See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/253799044

Development of a 3D Electromagnetic Model for Eddy Current Tubing

Inspection: Application to Steam Generator Tubing

Article · February 2004

DOI: 10.1063/1.1711640

CITATIONS READS

23 204

4 authors, including:

Denis Prémel Thierry Sollier

Atomic Energy and Alternative Energies Commission Institut de Radioprotection et de Sûreté Nucléaire (IRSN)
86 PUBLICATIONS 383 CITATIONS 95 PUBLICATIONS 202 CITATIONS

SEE PROFILE SEE PROFILE

Valérie Maillot
Andra
21 PUBLICATIONS 242 CITATIONS

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

NDT on AIR European project View project

Corrosion and mechanical behavior of metallic components of nuclear waste repositories View project

All content following this page was uploaded by Denis Prémel on 30 September 2014.

The user has requested enhancement of the downloaded file.

DEVELOPMENT OF A 3D ELECTROMAGNETIC MODEL FOR
EDDY CURRENT TUBING INSPECTION: APPLICATION TO
STEAM GENERATOR TUBING

G. Pichenot1, D. Prémel1, T. Sollier1, V. Maillot2

1
C.E.A. Saclay, 91191 Gif-sur-Yvette, France
2
I.R.S.N, 92262 Fontenay-aux-Roses, France

ABSTRACT. In nuclear plants, the inspection of heat exchanger tubes is usually carried out by using
eddy current nondestructive testing. A numerical model, based on a volume integral approach using the
Green’s dyadic formalism, has been developed, with support from the French Institute for Radiological
Protection and Nuclear Safety, to predict the response of an eddy current bobbin coil to 3D flaws
located in the tube’s wall. With an aim of integrating this model into the NDE multi techniques
platform CIVA, it has been validated with experimental data for 2D and 3D flaws.

INTRODUCTION

The inspection of heat exchanger tubes is usually carried out by using eddy current (EC)
nondestructive testing. This technique, based on the analysis of changes in the impedance of
one or more coils placed inside the tube, is used to detect and characterize possible flaws or
anomalies in the tube. In this context, for bobbin coils moving on the axis of the tube, many
works have been done on the development of axisymmetric models in order to predict EC
signals for circumferential flaws [1, 2, 3]. Others works, based on the Green’s dyadic
formalism [4, 5], are focused on 3D volumetric flaws by using volume integral models. In
most investigated models, the axis of the driving coil is assumed to be coaxial with the axis
of the inspected tube. Some recent works consider the EC problem of a conducting tube
inspected by an eccentric circular coil [6, 7, 8, 9]. However, these works highlight the
computation of the induced field in the tube’s wall or the impedance of the coil without
considering any flaw.
This paper describes the progress in developing a 3D computer code based on the
volume integral method which has the capability to predict quickly the response of an eddy
current probe to 3D flaws. This model is suitable to predict the probe response when the
probe moves straight along the tube axis. The model can also be used to quantitatively
evaluate perturbation factors on the eddy current signal such as the probe wobble (off-axis
response for a non-centered bobbin coil configuration). The model gives the eddy currents
distribution within the tube’s wall and the changes in self and mutual inductance due to 3D
flaws. 3D flaws are described as a local variation of conductivity which may vary by the
shape, the size and the place in the tube’s wall.
 z z
C

0.5 mm

FIGURE 1. Schematic of the two bobbin coils probe used to inspect a conducting tube. The tube, made of
inconel 600 (conductivity of 1 MS/m), has an inner radius of 9.84 mm and a thickness of 1.27 mm. The probe
consists of two identical coils excited with a time harmonic drive current in phase in both coils (additive flux
mode). The inner radius of the coils is 7.83 mm, the outer radius is 8.5 mm, the height is 2 mm, the number of
turns is 70, the gap between the coils is 0.5 mm. The probe may moves straight along the tube axis (at left) or
it may be non centered on the tube axis considering an eccentricity C (at right).

The probe can as well operate in absolute mode as in differential mode or pseudo
absolute mode with additive or subtractive flux. To illustrate the matter, the configuration
considered in this paper is given in Figure 1.
With an aim of integrating this model into the NDE multi techniques platform CIVA for
an industrial use, the model has been validated with experimental data for 2D and 3D flaws.

THE VOLUME INTEGRAL FORMULATION

The inspected workpiece is a linear, isotropic, non-magnetic and conducting circular tube
(permittivity , permeability , conductivivity ) placed along the z-axis of a cylindrical
coordinate system . The probe is located inside the tube (region D1) and excited by a
time harmonic current of amplitude and angular frequency . A 3D flaw is considered
within the tube's wall (Region D2) by local changes in the conductivity . The electric
field in region D2 can be separated in two vectors and the
forward problem is therefore treated in two steps. The first vector represents the incident
field due to the current sources when there is no flaw while the second one is the perturbed
field due to the presence of the flaw. So, the first step consists in computing the
incident field in region D2 due to the currents sources when there is no flaw, taking
into account or not the eccentricity of the probe. The flaw problem can be handled in a
similar way by considering the flaw is equivalent to a fictitious source of current
which can be seen as a current dipole density. The perturbation
field within the flaw domain satisfies the Helmholtz equation:

(1)

The integral solution of this equation involves the electric-electric dyadic's kernel
where the subscripts i and j denote respectively the region of observation and
source. For canonical geometries, such as cylindrically layered media, the dyadic Green’s
functions can be found in explicit analytical expression [10]. The total internal electric field
in region D2 is obtained by the superposition of the incident field and the perturbed
field :
 (2)

A discrete solution of this integral equation is obtained by application of a Galerkin

variant of the method of moments involving a suitable discretization of the flaw domain.
The incident electric field may be also expressed by an integral equation [10]:

(3)

where is the dyadic Green’s functions corresponding to a source in region D1

and the field observed in the region D2. When the probe moves straight along the tube axis,
the incident field calculation is straight forward by using Dodd and Deeds results valid for
configuration with axial symmetry [1]. On the other hand, when the probe is non centered
on the tube axis, the expressions of the current density in Equation (3) has to be written in
the cylindrical system related to the tube axis [8].
By using the reciprocity principle which involves the incident electric field and the
internal electric field in region D2, the changes in self and mutual inductances due to
3D flaws is given for each coil by:

(4)

where is the incident field only produced by the coil and is the internal field
produced when the tube is only excited by the coil .

EXPERIMENTAL SETUP

As the changes in self and mutual inductances are very small, the measurements are
usually done by using a Wheatstone bridge. The probe operates in differential mode at
100 kHz, 240 kHz and 500 kHz and in pseudo absolute mode at 100 kHz.

DIFFERENTIAL OPERATING MODE

In differential operating mode, the system is excited by a voltage while the quantity
which is measured is the voltage .
As it has already been mentioned by Dodd and Deeds [1], in this mode the experimental
setup has no significant influence on the measurements and the measurement is
proportional to difference of the changes of impedance due to the flaw:

(5)

PSEUDO ABSOLUTE OPERATING MODE

In pseudo absolute operating mode, the system is excited by a voltage while the
quantity which is measured is the terminal voltage (denoted ) of one of the two coils.

By considering only the two coils, the expression of is given by:

 (6)

RESULTS

In order to validate this numerical model, the simulated results were compared with
experimental data. The EC signals are normalized with a 1 mm width 40% through wall
outer groove (OG40). The simulated results have been validated with experimental data
with various circumferential defects (Figure 2) as the 1 mm width 10% through wall inner
groove (IG10).
As these flaws are circumferential defects, these configurations are 2D problems when
the probe is centered in the tube. The results obtained with the 3D model in these
configurations can thus also be compared with data obtained with a previous 2D computer
modeling tool CIVA [3].
The results obtained with the 3D model have been validated with experimental data with
various 3D defects by always normalizing the EC signal with the OG40 flaw.

OG40

IG10

FIGURE 2. Circumferential OG40 and IG10 flaws: schematic and EC signal plane diagram for a centered
probe operating at 100 kHz in differential mode (‘’ experimental data, ‘•’ 2D model, ‘+’ 3D model).
 z

FIGURE 3. Transversal TN100 notch: schematic and EC signal plane diagram for a centered probe operating
at 100 kHz in differential mode (‘’ experimental data, ‘+’ 3D model).

Figure 3 displays the experimental and simulated EC signal plane diagram obtained with
a transversal notch at 100 kHz. This notch (TN100) is a 100% through wall flaw (
) of 0.113 mm length ( ) and of 82° angular
extension ( ).
Figure 4 displays the EC signal plane diagram obtained with a longitudinal notch at
100 kHz. This notch (LN54) is a 54% through wall outer flaw ( ) of 10 mm
length ( ) and of 0.6° opening ( ).

FIGURE 4. Longitudinal LN54 notch: schematic and EC signal plane diagram for a centered probe operating
at 100 kHz in differential mode (‘’ experimental data, ‘- -’ 3D model).
FIGURE 5. Circumferential OG40 flaw: EC signal plane diagram for a centered probe operating at 100 kHz
in pseudo absolute mode (‘’ experimental data, ‘+’ 3D model).

Figure 5 displays the experimental and simulated EC signal plane diagram obtained at
100 kHz with the OG40 circumferential flaw for a centered probe operating in pseudo
absolute mode.
Figure 6 displays the results obtained with the OLG30 circumferential groove and the
longitudinal LN54 notch. The OLG30 flaw is a 30% outer through wall groove of 20 mm
length ( ).

FIGURE 6. Circumferential OLG30 (left) and longitudinal LN54 (right) flaw: EC signal plane diagram for a
centered probe operating at 100 kHz in pseudo absolute mode (‘’ experimental data, ‘+’ 3D model).
FIGURE 7. Transversal TN100 flaw: EC signal plane diagram for a centered probe operating in pseudo
absolute mode (left) at 100 kHz and in differential mode (rigth) at 500 kHz (‘’ experimental data, ‘+’ 3D
model).

Figure 7 displays the results obtained with the longitudinal TN100 notch. The simulated
results were also compared with experimental data at other frequencies. The observed
agreements between the results provided by the 3D model and the experimental data are as
good as those presented here (for example at 500 kHz, see Figure 7).
Finally, Figure 8 presents typical impedance plane diagram for a non centered probe
operating in differential mode for three values of the eccentricity C obtained with a
transversal notch. This 3D flaw is a 20% through wall outer notch of 0.1 mm length (
) and of 90° opening ( ).

FIGURE 8. Transversal flaw: impedance plane diagram for a non centered probe operating at 100 kHz in
differential mode for three values of eccentricity: C = -0.3, C = 0 mm and C = +0.3 mm.
 CONCLUSION

A 3D model for eddy current tubing inspection has been developed and a fast numerical
code has been implemented. The computing times for the presented EC signal plane
diagrams vary from few minutes for 2D configuration up to half an hour for 3D
configuration with a 1GHz/1GB PC.
The code may be used to model EC tubing inspection and to study the variations of the
probe eccentricity. This 3D model will be integrated in the nondestructive multi techniques
platform CIVA.
Good agreements have been observed between experimental and simulated data as well
in 2D configurations as in 3D configurations. Further validations will be done considering
other flaws (holes …).
The 3D model can be generalized to solve other geometries, in particular, the case of an
arbitrary shaped and positioned probe placed outside [8] or inside the tube by changing the
incident field computation. Work is presently in progress to simulate, for instance, and eddy
current tubing inspection with a pancake coil whose the axis of revolution is perpendicular
to the axis of the tube.

REFERENCES

1. Dood, C. V. and Deeds, W. E., International Journal of Nondestructive Testing 1, 29-90

(1968).
2. Sabbagh, H. A. and Sabbagh, L. D., “Development of a System to Invert Eddy-Current
Data and Reconstruct Flaws”, in Review of Progress in QNDE Vol. 2, edited by D. O.
Thompson and D. E. Chimenti, Plenum, New York, 1983, pp. 1555-1567.
3. Berthiau, G. and de Barmon, B., “MESSINE, an Eddy Current Parametric Model for
Flaw Characterization”, in Review of Progress in QNDE Vol. 18, edited by D. O.
Thompson and D. E. Chimenti, Plenum, New York, 1999, pp. 501-508.
4. Sabbagh, H. A. and Sabbagh, L. D., IEEE Trans. Magnet. MAG-22, 282-291 (1986)
5. Monebhurrun, V., Duchêne B. and Lesselier D., “Eddy Current Characterization of 3D
Bounded Defects in Metal Tubes Using a Wavefield Integral Formulation Modeling”, in
Nondestructive Testing of Materials, edited by R. Collins, W.D. Dover, J.R. Bowler and
K. Miya, London, 1995, pp. 195-202.
6. Dyakin, V. V., Russian Journal of Nondestructive Testing 33, 143-153 (1997).
7. Sakaji, N. M., J. Appl. Phys. D 33, 2239-2248 (2000).
8. Micolau, G., Pichenot, G., Lambert, M., Lesselier, D., Premel, D., “Three-Dimensional
Electromagnetic Field in a Conductive Cylinder at Eddy-Current Frequencies”, in
ENDE’2002 Workshop Proceedings (to appear).
9. Theodoulidis, T. P., Res. Nondestr. Eval. 14, 111-126 (2002).
10.Chew, W. C., Waves and Fields in Inhomogeneous Media, IEEE Press, Piscataway,
1995, Chapters 3, 7 and 8.
11.Li, Y., Zhang, Z., Sun, Y., Udpa, L. and Udpa, S., “Numerical Simulation Results for the
Eddy Current Benchmark Problem”, in Review of Progress in QNDE Vol. 21, edited by
D. O. Thompson and D. E. Chimenti, Plenum, New York, 2002, pp. 1902-1908.
12.Tian, Y., Li, Y., Udpa L. and Udpa, S., “Simulation of the World Federation’s Second
Eddy Current Benchmark Problem”, in Review of Progress in QNDE Vol. 22, edited by
D. O. Thompson and D. E. Chimenti, Plenum, New York, 2002, pp. 1816-1823.

View publication stats