Materials Letters 58 (2004) 2102 – 2104

www.elsevier.com/locate/matlet

Remote field eddy current testing for detection of stress corrosion

cracks in gas transmission pipelines
D. Kim a,*, L. Udpa b, S. Udpa b
a
Iowa State University, 303 Durham Center, Ames, IA 90089, USA
b
Michigan State University, 2120B Engineering, East Lansing, MI 48824, USA
Received 20 August 2003; accepted 5 January 2004

Available online 6 February 2004

Abstract

Magnetic flux leakage (MFL) pigs are traditionally used for the detection of gross corrosion on steel pipelines used for the transmission of
natural gas. Alternative nondestructive evaluation (NDE) modalities are required for the detection of stress corrosion cracking (SCC), which
tends to exist in colonies oriented axially along the length of the pipeline. This paper describes the use of multi-phase rotating magnetic fields
in the remote region of the probe as a possible SCC detection mechanism. Details of a prototype pig and test rig are given and the challenges
associated with the finite element modeling of the device are discussed. Initial experimental results show that this novel NDE modality is
sensitive to axially oriented tight cracks.
D 2004 Elsevier B.V. All rights reserved.

PACS: 81.70.Ex

Keywords: Magnetic materials; Defects; Nondestructive evaluation; Remote field; Eddy current; Stress corrosion cracks

1. Introduction cracks. This paper describes the use of poly-phase rotating

Magnetic flux leakage (MFL) pigs are widely used in
the gas industry for the detection of gross corrosion in gas
transmission pipeline [1]. Such devices are used because
the corrosion presents a strong discontinuity to the axially
oriented magnetic flux, giving rise to strong leakage field
signals easily detectable by the on-board Hall element
sensors. Axially oriented cracks are not so easily detectable
with conventional MFL equipment, however, as such
defects do not have a large dimension in a direction
orthogonal to the axially oriented excitation flux. Alterna-
tive MFL modalities must be sought if MFL pigging is to
be used as the detection mechanism for axially oriented
cracks such as stress corrosion cracking (SCC) colonies.
Several alternative modalities have been suggested by the
authors’ research group [2], including the use of velocity-
induced currents in the pipeline wall, which have a
primary orientation orthogonal to axially oriented tight
magnetic fields to set up currents and magnetic fields
having components that are oriented orthogonal to axially
oriented SCC colonies. Measurements are made in the
remote field region of the magnetic circuit some distance
away from the exciter where the induced eddy current
fields are decaying inwards from the outside diameter [3].
The following sections of this paper describe the modeling
challenges to be faced in simulating the rotating field
phenomenon in the remote field region. Details are given
of an experimental test rig, and results are described which
show this type of pigging is sensitive to axially oriented
artificial cracks.
2. Simulation challenge
The governing equations for this problem are given by
the Maxwell – Ampere law:

* Corresponding author. Present address: Samsung Electronics Co.

Z Z
Ltd., P.O. Box 105, Suwon, Kyungki-Do, 442-742, South Korea.
E-mail address: drdw.kim@samsung.com (D. Kim).
lH̄˙dl ¼ J̄ ds

0167-577X/$ - see front matter D 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.matlet.2004.01.006
 D. Kimcor et al. / Materials Letters 58 (2004) 2102–2104 2103

and the Maxwell– Faraday law:

Z Z
l Ē dl ¼ jxl H̄ ds

with the following interface conditions between the media:

Ht1 ¼ Ht2

Bn1 ¼ Bn2

Et1 ¼ Et2

Fig. 1. Experimental setup of RFEC system.

and constitutive relations:

J̄ ¼ rĒ
B̄ ¼ lH̄
Previous work based on finite element analysis [4,5] has
been concerned with modeling tight cracks in the remote
field region of a conventional MFL pig in order to
determine the sensitivity of circumferentially oriented pipe
wall currents to SCC. Modeling of the poly-phase fields
will need to incorporate the effects of non-linearity of pipe
wall material properties, sensor velocity in the axial direc-
tion and time-varying fields in circumferential direction.
The mesh generated must model the remote field region
and presence of tight, zero volume cracks in the pipe wall.
Consequently, a full fledged model of the inspection
geometry will result in large finite element mesh with
iterative procedures for dealing with non-linearity and time
variations.
3. Experimental setup
The experimental setup consists of a three-phase power
supply, a wound induction motor rotor, a pick-up coil, a
lock-in-amplifier (LIA), a PC-based motion controller and a
personal computer. In order to perform the experiment, four
different saw-cut machined defects were made on the outer
wall of a clean 215.9-mm-diameter pipe. Each of the defects
is 25.4 mm long and of 80%, 60%, 40%, 20% depth of pipe
wall thickness (4.7625 mm). The dimensions of these
defects are given in Table 1.
Table 1
Dimensions of defects [mm]
Defects 80% 60% 40% 20%
Depth 3.81 2.857 1.905 0.952
After determination of the extent of the remote field
region, the location of the excitation coil is identified.
Then, the pick-up coil, which is a differential sensor in
this experiment, is located 990 mm away from the center
of the induction motor rotor. Fig. 1 shows the experimen-
tal setup. The induction motor rotor (excitation coil) is
powered by a three-phase power supply, and a pick-up
coil is located on a circular disc, which is capable of
circumferential movement inside the pipe. The disc is
controlled by a PC-controlled motor. The PC-based mo-
tion controller (Arrick, MD-2) controls the circumferential
movement of the disc, which contains the pick-up coil.
The signal from the pick-up coil is pre-amplified with
appropriate gain and the pre-amplified signal goes into a
LIA (Ithaco 3981), which is used to measure the phase
and magnitude of the low amplitude induced signal. The
LIA is composed of the remote signal input module,
which is connected externally to a PC and main board
installed inside the PC. The reference frequency comes
from the three-phase power supply and goes into the LIA
main board. The outputs of LIA are phase and magnitude
signals, and those values are saved in the data acquisition
system.
4. Results
The pick-up coil senses the electromagnetic field that
penetrates through the pipe wall from OD to ID in the
remote field region. The RFEC scanning is done in a two-
dimensional manner. First, the disc that contains the pick-up
coil moves circumferentially and then axially.
Fig. 2 shows the scanned results for each of the four
defects. The x- and y-axis in this figure stand for inches
and degrees in axial and circumferential directions, re-
spectively. The phase signal has been de-trended to
remove its linear trends and the derivative of the phase
2104 D. Kimcor et al. / Materials Letters 58 (2004) 2102–2104
Fig. 2. Detection of phase signal for defects: (a) 80%, (b) 60%, (c) 40%, (d) 20% depth.
is filtered by a wavelet de-noising scheme. A boundary
detection algorithm is employed to estimate the edges of
the defects.
5. Concluding remarks
Initial studies have shown that the rotating magnetic field
pig is sensitive to axially oriented tight cracks. Future work
calls for experimental data to be gathered from runs on the
recently constructed pipeline testing facility that contains
pipeline sections having axially oriented SCC colonies.
Design studies will then be made for a full size prototype
pig for field use.
References
[1] L. Udpa, S. Mandayam, S.S. Udpa, Y.S. Sun, W. Lord, Developments
in gas pipeline inspection technology, Materials Evaluation 54 (4)
(1996 April) 467 – 472.
[2] G. Katragadda, J.T. Si, W. Lord, Y.S. Sun, S.S. Udpa, L. Udpa, Alter-
native magnetic flux leakage modalities for pipeline inspection, IEEE
Transactions on Magnetics 32 (3) (1996 May) 1581 – 1584.
[3] W. Lord, S. Sun, S.S. Udpa, S. Nath, A finite element study of the
remote field eddy current phenomenon, IEEE Transactions on Mag-
netics 24 (1) (1988 January) 435 – 438.
[4] G. Kobidze, W. Lord, Tight crack modeling for the finite element
simulation of inspection tools in pipelines, Materials Evaluation 56
(10) (1998 October) 1223 – 1226.
[5] S. Yang, Y. Sun, L. Udpa, S.S. Udpa, W. Lord, Application of pertur-
bation methods in finite element analysis of stress corrosion cracking,
IEEE Transactions on Magnetics 36 (4) (2000 July) 1714 – 1718.