Eddy Current Testing for Heat Exchanger Tube Inspection

Why Heat Exchanger Tubes Fail Quietly Until They Don't

A shell-and-tube heat exchanger rarely announces its problems. A tube can lose 60% of its wall thickness to erosion-corrosion at a baffle and still hold pressure right up until a plant trip stresses it past the remaining margin. By the time an operator sees a temperature excursion on the process side or a chloride spike downstream, the failure has already happened. This is precisely why eddy current testing (ECT) became the workhorse volumetric method for tube bundles in refineries, petrochemical plants, and power generation: it is fast enough to inspect thousands of tubes during a turnaround window, and sensitive enough to catch the early-stage wall loss that visual and hydrostatic testing miss entirely.

This piece is written for the inspectors doing the work and the reliability engineers who have to sign off on the plugging list afterward. We will walk through the physics that make eddy current effective on non-ferromagnetic tubing, where it breaks down on ferromagnetic alloys, the probe technology you actually choose between on a job, calibration requirements per ASME Section V, and the decision tree between ECT, remote field testing (RFT), internal rotary inspection system (IRIS), and ultrasonic testing (UT) for a given tube bundle.

The Physics: Why Eddy Currents Detect What They Detect

When an alternating current flows through a bobbin coil positioned inside a tube, it generates a primary magnetic field. That field induces circulating eddy currents in the tube wall directly beneath the coil. Those eddy currents, in turn, generate their own secondary magnetic field that opposes the primary field — this is Lenz's law doing the work. The coil's impedance changes in response to that secondary field, and any change in the tube's electrical conductivity, magnetic permeability, or wall thickness alters the eddy current flow pattern, which shows up as a measurable impedance change at the coil.

A defect — a pit, a wall-thinning zone, a crack — disrupts the eddy current path locally. The coil sees that disruption as a shift in both the magnitude and phase angle of its impedance signal, plotted on an impedance plane display. The phase angle is the critical piece: for a given tube alloy, frequency, and fill factor, the phase angle of the signal correlates to defect depth as a percentage of wall thickness, while amplitude roughly correlates to defect volume. An experienced analyst reads the "signature" of a defect — its phase rotation and the shape of the Lissajous pattern — to distinguish OD pitting from ID pitting, a baffle-support indication from a genuine defect, and a real flaw from a probe wobble or dent-induced permeability change.

The depth of penetration of eddy currents into the tube wall is governed by the standard depth of penetration equation, which is inversely proportional to the square root of frequency, conductivity, and permeability. This is why test frequency selection is not arbitrary — a higher frequency (say 400 kHz on a 0.049-inch admiralty brass condenser tube) gives excellent resolution for OD-side and near-surface defects but limited penetration, while a lower frequency (around 30-50 kHz) is needed to see through the full wall of a thicker titanium or stainless steel tube. Most field procedures run a multi-frequency mix — a primary test frequency plus one or two auxiliary frequencies used for mixing out unwanted signals like support plate indications or probe wobble.

Where Conventional ECT Excels — and Where It Fails

Conventional bobbin-coil eddy current testing works best on non-ferromagnetic, electrically conductive tubing: admiralty brass, 90/10 and 70/30 copper-nickel, aluminum brass, stainless steel (300 series), titanium, and Inconel/Incoloy in some configurations. These alloys have low or negligible magnetic permeability, so the eddy current field is not dominated and scattered by magnetic domain effects. On these tube materials, a well-calibrated bobbin probe reliably detects general wall loss, localized pitting, and through-wall holes down to roughly 20% wall loss for a comparably-sized calibration defect, with better sensitivity achievable using rotating or array probes for smaller, more localized flaws.

The method breaks down on ferromagnetic tubing — carbon steel, 400-series stainless, and duplex grades with significant ferrite content. In these alloys, magnetic permeability variation swamps the conductivity-based signal; a bobbin coil sees permeability noise orders of magnitude larger than the signal from a real flaw, making conventional ECT effectively blind on ferromagnetic tube walls thicker than a few hundredths of an inch. This is the single most common mistake I see junior technicians make when scoping a job: assuming ECT is a universal tube-inspection method rather than checking the tube alloy against the ASME calibration standard first.

For ferromagnetic tubes, two workarounds exist:

• Remote Field Testing (RFT): Uses a low-frequency exciter coil and a detector coil separated axially, relying on the field that travels twice through the tube wall (once out, once back) rather than direct coupling. RFT is largely immune to permeability noise and is the standard method for carbon steel feedwater heater and air-cooled heat exchanger (ACHE) tubes, though its resolution and speed are lower than bobbin ECT.
• Partial Saturation Eddy Current Testing (PSECT/Saturation ECT): Uses a strong DC or permanent magnet to locally saturate the ferromagnetic material near the probe, suppressing permeability variation enough to allow a conventional-style eddy current signal to be read. This is common on duplex stainless and ferritic stainless condenser tubing where RFT's lower resolution is not acceptable.

Choosing the wrong technique for the alloy does not just waste a shift on the bundle — it produces a report that looks complete but has effectively zero probability of detection (POD) for real degradation. This is a recurring root cause when a tube fails between inspection intervals despite a "clean" prior ECT report on file.

Probe Selection: Bobbin, Rotating, and Array

Choosing a probe is a tradeoff between speed, resolution, and defect orientation sensitivity.

• Bobbin coil (absolute or differential): The standard screening probe for full-bundle surveys. A differential bobbin coil compares two adjacent coil windings against each other, which cancels out slow, gradual changes (like general wall thinning or temperature drift) and makes the probe extremely sensitive to sharp, localized defects — but it can miss long, gradually tapering wall loss because there is nothing for the two coils to differentiate against. An absolute coil, by contrast, compares the signal to a fixed reference and is better at detecting gradual, longer-length degradation such as general erosion zones, at the cost of being more susceptible to drift and lower sensitivity to small pits. Most field procedures run both configurations, or an absolute/differential hybrid coil, specifically to cover both defect morphologies in one pass.
• Rotating probe / RPC (Rotating Pancake Coil): A pancake-style coil mounted on a rotating head that spirals along the tube's ID as it advances, giving true circumferential and axial resolution. RPC is the go-to follow-up technique when a bobbin coil flags an indication that needs characterization — particularly for axial or circumferential cracking, ID stress corrosion cracking (SCC) near tube-to-tubesheet joints, or discriminating a crack-like indication from a volumetric pit. The tradeoff is speed: RPC inspection runs at a fraction of bobbin coil throughput, so it is used selectively rather than on 100% of a bundle.
• Array probes (Array-based ECA): Multiple small coil elements arranged circumferentially around the probe head, electronically multiplexed to give near-RPC resolution at closer to bobbin-coil speed, without the mechanical rotation. Array probes have become the preferred choice for tube-to-tubesheet joint inspection and for U-bend regions where mechanical rotation is difficult, and are increasingly specified for high-consequence bundles (nuclear-adjacent steam generators, high-pressure feedwater heaters) where both speed and circumferential sensitivity matter.

Calibration: The ASME Section V Reference Standard

Every ECT tube inspection procedure is built around a calibration (reference) standard manufactured to match the tube being inspected in outer diameter, wall thickness, and alloy — per ASME Boiler and Pressure Vessel Code, Section V, Article 8 (Eddy Current Examination Methods for Nonferromagnetic Tubing), with ASTM E243 as the underlying standard practice for eddy current examination of copper and copper-alloy tubing, and ASTM E2096 for the general inservice inspection of nonferromagnetic heat exchanger tubes.

A typical ASME calibration tube includes a defined set of artificial discontinuities: four flat-bottom holes (FBH) at 20%, 40%, 60%, and 100% (through-wall) of nominal wall thickness, a circumferential groove representing OD wall loss, and a drilled hole simulating a through-wall pit. The technician runs the probe through this standard before, periodically during, and after the production run, normalizing the phase angle so that the 100% through-wall hole reads at a fixed reference angle (commonly 40 degrees on the impedance plane display) and setting gain so that signal amplitude from the reference defects falls within the procedure's specified range. Any calibration drift beyond the procedure's tolerance — typically checked every 2 to 4 hours or per a fixed number of tubes — requires re-calibration and, in most quality programs, re-inspection of tubes scanned since the last valid calibration.

This calibration discipline is also where digital recordkeeping earns its keep. On a bundle with 2,000+ tubes, tracking which tubes were scanned under which calibration block, with what probe serial number, at what frequency set, is a data management problem as much as an NDT problem — and it's exactly the kind of traceability that structured NDT reporting software is built to enforce automatically rather than leaving it to a technician's handwritten log sheet that gets transcribed at 2 a.m. before a turnaround deadline.

Common Failure and Degradation Mechanisms in Tube Bundles

Knowing what you are looking for shapes how you set up the exam. The mechanisms below account for the large majority of tube plugging decisions across refinery and power-plant bundles:

• General wall loss / erosion-corrosion: Gradual thinning from high-velocity flow, often worst at tube inlet ends or where flow-induced turbulence concentrates — classic in cooling water and seawater-cooled condensers.
• Pitting corrosion: Localized, often deep, OD or ID pits from chloride attack, stagnant water, or microbiologically influenced corrosion (MIC) during idle/layup periods. Admiralty brass and copper-nickel condenser tubes are particularly susceptible under poor water treatment.
• Baffle wear / fretting: Mechanical wear where a tube passes through a baffle plate, driven by flow-induced vibration. Shows up as a sharp, localized signal right at the baffle location — one reason baffle plate positions are mapped before the exam so real defects aren't confused with support-plate signal artifacts.
• Denting: Localized ID diameter reduction from corrosion product buildup (magnetite) at the tube-to-tubesheet or tube-to-support-plate crevice, historically a major issue in recirculating steam generators. Denting itself changes local permeability and can mask or mimic other signals, which is why dent-affected regions typically get flagged for RPC follow-up rather than a bobbin-only call.
• Stress corrosion cracking (SCC), ID and OD: Intergranular attack (IGA) and intergranular stress corrosion cracking (IGSCC) at tube-to-tubesheet crevices and U-bends, historically associated with caustic and chloride environments in nuclear and fossil steam generator tubing. These crack-like indications are the primary reason bobbin-only screening is inadequate for high-consequence bundles — cracks are notoriously under-called by bobbin coils and require RPC or array follow-up for reliable characterization.
• Deposit-induced corrosion: Underdeposit corrosion beneath scale, biofouling, or corrosion product buildup, common in cooling water service with inadequate chemical treatment. Deposits can also attenuate or distort the eddy current signal itself, which is why cleaning (mechanical or hydrolasing) before ECT is a prerequisite step in nearly every field procedure, not an optional one.

ECT vs. IRIS vs. RFT vs. UT: Choosing the Right Method

No single volumetric method covers every tube material and defect type, and picking wrong is a common — and expensive — mistake on a turnaround scope.

• Eddy current (bobbin/RPC/array): Fastest method, ideal for non-ferromagnetic tubing (admiralty brass, copper-nickel, stainless, titanium). Excellent for general wall loss and pitting screening across full bundles; RPC/array needed for crack characterization.
• Remote Field Testing (RFT): The default for carbon steel and other ferromagnetic tubing where conventional ECT can't overcome permeability noise. Slower than bobbin ECT and generally lower resolution, but it is the practical, code-recognized option for ferromagnetic bundles like many air-cooled heat exchangers and feedwater heaters.
• Internal Rotary Inspection System (IRIS): An ultrasonic method using a rotating mirror to direct a focused UT beam circumferentially around the tube ID, measuring actual wall thickness rather than an impedance-derived estimate. IRIS is material-independent (works on ferromagnetic and non-ferromagnetic tubing alike) and gives a direct, highly accurate thickness measurement, which makes it the preferred confirmation method for verifying and sizing indications flagged by ECT or RFT — but it requires a couched, water-filled tube and runs far slower than eddy current, so it's typically applied to a sample percentage rather than a full bundle.
• Ultrasonic testing (UT), conventional or guided-wave: Point-by-point conventional UT thickness gauging is used selectively, often externally on accessible tube runs or ACHE finned tubes where ECT probe access isn't practical; guided-wave UT can screen longer accessible tube lengths for gross wall loss but doesn't match ECT's resolution for pitting.

In practice, most turnaround scopes run ECT (or RFT, per alloy) as the 100% screening tool across the bundle, then follow up flagged tubes with IRIS or RPC for sizing confirmation before a plugging decision is made — a two-tier approach that balances speed against the accuracy needed to defend a plug/leave-in-service call to a reliability engineer or code inspector.

POD, Inspection Intervals, and Plugging Criteria

Probability of detection for eddy current testing is not a fixed number — it depends on defect type, size, tube material, probe type, frequency selection, and analyst experience, which is precisely why procedure qualification (often against a known-defect pull tube or a POD study referencing EPRI guidelines for steam generator and condenser tube inspection) matters as much as the raw equipment capability. A bobbin coil calibrated and run correctly can reliably detect volumetric wall loss down to roughly 20% through-wall on a clean OD reference defect, but real-world POD for small, tight ID cracks under a bobbin-only exam is materially lower — which is the technical justification for RPC/array follow-up rather than a matter of preference.

Inspection intervals are typically set by a combination of TEMA (Tubular Exchanger Manufacturers Association) standard guidance, API recommended practices (notably API 573 for fired heater tubes and general inspection philosophy that extends to exchanger bundles), and a plant's own risk-based inspection (RBI) program, which weighs known degradation rate against consequence of failure. A cooling water condenser with a documented history of MIC pitting might see annual or biennial ECT surveys, while a clean, dry, non-corrosive service exchanger might run on a 4- to 6-year turnaround-aligned interval. The RBI-driven interval, not a blanket rule, is the modern standard of care — and it depends on having reliable historical inspection data to trend wall-loss rates tube by tube, which is where digitized, structured inspection records (rather than static PDF reports) materially improve the next interval-setting decision.

Plugging criteria are generally set as a percentage of remaining wall thickness relative to the tube's minimum required wall for the design pressure and temperature, per TEMA tube gauge tables and the exchanger's original design calculations — commonly in the 40-50% remaining wall range as an owner-user retirement threshold, though the exact number is set by the plant's engineering standard and code case, not by a universal number in the code itself. A tube flagged below the plugging threshold, or one with a crack-like indication that can't be reliably sized and dispositioned as volumetric, gets mechanically plugged rather than left in service — a conservative call, but the right one given that a single tube failure during operation can propagate to a tube-to-tubesheet leak, cross-contamination between shell and tube fluids, or in the worst case a tube rupture event.

Building the Program: Data, Training, and Repeatable Execution

A single bundle inspection is a snapshot. The value compounds when tube-by-tube signal data, wall-loss trends, and plugging history are tracked across multiple turnarounds so an analyst reviewing this year's data can see that Tube R14C22 has been losing wall at a consistent rate since the last three outages — a trend invisible if each inspection lives in a standalone PDF. This is the case for treating tube inspection data as a managed asset: an Atlantis NDT ERP deployment ties inspection history, technician certifications, and equipment calibration records to the asset register so a reliability engineer can pull five turnarounds of ECT trend data for a single exchanger in minutes rather than digging through archived reports. For teams visualizing exchanger condition alongside piping and vessel data on a live asset model, a digital twin platform can overlay tube-map wall-loss data directly onto the 3D exchanger model, making it immediately visible where degradation is clustering across the bundle.

None of this software matters without qualified analysts behind the probe. Eddy current signal interpretation is a trained skill — distinguishing a genuine crack signature from a support-plate artifact or probe wobble takes hundreds of hours of supervised analysis before an analyst should be signing off on a plugging list independently. NDT training programs built around ASNT SNT-TC-1A guidelines establish that structured path from Level I data collection through Level II independent analysis and reporting. For organizations building or auditing their written practice, qualification records, and procedure technical basis — particularly where a POD claim or interval decision needs to be defensible to a client or a jurisdictional inspector — ASNT Level III consulting provides the procedure review and personnel qualification oversight that keeps an in-house ECT program audit-ready rather than scrambling before the next turnaround.

Eddy current testing on heat exchanger tubing sits at an intersection most NDT methods don't: it has to be fast enough to clear thousands of tubes in a compressed outage window, and precise enough that the plugging decisions built on it hold up against both engineering scrutiny and the cost of getting it wrong. Understanding the physics behind the phase angle, matching the probe and technique to the tube alloy, and running calibration discipline exactly to ASME Section V Article 8 are what separate a screening exercise from an inspection program a plant can actually trust its next run cycle to.

Whether you're scoping the next bundle survey, building out an in-house ECT program, or looking for support on procedure qualification and reporting, here's where to go next:

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.