Fukutomi 2001
Fukutomi 2001
Fukutomi 2001
www.elsevier.com/locate/ndteint
Abstract
Unlike the impedance plane analysis form of common eddy current testing (ECT), the remote ®eld eddy current (RFEC) technique is a
through-transmission effect that reduces problems such as lift-off normally associated with ECT. In the inspection of steam generator (SG)
tubes, the real issue is to detect the minute cracks growing up from the outside. However, using ECT, it is considered infeasible to accurately
®nd them from the inside because of the limitations of penetration of eddy currents. This paper describes a ®nite-element approach to the
solution of time-harmonic electromagnetic ®elds for the RFEC technique based on a magnetic vector potential and an electric scalar
potential. A comparison is made of experimental and ®nite-element predictions of electromagnetic phenomena under the inspection of
non-magnetic tubes. For the cracks outside demanding high sensitive and precise measurements in the SG tube inspection, numerical results
are given for parameters to design a RFEC probe. q 2001 Elsevier Science Ltd. All rights reserved.
Keywords: Eddy current testing; Remote ®eld eddy current techniques; Cracks; Non-magnetic tubes; Transition zones; Finite-element methods
1. Introduction attenuated direct ®eld. The region where the direct and indirect
®elds are of comparable magnitude and can interfere destruc-
The remote ®eld eddy current (RFEC) technique is a well tively is termed the transition zone. At approximately 2±2.5
established nondestructive evaluation (NDE) method for outside diameters away from the exciter, the rediffusing
detecting corrosion and wall losses in ferromagnetic external ®eld will dominate the internal ®eld at the detector.
conducting tubes. It is used particularly in heat exchanger This region is termed the remote ®eld. On the other hand, the
and small diameter pipes since it detects internal and exter- region that is close to the exciter is termed the near ®eld.
nal discontinuities with approximately equal sensitivity, and Research has recently been done to study RFEC phenom-
the probes ®t small space requirements. The remote ®eld ena and its applications by using analytical solutions [2] or
phenomenon is based on the through-wall indirect coupling axisymmetric ®nite-element modeling [3]. The RFEC
energy path that exists for a low frequency ®eld generated in problems require larger domains and ®ner meshes because
a straight tube [1]. This indirect path contains two effective electromagnetic ®elds far from exciting coils are the most
transits of the pipe wall. One transit is from the inside out, important to predict the phenomena and signals, and the
the other from the outside back in, making it an attractive RFEC signal changes due to defects are fairly smaller
NDE method for wall metal losses. than those of eddy current testing (ECT) in general. For
The typical con®guration consists of two internal the reduction of computer resources, an approach has been
solenoids, one exciter coil and one detector. At low frequen- proposed in which small important parts including the
cies, where the wall thickness to skin depth ratio is of the defects, basically three-dimensional parts, are under consid-
order of one, the ®eld diffuses out rapidly through the tube eration, and some data from the corresponding axisym-
wall with some attenuation and phase shift. This external metric problems are used as the boundary conditions [4].
®eld is guided along the outer surface and then rediffuses As far as the authors know, there has been no report of
back into the tube, where it combines with the rapidly success in three-dimensional numerical simulation, because
of the necessity of very high accuracy.
* Corresponding author. Tel.: 181-22-217-5248; fax: 181-22-217-5248. Using ECT, the steam generator tubes, which are non-
E-mail address: takagi@ifs.tohoku.ac.jp (T. Takagi). magnetic in nuclear power plants, are inspected for the
0963-8695/01/$ - see front matter q 2001 Elsevier Science Ltd. All rights reserved.
PII: S 0963-869 5(00)00026-8
18 H. Fukutomi et al. / NDT&E International 34 (2001) 17±23
Fig. 2. Measured RFEC signals as a function of the distance between the exciter and detector versus numerical predictions. The third line (3D,N) represents the
results from the node-based ®nite elements, and the fourth line (3D,E) represents those from the edge-based ®nite elements. (a) On the sides of the valley in the
transition zone between the near and remote ®eld zones, the tendencies that signal traces decrease are different. (b) In the transition zone there are sharp drops
in phase.
models are set: one is for a comparison between the exper- Similar to Fig. 2, numerical predictions of the internal
imental measurements [13] and numerical predictions, and ®eld are taken as the detector is pulled away from the
the other is for the applications to SG tubes. The straight stationary exciter at 200 kHz (skin depth: 1.13 mm). Fig.
tubes are fabricated from brass and nickel-base superalloy, 3 shows the comparison in model 2 between the axisym-
which realistic SG tubes are made from. Each probe consists metric version and three-dimensional version with the
of an exciter and a detector coil, which have the same edge-based elements, and these results are also in excellent
dimensions. Table 1 summarizes the important dimensions, agreement in the SG tube model. At this frequency, when
material properties, and test conditions. skin depth approximately corresponds to the tube wall thick-
In Fig. 2, the induced voltage changes are plotted as the ness, the log magnitude falls rapidly with a near straight line
detector moves with the exciter ®xed, to show the agreement in the direct ®eld (0 to 0.9 diameters from the exciter), and
between the experimental measurements and the numerical dips are observed; the phase promptly drops in the same
predictions of the axisymmetric and three-dimensional ®eld. This can be explained by the fact that the decline of
versions in model 1. This good agreement implies the magnetic ¯ux directly transmitted from the exciter is much
same degree of accuracy in predicting the eddy-current faster than that outside the tube in the axial direction.
response due to a defect or a crack. By the four-way sym- Therefore, indirect magnetic ¯ux is dominant in the remote
metry, only a quarter of the volume is discretized in three- ®eld zone inside the tube. To comprehend the RFEC
dimensional problems, and Table 2 summarizes the phenomena, it is also worth mentioning that a Poynting
computational costs. This ®nite-element code is considered vector real energy plot shows that the transition zone
useful for three-dimensional problems. occurs where outwardly directed energy from the exciter
meets inwardly directed energy from the outside tube in
Fig. 4. Although it is known that the distance between
Table 2
Computational costs of the ®nite-element code in three-dimensional models an exciter and a transition zone, in general, is 2±
2.5 diameters, it appears in a diameter from the exciter
Model 1 Model 2 in this case.
Node-based Edge-based Edge-based
The signal traces associated with a circling circumferen-
tial outer defect, which are non-through, 2 mm wide and
Elements 24.192 22,176 20% deep through the wall, are shown in Fig. 5 and are
Nodes 28,250 25,810 computed by axisymmetric version. In this ®gure, the indi-
Edges 80,545 73,669
cations occur when the exciter and detector pass the defects.
Unknowns 71.706 65,296 59,637
Memory (MB) a 220 200 200 A circumferential defect gives us two approximately same
CPU time (s) b 1836 1375 1302 high peaks in magnitude with the exciter and detector
a
passages. In this model, the same solution is obtained
The computation is carried out with double precision.
b
The computer is from Visual Technology with DEC alpha A21164,
using the node-based ®nite elements. The voltage changes
whose clock frequency is 533 MHz. The acceleration factors of the ICCG due to the defects mentioned above are plotted in Fig. 6,
method are 1 in node-based elements and 1.02 in edge-based elements. following impedance plane analysis form of ECT. The
Convergence tolerance is 1 10 27. maximum value and its phase changes due to the defect
20 H. Fukutomi et al. / NDT&E International 34 (2001) 17±23
Fig. 3. Numerical predictions of RFEC signals as a function of the distance between the exciter and detector at 200 kHz.
Fig. 4. Axisymmetric ®nite-element prediction of Poynting vector real energy directions under the RFEC phenomena in the non-magnetic tube.
are de®ned as: defects (an inner defect (ID) and an OD) were predicted
q with various coil intervals and excitation frequencies. Fig.
jV j DV x2 1 DVy2 ;
3 7 shows the ratio of the OD to ID in magnitude and differ-
ences of the phase changes between them. It can be seen that
the signal changes of the OD are discernibly higher than
/V tan21 DVy =DVx £ 180=p;
4 those of the ID when detector is in and around the transition
zone, as shown in Fig. 4. To be exact, the volume of the two
where DVx and DVy are the real and imaginary parts of the
defects is slightly different but the difference can be
changes, respectively.
neglected in this comparison. Irrespective of preferable
To detect outer defects (ODs) with higher sensitivity,
high sensitivity, due to noises linked to lift-off variations
maximum changes due to two 40% deep and 1 mm wide
or magnetic inhomogeneities, the use of the transition zone
is generally avoided. However, SG tubes are non-magnetic
Fig. 7. Magnitude rates and phase changes of maximum signals of an ID and an OD, which are 2 mm wide and 40% deep, Si and So, with the various
frequencies and coil intervals. The rate and change are de®ned as uSo u=uSi u and arg(So) 2 arg(Si), respectively. The OD is detected with higher sensitivity when
the detector is in the transition zone, and then the difference of the phase changes appears between the maximum signals of the ID and OD.
Fig. 8. RFEC signals as a function of the detect depth due to the circumferential IDs, a through detect and the ODs, 2 mm wide, at 200 kHz with coil intervals
of 15 (transition zone) and 30 mm (remote ®eld zone). (a) There is an advantage that ODs are detected to be more sensitive than IDs, and the phase lag is useful
to know defect depths. (b) With approximately equal sensitivity to the depths, the phase changes do not depend on the depth, and the phase corresponding to
the maximum magnitude remains constant.
22 H. Fukutomi et al. / NDT&E International 34 (2001) 17±23
Fig. 9. RFEC signals as a function of the detect depth due to inner, through, and outer pits, 2 mm length and width, at 200 kHz with a coil interval of 15 mm.
The same characteristics are observed on the condition that the detector is ®xed in the transition zone as in the case of the circumferential defects.
and have smooth surfaces inside, and the there is no need to tion constructed with a magnetic vector and an electric
take the velocity effect by the probe motion system into scalar potential was used in the axisymmetric and three-
account. Therefore, putting a detector in the transition dimensional problems, to evaluate the detectability. Both
zone is expected to be useful in terms of the detection of numerical predictions were compared for the purpose of
small outer defects such as cracks. As the tube wall becomes accuracy of the numerical simulations tested in the model
thicker, the same characteristics are observed at a lower simulating the SG tube inspection. Since they are in excel-
frequency where the wall thickness to the skin depth ratio lent agreement, the use of the simulation is possible and,
is of the order of one. what is more, valid to study the phenomena and perfor-
The RFEC signals obtained by means of the axisym- mance under the test. From the results of the numerical
metric and the three-dimensional models are compared at simulation, the common RFEC techniques are expected to
200 kHz for defects with various depths and for two loca- detect outer cracks as sensitively as inner cracks when they
tions of the detector: in the typical remote ®eld and transi- are applied to the non-magnetic tubes such as steam gener-
tion zone. The defects are 2 mm wide, and 20, 40, 60, 80 and ator tubes in nuclear power plants. Moreover, this paper
100% deep. Fig. 8 shows the maximum changes in magni- demonstrates that when putting a detector in a transition
tude due to the defects as a function of the defect depth. The zone, signals to an outer defect are higher than those to an
results demonstrate good agreement and it is distinctive that inner defect with the same depth and volume, and the depths
the correlation between the changes and depths is approxi- are classi®ed by their signal phase information, as it is
mately linear although the changes due to the outer defects usually used in ECT.
from typical ECT exponentially decline, not in direct
proportion to their depths. Unlike Fig. 8(b), the signal
changes due to the ODs are larger than those of the IDs in Acknowledgements
Fig. 8(a) at the frequency as the wall thickness to the skin
depth ratio is of the order of one. As for the phase, while the This study was supported in part by the Research
changes re¯ect the depths in Fig. 8(a), the changes approxi- Committee on the Nondestructive Evaluation Technology
mately remain constant in Fig. 8(b). This means setting the by Eddy Current Testing of the Japan Society of Applied
detector in the transition zone makes it possible to know Electromagnetics and Mechanics through a grant from 5
defect depths from acquired signals as well as typical PWR utilities and Nuclear Engineering Ltd., and the
ECT. For practical purposes, RFEC signals must be Grant-in-Aid for COE Research (11CE2003). The authors
predicted from arbitrarily shaped defects in three-dimen- would like to thank Prof Hiroshi Hoshikawa and Dr Kiyoshi
sional problems. The same tendency observed in Fig. 8 is Koyama of Nihon University in Japan for their advice and
observed with pit-type defects of 2 mm length and width, as encouragement for this work.
shown in Fig. 9.
References
4. Summary [1] Schmidt TR. The remote ®eld eddy current inspection technique.
Mater Eval 1984;42(2):225±30.
For an improvement in detectability of outer cracks, the [2] Haugland SM. Fundamental analysis of the remote-®eld eddy-current
advantages of the RFEC techniques that stems from the effect. IEEE Trans Mag 1996;32(4):3195±211.
[3] Lord W, Sun YS, Udpa SS, Nath S. A ®nite element study of the
through-transmission effect was used for the inspection of remote ®eld eddy current ®eld. IEEE Trans Mag 1988;24(1):435±8.
non-magnetic tubes. The experimentally validated ®nite- [4] Lin HY, Sun YS. Application of zoom-in technique in 3D remote ®eld
element code solving an electromagnetic governing equa- eddy current effect computation. IEEE Trans Mag 1990;26(2):881±4.
H. Fukutomi et al. / NDT&E International 34 (2001) 17±23 23
[5] Dodd CV, Deeds WE. Analytical solutions to eddy-current probe-coil [10] Fukutomi H, Huang H, Takagi T, Tani J. Identi®cation of crack
problems. J Appl Phys 1968;39(6):2829±38. depths from eddy current testing signal. IEEE Trans Mag
[6] Palanisamy R, Lord W. Finite element modeling of electromagnetic 1998;34(5):2893±6.
NDT phenomena. IEEE Trans Mag 1979;15(6):1479±81. [11] Takagi T, Uesaka M, Miya K. Electromagnetic NDE research activ-
[7] Ida N, Lord W. A ®nite element model for three-dimensional eddy ities in JSAEM. In: Takagi T, Bowler JR, Yoshida Y, editors. Elec-
current NDT phenomena. IEEE Trans Mag 1985;21(6):2635±43. tromagnetic nondestructive evaluation, IOS Press, 1997. p. 9±16.
[8] Takagi T, Hashimoto M, Fukutomi H, Kurokawa M, Miya K, Tsuboi [12] Takagi T, Huang H, Fukutomi H, Tani J. Numerical evaluation of
H, Tanaka M, Tani J, Serizawa T, Harada Y, Okano E, Murakami R. correlation between crack size and eddy current testing signal by a
Benchmark models of eddy current testing for steam generator tube: very fast simulator. IEEE Trans Mag 1998;34(5):2581±4.
experiment and numerical analysis. Int J Appl Electromagn Mater [13] Isobe M, Iwata R, Nishikawa M. High sensitive remote ®eld eddy
1994;5:149±62. current testing by using dual exciting coils. In: Collins R, Dover WD,
[9] Fukutomi H, Takagi T, Tani J, Hashimoto M. Numerical evaluation of Bowler JR, Miya K, editors. Nondestructive testing of material, IOS
ECT impedance signal due to minute cracks. IEEE Trans Mag Press, 1997. p. 145±52.
1997;33(2):2123±6.