What is the smallest defect detectable with ACFM?

Summary

In general the smallest defect detectable with ACFM is in the range 2mm long by 0.3mm deep to 10mm long by 1mm deep. This depends on the material, the surface condition and the type of probe used.
A better question to ask of all NDT techniques, though, is "What is the largest defect that can be missed?".

Detail

ACFM was designed to reliably detect and size in-service defects with a low number of false calls. It was designed to work on rough or corroded surfaces, or through protective coatings. It was not designed to be an ultra-sensitive detection technique.

All NDT techniques miss some defects. With an ideal technique, the probability of detection should increase with increasing defect depth. ACFM achieves this because unlike some techniques, the deeper a defect is, the larger the ACFM signal becomes.

The results of extensive POD trials carried out over the years have demonstrated that ACFM can reliably detect surface-breaking cracks of 10mm long by 1mm deep at manual multi-pass weld toes, or 5mm long by 0.5mm deep on good surfaces. More importantly, however, these same trials have shown that other techniques, while finding some smaller defects, can also miss larger ones. The most important question for an NDT technique is not "What is the smallest defect it can find?", but rather "What is the largest defect it can miss?".

The use of micro pencil probes and higher frequencies provides a more sensitive ACFM signal In this case defects as small as 2mm long or 0.3mm deep can be detected in ferritic steel (about 4mm by 0.5mm in non-ferrous metals). However this sensitivity comes at the expense of a higher response to other influencing parameters such as surface roughness or lift-off. As a result the minimum defect sizes can only be achieved on smooth uncoated surfaces.

Can ACFM detect sub-surface defects or porosity?

Summary

Generally not on ferrous metals, but sub-surface defects can be detected in non-ferrous metals if they are near the surface.

Detail

The a.c. skin effect means that the currents induced in the ACFM technique are confined to a thin layer at the surface of the material. ACFM can only detect defects lying within this layer where they perturb the current and so can be detected.
The thickness of this layer (the "skin-depth") is inversely proportional to the square root of the magnetic permeability of the metal so is much smaller for ferritic steel than for non-ferrous metals. The skin depth in ferritic steel is too small to allow detection of sub-surface defects, but the larger skin depth in non-ferrous metals makes it possible to detect defects that do not break the top surface.

The signal from a sub-surface is much less "sharp" than that from a surface breaking defect. To be detected sub-surface defects must disturb a significant portion of the current flow. To achieve the required disturbance in the flow a defect must extend to within about half the skin depth of the top surface and have a through thickness height of at least half the skin depth. Standard low frequency probes should be used if sub-surface defect detection is required.
The skin depth in low conductivity metals (such as stainless steel, titanium, nickel alloys, bronzes etc.) is around 5-8mm at 5kHz, whereas the skin depth in high conductivity metals (such as aluminium, copper and tungsten) is around 1-2mm. The form of the signal from a sub-surface defect depends on the relative sizes of the remaining ligament and the defect length, but usually results in an upward-moving Bx (and consequently an upward butterfly loop).

Volumetric defects, such as corrosion pitting or porosity, give much weaker signals than planar defects so it is not recommended that ACFM be used to detect sub-surface porosity.

Does ACFM work on forgings and castings?

Summary

Yes, ACFM does work on forgings and castings.

Detail

ACFM detects and sizes planar defects in any metal, regardless of how a defect is formed. The technique is relatively insensitive to surface roughness, so forgings and castings can be inspected as easily as machined or welded components.
Forgings and castings often have large areas ideally suited to array probes and consideration must be made for in-service defects that can initiate from features such as pores or laps.
Welded or machined components often have discrete stress raisers that need to be inspected, limiting the use of array probes, but don't suffer the same form of in-service defects common to forged or cast components.

Can ACFM measure the through-thickness depth of an inclined crack?

Summary

No, ACFM measures the crack path length. If the crack is inclined to the surface this length is greater than the through-thickness penetration.

Detail

The disturbance in the current flow, and hence the size of the signal, is related to the extra path length, i.e. the length down the crack face. In ferritic steel the skin depth is small compared to the crack depth, so no information can be gained on crack inclination. In thick-skin materials there is an asymmetry in current density either side of an inclined crack that, in principle, could give information on crack inclination (and hence through-thickness depth), but this is very difficult to measure and interpret in practice.

Can ACFM work through metallic coatings, scale or rust?

Summary

Yes, ACFM will work through metallic coatings, scale and rust.

Detail

ACFM generally has no problem working through rust, surface oxides or other low conductivity layers. The technique also works through thin uniform metal coatings such as zinc galvanising, even if the crack does not penetrate the coating. However, some problems can occur with non-uniform metal coatings such as manually applied flame-sprayed aluminium. Such coatings can produce strong background variations in the signals due to overlapping coats and differences in thickness which can mask signals from any defects present. In such cases, it is advisable to inspect coated areas away from any anticipated cracks to determine the nature of the background variations (if any). The background variations are typically fairly constant over a wide area, so localised signals from a defect can be picked out by comparing adjacent scans (recommend using an array probe for this).

Can ACFM detect cracks which extend the full width across a plate?

Summary

Yes, ACFM can detect cracks which extend the full width across a plate.

Detail

The usual (and most reliable) method for confirming the presence of a defect is to obtain a complete loop in the butterfly plot, which requires scanning a probe along the complete defect, including passing over both ends. For cracks that run up to an edge, this procedure is modified, so that only half a loop is expected. Obviously, if the defect runs completely across a plate, or full circumference around a pipe, there will be no defect end signals in Bz, and, if the defect is of uniform depth, there will be little change in Bx signal. In order to detect such defects, the inspection procedure requires a scan across the expected line of the defect. If a defect is present, a sharp dip in the Bx signal will be seen. This transverse scan provides the Bx background and minimum values for sizing purposes.

In practice, long fatigue defects, especially at weld toes, normally consist of several cracks which have coalesced together. This process gives rise to local crack bridging and line contacts which give strong ACFM signals during the normal inspection scan, significantly aiding detection.

Can ACFM detect transverse cracks?

Summary

Yes, ACFM can detect transverse cracks.

Detail

The directional nature of the induced current in ACFM means that it will not flow across a crack that runs transverse to the scan direction. In practice an upward going butterfly is often obtained from a transverse defect (due to flux leakage effects), however because this cannot be relied on, and only occurs in ferritic steel, the procedure contains an instruction to make a scan with the probe rotated 90 degrees, in order to detect transverse defects. Once detected, such cracks are sized by scanning the probe along the crack.
Detection and sizing of transverse defects is made quicker and simpler with use of an array probe which contains two energising fields at right angles to each other, so that defects in any orientation are detected in one pass.

Tech Support / Technique    Next