Piping Inspection (API 570): What Triggers an Internal Inspection

Why "Internal" Inspection Is the Wrong First Question

Every API 570 inspector eventually gets the same phone call: a plant reliability engineer wants to know when a piping circuit is "due" for an internal inspection. The honest answer is that API 570 doesn't schedule internal inspections on a calendar the way it schedules external visual checks. It schedules them on corrosion behavior, and corrosion behavior is what forces the question of whether you can inspect from outside the pipe at all, or whether you have to get inside it. Before triggering an internal inspection, an experienced inspector first asks whether the degradation mechanism is even visible or measurable externally. Internal corrosion, erosion, and certain cracking mechanisms hide from the outside; external corrosion, coating failure, and insulation damage do not. That distinction drives everything that follows.

API 570, the Piping Inspection Code covering in-service inspection, rerating, repair, and alteration of metallic and FRP piping systems, sorts piping into three classes based on consequence of failure: Class 1 (highest risk — hydrocarbon/toxic service, high pressure, or governed by jurisdictional regulation), Class 2 (moderate risk), and Class 3 (lowest risk, typically utility or low-hazard systems). Class assignment sets the maximum inspection intervals for thickness measurements at CMLs (condition monitoring locations) and for external visual inspection, so it materially affects the interval math discussed below. A circuit's class is not a formality — it's the first fork in the decision tree that determines whether an internal inspection is even on the table for this cycle or the next.

The Interval Math: RL/2 and the 10-Year Ceiling

API 570 Section 6 sets the governing rule for internal (or on-stream, where permitted) inspection intervals: the thickness-measurement interval is the lesser of half the remaining life (RL/2) or the class maximum in the code's table: 5 years for Class 1, 10 years for Class 2 and 10 years for Class 3 (3 years for injection points; Class 4 optional). External visual maximums are 5, 5 and 10 years for Classes 1, 2 and 3. When remaining life is under four years, the interval may be the full remaining life up to two years. An RBI assessment per API RP 580 may be used to set intervals instead. Remaining life is calculated from the corrosion rate derived from thickness readings at condition monitoring locations, using the standard formula:

RL = (t_actual − t_min) / corrosion rate

Where t_min is the minimum required thickness per ASME B31.3 (or B31.1 for power piping, B31.8 for gas transmission) pressure design calculations, and the corrosion rate is derived from either the Long-Term Corrosion Rate (LTCR, using original or previous inspection data) or the Short-Term Corrosion Rate (STCR, using the two most recent readings), whichever governs per the inspector's judgment of which better represents current conditions. A circuit with 0.250 in nominal wall, 0.140 in t_min, currently reading 0.198 in, and a corrosion rate of 4 mpy (mils per year) has a remaining life of (198−140)/4 = 14.5 years, which for a Class 2 or Class 3 circuit sets a maximum interval of 7.25 years (a Class 1 circuit would be capped at 5 years), rounded down to whatever scheduling cadence the plant runs (commonly the next turnaround inside that window, not necessarily exactly 7.25 years out).

The class maximum applies regardless of how favorable the corrosion rate looks. I've had reliability engineers push back on this — "the pipe's barely corroding, why can't we go 15 years?" — and the answer is that, without an RBI assessment, API 570 does not let calculated remaining life alone extend the thickness interval past the table maximum (5 years for Class 1, 10 years for Classes 2 and 3). The code assumes that any metal loss mechanism can change behavior (a new corrosive stream, a process upset, MIC taking hold in a previously benign line) faster than a decade-plus gap would catch it.

The Real Triggers: What Actually Forces an Internal Look

In practice, five conditions trigger (or accelerate) an internal inspection independent of the RL/2 clock:

• CUI indication at insulation breaches or low points. Corrosion Under Insulation is the single most common reason a "healthy" external circuit suddenly needs internal verification. Wet insulation, damaged jacketing, or a CUI risk-based inspection (RBI) flag at temperatures between 25°F and 250°F (the classic CUI window, with the worst case around 140-200°F for carbon steel) pushes the circuit onto an accelerated schedule regardless of where it sits in its RL/2 cycle.
• Process change or new corrosive species. A refinery switching crude slates, introducing a higher-TAN (total acid number) feedstock, or increasing H2S partial pressure in a sour water stripper circuit invalidates the historical corrosion rate. API 570 requires re-evaluation of the interval whenever process conditions materially change, and often the only way to validate the new rate is a UT survey that, if it reveals unexpected metal loss, escalates to internal inspection or RT/UT profile mapping.
• External indications that can't be resolved by surface NDT. Bulging, blistering, visible deformation, or a positive result on a screening technique (pulsed eddy current, guided wave UT) that suggests internal wall loss but can't be quantified from outside forces entry — either a direct visual internal inspection at an opened flange/fitting, or profile radiography, or ultrasonic thickness mapping via multiple CMLs.
• Repair or alteration history. Any circuit that has undergone a repair per API 570 Section 8 (weld overlay, sleeve, hot tap) typically gets a follow-up internal or on-stream verification inside a shortened interval to confirm the repair is holding and the surrounding base metal hasn't continued degrading at the pre-repair rate.
• Missed or overdue CML readings. If a circuit's thickness monitoring lapses past its due date — common on hard-to-access elevated piping or double-jacketed lines — the inspector has no current data to calculate a valid remaining life, and the conservative default is to treat the circuit as due for inspection now, not to extrapolate forward from stale numbers.

On-Stream vs. Internal: When Substitution Is Legitimate

A frequent point of confusion, even among certified inspectors, is that API 570 allows on-stream inspection (external UT thickness surveys at established CMLs) to substitute for a physical internal inspection under defined conditions. This substitution is not automatic — it requires that the piping circuit's corrosion mechanism be uniform, predictable, and reliably detectable from the outside surface. Localized corrosion, pitting, or mechanisms like naphthenic acid corrosion (which attacks preferentially at high-velocity elbows and reducers) are poor candidates for on-stream substitution because a UT grid from the outside can miss a pit cluster that a direct internal visual or profile RT would catch. Class 1 piping in cyclic or high-consequence service is held to a stricter standard for justifying substitution; the inspector of record has to document why on-stream data is equally reliable as direct internal examination, not just more convenient.

This is where the quality of your CML program pays for itself. A circuit with a robust, well-distributed CML population — placed at elbows, tees, low points, and areas of known turbulence per API 570 Appendix B guidance — gives you defensible on-stream data. A circuit with three CMLs on straight runs and nothing at the elbows gives you a false sense of security and, eventually, a surprise leak at a point you never measured.

CUI and the 570/RP 583 Overlap

Because CUI is such a dominant driver of unplanned internal inspections, it's worth separating the two codes inspectors juggle here. API 570 governs the piping inspection program itself — classes, intervals, CML methodology, documentation. API RP 583 (Corrosion Under Insulation and Corrosion Under Fireproofing) provides the risk-ranking methodology for identifying which insulated circuits are most likely to be hiding CUI damage, based on temperature, insulation type, jacketing condition, geographic exposure (coastal/high-humidity sites score worse), and piping configuration (dead legs, low points, and pipe supports are classic CUI traps). A plant running a mature RP 583 program will typically catch CUI-driven internal inspection triggers years before a circuit would have hit its RL/2 date on paper, because the risk score — not the corrosion rate calculation — is what pulls insulation for inspection.

For a facility without a formal RP 583 program, the practical substitute is a disciplined visual walk-down focused on insulation jacketing seams, caulking failure, sagging insulation, staining, and rust bleed-through — all of which are cheap to spot and expensive to ignore. A single missed CUI pinhole leak on a Class 1 hydrocarbon line is the kind of event that ends up in an incident investigation report, not a maintenance log.

Documentation That Survives an Audit

Every internal inspection trigger decision needs a paper trail that would hold up if a jurisdictional authority, insurance auditor, or corporate PSM (Process Safety Management) reviewer asked "why did you inspect this circuit when you did, and why didn't you inspect that one." The record should tie together: the circuit's class and service, the governing t_min calculation and its code basis (B31.3 edition and addenda used), the LTCR/STCR determination and which one governed, the calculated RL and resulting interval, any RBI or RP 583 risk score that accelerated the schedule, and the inspector of record's signature and API 570 certification number. Fragmented records — CML data in one spreadsheet, RBI scores in a separate system, inspection reports in PDF folders with no cross-reference — are the single biggest reason facilities fail to catch an overdue circuit until it's already leaking. This is exactly the gap that a purpose-built inspection management platform closes: when CML history, corrosion rate trending, RBI flags, and inspection due dates live in one system instead of three, the "is this circuit due" question answers itself instead of requiring someone to reconstruct it from memory during a turnaround crunch. Atlantis NDT's inspection management ERP was built around this exact workflow — tracking CML populations, auto-calculating remaining life against the RL/2 rule, and flagging circuits before they go overdue rather than after.

A Worked Example: Sour Water Stripper Overhead Line

Consider a 6-inch, Schedule 40, carbon steel overhead vapor line off a sour water stripper — a classic Class 1 circuit given the H2S content and public/personnel exposure risk. Nominal wall is 0.280 in, calculated t_min per B31.3 is 0.156 in. The last two UT surveys, three years apart, show thickness dropping from 0.225 in to 0.201 in — a short-term corrosion rate of 8 mpy, notably higher than the long-term average of 5.2 mpy calculated from original construction thickness. Because STCR exceeds LTCR and the difference is significant, the inspector uses the more conservative (higher) rate. RL = (0.201−0.156)/0.008 = 5.6 years. Internal (or qualifying on-stream) inspection interval = RL/2 = 2.8 years, well inside the class maximum and driven entirely by the accelerating corrosion trend. If this circuit also shows CUI risk factors — say it's insulated for personnel protection and running near the CUI temperature window at points where steam tracing has failed — RP 583 scoring could pull the interval in further, independent of the UT-derived remaining life.

Building the Skill, Not Just Following the Checklist

API 570 gives inspectors a rule-based framework, but the judgment calls — which corrosion rate governs, whether on-stream substitution is defensible, how aggressively to weight an RP 583 score against a clean RL/2 calculation — are exactly where formal training and mentorship matter. New Level II inspectors often apply the code mechanically: calculate RL, divide by two, schedule it, move on. Experienced Level III inspectors read the circuit's history, service, and configuration first, and use the calculation to confirm or challenge that read. That's a skill built through structured training, supervised inspection hours, and exposure to circuits that don't behave the way the textbook example does. Atlantis NDT's NDT training and certification programs, built on ASNT SNT-TC-1A guidelines, and our ASNT Level III consulting services exist specifically to close that gap — whether that means auditing an existing piping inspection program, training a Level II team on interval logic, or serving as the outside Level III of record for a facility that doesn't have one on staff.

Where This Fits Into a Broader Mechanical Integrity Program

API 570 internal inspection triggers don't operate in isolation. They feed into and draw from the plant's broader RBI program (API 580/581), the fixed equipment mechanical integrity element of PSM (29 CFR 1910.119 in the US), and increasingly, digital asset models that let engineers visualize where CML data, corrosion trends, and inspection history sit relative to the physical piping run. A digital twin platform that overlays live thickness-monitoring data and inspection due dates directly onto a 3D model of the unit turns "which circuits are approaching their RL/2 date" from a spreadsheet query into a visual walk-down anyone on the team — not just the inspector who built the spreadsheet — can understand at a glance. For plants managing hundreds of circuits across multiple units, that visibility is often the difference between a proactive turnaround scope and a reactive shutdown.

The bottom line for any inspector or reliability engineer working through an API 570 program: the internal inspection trigger is never just "the calendar says it's time." It's the RL/2/10-year math, checked against CUI and RBI risk scores, checked against process changes, checked against repair history, and checked against whether your CML population is actually good enough to trust an on-stream substitution. Get any one of those wrong and you either waste a turnaround opening a circuit that didn't need it, or worse, leave one closed that did.

Atlantis NDT Products & Services

Atlantis NDT pairs field expertise with software: NDT inspection management software — Atlantis ERP, a digital twin platform for asset integrity, and NDT reporting software. Build your team with NDT training & certification (ASNT SNT-TC-1A) and ASNT certification pathways, or bring in ASNT Level III consulting. Affordable, accessible, fully customizable — book a free consultation.