12 Journal of Material Science and Technology Research, 2021, 8, 12-18

Numerical Model of Eddy Current Inspection with DC Magnetic

Field Associated

J.V. Rocha1, C. Camerini1,*, R.W.F. Santos2, V.M. Silva1, Lucas B. Campos1 and
G.R. Pereira1

1
Laboratory of Non-Destructive Testing, Corrosion and Welding, Department of Metallurgical and Materials
Engineering, Federal University of Rio de Janeiro, Brazil
2
Petrobras Research and Development Center, Rio de Janeiro, Brazil
Abstract: Most non-destructive techniques can be well represented in a virtual environment, in particular, eddy current
testing (ECT) simulation is a useful and well-established tool to predict and represent real inspection situations permitting
testing customization in a fast, cheap and efficient way. Conventional ECT generally works with low-intensity magnetic
fields, however, for advanced variations of the technique, where external DC magnetic fields can be applied to locally
decrease the magnetic permeability, there is no Finite Element Method (FEM) packages available to deal with such
nonstandard model. Many authors [1] and [2] have presented this ECT solution for different industrial applications using
external DC magnetization to carry nonlinear ferromagnetic materials to the saturation level of the magnetization curve to
increase the ECT depth penetration. In general, ECT modelling calculation is benefited by properties of steady-state
regime where all magnetic fields are oscillating at the same frequency not permitting through multi-frequency calculation.
The present work proposes a simulation solution for such a case where DC magnetic field is associated with ECT. A
theoretical model is presented together with experimental results validation.

Keywords: Eddy current testing, External DC magnetic field, Material inspection.

1. INTRODUCTION model. In general, ECT modelling calculation is

benefited by properties of steady-state regime where
Eddy current testing (ECT) is a widely used non-
all magnetic fields are oscillating at the same frequency
destructive technique to evaluate materials in many
not permitting through multi-frequency calculation.
industries scenarios. Aircraft, nuclear and
Many scientific efforts have been deposited to
petrochemical industries are examples where ECT
overcome the limitations of the existing ECT FEM
plays a key role in ensuring components integrity. The
packages. Cuihua Tian et al. computationally evaluated
method is based on the detection of coil impedance
demagnetization in NdFeB permanent magnets present
change due to eddy currents induced on the test
in saturated core fault current limiter (SCFCL) proved
specimen [3-7]. The presence of a defect modifies the
by eddy currents [12]. For this, a finite element analysis
eddy current pattern which gives rise to a field
was performed in order to obtain the transient
perturbation and hence changes the coil impedance.
configuration of the SCFCL, then the expression of the
From the industry point of view, it is relevant to have flow density was obtained by adjusting the curve and,
a dedicated simulation tool to reliably represent real finally, the losses were calculated using the analytical
inspection situations in a virtual environment. In this model. Works related to the inspection area by non-
way, several works have addressed methodologies of destructive techniques presented by Fabrice Foucher
computational analysis aiming at the validation of et al. [13], Zhiyang Deng et al. [14] and Satoru Horai et
experiments and customization of ECT sensors [8-11], al. [15] show the use of finite element analysis methods
resulting though in a few FEM commercial software for for the case of eddy currents with magnetic saturation
the standard ECT application. However, conventional through an external DC field. These were developed by
ECT generally works with low-intensity magnetic fields separating the problem into two stages, in the first it
and for advanced variations of the technique, where analyzed only the interaction of the DC field with the
external DC magnetic fields can be applied to locally material in order to obtain the distribution of magnetic
decrease the magnetic permeability, there are no FEM permeability. Then eddy current inspection simulations
packages available to deal with such a nonstandard were performed for the permeability settings obtained
in the first stage. However, due to the focus of the
works being the results of the inspection of the
*Address correspondence to this author at the Laboratory of Non-Destructive simulation, it is not presented in details of the
Testing, Corrosion and Welding, Department of Metallurgical and Materials
Engineering, Federal University of Rio de Janeiro, Brazil; mathematical formulation of the methodology.
E-mail: [email protected]

E-ISSN: 2410-4701/21 © 2021 Zeal Press

Model of Eddy Current Inspection with DC Magnetic Field Journal of Material Science and Technology Research, 2021, Vol. 8 13

The present work proposes a simulation solution for 2.1. Magneto-Static

such a case where DC magnetic field is associated
with conventional ECT. A theoretical model is The FEM simulation magnetostatic model can be
presented together with experimental results validation. divided into three regions for a better understanding.
ECT in combination with DC magnetic field improves The first one is the region of interest, which includes
the NdFeB magnets, the inspection sample, and a U-
ferromagnetic materials inspection. Eddy currents have
shape metal to close the magnetic circuit, all have
a small penetration depth in ferromagnetic materials
nonzero conductivity, permittivity, or permeability. The
due to high relative magnetic permeability, this limited
second space is the air surrounding the whole model,
penetration is known as the “skin effect” [16, 17] and it
which is considered as free space. The last one is the
can be reduced by imposing a DC magnetization which
outer boundaries. As with all electromagnetic
locally reduces the magnetic relative permeability of the modelling, the basis comes from Maxwell’s Equations
material and increases the eddy current depth [17].
penetration, permitting inspection of thicker
components [18]. Besides the enlargement of depth From Ampere:
penetration, the imposed DC magnetic field lines in
case of a defect have a higher concentration density in
the remaining wall thickness which consequently (1)
changes the relative permeability in the area and
changes the eddy current field lines. Figure 1 presents From Faraday:
the experimental set-up schematic consisted of a
magnetic circuit employing two permanent magnets,
the test piece, a U-shaped piece of low carbon steel to (2)
close the magnetic circuit and the ECT probe between From Gauss:
the magnets. In case of a defect the magnetic field
lines are concentrated in the remaining wall, increasing
(3)
the magnetic flux and changing the eddy current field
lines.
(4)
2. EQUATIONS
The fields B and H have a non-linear relationship
To simulate ECT with DC magnetic field associated with each other so that the equality is founded through
is important to understand the governing equations of µ which is also dependent on the field (Eq. 5):
each situation separately. The DC magnetic field is
calculated by using magnetostatic equations present in (5)
section 2.1 whereas ECT modelling is benefited by
The magnetic scalar potential is defined according
properties of the steady-state regime shown in section
to equation 6, where the ψ represents the (total) scalar
2.2. potential.

Figure 1: Experimental set-up schematic of ECT with DC magnetic field associated.

14 Journal of Material Science and Technology Research, 2021, Vol. 8 Rocha et al.

The solution of this linear diffusion equation when

(6)
the excitation is sinusoidal can be achieved in terms of
In one dimension: A, solving the Eq. 10 taking care of the contour
conditions. With the values of A calculated, it is
possible to obtain many electromagnetic features, such
as: energy dissipation, magnetic flux density and
(7) impedance variation of the eddy current coil [19].

Combining Eq. (6), Eq. (5) and Eq. (4): 3. MODEL DEFINITION

It is not possible to deal with magneto-static and

(8) ECT Steady State equations at the same simulation
model. As noticed by the equations, the magneto-static
This is the equation used for modelling model does not consider oscillating magnetic fields
Magnetostatics problems in the absence of electric while ECT Steady State does not support multi-
currents. Finding ψ allows B and H to be determined. frequency calculation. To solve this conflict two
solutions were considered, the time Domain Solution
2.2. Steady State
and Superposition solution.

Starting with Maxwell’s equations as in equations 1- 3.1. Time Domain Solution

4. In a linear and isotropic environment in which the
physical properties, such as, magnetic permeability µ, The first approach to solve this problem is to use a
electric conductivity σ and permittivity ϵ do not depend time-domain solution instead of a time-harmonic
on the direction, it is possible to assume some relations solution. That way it is possible to consider as many
like Eq. (5), D = ϵE and the current density can be frequencies as necessary in the same solution, but on
written as J= σE. the other hand, many calculations must be made
incrementing time with a small-time step with at least
After manipulating the Maxwell equations in its twenty points per cycle in order to have a sufficient
differential derivative form, one can reach the sample rate. This approach is widely used in Pulsed
differential diffusion that governs the pattern of eddy Eddy Current (PEC) simulation, as the rectangular
currents in conducting materials (Eq. 09) [19]. pulse is a combination of several frequencies [20].
Combine that and the high frequencies that are usually
used in ECT and the simulation process becomes
computationally slow and expensive. It is worth
(9)
mentioning that the main objective of this paper is to
Where: work around this limitation and propose a new solution
using superposition of just two simulations, firstly a
µ: magnetic permeability [H/m]. magneto-static and then an ECT steady state.

A: magnetic potencial vector [Weber/m]. 3.2. Superposition Solution

2
Js: inducted current vector [A/m ]. In this alternative solution the main goal is to
achieve the same results as the transient solution, but
σ: electrical conductivity [S/m]. with a less computational resource. In order to do it,
some assumptions must be made. First, both the
For the linear case, when the coil is excited by a solutions must have a linear relationship between
sinusoidal wave, the equation 09 can be reduced to: other, second, the oscillating field excited by the coil
must be sufficiently small that it can be considered as a
perturbation that does not affect the magnetic
(10) saturation of the sample. This solution idea comes
directly from the electric circuit and linear system
Where: theories where it is possible to compute multiple
sources contributions by calculating each one
ω: angular frequency [rad/s].
individually and just after all computations combine
Model of Eddy Current Inspection with DC Magnetic Field Journal of Material Science and Technology Research, 2021, Vol. 8 15

them, such approach was also applied in previous ECT Because of the symmetry of the model, it is possible
papers [21]. Consequently, it is possible to import the to solve just one-quarter of the structure. The tangential
permeability distribution from the static model to the magnetic boundary condition is applied to the XY plane
dynamic steady-state model. and the normal magnetic boundary condition to the YZ
plane. Magnetic saturation causes the permeability in
4. MODEL DEFINITION the duplex steel to be non-uniform, as shown in Figure
3. The contours are obtained by plotting component
Opera 3D from Vector Fields Cobham was the
B/H (CGS units).
simulation software and the model to be solved is a
horseshoe type magnet with permanent magnet NdFeB The quality of the duplex steel is accessed via ECT,
poles and SAE 1020 steel. The plate, testing sample, by applying an AC field to the sample using a coil. As
of duplex stainless steel completes the magnetic pointed before, this perturbation does not affect the
circuit, as shown in Figure 2. magnetic saturation of the duplex sample.
Consequently, it is possible to use the information of
the magnetic permeability in the saturation state given
by the magnetostatic solution.

The duplex sample plate in the steady-state solution

must have the same geometry like the one in the
statics solution. After the mesh is done in this sample it
is possible to create a table that contains x, y and z
coordinates for all the element centroids for the whole
mesh. These coordinates will serve as inputs values for
the permeability table that is going to be used for the
steady-state solution. It is possible to see the magnetic
permeability values imported to the Steady State model
in Figure 4, note that because there is just one
permeability value for each mesh element the image is
more pixeled than in Figure 3.

After all the calculations are done, it is possible to

Figure 2: Horseshoe magnet setup for statics solving.
see the induced current pattern as in Figure 5. The

Figure 3: Horseshoe magnet solved in non-linear Statics.

16 Journal of Material Science and Technology Research, 2021, Vol. 8 Rocha et al.

Figure 4: Imported magnetic permeability distribution.

results are obtained by analyzing the impedance of the

coil that changes when it approaches the conductive
media that is being tested. To represent the coil (12)
approach to the testing piece three simulations with
liftoff (distance of the sensor from the testing sample) Where:
of 0.5, 2 and 5 millimetres have been made, as can be
seen in Figure 6. Jf and J are the current density field with and without
the defect, respectively.
Once the impedance is a complex number, the
resistance, real part, is computed with the Joule Losses Bf and B are the magnetic induction field with and
(Equation 11) and the inductive reactance, imaginary without the defect, respectively.
part, is computed with the magnetic energy (Equation
I is the applied electric current.
12) both calculations are done within the whole
meshed domain, as in [22]. ω is the angular frequency.

Previous experimental results [2] showed that this

variation of the ECT technique (with external DC
(11)

Figure 5: Eddy current pattern after solving.

Model of Eddy Current Inspection with DC Magnetic Field Journal of Material Science and Technology Research, 2021, Vol. 8 17

The reference samples have been solubilized at

1120° C for 1 hour and water quenched in order to
contain approximately 50% of ferrite and 50% of
austenite and is indicated in the graphic legend as
50/50. Further heat treatments were conducted to
obtain unbalanced microstructures with 80% and 70%
of ferrite and 20% and 30% of austenite, and 8.3% of
sigma phase. The heat treatments parameters to
obtain each of the tested microstructure are detailed by
Camerini et al. [2].

Figure 6: Three different liftoff of 0.5, 2 and 5 millimeters to

simulate the probe approaching the specimen.

magnetization) was capable of sorting through different

types of duplex stainless steel whereas the
conventional ECT could not. Unfortunately, it is not
possible to validate the results in a numerical manner.
This happens because the eddy current equipment
used to make the measures uses digital tools to
enhance the signal, such as: digital filters, overall gain
and vertical gain. The way these tools are utilized is not
explicit in the equipment datasheet. In this way, the
comparison between the results obtained by the Figure 8: Experimental results of duplex stainless-steel
sorting using ECT with external DC magnetic field.
simulation and by the experimental measures was
made qualitatively through the morphology of the 5. CONCLUSION
signals and in the ability to distinguish the different
specimens. In this way it was just possible to validate Based on the advantages of use ECT associated
the results by comparing them qualitatively. The with DC magnetization, a novel numerical model
agreement between the simulated and experimental solution was presented to simulate and predict the
results is good enough as can be seen in Figures 7 and interaction between the alternate and constant
8. The simulated results as well as the experimental magnetic field. The results showed the magnetic
results could sort the different types of duplex stainless permeability variation caused by the imposed DC
steel successfully. magnetic field and using the superposition solution
ECT inspection could be evaluated in the presence of
external DC magnetic fields.

The simulation results presented the possibility to

evaluate metallurgical changes in the stainless steel
samples combining ECT and DC magnetization,
therefore the solution can be likewise applied for
geometrical defects to identify metal losses of the
corrosion process.

REFERENCES

[1] Reber K, Boenisch A. Advanced Electromagnetic

Technology and Tools for Subsea Inspection, Subsea
Pipelines Integrity Management Conference 2012.
[2] Camerini C, Sacramento R, Areiza M, Rocha A, Santos R,
Rebello J, Pereira G. Eddy current techniques for super
duplex stainless steel characterization, Journal of Magnetism
Figure 7: Simulation results. and Magnetic Materials 2015; 388: 96-100, ISSN 0304-8853.
https://doi.org/10.1016/j.jmmm.2015.04.034
18 Journal of Material Science and Technology Research, 2021, Vol. 8 Rocha et al.

[3] Xie S, Duan Z, Li J, Tong Z, Tian M and Chen Z. A novel [12] Tian C, Zhong Y, Wei L, Lei Y, Chen B, Gao Y and Yuan J. A
magnetic force transmission eddy current array probe and its coupled method for evaluating eddy current loss of NdFeB
application for nondestructive testing of defects in pipeline permanent magnets in a saturated core fault current limiter.
structures. Sensors and Actuators A: Physical 2020; 309: IEEE Transactions on Magnetics 2017; 53(6): 1-4.
112030. https://doi.org/10.1109/TMAG.2017.2664828
https://doi.org/10.1016/j.sna.2020.112030 [13] Foucher F, Kalai A, Kelb W, Ramadan S and Delemontez J.
[4] Zhao Y, Qi P, Xie Z, Bai P, Chen HE, Xie S, and Chen Z. A A modeling study of the SLOFECTM Eddy Current system. In
new array eddy current testing probe for inspection of small- 19th World Conference on Non-Destructive Testing 2016.
diameter tubes in Tokamak fusion devices. Fusion [14] Deng Z, Sun Y, Kang Y, Song K and Wang R. A
Engineering and Design 2020; 157: 111627. permeability-measuring magnetic flux leakage method for
https://doi.org/10.1016/j.fusengdes.2020.111627 inner surface crack in thick-walled steel pipe. Journal of
[5] Zhao Y, Qi P, Xie Z, Bai P, Chen HE, Xie S and Chen Z. A Nondestructive Evaluation 2017; 36(4): 1-14.
new array eddy current testing probe for inspection of small- https://doi.org/10.1007/s10921-017-0447-z
diameter tubes in Tokamak fusion devices. Fusion [15] Horai S, Hirata K & Niguchi N. Flux-focusing eddy current
Engineering and Design 2020; 157: 111627. sensor with magnetic saturation for detection of water pipe
https://doi.org/10.1016/j.fusengdes.2020.111627 defects. International Journal of Applied Electromagnetics
[6] James R, Faisal Haider M, Giurgiutiu V, and Lilienthal D. A and Mechanics 2016 52(3-4): 1231-1236.
simulative and experimental approach toward Eddy current https://doi.org/10.3233/JAE-162099
nondestructive evaluation of manufacturing flaws and [16] Rifai D, Abdalla AN, Khamsah N, Aizat M and Fadzli M.
operational damage in CFRP composites. Journal of Subsurface defects evaluation using eddy current testing.
Nondestructive Evaluation, Diagnostics and Prognostics of Indian J. Sci. Technol 2016; 9: 10-17485.
Engineering Systems 2020; 3(1). https://doi.org/10.17485/ijst/2016/v9i9/88724
https://doi.org/10.1115/1.4044722
[17] Jackson JD. Classical electrodynamics 1999.
[7] Sophian A, Tian GY, Taylor D and Rudlin J. Electromagnetic
and eddy current NDT: a review. Insight 2001; 43(5): 302- [18] Bönisch A, Dijkstra FH and de Raad JA. Magnetic flux and
306. SLOFEC inspection of thick walled components. In Proc.
15th World Conference on Nondestructive Testing 2000; 15:
[8] Hollaus K and Schöberl J. Some 2-D multiscale finite- 1-8.
element formulations for the eddy current problem in iron
laminates. IEEE Transactions on Magnetics 2018; 54(4): 1- [19] Ida WLN, Palanisamy R. Eddy Current Probe Design Using
16. Finite Element Analysis, Materials evaluation 1983; 41: 1389-
https://doi.org/10.1109/TMAG.2017.2777395 1493.
[9] Oh S, Choi G, Lee D, Choi M and Kim K. Analysis of Eddy- [20] Mengbao Fan, Pingjie Huang, Bo Ye, Dibo Hou, Guangxin
Current Probe Signals in Steam Generator U-Bend Tubes Zhang, Zekui Zhou. Analytical modeling for transient probe
Using the Finite Element Method. Applied Sciences 2021; response in pulsed eddy current testing, NDT & E
11(2): 696. International 2009; 42(5): 376-383.
https://doi.org/10.3390/app11020696 https://doi.org/10.1016/j.ndteint.2009.01.005

[10] Barrarat F, Rayane K, Helifa B and Lefkaier IK. [21] Villone F. Simulation of thin cracks with finite resistivity in
Characterization of subsurface cracks in eddy current testing eddy current testing, IEEE Transactions on Magnetics 2000;
using machine learning methods. International Journal of 36(4): 1706-1709.
Numerical Modelling: Electronic Networks, Devices and https://doi.org/10.1109/20.877771
Fields 2021; e2876. [22] Santandrea L and Le Bihan Y. Using COMSOL-multiphysics
https://doi.org/10.1002/jnm.2876 in an eddy current non-destructive testing context. In
[11] Li S and Cui X. An edge-based smoothed finite element Proceedings of the COMSOL Conference 2010.
method for nonlinear magnetostatic and eddy current
analysis. Applied Mathematical Modelling 2018; 62: 287-302.
https://doi.org/10.1016/j.apm.2018.06.003

Received on 18-03-2021 Accepted on 12-05-2021 Published on 16-07-2021

DOI: https://doi.org/10.15377/2410-4701.2021.08.2

© 2021 Rocha et al.; Zeal Press.

This is an open access article licensed under the terms of the Creative Commons Attribution Non-Commercial License
(http://creativecommons.org/licenses/by-nc/3.0/) which permits unrestricted, non-commercial use, distribution and reproduction in any medium,
provided the work is properly cited.