, the rating and repair rules. Where an owner's written inspection programme invokes 574 by name, however, the practice becomes auditable against that programme. Auditors regularly write findings not because 574 was breached, but because a procedure cited 574 and then did something different without justification.

What is the minimum thickness for carbon steel piping under API 574?

There is no single number. API 574 tabulates default minimum structural thicknesses for carbon and low-alloy steel pipe — around 0.07 in for NPS 2 rising to roughly 0.13 in at NPS 20 — which represent the wall needed to carry the pipe's own weight and normal handling loads on conventional spans. That value is a floor, not the answer. The governing minimum is whichever is greater: the pressure design thickness from the construction code, or the structural minimum.

Why did TMLs become CMLs, and does the change matter?

API 570 renamed thickness measurement locations as condition monitoring locations because thickness is only one condition being monitored. A CML may be a UT grid, a radiographic profile shot, a hardness survey point, or an external visual station for corrosion under insulation. The change matters in an audit because a programme that records only thickness at points chosen for accessibility cannot demonstrate that it monitors the damage mechanisms actually credible for the circuit.

Do ultrasonic thickness readings have to be corrected for metal temperature?

Yes, once the metal is meaningfully above the temperature at which the instrument was calibrated. Sound velocity in steel falls as temperature rises, so an instrument calibrated at ambient over-reads a hot component by roughly one percent per 100°F of temperature rise. On a 0.300 in wall at 500°F that is about 0.012 in of phantom metal — enough to hide a full inspection interval of real corrosion when the same error repeats each survey.

How far does an injection point inspection circuit extend?

Injection point circuits are treated as their own circuit because mixing, flashing and localised corrosion concentrate immediately downstream. The circuit conventionally begins at least 12 inches or three pipe diameters upstream of the injection point, whichever is greater, and runs downstream to the second change in flow direction past the injection point. Inside that boundary, CML density is raised and readings are taken on a grid rather than at single points, because the metal loss is local and a single point will miss it.

Does API 574 cover small-bore piping and deadlegs adequately?

It addresses both, but a thickness-led programme built on 574 will not by itself catch the way they usually fail. Small-bore piping most often fails by vibration-induced fatigue at socket welds and unsupported branch connections — a crack, not a thin wall. Deadlegs corrode hardest at the stagnant end, which is the part nobody scaffolds. Both require the inspection plan to specify what is examined and how, not merely where thickness is taken.

Request a consultation

API 574 vs API 570: code versus recommended practice

Direct answer: API 570 is the code. It sets the requirements, written as "shall", for inspecting, rating, repairing and altering in-service metallic piping and its pressure-relieving devices, and it is the document a regulator, auditor or insurer will hold you to. API RP 574 is a recommended practice. It explains how to inspect pipe, valves (other than control valves) and fittings so that the API 570 requirements can be met, but it does not, by itself, impose obligations. Use 570 to decide what must be done and when; use 574 to decide how to do it well.

The two documents were revised together. Both are now in their fifth editions, released in early 2024, and API announced them as a pair. Standards resellers list RP 574's fifth edition as February 2024 with an Addendum 1 issued in 2025; confirm which addendum your site has adopted, because a program written to the base edition may not reflect addendum changes. API 570's fifth edition is likewise dated February 2024. The rest of this section maps the division of labour between the two, which is where most program and audit confusion starts.

Side-by-side comparison

The table below summarizes the roles of the two documents and the neighbouring standards that a piping program normally uses with them. It describes each document in our own words; it does not reproduce code text.

QuestionAPI 570 (code)API RP 574 (recommended practice)
Legal and contractual weightRequirements ("shall"); referenced in US federal and state regulation and widely treated as RAGAGEP for PSM mechanical integrityGuidance ("should"); becomes binding only if your program, contract or jurisdiction adopts it
Main readerOwner-user, API-certified piping inspector, inspection supervisor, repair organizationPiping inspectors, NDE technicians, inspection planners, corrosion and materials specialists
Sets inspection intervalsYes, by piping class, with an RBI alternativeNo; discusses how to gather data that supports intervals
Piping classificationYes (Classes 1 to 4 by consequence)No
Corrosion rate and remaining lifeRequires the calculations and their useExplains measurement quality, data treatment and practical pitfalls
Inspection techniquesStates which inspections are requiredDescribes components, typical damage, and how to apply UT, RT, visual and other techniques
Repairs, alterations, reratingRequirements, including authorization by the piping inspectorGuidance on evaluating results, follow-up actions, repairs and rerating (expanded in the fifth edition)
Fifth-edition headline changesInspection Supervisor role; stronger pressure-relief device requirements; review and deferral of inspection repair recommendationsMore guidance on selecting inspection activities and locations; stainless steel tubing in chloride service; updated evaluation, FFS, rerating, repair and remediation guidance
ExclusionsDefined scope with listed excluded and optional systemsSpecialty items such as control valves, instrumentation and furnace tubulars are outside its component guidance

The short version: if a sentence in your inspection procedure says "must", check that it traces to API 570, to the owner's own standard, or to a regulation. If it traces only to RP 574, it is your company's choice to make it mandatory, which is legitimate, but it should be a conscious choice.

What API 570 hands off to RP 574

API 570 repeatedly refers readers to RP 574, together with API RP 571 for damage mechanisms, rather than repeating technique guidance. Reviewing the fourth edition (2016) text, those cross-references covered, among other things:

  • Condition monitoring location (CML) selection. The code requires CMLs placed where the expected damage will show up; RP 574 and RP 571 describe where specific mechanisms concentrate, which is what makes a CML layout defensible.
  • Soil-to-air interfaces. The code requires attention to the zone where buried piping emerges; RP 574 describes the practical inspection approach.
  • Small-bore piping design issues. The code sets which small-bore lines must be inspected; RP 574 discusses the design features that drive failures. Our guide to small-bore and auxiliary piping under API 570 covers the mandatory/optional split.
  • Buried piping inspection methods. The code sets buried piping requirements and intervals; RP 574 describes techniques for evaluating buried lines.
  • Non-metallic piping. Guidance on inspecting fibre-reinforced plastic piping was pointed toward RP 574.

Readers working from the fifth edition should confirm that each cross-reference still exists and points to the same RP 574 section; renumbering between editions is common. The practical lesson is unchanged: a piping program that cites API 570 but has never been checked against RP 574 is usually weakest in exactly these handed-off areas, CML placement above all.

Where B31.3, API 571 and API 579 fit on the same map

Piping programs usually involve five documents, and it helps to know which question each one answers.

  • ASME B31.3 (construction). How process piping is designed, fabricated, examined and tested before it enters service, including weld acceptance criteria for new construction. API 570 points back to the construction code for repair and alteration welding principles. See ASME B31.3.
  • API 570 (in-service code). What must be inspected, how often, by whom, and what is required for repairs, alterations and rerating.
  • API RP 574 (in-service practice). How to carry out piping inspection on pipe, valves and fittings in a way that gives reliable data.
  • API RP 571 (damage mechanisms). What damage to expect in which service, and which inspection approach suits each mechanism. See API 571.
  • API 579-1/ASME FFS-1 (fitness-for-service). How an engineer assesses whether a damaged component can stay in service. This is an engineering assessment the owner commissions from qualified engineers; RP 574 discusses when it is used, and Atlantis does not offer it.

When a finding comes in below the minimum required thickness, the sequence normally runs: RP 574 practices make sure the reading is real, API 570 decides what action is required, and API 579-1 is one route the owner may use to justify continued operation. Mixing up these roles is the most common reason audit findings are disputed.

Applying both documents in one program: a worked example

Consider a Class 2 hydrocarbon circuit in a refinery crude unit, described qualitatively. API 570 sets the circuit's class, requires CMLs, requires a thickness-measurement interval and an external visual interval tied to that class (or to an RBI analysis), and requires corrosion rate and remaining life calculations from the readings. That is the compliance skeleton.

RP 574 then shapes how the work is carried out. It informs where CMLs should sit on elbows, tees, reducers and downstream of control valves; how readings should be taken so the same physical point is measured each time; how temperature affects UT readings on hot lines; how to treat outliers before they distort a corrosion rate; what to look for on valves and flanges that are not part of the thickness grid; and how to record the results. In the fifth edition it also adds guidance on choosing inspection activities and locations, and on stainless steel tubing in chloride service, which is a common source of unexpected leaks on instrument and analyzer tubing.

The owner's API-certified piping inspector reviews the data against the API 570 requirements, calculates remaining life, and sets the next inspection date. If readings fall below the minimum required thickness, the inspector and engineer decide on repair, replacement, rerating or an FFS assessment. RP 574's expanded fifth-edition guidance on evaluation, follow-up actions, rerating and repairs supports that decision but does not make it. For the rate calculation itself, see our explainer on long-term and short-term corrosion rates.

NDE techniques RP 574 discusses, and what each is good for

RP 574 describes the techniques an inspector uses to gather piping data. The summary below is our own, matched to the work NDE contractors typically perform inside an API 570 program.

TechniqueTypical use in a piping programKey limitation
UT thickness (spot readings)CML thickness for corrosion ratesMisses localized pitting between points
UT scanning and corrosion mappingLocalized corrosion at elbows, injection points, deadlegsSlower; needs access and surface preparation
Profile radiographySmall-bore lines, threaded connections, insulated linesRadiation controls; sizing accuracy depends on technique
Phased array UTWeld cracking, encoded corrosion mapsProcedure must be qualified for the geometry
Guided wave screeningLong runs, insulated or buried piping, road crossingsScreening only; call-outs need follow-up
Visual examinationExternal condition, supports, coatings, insulation, leaksSurface only

Method choice should follow the damage mechanism, not the other way round. Our method selection by damage mechanism guide walks through that logic, and the corrosion mapping and guided wave testing pages explain the two most common upgrades to a spot-reading program.

Common mistakes when using API 570 and RP 574 together

  • Citing RP 574 as the source of a mandatory interval. Intervals come from API 570 or from an RBI analysis under it, not from the recommended practice.
  • Ignoring RP 574 because it is "only guidance". Auditors and incident investigators routinely ask whether recognized practice was followed. A CML layout that ignores the practice's guidance on where damage occurs is hard to defend.
  • Running an old RP 574 edition against a new API 570. The two were updated together; programs that update one and not the other end up with mismatched terms and references.
  • Missing addenda. Addenda change text without a new edition number. Check which addendum your procedures cite.
  • Applying RP 574 component guidance to control valves or instrumentation. Those are outside its component scope; use the manufacturer's guidance and the owner's standards.
  • Treating stainless tubing as low-risk. The fifth edition's added guidance on stainless tubing in chloride service exists because chloride pitting and cracking on tubing is easy to overlook.

Regulatory overlay: United States and Canada

In the US, OSHA's Process Safety Management standard (29 CFR 1910.119) requires inspection and testing of process equipment, including piping, under recognized and generally accepted good engineering practices. API 570 is commonly used as that practice for in-service piping, and API notes that its inspection codes are incorporated in federal and state regulation. RP 574 is not normally cited by regulation directly, but a facility that adopts it into its mechanical integrity procedures will be audited against what it adopted. EPA's Risk Management Program (40 CFR Part 68) has a parallel mechanical integrity element. Confirm your facility's specific obligations with your jurisdiction.

In Canada, pressure piping falls under provincial regulators. In Alberta, ABSA administers the pressure equipment rules and owner integrity management programs; in Ontario, TSSA administers the boilers and pressure vessels regulation, which adopts CSA B51. Provincial requirements may reference or accept API 570 and RP 574 practices within an owner's program, so confirm with the regulator in your province.

How Atlantis supports API 570 programs that use RP 574 practices

Atlantis performs the NDE that an API 570 program calls for, using the techniques RP 574 describes: UT thickness and corrosion mapping, profile radiography by crews licensed where the work is done, PAUT, guided wave screening with follow-up, and visual examination. ASNT-certified technicians work under ASNT Level III oversight, to written procedures. We deliver results to the owner's API-certified piping inspector, who remains the inspector of record and sets intervals and repair decisions. See piping circuit and CML inspection. Request a quote for piping NDE; we reply within one business day.

More questions about API 574

What does API 574 cover?

RP 574 covers inspection practices for piping system components: pipe, valves other than control valves, and fittings, including their typical damage, how to inspect them, and how to evaluate and record the results. It supports API 570.

What is the latest edition of API 574?

The fifth edition, released in early 2024 alongside API 570's fifth edition. Resellers list an Addendum 1 issued in 2025. Check api.org or your standards subscription for the current status.

Is API 574 a code or a standard?

It is a recommended practice. API 570 is the code. A recommended practice gives guidance; it becomes binding when an owner, contract or jurisdiction adopts it.

What is the difference between API 570 and API 574?

API 570 sets requirements: classes, intervals, calculations, repairs and authorization. RP 574 explains how to carry out piping inspection so that those requirements are met with reliable data.

Does API 574 set piping inspection intervals?

No. Intervals come from API 570, either by piping class or through a risk-based inspection analysis performed under it.

Does API 574 cover control valves?

No. Control valves are outside its component guidance, along with specialty items such as instrumentation and furnace tubulars.

What changed in the fifth edition of API 574?

API lists added guidance on selecting inspection activities and locations, stainless steel tubing in chloride service, and updated guidance on evaluating results, follow-up actions, fitness-for-service, rerating, repairs and remediation.

Do NDE technicians need API certification to work on an API 570 program?

The API 570 certification applies to the authorized piping inspector. NDE technicians are normally qualified under the owner's written practice, commonly ASNT SNT-TC-1A; see ASNT Level III training.

Have a piping campaign coming up? Tell us the scope or ask about corrosion mapping and profile RT.

Sources: API announcement of API 570 and RP 574 fifth editions (api.org); API 570 4th edition (2016) cross-references reviewed; standards-reseller listings for edition and addendum dates. Summary only; confirm against the current editions.

Next step: UT Level 1 course

The corrosion rates and remaining-life figures that API RP 574 works from come from ultrasonic thickness readings. The UT Level 1 course covers straight-beam thickness measurement, calibration on reference blocks and recording readings to a written procedure. Completing a course is not certification: under an employer-based scheme such as SNT-TC-1A, your employer certifies you after its own examinations.

For employers: ASNT Level III consulting covers procedures, the written practice, qualification examinations and audit support for your technicians.