Click for:
 | TRECSCAN
Training Course |
| Commercial Licensees |
| Essential Factors for Transient Eddy-Current Detectability pdf (3.5Mb) |
Figure 1. TRECSCAN® being used with an ANDSCAN® arm and software.
Transient eddy-current NDE using Hall sensors now has recognised potential for detection of corrosion and cracks in ageing aircraft fleets. There are significant benefits to be realised from the use of transient eddy-currents in terms of inspection time and ease of acquisition and analysis of the data. Large areas of structure incorporating multiple variations in thickness can be scanned without the need for probe or set-up changes. In addition, the use of a Hall sensor rather than a coil as a field detector improves the spatial resolution and the detectability of deep defects.
The authors have investigated the effects of new probe designs on defect detectability at depths up to 11 mm in aluminium. As a result they have developed a methodology for assessing the most appropriate probe for a particular inspection. In addition, further work is presented on advanced analysis techniques for defect depth determination.
A particularly difficult inspection case study is presented where mushroom-headed fasteners over a four-layer structure prevents adequate inspection for corrosion using other techniques. This proved to be an ideal application for transient eddy-currents, despite the enforced lift-off and associated reduction in signal-to-noise ratio.
Several recent publications have described the transient eddy-current method and its variations, as well as the advantages or otherwise of using coil sensors compared with direct magnetic field sensors such as Hall, GMR, or SQUID sensors [1-9].
The TRECSCAN® system used for this work was developed by QinetiQ Ltd (formerly the Defence Evaluation and Research Agency, DERA) in collaboration with the Defence Science and Technology Organisation, Melbourne, Australia [2-4]. Eddy current pulses are generated using a coil and the component of the resultant magnetic field perpendicular to the specimen surface is measured using a Hall sensor on the axis of the coil [1-3]. The field generation and reception components of the TRECSCAN® system have been described previously [9] and are beyond the scope of this paper.
Previous work by the authors has discussed the various analysis methods appropriate to the detection and characterisation of corrosion and cracks, including algorithms to remove the effects of lift-off, plate separation and edges, together with an initial evaluation of the capabilities and limitations of such a system [8-9].
A set of versatile specimens was designed for determining the defect detection and characterisation capabilities of transients for both corrosion and cracks at a range of depths in the structure [9]. These specimens simulate hidden interfacial corrosion but do not attempt to simulate exfoliation corrosion, although this is also a major ageing-aircraft problem. For the current work, four different probes were used: small, medium and large ferrite-cored probes and a large air-cored pancake coil [9].
The spatial and depth resolution with which corrosion-induced metal loss within a multi-layer structure can be characterised depends on the spatial characteristics of the eddy-current field and the location of the metal loss in the field. In order to evaluate a transient eddy-current system it is necessary to investigate the dependence of the corrosion detectability on four variables: defect depth, in-plane defect size (extent), metal loss, and the probe characteristics. It is also necessary to set a criterion for reliable defect detection. The current work used a signal threshold criterion for defect detectability equal to three times the ‘smoothed noise level’ calculated for each scan with a particular probe based on the trend in noise levels as a function of time in the transient signal. A comparison of minimum defect detectability for large defects is shown in Figure 3 for the four probes as a function of defect depth (using data from 30 mm defect diameters). The detectability for small defects was also measured but space does not permit the results to be presented here [9].
Figure 3. Minimum detectable metal loss for large defects (30 mm diameter) with lines of best fit. In each case, the machined defects were in the near side of the back layer of the stack of plates. The curve corresponds to 1% of defect depth.
The medium ferrite-cored probe appears to be the best of the four over the whole depth range measured. The small ferrite-cored probe has poor sensitivity to large defects at all depths, and the large ferrite-cored probe has poor sensitivity to shallower defects.
Optimum sensitivity to cracks from fastener holes is normally obtained by looking for symmetry variations around a fastener. A probe with a symmetrical response is therefore needed. An approach taken by the authors is to produce a scan of a region containing fasteners, using TRECSCAN® with a symmetrical probe, and capture data at a 0.5 mm resolution around the region of each fastener. Then it is possible to investigate the symmetry variations by post-processing the scanned image. Software was developed that uses the rotational symmetry of the fastener to find its centre and then plots any deviations from rotational symmetry as large colour swings. Hence cracks show up as dark regions on one side of the fastener.
Figure 4. Crack detectability in terms of crack cross-section for cracks emanating from countersunk fastener holes with installed aluminium fasteners. The horizontal error bars indicate the extent of the crack in the thickness direction.
Measurements were made for cracks ranging from 1 to 6 mm in length, in 1.5 mm and 4 mm thick sheets in stacks to a depth of 13 mm. The results shown in Figure 4 suggest that the large ferrite-cored probe is the best for detecting cracks at fastener holes within 7 mm of the surface. However, the large size of this probe would considerably reduce its capability in regions with closely-spaced fasteners and edges. Beyond 7 mm depth the medium ferrite probe appears to be more sensitive.
Once defects have been detected and their type distinguished it is desirable to know their depth within the structure. A method for measuring the time to the peak of a balanced transient can be used to produce simple time-of-flight scans, Figure 5, which can be related to depth in the structure via a previous calibration. From this information, it can be determined in which layer the defect is. Also shown is the relationship between the time-to-peak measurements and the actual 1/e defect depth (ie the depth corresponding to 1/e through the thickness of the defect) for the machined metal loss.
Figure 5. Scan using the medium ferrite-cored probe (left) showing time-of-flight to the peak of the transient for 0.75 mm machined metal loss at the back of the top 1.5 mm thick layer (top-left circular defect) and then gradually descending 0.75 mm further per defect in a clockwise direction within a 12 mm thick multi-layered structure. The time-to-peak values are related to depth in the structure for three of the probes (right).
From the theoretical treatment of the ideal situation with spatial frequencies approaching zero [11], a parabolic relationship would be expected between time-to-peak and actual depth. However, the effect of spatial frequency distribution, current rise time, filters etc on the measured signal will affect the relationship between time-to-peak and actual defect depth in a way that is poorly understood at present. Consequently, the data in Figure 5 is fitted assuming a general quadratic relationship. In spite of this, use of the graphs in Figure 5 as calibration curves would be sufficient to indicate which layer a defect occurs in, for a similar thickness and conductivity of metal.
A difficult inspection requirement exists on the Sea King helicopter boat hull. There is a known possibility of corrosion occurring in the structure, which can be up to four layers of approximately 1 mm thick aluminium alloy (see Figure 6). An added complication is numerous closely-spaced mushroom-headed fasteners and edges. Conventional eddy-current methods found considerable difficulty in distinguishing between corrosion and the multitude of structural variations. The close fastener spacing and edges limit the in-plane dimensions of the field that can be used and the heads of the fasteners introduce 1.6 mm of lift-off if the probe is to glide over them. TRECSCAN uses the method of Burke et al [4] to measure total thickness of metal in the structure, resulting in the removal of plate separation effects - an advantage for this inspection.
The evaluation of the transient eddy-current technique involved both defect detectability studies and also real inspections of structure underneath real helicopters during maintenance. These practical trials were invaluable in highlighting the problems of maintaining a constant lift-off.
The lift-off was expected to reduce the signal strength and therefore the signal-to-noise ratio obtained for a defect would decrease. Defect detectability was shown experimentally to reduce by a factor of two on the introduction of 1.6 mm of lift-off. After testing various brushes for maintaining lift-off, the most effective was a shoe brush with clumps of bristles spaced a few millimetres apart. This worked well because each fastener only pushed one clump of bristles aside and the others were unaffected.
Figure 6. Diagram of the Sea King keel beam and surrounding structure. The wedge-shaped structure (shown in black) allows the probe to slide easily without rocking.
Known amounts of thinning (50%, 20% and 10% loss of a layer’s thickness) were machined into representative structures. In evaluating the scans obtained on both the real aircraft and the specimens the most critical aspect was the removal of lift-off variations and then subtraction of the edge effects from the image [9]. Although these processes were successful (see Figure 7), further work still needs to be done on the edge subtraction algorithms to inprove the ability to cope with such closely-spaced fasteners.
Figure 7. Thickness change scan of the specimen with five regions of metal loss (50% of a layer). They are in (left to right) back of 4th layer (x = 330 mm), back of 3rd layer (x = 420 mm), top of 3rd layer (x = 500 mm), top of 2nd layer (x = 580 mm), and back of 1st layer (x = 660 mm). The defect on the right, at x = 760 mm, is a large drainage hole and two smaller holes in the real keel beam (4th layer).
Analysis of the 20% loss scans indicated that 10% metal loss in a layer would be approaching the noise level due to the fasteners. This was confirmed when the 10% loss specimen was scanned - although the first-layer 10%-loss defect could be discerned, the deeper defects were really obscured by noise from the fasteners.
The capabilities and limitations of transient eddy-currents have been evaluated at depths down to 11 mm in aluminium alloy for corrosion and crack detection. Using the medium ferrite-cored probe, 0.09 mm of metal loss could be detected at depths down to 7 mm, equivalent to 1% loss, and 0.01 mm at 1.5 mm depth. However, at 11 mm depth a loss of 0.5 mm was just detectable.
For detecting cracks at fastener holes, the large ferrite-cored probe proved the best within 7 mm of the surface, detecting crack areas of 2.5 mm2 at 3 mm depth and 6 mm2 at 6 mm depth. However, the large size of this probe would considerably reduce its capability in regions with closely spaced fasteners and edges. Beyond 7 mm depth the medium ferrite probe appears to be more sensitive, detecting crack areas of 10 mm2 at 7.5 mm depth (crack in a plate at 5.5-9.5 mm depth) and 20 mm2 at 11.5 mm depth (crack in a plate at 9.5-13.5 mm depth).
Finally, a case study
involving corrosion detection in a four-layer structure (4.5 mm thick)
with a multitude of fasteners and nearby edges has been described. 2.5% total
metal loss could be detected in the first layer and 5% at a depth of 4 mm.
The limitation was the noise due to fasteners which was at roughly the 2.5%
level. The medium ferrite-cored probe proved the best for this corrosion detection
case.
The authors gratefully acknowledge the contributions to this work from Dr David Harrison, Mr Tim Jarman, Miss Hazel Hung and Mr Lyn Jones of QinetiQ Ltd and Dr Steve Burke, Miss Cayt Harding and Mr Mark Taylor of AMRL. Part of this work was funded by the Ministry of Defence.
1. D.
J. Harrison,
“Progress in the detection of cracks under installed fasteners using eddy
currents.” AGARD Conference Proceedings No 462, Impact of Emerging NDE/NDI Methods on Aircraft Design Manufacture and
Maintenance, Brussels, (1989).
2. D. J. Harrison, Eddy-current inspection using Hall sensors and transient excitation, Defence Research Agency Tech Report DRA/SMC/TR941008,
DRA Farnborough (1994).
3. D. J. Harrison, in Nondestructive Testing of Materials, Studies in Applied
Electromagnetics and Mechanics, Vol 8, eds. R. Collins, W. D. Dover, J.R.
Bowler and K. Miya, (IOS Press, Amsterdam, 1995), pp. 115–124.
4. S.K. Burke, G.R. Hugo, and D.J. Harrison, in Review
of Progress in QNDE, Vol 17A, eds. D. O. Thompson and D. E. Chimenti,
(Plenum, New York, 1998), pp. 307–314.
5. W.W. Ward III and J.C. Moulder, in Review of Progress in QNDE, Vol 17A, (1998), pp. 291–298.
6. J. A. Bieber et al, in Review of Progress
in QNDE, Vol 17A, (1998), pp. 315–322.
7. S. Gigučre, B.A. Lepine and J.M.S. Dubois, “Pulsed eddy-current (PEC)
characterization of material loss in multi-layer structures.” Canadian Aeronautics
and Space Journal, Vol. 46, No. 4, (2000), pp. 204-208.
8. G. R. Hugo and
D. J. Harrison, in Review of Progress
in QNDE, Vol 18B,
eds. D. O. Thompson and D. E. Chimenti, (Kluwer Academic/Plenum Publishers,
1999), pp. 1401-8.
10. D J Harrison, “The characterisation
of cylindrical eddy-current probes in terms of their spatial frequency spectra”.
IEE Proceedings;
Science, Measurement and Technology (SMT), Special Issue on Non-Destructive
Testing and Evaluation, Vol 148, No 4, 2001.
©
Copyright QinetiQ Ltd, 2001. Published with the permission of QinetiQ Ltd
and the Australian Defence Science and Technology Organisation.
ANDSCAN and TRECSCAN are Registered Trademarks of QinetiQ Ltd.
Author: R A Smith.
Copyright © 2004 QinetiQ Ltd. All rights reserved.
Revised: March 04, 2009
