ACFM and crack detection through coatings

Alternating Current Field Measurement (ACFM) is an electromagnetic nondestructive testing (NDT) method that both detects and sizes cracks. It works by measuring how a surface-breaking crack disturbs a uniform alternating current field induced in a conductive material. Its greatest practical strength is that it needs no paint stripping — up to a certain coating thickness — and still estimates both the length and the depth of a crack in a single scan. That combination makes it a preferred method on offshore platforms, on painted welds and on structural connections with difficult access. This article covers how ACFM works, the logic behind its sizing, how it differs from MT and ET, its field applications and its limits, from an NDT Level III point of view.

1. How it works: a uniform field disturbed by a crack

An exciting coil inside the ACFM probe induces a uniform alternating current that flows in one direction just below the conductive surface under examination. The current is induced without contact and, because of the skin effect, circulates in a very thin surface layer. Where the surface is smooth and free of discontinuities, this current flows in straight parallel lines and the magnetic field it produces is uniform.

When a surface-breaking crack lies perpendicular to the current path, two things happen: part of the current has to flow around the ends of the crack, and part of it runs down the crack face and back up to the surface. This redirection concentrates current at the crack ends and thins it out at the crack center. The disturbance changes the magnetic field above the surface in two characteristic components:

Bx (the component along the current flow direction): Because current density drops along the crack, Bx shows a trough (minimum) at the crack center. The depth of that trough relates to the depth of the crack.
Bz (the component normal to the surface): Because current concentrates at the crack ends, Bz produces a positive peak at one end and a negative peak at the other. The distance between the two peaks gives the length of the crack.

This is where the elegance of ACFM lies: both components are captured at the same time, one encoding length, the other depth.

2. The butterfly plot and reading the signal

As the probe scans along the full length of a crack, the Bx and Bz signals take a characteristic shape. Plotted against each other (Bx versus Bz), they form a closed loop known in the field as the butterfly plot. For the inspector this plot is the strongest visual tool for separating a real crack from noise: a genuine surface-breaking crack produces a clean, symmetric, closed butterfly, while surface roughness, changes in geometry or lift-off variation (probe-to-surface distance) leave scattered, open or asymmetric traces.

In practice the inspector reads three things together: the trough in the Bx time/position signal, the twin peaks in Bz, and the butterfly the two produce. Classifying an indication as a crack typically requires a clear peak-and-trough pair in Bz and a corresponding drop in Bx. One without the other usually points to geometry or a material effect and calls for careful interpretation.

3. Examination without stripping paint: the decisive advantage

The feature that changes ACFM's field economics is that the examination runs without removing nonconductive coating. Because the exciting field is induced and the measurement is contactless, a thin nonconductive layer such as paint or epoxy between the probe and the conductive metal behaves as lift-off, and calibration accounts for that stand-off.

Keep two cases apart here, because this is exactly where the method is oversold. Detection (seeing whether a crack exists) and sizing (depth estimation in particular) do not tolerate the same coating thickness. With generic probes on nonconductive coating, detection is typically possible up to roughly 5–10 mm of coating, while reliable depth sizing with most standard probes is limited to about 4 mm (some special U-type probes push that limit a little). Collapse the two into a blanket "it works through paint" claim and you lose your safety margin in the field.

The field payoff is still large: the strip, blast, reinstate the coating cycle that MT or PT demands disappears. Stripping the paint from a welded connection on an offshore platform, erecting scaffolding, running the examination and then restoring corrosion protection costs days and serious money. ACFM can scan the same connection with the paint in place, often in a single operation by diver or rope access.

Conductive coatings are a different case. Galvanized (zinc) or metallized coatings conduct; the coating itself carries part of the induced surface current, so it does not behave as passive lift-off the way paint or epoxy does. As it gets thicker, a conductive coating can suppress and mask the indication from the crack beneath it; simulation work shows that indications in conductive zinc coating can become undetectable at surprisingly modest layer thicknesses. Conductive metal coatings therefore call for separate engineering assessment, a separate probe and a separate approach; the rough lift-off rule that applies to paint does not carry over.

Finally, lift-off is not free: as the stand-off grows, signal amplitude falls, the noise floor rises and the uncertainty of the depth estimate increases. So where the coating thickness is known, account for it in calibration, and interpret depth values more cautiously on thick or irregular coatings.

4. Sizing: estimating length and depth

What separates ACFM from a simple go/no-go scanner is its sizing capability, and the physics behind that capability is sound.

Length. Read the crack length from the positions of the positive and negative peaks in the Bz signal. Because these two peaks correspond to the points where current flows around the crack ends, the distance between them relates directly to the physical crack length. Length measurement is generally reliable and relatively tolerant of probe orientation.

Depth. Depth comes mainly from the depth of the trough in the Bx signal: the deeper the crack, the more pronounced the thinning of current at its center, and the larger the drop in Bx. The measured Bx and Bz amplitudes are converted to a depth value through the method's theoretical crack model (the ACFM "unfolding" model that originated at University College London and was later commercialized by TSC) and, above all, through its calibration charts. A key advantage of ACFM is that its largely model-based depth estimate reduces the need for a separate reference-crack calibration for every material and geometry.

Set your field expectations accordingly. Under good conditions, ACFM depth estimation carries a condition-dependent uncertainty ranging from sub-millimeter to a few millimeters. No absolute figure is standardized, because the number depends on the probe, the coating and the geometry. At that level the result is strong input data for fracture mechanics and fitness-for-service (FFS) assessment. Absolute depth accuracy, however, remains sensitive to lift-off, crack profile (cracks that are not semi-elliptical, that are branched or that are closed), nearby geometry and probe alignment. For critical accept/reject or remaining-life decisions, confirm the ACFM depth with a method such as focused UT or TOFD.

5. Comparison with MT

MT (magnetic particle testing) and ACFM often go after the same job — finding surface cracks in a ferromagnetic weld — but their strengths differ.

Preparation: MT generally needs a clean surface free of thick paint; ACFM works through nonconductive paint. On offshore and coated structures this is ACFM's single biggest advantage.
Sizing: MT shows the length of a crack on the surface but gives no practical information about depth. ACFM produces a depth estimate, which can be decisive in accept/reject and remaining-life decisions.
Material: MT works only on ferromagnetic materials. ACFM can also be used on non-ferromagnetic materials, provided they are conductive and the probe and calibration suit them.
Sensitivity and speed: MT gives very high sensitivity and an immediate visual indication on open, shallow surface cracks; on a freshly ground clean surface it is still excellent at catching very small cracks. ACFM scans more slowly and works point by point, but the result is recordable, auditable digital data.
Records: An MT indication is transient and leaves no permanent numerical trace beyond photographs and the report; ACFM records every scan digitally, which is valuable for repeatable in-service monitoring (tracking the growth of the same crack over time).

The practical rule: MT for a fast surface scan on a clean, unpainted weld; ACFM for a painted or hard-to-reach connection where depth information is needed or the joint is monitored in service.

6. Comparison with ET (eddy current)

As an electromagnetic method, ACFM belongs to the same family as ET (eddy current testing), but several differences separate them.

Lift-off sensitivity: Conventional ET is highly sensitive to small changes in lift-off, a major noise source on painted or rough surfaces. ACFM's uniform-field approach and model-based interpretation tolerate lift-off far better — which is why examination through thick paint is ACFM's natural territory.
Calibration: ET usually needs careful calibration with reference standards for every combination of material, coating and geometry. ACFM's model-based depth approach reduces that dependency.
Sizing: ET mainly gives detection and comparative magnitude; reliably reporting the depth of a surface crack is not conventional ET's strong suit. ACFM targets depth numerically.
Conductivity and material: Both methods require conductive material. ET is very common on non-ferromagnetic conductors (stainless steel, aluminum, copper) and in applications such as heat exchanger tubes; ACFM stands out on structural welds and sizing on painted connections.

In short, ET wins on fine-tuned scanning and surface or tube examination of conductive material; ACFM wins on crack sizing through paint.

7. Offshore and weld applications

ACFM is still strongest where it was born, in the offshore industry. Node connections in the steel jackets of fixed platforms — particularly fatigue cracks in tube-to-tube welds — are its classic target. These connections are painted, they are underwater and stripping the paint is not practical; that need produced underwater ACFM probes that a diver, an ROV or a manipulator can operate subsea. Run the probe along the weld toe, the line where fatigue cracks are most likely to start.

Several field practices are decisive in weld applications:

Orientation: ACFM is most sensitive to cracks perpendicular to the exciting current. Because weld toe cracks generally run parallel to the weld, position the probe along the weld so the current crosses it at right angles. If cracks with other orientations are suspected, change the scan direction.
Following the toe line: Keep the probe on a constant line along the weld toe; drifting off the line reduces sensitivity and generates false geometry signals.
Geometry signals: Breaks in profile such as the weld cap, the weld toe and transitions between weld passes produce signals of their own; an experienced inspector knows how to separate these geometry traces from a real crack butterfly.
Recordability: Every scan is recorded digitally, so the same node can be rescanned periodically over the life of the platform and crack growth tracked. That traceability is critical value for fatigue management.

Beyond offshore, ACFM is used wherever structures are painted, structural and fatigue-loaded: bridges, cranes and lifting equipment, wind turbine tower welds, pipeline and pressure vessel welds. One distinction matters on pressure equipment: ACFM is not among the code-recognized surface examination methods for new construction under ASME Section VIII, Div. 1 (MT — Mandatory Appendix 6; PT — Mandatory Appendix 8). On pressure vessels, ACFM typically belongs to in-service examination and to providing input data for fitness-for-service assessment (API 579-1 / ASME FFS-1); code acceptance for new construction runs through MT and PT.

8. Limits and points to watch

ACFM is powerful but it is not a cure-all, and knowing its limits protects you from a false "clean" report.

Surface-breaking cracks only: ACFM is a surface method. Because of the skin effect the current flows only in a thin surface layer, so it cannot detect internal (embedded) discontinuities that do not break the surface. Volumetric discontinuities call for UT or RT.
Conductive material only: The material must be conductive for current to be induced. ACFM does not work on nonconductive materials such as composites, ceramics or plastics. (It need not be ferromagnetic; conductivity is enough, but select the probe and calibration to suit the material.)
Conductive coatings and heavy marine growth: Conductive metal coatings (galvanizing, metallizing) do not behave as simple lift-off, can suppress the indication and require separate assessment. Heavy and irregular marine growth must be cleaned off before the examination (water jetting or scraping); it creates a large, irregular lift-off and corrupts depth estimation in particular.
Orientation dependence: Cracks running parallel to the current do not divide the current and are easily missed; multi-directional scanning is essential in critical areas.
Depth uncertainty: Coating thickness, closed or branched crack profiles, probe alignment and nearby geometry can all corrupt the depth estimate; confirm it for critical decisions.
Geometry noise: Sharp corners, section changes and the weld profile generate false signals; interpretation takes experience.
Lower limit on small or shallow cracks: For very shallow or very short cracks the signal approaches the noise floor; on a freshly ground clean surface MT may be more sensitive at those sizes.
Personnel: The reliability of the result depends tightly on the inspector's butterfly interpretation and field discipline; the method does not hand you an automatic pass/fail.

9. Related standards

ASTM E2261 / E2261M: Standard Practice for Examination of Welds Using the Alternating Current Field Measurement Technique — the standard practice for ACFM examination of welds in any metallic material; it targets baseline and service-induced surface-breaking discontinuities.
ASME BPVC Section V, Article 15 (ACFMT): a dedicated article reserved for ACFM (Alternating Current Field Measurement Technique); it sets out the method requirements in subarticles beginning with the T-1510 scope. This article is distinct from Article 8, which covers eddy current; do not confuse the two. Article 15 typically covers linear discontinuities of about 6 mm (1/4 in.) and longer.
ISO 9712: qualification and certification of NDT personnel (Level I/II/III) — the general competence framework. Note that ACFM does not appear as a separate method code in ISO 9712 (the methods are ET, MT, PT, RT, UT, VT and so on); ACFM-specific qualification usually runs under ET or through employer- or scheme-based routes (SNT-TC-1A, or a PCN special appendix).
Offshore framework: the correct bases for offshore ACFM work are ASTM E2261 and ASME Section V, Article 15. Industry practice also cites equipment manufacturer (TSC/Eddyfi) application procedures, operator or project specifications, class society rules (DNV, Lloyd's) and the historical HSE offshore guidance. (API RP 2X governs only UT and MT in offshore structural fabrication; it does not cover ACFM and should not be cited as an ACFM reference.)
Acceptance criteria: evaluate indications for acceptance or rejection against the construction or in-service code that governs the examination (the relevant welding, pressure vessel or offshore code). ACFM sets no acceptance criteria of its own; it feeds sizing data into that code.

If you are not certain of a standard number or its edition, always work from the edition called out in the project specification that governs the examination.

From the field

The most common mistake in ACFM is calling every signal a crack without separating the butterfly from noise. The reverse happens just as often: a real butterfly gets written off as a geometry signal.

The second most common mistake is probe alignment. Drift a few millimeters off the weld toe and you lose sensitivity and pick up false indications. Run the probe slowly and steadily on a constant line.

Always know the coating thickness and reflect it in your calibration. The comfort of "it works through paint" quietly corrupts your depth estimate as the coating gets thicker, and the thickness tolerated for detection is greater than the thickness tolerated for reliable sizing. Never mistake a conductive metal coating such as galvanizing for harmless paint-like lift-off; it can suppress the indication altogether.

Never present a depth value as absolute truth. Near an accept/reject limit, confirm it with focused UT or TOFD. ACFM is strongest at first-pass sizing and in-service monitoring, but the last word usually belongs to a second method alongside it.

Finally, remember that ACFM is a surface method. It sees nothing that does not break the surface, so an "ACFM clean" report is no guarantee of volumetric integrity. The best ACFM examination runs in the right direction, on a constant line, with a known coating, and is sensibly confirmed by a second method. Discipline and interpretation decide the result, not the equipment.

Take it to the field: To keep these NDT methods, standard references and field steps in your pocket — completely offline and free — take a look at the Doawise NDT Guide app.

At DoaWise we provide nondestructive testing services, ACFM included, to international standards — from offshore node connections to painted welds and in-service crack monitoring — producing results that can be sized, recorded and audited.