Download as pdf or txt
Download as pdf or txt
You are on page 1of 6

444 IEEE SENSORS JOURNAL, VOL. 17, NO.

2, JANUARY 15, 2017

Pulsed Eddy Current Inspection of Wall Loss


in Steam Generator Trefoil Broach Supports
Sarah G. Mokros, Peter Ross Underhill, Jordan E. Morelli, Senior Member, IEEE, and Thomas W. Krause

Abstract— Inspection for corrosion in steam generator (SG)


tube support structures is a significant component of monitoring
SG condition, since support degradation can lead to SG tube
flaws, and thereby further SG deterioration. A pulsed eddy
current probe was developed to augment SG inspection at
carbon steel trefoil broach supports. The probe demonstrated
capability to measure far side wall loss as small as 20% of
the original 2.7-mm-thick broach support ligament from within
the Alloy 800 SG tube. A power law fit of the data was found
to describe signal variation at later times under conditions of
varying wall thickness and lift-off. The associated power law
exponent varied linearly with wall thickness and was observed
to be largely independent of lift-off and angular orientation of
the probe, identifying it as a potential parameter for ligament
wall thickness monitoring. Voltage response integrated over time
also provided a low-noise means of measuring the wall loss when
lift-off variation and misalignment were not present. Fig. 1. Sample #1 - trefoil broach support plate structure with no wall loss
and a single SG tube. As labeled, the land is the point of contact between
Index Terms— Pulsed eddy current, broach support, steel wall broach support ligament and SG tube.
loss, steam generator.

I. I NTRODUCTION that separate the trefoil holes, containing the SG tubes, from
the flow regions.
S TEAM generators (SGs) are a key component of thermal
power reactors, including CANDU nuclear reactors.
SGs consist of thousands of tubes that transport heat generated
Steam generators are subject to a number of degradation
modes, including cracking, denting or fretting, which often
in the reactor core and convert that heat into usable energy [1]. arise in the vicinity of support structures, and if serious
Support structures, typically made of carbon or stainless steel, enough can require plugging or removal of SG tubes [3]. Other
are located at intervals along the SG and are used to brace common issues such as flow assisted corrosion or tube vibra-
tubes [2]. Support structures are of various types, such as tions may cause wall loss in support structure ligaments [4].
lattice bar [1], trefoil or quatrefoil shaped holes [3]. These Characterization of the SG tube and support structure condi-
structures prevent SG tube vibration and the associated risk tion can be crucial for monitoring SG life and, when combined
of SG tube fretting, while allowing water to flow past the with preventative maintenance programs, can be used to extend
tubes [3]. An example of a simulated trefoil broach support SG life.
plate (BSP) structure, with only a single SG tube passing Eddy current testing (ECT) is currently employed to per-
through it, is shown in Figure 1. Also shown in Figure 1 are the form the bulk of SG tube inspection [5], [6]. ECT is successful
lands, which are the SG tube support points, and the ligaments at identifying integrity issues such as SG tube fretting [7] and
cracking, but is challenged at identifying flaws that may occur
Manuscript received August 20, 2016; revised October 19, 2016; accepted in the presence of ferromagnetic materials, including steel
November 15, 2016. Date of publication November 23, 2016; date of current support structures [8], as well as in cases where overlapping
version December 20, 2016. This work was supported in part by the University
Network of Excellence in Nuclear Engineering and in part by the Natural degradation modes are present [9].
Sciences and Engineering Research Council of Canada. The associate editor PEC is an emerging technology that has been successful at
coordinating the review of this paper and approving it for publication was detecting many types of flaws that may appear in conducting
Prof. F. Reverter.
S. G. Mokros and T. W. Krause are with the Department of Physics, structures. Horan et al. [10] were able to identify cracks in
Royal Military College of Canada, Kingston, ON K7K7B4, Canada (e-mail: aluminum structures at large lift-off. PEC, when combined
sarahmokros@gmail.com; thomas.krause@rmc.ca). with statistical methods such as principal components analysis,
P. R. Underhill is with the Department of Mechanical Engineering,
Royal Military College of Canada, Kingston, ON K7K7B4, Canada (e-mail: has demonstrated a capability to detect cracks in the sec-
ross.underhill@rmc.ca). ond layer of multilayer aluminum structures [6], [7]. Recent
J. E. Morelli is with the Department of Physics, Engineering Physics and work on PEC steam generator inspection has investigated the
Astronomy, Queen’s University, Kingston, ON K7L3N6, Canada (e-mail:
morelli@physics.queensu.ca). detection and characterization of multiple failure modes in
Digital Object Identifier 10.1109/JSEN.2016.2631541 Alloy 800 at baffle plates including fretting [8], [9]. PEC has
1558-1748 © 2016 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.
See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.

Authorized licensed use limited to: King Mongkut's University of Technology North Bangkok. Downloaded on January 25,2021 at 09:37:30 UTC from IEEE Xplore. Restrictions apply.
MOKROS et al.: PULSED EDDY CURRENT INSPECTION OF WALL LOSS IN SG TREFOIL BROACH SUPPORTS 445

also been used to characterize low-energy impacts and inserted Crude estimates for solutions to the diffusion equations are of
defects in carbon fiber reinforced polymer laminates [5] and the form [20]:
corrosion of mild steels [11]. Chen et al. [12] have applied ∂B
the Hilbert transform for feature extraction in order to attain ∼ B/τ and ∇ 2 B ∼ B/2 . (2)
∂t
robust and rapid identification of different defects.
where τ is the characteristic diffusion time and l is a char-
In contrast to ECT, which uses sinusoidal voltage excitation
acteristic length. In combination with equation 1, an approxi-
of coils, pulsed eddy current (PEC) utilizes a square pulse
mate expression for the parameters that describe the diffusion
to excite the drive coil. The transient voltage response of
time, τ , of eddy currents into the target material, may be
pickup coils can be interpreted as consisting of a series of
expressed as [20]:
discrete frequencies, while the approach to direct current (DC)
excitation in the pulse provides magnetization of ferromagnetic τ ∼ μ σ 2 (3)
materials, which further enhances pickup coil response [13].
Note that l is dependent on the geometry of the system under
While higher frequency components of the induced signal
which the measurements are being performed. Generally, time
form close to the surface of the material at early times, lower
domain solutions for pulse excitations are of the type [21]:
frequency components, which are present at later times where
approach to a constant voltage arises, are capable of penetrat- B = f (e−t /τ ) = A1 e−t /τ1 + A2 e−t /τ2 + A3 e−t /τ3 + . . . (4)
ing deeper into conducting and ferromagnetic materials [14].
The detection of wall loss in ferromagnetic structures has are obtained. The time-dependent change of magnetic field,
been explored using pulsed eddy current. Results of experi- integrated over the area, a, and number of turns, N,
mental work on ferromagnetic tubes demonstrate that the PEC in a pick-up coil, results in an induced voltage as given
technique provides an improvement over present EC tools for by Faraday’s law [22]:

NDT of ferromagnetic tubes [15]. Spies [16] and Lara [17] ∂B
V = −N · da. (5)
used PEC to measure remaining wall thickness in conduc- ∂t
tive ferromagnetic containers. More recently, Xu et al. [18] In this event equation 4 becomes:
demonstrated a technique to measure wall thickness from the
outside of ferromagnetic pipes through a layer of insulation. V = C1 e−t /τ1 + C2 e−t /τ2 + C3 e−t /τ3 + . . . (6)
Angani et al. [19] were capable of detecting localized wall where the Ci are constants incorporating the integrated
loss on stainless steel pipes through 25 mm of insulation. response over the turns area of the coil and time derivative
A PEC probe has been developed to simultaneously measure of B in equation 4. In complex systems when multiple similar
tube-to-support gaps and baffle plate hole inner diameter (ID), exponential decay processes are present, the time-dependent
simulating corrosion, in Alloy-800 SG tube within SS410 sup- decay has been observed to take on the form of a power law
ports [8], [19]. Depth sizing of rectangular frets was demon- given by [23]:
strated as being independent of variations in SS410 hole inner
diameters (IDs) for centered tubes [8]. The probe featured V = At −b , (7)
a single drive coil wound coaxially with the SG tube axis
where A is the power law coefficient and b is the power
and four or eight surface pick-up coils with axes mounted
law exponent. In the complex geometry of the trefoil broach
perpendicular to that of the drive coil [8].
support multiple relaxation times will be present and there-
However, this probe design was not configured to examine
fore, the behaviour of the power law time dependence
the threefold symmetry of trefoil broach supports. In addition,
described by equation 7 may be anticipated. Previous work
the probe was not investigated for detection of wall loss on the
by He et al. [11] also used power law functions to fit PEC
far side of support structure ligaments from within the SG tube.
response to early corrosion in coated and uncoated mild
This work uses laboratory measurements to investigate a PEC
steels. This work was focused on thin steel samples with PEC
probe design adapted to inspect remaining wall thickness
measurements performed in the absence of large lift-offs [11].
in the ligaments of ferromagnetic trefoil broach supports.
Two signal analysis methods are developed to provide a means
III. E XPERIMENTAL T ECHNIQUE
of measuring the amount of ligament wall loss on the far side
of the broach support land. One of the methods, based on a A. PEC Probe and Data Acquisitions
power law fit, demonstrates sensitivity to wall loss while being The PEC probe consists of a single drive coil wound
largely independent of lift-off and angular orientation of the coaxially with the SG tube axis, modified from that presented
probe, identifying it as a potential candidate for ligament wall elsewhere [8], [16], with a configuration of six surface pick-up
thickness monitoring. coils mounted with axes perpendicular to that of the drive
II. T HEORY coil as shown in Figure 2. Figure 2(a) shows a schematic
of the probe design and Figure 2(b) the experimental probe.
The expression for diffusion of the magnetic fields, B, into Three pick-up coils, placed at 120° intervals around the probe
a good conductor, with isotropic conductivity, σ, and perme- axis, are mounted above the drive coil and three below as
ability, μ, is given by [19]: shown in Figure 2. This configuration was chosen to align with
∂B 1 2 the lands of the trefoil broach support structure for relative
= ∇ B (1)
∂t μσ comparison of signals between neighbouring lands. The probe

Authorized licensed use limited to: King Mongkut's University of Technology North Bangkok. Downloaded on January 25,2021 at 09:37:30 UTC from IEEE Xplore. Restrictions apply.
446 IEEE SENSORS JOURNAL, VOL. 17, NO. 2, JANUARY 15, 2017

Fig. 3. Broach support structures with simulated flaws representing: (a)


40%, 75% and 0% ligament wall loss (Sample 2), (b) 20%, 0% and 60%
wall loss (Sample 3) and (c) 0%, 100% and 0% wall loss (Sample 4).
Fig. 2. (a) Schematic and (b) experimental probe design used for inspection
of broach support structures.

can accommodate 3 additional coils on each row for improved


coverage. The addition of a 30° angular offset between the
front and back rows could also be used to further enhance
coverage of the probe configuration.
The drive coil was excited by a 9V square pulse 4 ms in
duration with a rise time of 5 μs and a 50% duty cycle.
The drive coil resistance of 7.4 , in series with the output
impedance of the pulse circuit of 0.3 , and drive coil
inductance of 232 μH results in a circuit rise time of τC =
L/R = 30 μs for which the rising current may be estimated
as [22]
V
I (t) = (1 − e−t/τC ). (8)
R
PEC pickup coil responses were collected separately and Fig. 4. Semi-log plot of experimental pick-up coil response under various
wall loss conditions. Solid lines indicate best fit to power law.
amplified by a factor of 10000. Shielded twisted pair cable
carried signals from the probe to a National Instruments (NI)
TABLE I
6356 USB Data Acquisition (DAQ) board where they were
C ARBON S TEEL S AMPLES W ITH FAR S IDE WALL L OSS (%)
digitized at 1 MHz per channel. The DAQ was connected to
a computer running a LabView program, which was used to
set the pulse parameters and record the signals.

B. Samples
Four broach support samples were water jet cut out of a
25.4 mm (1”) thick ferromagnetic A516 Carbon Steel plate.
Sample 1, shown in Figure 1, had no flaws. Three remain-
ing broach support plate samples had flaws machined into
ligaments to simulate varying amounts of wall loss between
0 and 100% on the far side of the lands, as shown in Figure 3.
Table 1 summarizes relative ligament wall loss and location
in each sample.
For the case where a ligament has been completely corroded corresponded to a 0.25 mm normal increment, up to 1.5 mm
away, as represented in Figure 3(c), the tube is free to from the land, which was at 60° to the line of motion.
move off-centre and away from other lands in the trefoil.
Alternatively, a land area may also loose material from the IV. R ESULTS AND D ISCUSSION
near side. Both of these cases could have a significant effect
on the pick-up coil response due to increased lift-off with A. Effects of Wall Loss
respect to the remaining structure. To experimentally inves- Wall loss on the far side of broach support lands was
tigate this effect, the tube was fixed to a micrometer that detected by observing the long transient decay times of
could translate the probe at desired increments away from pick-up coil responses. Figure 4 shows the experimental results
an unflawed land [9]. Measurements were taken at a 0.5 mm for pick-up coils located at lands in Figure 1, for no wall
interval into the gap left by the removed ligament in Sample 4, loss (0%), and at laboratory simulated wall loss locations
from 0 mm (touching the BSP land) up to 3 mm, which of 20%, 40%, 60% and 75% as listed in Table I.

Authorized licensed use limited to: King Mongkut's University of Technology North Bangkok. Downloaded on January 25,2021 at 09:37:30 UTC from IEEE Xplore. Restrictions apply.
MOKROS et al.: PULSED EDDY CURRENT INSPECTION OF WALL LOSS IN SG TREFOIL BROACH SUPPORTS 447

Fig. 5. Power law coefficient values, A, from power law best fit of the data. Fig. 7. Experimental integration of pick-up coil response as a function of
Dashed curve is polynomial best fit to the experimental results. wall loss. Dashed curve is a quadratic polynomial best fit to the data.

10 times. Large uncertainty is present for the 75% wall loss,


possibly due to a departure from power law behavior for the
thinnest ligament as well as higher noise in the signal as also
observed in Figure 4. Uncertainty in the fit parameter is again
shown to one standard deviation.
A numerical integration of the pick-up coil signal can also
be used to determine the wall loss and is shown in Figure 7.
The dashed curve is a quadratic polynomial best fit to the data.
By performing the integration, the majority of the noise that
arises in the experimental results at later times and at lower
signal amplitudes is suppressed. Combining integration with
the power law coefficient A and exponent b could provide
a robust technique to characterize the wall loss on the far
side of broach support plate ligaments under constant lift-off
conditions with probe centered over the land.
Fig. 6. Power law exponent values extracted from power law line of best It can be observed in Figure 4 that as the amount of
fit. Dashed curve is a linear best fit to the data.
wall loss is increased, the transient decay time decreased.
This result is consistent with that expected from Equation 3,
B. Quantification of Wall Loss if the characteristic term l is associated with wall thickness.
Small deviations from the dashed curve and larger uncertainty
Experimental pick-up coil signal responses were fit with a
values for 60% wall loss ligament may be attributed to a
power law expression as expressed by equation 7, with the two
20% variation in the wall loss between top and bottom of
fit parameters, A, the power law coefficient and b, the power
the sample. Under steam generator inspection conditions non-
law exponent. The power law was fit to the full range of wall
uniform corrosion of the broach support structures and the
loss results shown in Figure 4. Determining coefficient, A,
presence of corrosion products such as magnetite may be
and exponent, b, of the power law line of best fit could be
anticipated [25]. In this event deviations from the more ideal
used as a technique to characterize wall loss on far side of
conditions observed here may arise. However, PEC probe
the land and form the basis for a calibration under inspection
interaction will still be with average wall loss in the vicinity of
conditions.
the drive coil. Recently, magnetite fouling has shown to have
Figure 5 shows the power law fit coefficient,
a negligible impact on measured decay rate of the PEC probe
A, in Equation 7 as a function of wall loss. Uncertainty
response [25].
to one standard deviation was calculated using Microsoft
Excel 2010 LINEST function [24] applied to the originally
acquired data. Large uncertainty is present for the sample C. Effects of Lift-Off
at 60% wall loss. This may be attributed to the observed 20% Lift-off of the probe from one land may arise if neighboring
wall thickness variation (taper from 1.0 to 1.3 mm thickness lands on the nearside have undergone corrosion or loss of a
as per Table I) in the 60% wall loss ligament, Sample 3. ligament has occurred. Figure 8 shows the experimental probe
Figure 6 shows that the power exponent, b varies linearly response as a function of lift-off distance from the land of
with wall loss. The 100% wall loss power law exponent is not a no wall loss ligament in the 100% wall loss Sample 4
included due to its increase in amplitude by approximately shown in Figure 3(c). Experimental results were obtained at a

Authorized licensed use limited to: King Mongkut's University of Technology North Bangkok. Downloaded on January 25,2021 at 09:37:30 UTC from IEEE Xplore. Restrictions apply.
448 IEEE SENSORS JOURNAL, VOL. 17, NO. 2, JANUARY 15, 2017

Fig. 10. Power law coefficient (A) and exponent (b) as the alignment of the
pick-up coil was rotated. Power law coefficient A ( ) and exponent b () are
fit with cubic spline and linear best fit, respectively.
Fig. 8. Effect of lift-off on pick-up coil response when inspecting an
unflawed broach support plate land. Solid curves are best fits to the power extracted from experimental results collected as the probe was
law, Equation 7. rotated inside Sample 2, at the unflawed land. The probe
was rotated from 0° to 60° in 15° increments (in the current
probe configuration 60° is the maximum angular separation
of a pick-up coil from the centre of a land in the current
probe configuration). The power law exponent shows that
even at large angular displacements, wall loss may still
be measured. A maximum 5% relative change in b occurs
between 0 and 15°. In contrast, the variation of power law
coefficient, A, indicates that the angular displacement has
a large effect on this parameter. Addition of pick-up coils
between the present coils (3 additional coils in a row) could
be used to further reduce potential misalignment with a given
land and improve coverage for measurement of remaining wall
thickness.
V. C ONCLUSION
A PEC probe was modified from a previously existing
design [8] to include six surface pick-up coils with spacing
Fig. 9. Comparison of A ( ) and b () extracted from power law fit of
Equation 7 for pick-up coil response when lift-off increases. Both A and b
to match the three-fold symmetry of trefoil broach supports.
data have been best fit with a line. Experimental measurements for various conditions of wall loss
in the 2.7 mm thick broach support ligaments demonstrated
0.43 mm interval as the pick-up coil was moved away from the an ability to detect wall loss of as little as 20%. Two methods
land. Solid curves are a best fit to the power law expression, of characterizing the amount of wall loss were investigated.
equation 7, and demonstrate that the power law still continues A two-parameter power law fitting function was found to
to provide a good fit of the experimental results even at higher describe the pick-up coil responses. Power law exponents were
lift-offs. A decrease in signal response is observed as the probe found to be largely independent of lift-off up to 1.5 mm
is translated further from the land. Figure 9 shows a plot and angular orientation up to 60° from the ligament centre.
of the power law coefficient A and the power law exponent The power law exponent varied linearly with wall loss and was
b as a function of lift-off. Error bars are to one standard observed to be robust against variations in lift-off and angle.
deviation [23]. It can be seen that while the coefficient A Therefore, the power-law fit has potential as a means to char-
decreases linearly as the lift-off is increased, b only changes by acterize wall loss based on in-situ measurements. A second
10% for up to 1.5 mm of lift-off, suggesting that an estimate method that integrated the pick-up coil response over time was
of the percentage wall loss could still be extracted, even under also examined. The integrated signal response could be used to
variable lift-off conditions. manage noise contributions arising at low signal amplitudes at
later times, if lift-off variations or misalignment of the pick-up
coil with the land were not present.
D. Effects of Angular Variation
Effects due to misalignment of the pick-up coil with ACKNOWLEDGMENT
respect to the centre of the land were also investigated, since The first author would like to thank Vijay Babbar, Jeremy
probe orientation may not be known during SG inspection. Buck and Brian Lepine at Canadian Nuclear Laborato-
Figure 10 shows the power law coefficient A and exponent b ries (CNL) and Dr. Sean Sullivan from Ontario Power

Authorized licensed use limited to: King Mongkut's University of Technology North Bangkok. Downloaded on January 25,2021 at 09:37:30 UTC from IEEE Xplore. Restrictions apply.
MOKROS et al.: PULSED EDDY CURRENT INSPECTION OF WALL LOSS IN SG TREFOIL BROACH SUPPORTS 449

Generation for useful discussion. The first author would also [24] Office Support. (Jul. 2015). LINEST Function. [Online]. Available:
like to thank Peter Snell for technical assistance. https://support.office.com/en-ca/article/LINEST-function-84d7d0d9-
6e50-4101-977a-fa7abf772b6d
[25] S. Mokros, J. Buck, P. R. Underhill, J. Morelli, and T. W. Krause,
R EFERENCES “Pulsed eddy current technology for steam generator tube support
structure inspection,” in Proc. 19th World Conf. Non-Destruct. Test.,
[1] R. L. Tapping, J. Nickerson, P. Spekkens, and C. Maruska, “CANDU Munich, Germany, Jun. 2016, pp. 1–8.
steam generator life management,” Nucl. Eng. Des., vol. 197, nos. 1–2,
pp. 213–223, 2000.
[2] N. J. Fisher, M. J. Olesen, R. J. Rogers, and P. L. Ko, “Simulation
of tube-to-support dynamic interaction in heat exchange equipment,”
J. Press. Vessel Technol., vol. 111, no. 4, pp. 378–384, Nov. 1989.
[3] D. R. Diercks, W. J. Shack, and J. Muscara, “Overview of steam gen-
erator tube degradation and integrity issues,” Nucl. Eng. Des., vol. 194,
no. 1, pp. 19–30, 1999.
[4] M. H. Attia, “Fretting fatigue and wear damage of structural components Sarah G. Mokros received the B.Sc. degree in engineering physics from
in nuclear power stations—Fitness for service and life management Queen’s University in 2013 and the M.A.Sc. degree from Queen’s University
perspective,” Tribol. Int., vol. 39, no. 10, pp. 1294–1304, Oct. 2006. and the Royal Military College of Canada, with a focus on nondestructive
[5] Y. He, G. Tian, M. Pan, and D. Chen, “Non-destructive testing of low- testing in steam generator tubes using pulsed eddy current inspection methods,
energy impact in CFRP laminates and interior defects in honeycomb finite-element modeling, and analytical modeling of eddy current response in
sandwich using scanning pulsed eddy current,” Compos. B, Eng., vol. 59, tubular structures. She is currently with Canadian Nuclear Laboratories, Chalk
pp. 196–203, Mar. 2014. River, ON, where she is involving in nondestructive testing of fuel channels.
[6] L. Obrutsky, J. Renaud, and R. Lakhan, “Overview of steam generator
tube-inspection technology,” CINDE J., vol. 35, no. 2, pp. 5–13, 2014.
[7] A. Sophian, G. Y. Tian, D. Taylor, and J. Rudlin, “A feature extraction
technique based on principal component analysis for pulsed eddy current
NDT,” NDT E Int., vol. 36, no. 1, pp. 37–41, 2003.
[8] J. A. Buck et al., “Pulsed eddy current inspection of support structures
in steam generators,” IEEE Sensors J., vol. 15, no. 8, pp. 4305–4312,
Aug. 2015.
[9] J. A. Buck, P. R. Underhill, J. E. Morelli, and T. W. Krause, Peter Ross Underhill received the B.Sc. degree in physics and maths
“Simultaneous multiparameter measurement in pulsed eddy current from Trent University, Peterborough, ON, Canada, in 1977, and the D.Phil.
steam generator data using artificial neural networks,” IEEE Trans. degree from York University, York, U.K., in 1981, which he attended as
Instrum. Meas., vol. 65, no. 3, pp. 672–679, Mar. 2016. a Commonwealth Scholar. He has 30 years of research and development
[10] P. F. Horan, P. R. Underhill, and T. W. Krause, “Real time pulsed eddy experience, both in industry and academia, in various aspects of materials
current detection of cracks in F/A-18 inner wing spar using discriminant research. His most recent work at the Royal Military College of Canada,
separation of modified principal components analysis scores,” IEEE Kingston, ON, has been in eddy current testing and development of pulsed
Sensors J., vol. 14, no. 1, pp. 171–177, Jan. 2014. eddy current systems for applications, including inspection of submarine hulls,
[11] Y. He, G. Y. Tian, H. Zhang, M. Alamin, A. Simm, and P. Jackson, aircraft components, and steam generator tubing.
“Steel corrosion characterization using pulsed eddy current systems,”
IEEE Sensors J., vol. 12, no. 6, pp. 2113–2120, Jun. 2012.
[12] T. Chen, G. Y. Tian, A. Sophian, and P. W. Que, “Feature extraction and
selection for defect classification of pulsed eddy current NDT,” NDT E
Int., vol. 41, no. 6, pp. 467–476, 2008.
[13] C. A. Stott, P. R. Underhill, V. K. Babbar, and T. W. Krause, “Pulsed
eddy current detection of cracks in multilayer aluminum lap joints,”
IEEE Sensors J., vol. 15, no. 2, pp. 956–962, Feb. 2015.
[14] D. R. Desjardins, G. Vallières, P. P. Whalen, and T. W. Krause, Jordan E. Morelli (M’01–SM’08) received the B.Eng. degree from the
“Advances in transient (pulsed) eddy current for inspection of multi- Royal Military College of Canada, Kingston, ON, Canada, in 1996, the
layer aluminum structures in the presence of ferrous fasteners,” in Proc. M.A.Sc. degree from the University of Windsor, Windsor, ON, Canada, in
AIP Conf., vol. 1430, Jul. 2012, pp. 400–407. 1998, and the Ph.D. degree from the University of Saskatchewan, Saskatoon,
[15] D. Vasić, V. Bilas, and D. Ambruš, “Pulsed eddy-current nondestructive SK, Canada, in 2003, all in electrical engineering. In 2003, he joined
testing of ferromagnetic tubes,” IEEE Trans. Instrum. Meas., vol. 53, the Department of Physics, Engineering Physics and Astronomy, Queen’s
no. 4, pp. 1289–1294, Aug. 2004. University, Kingston, ON, as an Assistant Professor. In 2010, he was
[16] B. R. Spies, “Transient electromagnetic method for detecting corrosion promoted to Associate Professor. His research has been concerned with
on conductive containers,” U.S. Patent 4 843 320 A, Jun. 27, 1989. artificial intelligence based optimization and control especially of magnetically
[17] P. F. Lara, “Transient electromagnetic method for detecting corro- confined plasma, electrical distribution systems, and recently eddy current
sion on conductive containers having variations in jacket thickness,” based nondestructive evaluation.
U.S. Patent 4 843 319 A, Jun. 27, 1989.
[18] Z. Xu, X. Wu, J. Li, and Y. Kang, “Assessment of wall thinning
in insulated ferromagnetic pipes using the time-to-peak of differential
pulsed eddy-current testing signals,” NDT E Int., vol. 51, pp. 24–29,
Jul. 2012.
[19] C. S. Angani, D. G. Park, C. G. Kim, P. Kollu, and Y. M. Cheong, “Dual
core differential pulsed eddy current probe to detect the wall thickness
variation in an insulated stainless steel pipe,” J. Magn., vol. 15, no. 4,
pp. 204–208, Dec. 2010. Thomas W. Krause received the B.Sc. degree in physics from the University
[20] H. C. Ohanian, “On the approach to electro-and magneto-static equilib- of Calgary, AB, Canada, in 1987, and the M.Sc. and Ph.D. degrees in physics
rium,” Amer. J. Phys., vol. 51, no. 11, pp. 1020–1022, Nov. 1983. from McMaster University, Hamilton, ON, Canada, in 1989 and 1992, respec-
[21] T. W. Krause, C. Mandache, and J. H. V. Lefebvre, “Diffusion of pulsed tively. In 1996, after working four years with the Applied Magnetics Group
eddy currents in thin conducting plates,” in Proc. Rev. Progr. Quant. at Queen’s University, he took a position at Canadian Nuclear Laboratories,
Nondestruct. Eval., vol. 975, 2008, pp. 368–375. developing nondestructive testing (NDT) technology for the nuclear industry.
[22] D. J. Griffiths, Introduction to Electrodynamics, 3rd ed. In 2006, he was with the Royal Military College of Canada. He has 24 years
Englewood Cliffs, NJ, USA: Prentice-Hall, 1999. of research and development experience on advanced NDT technology. He has
[23] W. J. Reed and B. D. Hughes, “Power-law distributions from exponential been a Professor for eight years, involved in research in pulsed eddy current,
processes: An explanation for the occurrence of long-tailed distributions eddy current, magnetic Barkhausen noise, and ultrasonic testing of materials
in biology and elsewhere,” Proc. JAMS, vol. 8, pp. 329–339, Mar. 2003. and structures.

Authorized licensed use limited to: King Mongkut's University of Technology North Bangkok. Downloaded on January 25,2021 at 09:37:30 UTC from IEEE Xplore. Restrictions apply.

You might also like