Corrosion Under Insulation (CUI) 2026 — API RP 583 Decoded

API RP 583 in plain English: what it is and what it asks owners to do

API RP 583, Corrosion Under Insulation and Fireproofing, is API's recommended practice for designing, maintaining, inspecting and mitigating external corrosion hidden under insulation (CUI) and under fireproofing (CUF). It applies to pressure vessels, piping, storage tanks and spheres. It covers external corrosion of carbon and low-alloy steels and external chloride stress corrosion cracking (ECSCC) of austenitic and duplex stainless steels. The current edition is the 3rd, published April 2026. It is guidance, not an inspection code. The binding interval and acceptance decisions still sit in API 510, API 570 or API 653 and with the owner's inspector.

That last point is the one most summaries of "API 583" get wrong, including earlier versions of this article. A recommended practice tells an owner-user how to understand a problem and what good practice looks like. It does not set a minimum thickness. It does not fix an inspection interval. It does not qualify NDT personnel, and it has no mandatory appendices or code cases the way an ASME construction code does. When a CUI finding has to be dispositioned, the inspector of record works from the in-service code that governs the equipment: API 510 for pressure vessels, API 570 for process piping and API 653 for aboveground storage tanks. RP 583 sits behind those codes as the reference on how CUI and CUF actually happen and how to find them.

API's own publication notice for the first edition gave the document four purposes. It helps owner-users understand how complicated CUI and CUF are. It explains the advantages and limitations of the NDE methods used to find the damage. It offers an approach to assessing CUI/CUF risk, meaning likelihood and consequence. And it gives guidance on designing, installing and maintaining insulation systems so the damage does not start. Everything else in the document supports one of those four aims.

Edition history: 2014 to the April 2026 third edition

RP 583 was first published in May 2014. The first edition ran to 88 pages. Distributor listings describe the third edition, released in April 2026, at about 130 pages. The scope is the same: CUI and CUF on pressure vessels, piping, tanks and spheres, carbon and low-alloy steel corrosion, and ECSCC of austenitic and duplex stainless steels. Catalogue descriptions of the 2026 edition describe its risk content as a "likelihood assessment" approach. That fits the first edition, whose informative Annex A gave examples of a qualitative likelihood assessment system.

This page does not reproduce clause numbers or tables from the 2026 edition. API documents are copyrighted. If your CUI procedure cites RP 583, buy the current edition from API or an authorised distributor and check every cross-reference against it. People searching for an "API RP 583 PDF" should know there is no legitimate free copy. Files labelled as free downloads are usually old editions or unauthorised copies, and neither is a sound basis for a site procedure.

The first edition stated that it did not cover atmospheric corrosion or corrosion at uninsulated pipe supports, though it did discuss corrosion at insulated supports. Bare pipe sitting on a beam therefore falls outside RP 583 and is dealt with as external corrosion under the piping code.

Where the damage happens: temperature, water and design details

An earlier version of this page gave a CUI band of "0-175°C". The figures more commonly cited from API 571 and API 570 are different. For carbon and low-alloy steel the band is roughly 10°F to 350°F (−12°C to 175°C). For austenitic and duplex stainless steel, ECSCC under insulation is associated with roughly 140°F to 400°F (60°C to 205°C). The two bands do not overlap neatly because the mechanisms differ. In carbon steel the problem is wall loss from wet, oxygenated water held against the metal. In stainless steel the problem is cracking driven by chlorides that concentrate as trapped water evaporates. Confirm the band your programme uses against the current editions of API 571 and the governing inspection code. Do not rely on a summary for it, this one included.

Temperature alone does not tell you which lines are at risk. RP 583 and API 571 both point to three conditions that widen the population:

• Cyclic or intermittent service. A line that normally runs above the band but cools during outages spends time wet and warm. Hot lines that spend time idle belong in the programme.
• Cold service below the dew point. Condensation forms continuously on the steel and the insulation holds it there. The first edition of RP 583 had a separate inspection subsection for piping operating below freezing, and separate tables for piping running below the dew point.
• Heat tracing and mothballing. Tracing leaks and tracing penetrations let water in. Equipment left insulated during a long shutdown can corrode faster than it did in service.

Water has to get in before any of this happens, which is why design details predict CUI better than pipe specifications do. The usual locations include:

• penetrations for nozzles, vents, drains and instruments
• insulation support rings on vertical vessels
• pipe supports and shoes where the jacketing is interrupted
• low points and the bottom of vertical runs
• terminations and missing end caps
• jacketing overlaps lapped the wrong way to shed water
• failed or missing caulking
• access panels cut and never resealed
• areas exposed to deluge, steam vents or cooling-tower drift

The first edition included photographs of exactly these failures. Its location tables, split by vessels, piping, tanks and spheres, are the most practical part of the document for anyone building a walk-down checklist.

Corrosion under fireproofing: the half of RP 583 most programmes skip

Search interest focuses on insulation, but the "F" in CUF matters just as much. Fireproofing goes on structural steel, vessel skirts, sphere legs and pipe-rack members. Water gets behind cracked, spalled or poorly sealed fireproofing and corrodes the steel underneath. The first edition illustrated a sphere-leg failure caused by CUF, which is the right example to keep in mind. The steel is structural, the fireproofing was installed for a fire case, and nobody looks behind it until something moves.

RP 583 covered the main fireproofing families: dense and lightweight concrete, lightweight cementitious products, and intumescent and subliming coatings. Each fails differently. Concrete cracks and separates at the top edge, which then acts as a water trap. Cementitious products absorb water. Coating systems blister or disbond. Inspection starts with a visual survey of the fireproofing for cracks, bulging, rust staining, separation at terminations and damage from impact or maintenance. Fireproofing is then removed at suspect locations so the steel can be examined and measured. Thickness and pit-depth data from those openings go to the structural or mechanical engineer, who owns the disposition.

When you scope a CUI campaign, ask whether fireproofed skirts, sphere legs and structural supports are in it. They are often left out because they belong to a different budget or a different discipline.

NDE methods for CUI and CUF: what each one sees and what it misses

The core of RP 583 is a frank comparison of inspection methods, and that is where it helps most in practice. The first edition covered guided wave testing, profile radiography (tangential and film-density), flash radiography, radiometric profiling, real-time radiography, pulsed eddy current, neutron backscatter and infrared thermography, plus pit-depth measurement once insulation is off. The document's message was that no single method does the whole job. Some tell you where water is, some tell you roughly where wall loss is, and only direct measurement tells you how much steel is left.

The first edition included a figure contrasting average and minimum wall thickness within a PEC footprint. It is the most important single idea for anyone buying CUI screening. A screening method that averages, like PEC, or one that reports change rather than thickness, like guided wave, is a ranking input. It decides where insulation comes off. The thickness used in a corrosion-rate or remaining-life calculation should come from direct measurement at that opening: contact UT, encoded corrosion mapping where loss is spread out, and pit gauging where it is localised. A programme that records screening averages as thickness readings has built a record that will not stand up to an inspector's review.

ECSCC on stainless steel needs a different approach. Thickness methods do not find cracking. Once insulation is removed at suspect locations, the surface is cleaned and examined with penetrant testing. Where chloride exposure is suspected, ask whether the insulation in service meets a low-chloride specification for stainless steel. ASTM C795 is the usual reference.

How RP 583 connects to API 510, API 570 and API 653

RP 583 tells you how CUI behaves and how to find it. The inspection codes tell you what an inspection programme must contain. In practice the connection works like this:

• API 570 (process piping, current 5th edition, February 2024) treats CUI as a specific external inspection concern. It identifies susceptible piping by temperature and service and gives guidance on how much NDE to apply at suspect locations by piping class. The piping inspector applies those provisions. RP 583 explains the mechanism and the methods behind them. See piping circuit and CML inspection for how circuit-level thickness data is organised.
• API 510 (pressure vessels) expects external inspections to consider CUI on insulated vessels in susceptible service. Vessel locations named in RP 583 include support rings, nozzles, manways, skirts and the bottom head of vertical vessels.
• API 653 (aboveground storage tanks) covers insulated tank shells, roof-to-shell junctions and the bottom course. Note that API 653 Addendum 4 (July 2025) revised the tank-bottom thickness measurement text, so check you are working to the current edition and addendum.
• API 571 is the damage-mechanism reference. Its CUI and chloride SCC entries describe the mechanism, affected materials and critical factors that RP 583 then expands on. See the API 571 damage mechanisms guide.

RP 583's risk content, and the older Annex A qualitative likelihood examples, are owner tools. They feed an owner's risk-based inspection programme under API 580/581, or a simpler susceptibility ranking. Likewise, when CUI damage is found below the code minimum, the owner may evaluate it under API 579 fitness-for-service. Those assessments belong to the owner's integrity engineers and the inspector of record. Atlantis does not perform RBI or FFS. We supply the measured NDE data that both depend on.

A worked example: building a CUI campaign on an insulated unit

Here is how an RP 583-informed campaign typically runs on a refinery unit with several hundred insulated piping circuits and a handful of insulated vessels.

1. Define the population. The owner lists insulated equipment and its operating temperature, flagging cyclic, idle, cold and heat-traced service. Anything outside the susceptible bands and not in those categories is documented as excluded, with the reason recorded.
2. Rank. The owner ranks circuits on susceptibility and consequence. That can be a formal risk method or a simple qualitative scheme like the examples in RP 583's annex. Previous CUI findings and jacket condition carry a lot of weight.
3. Walk down. Inspectors survey the jacketing, caulking, terminations, supports and low points. Each damaged area is photographed and tagged so the walk-down produces a location list.
4. Screen. On high-ranked circuits and walk-down locations, the NDE contractor applies the agreed screening: profile radiography on small bore and fittings, PEC on larger lines and vessel shells, and guided wave on long runs or inaccessible sections such as road crossings. Locations screened as clear are recorded as well as those that were not.
5. Open and measure. Insulation comes off at screening anomalies and at locations where visual evidence alone justifies it. The extent rule (how far either side of the indication) is agreed before mobilisation. The surface is cleaned and photographed, then measured with UT, corrosion mapping or pit gauge. Stainless locations get penetrant testing for ECSCC.
6. Report and hand over. Results go to the owner's inspector, organised by circuit and location. Each finding is reported with method, instrument and calibration, measured values, photographs and coverage. The inspector evaluates against the governing code and sets next actions and intervals.
7. Reinstate properly. Insulation, jacketing and caulking are reinstated to the owner's specification. Poor reinstatement is a common cause of the next cycle's CUI.

Mitigation: what actually prevents CUI (and what does not)

An earlier version of this page listed mitigation measures that apply to internal or buried corrosion rather than CUI: corrosion inhibitor injection, cathodic protection for buried equipment, post-weld heat treatment and alloy upgrades with more chromium or molybdenum. None of these is a CUI control. The measures RP 583 is built around are:

• Coatings under insulation, chosen for the service temperature and for wet, cyclic conditions. This includes thermal spray aluminium on carbon steel. See our TSA for CUI mitigation article.
• Insulation material selection, including water-resistant materials and low-chloride insulation on stainless steel.
• Jacketing and caulking detailed to shed water: correct overlaps, end caps, sealed penetrations and maintained caulk.
• Design that avoids water traps: fewer penetrations, supports designed for insulated pipe, and removing insulation that serves no process or personnel-protection purpose. Where insulation is there only for personnel protection, cages and guards are an alternative.
• Maintenance and quality control of the insulation system itself, treated as part of corrosion control rather than housekeeping.

How Atlantis supports an RP 583-based CUI programme

Atlantis performs the NDE that a CUI or CUF programme calls for. That includes visual surveys, profile and digital radiography by crews licensed where the work is, pulsed eddy current and guided wave screening, contact UT and corrosion mapping at openings, and penetrant testing for ECSCC on stainless. ASNT-certified technicians do the work under ASNT Level III oversight. Results go to your API-certified inspector, who remains the inspector of record and sets intervals and dispositions. Atlantis does not perform RBI, CUI risk ranking or FFS assessments; those stay with your integrity team. See corrosion under insulation inspection for the service scope.

Request a CUI inspection scope and quote, with a reply within one business day.

API RP 583 questions inspectors ask

What is the current edition of API 583?

The 3rd edition, published in April 2026. The first edition was published in May 2014. Buy the current edition from API or an authorised distributor and check your procedures against it.

Is there a free API RP 583 PDF?

No legitimate free copy exists. API standards are copyrighted. Free files online are usually old editions or unauthorised copies and should not be the basis of a site procedure.

Does API RP 583 set CUI inspection intervals?

No. It is a recommended practice. Intervals come from the governing inspection code (API 510, API 570 or API 653), the owner's inspection plan and the jurisdiction, as decided by the owner's inspector.

What does API RP 583 inspection involve?

It typically involves a susceptibility ranking, a visual survey of insulation and jacketing, non-intrusive screening (profile radiography, pulsed eddy current, guided wave, thermography or neutron backscatter for moisture), and targeted insulation removal with direct thickness measurement or crack detection at suspect locations.

What temperature range causes CUI?

For carbon and low-alloy steel, roughly 10°F to 350°F (−12°C to 175°C). For ECSCC in austenitic and duplex stainless, roughly 140°F to 400°F (60°C to 205°C). Cyclic, idle and cold-service equipment also needs attention. Confirm the ranges in the current API 571 and your governing code.

Does Atlantis perform CUI risk ranking or RBI?

No. Atlantis performs the NDE: screening, measurement and crack detection. The owner's integrity engineers and inspector own risk ranking, RBI and FFS.

Talk to us about CUI screening and follow-up NDE · Ask about corrosion under fireproofing inspection