applied

sciences
Article
Measurement System of Metal Magnetic Memory
Method Signals around Rectangular Defects of a
Ferromagnetic Pipe
J. Jesús Villegas-Saucillo 1 , José Javier Díaz-Carmona 1 , Carlos A. Cerón-Álvarez 2 ,
Raúl Juárez-Aguirre 2 , Saúl M. Domínguez-Nicolás 2 , Francisco López-Huerta 3 and
Agustín L. Herrera-May 2,4, *
1 Instituto Tecnológico de Celaya, Antonio García Cubas Pte. 600, Celaya, Guanajuato 38020, México
2 Micro and Nanotechnology Research Center, Universidad Veracruzana, Calzada Ruiz Cortines 455,
Boca del Río, Veracruz 94294, México
3 Facultad de Ingeniería Eléctrica y Electrónica, Universidad Veracruzana, Calzada Ruiz Cortines 455,
Boca del Río, Veracruz 94294, México
4 Maestría en Ingeniería Aplicada, Facultad de Ingeniería de la Construcción y el Hábitat, Universidad
Veracruzana, Calzada Ruíz Cortines 455, Boca del Río, Veracruz 94294, México
* Correspondence: leherrera@uv.mx; Tel.: +52-229-775-2000




Received: 29 December 2018; Accepted: 25 June 2019; Published: 2 July 2019


Abstract: Oil and gas pipeline networks require the periodic inspection of their infrastructure,
which can cause gas and oil leakage with several damages to the environment and human health.
For this, non-destructive testing (NDT) techniques of low-cost and easy implementation are required.
An option is the metal magnetic memory (MMM) method, which could be used for real-time
monitoring defects of ferromagnetic structures based on the analysis of self-magnetic leakage fields
distribution around each defect. This method only requires magnetic sensors with high resolution and
a data processing system. We present a measurement system of tangential and normal MMM signals
of three rectangular defects of an ASTM A-36 steel pipe. This system is formed by a magnetoresistive
sensor, an Arduino nano and a virtual instrumentation. The measured magnetic signals have
non-uniform distributions around the rectangular defects, which have small differences with respect
to the results obtained of a 2D magnetic dipole model. The size of each rectangular defect is related to
the amplitude and shape of its tangential and normal MMM signals. The proposed system could
be used for real-time monitoring of the size and location of rectangular defects of ferromagnetic
pipes. This system does not require expensive equipment, operators with high skill level or a special
treatment of the ferromagnetic samples.

Keywords: rectangular defect; magneto-mechanical memory method; metal magnetic memory

method; magnetic sensor; non-destructive testing; oil and gas pipeline

1. Introduction
The infrastructure for a gas and oil pipeline network must be constantly monitored using
non-destructive testing (NDT) techniques to assure its safe operation. Pipeline faults can cause oil and
gas leakage with severe damages in the environment and human health [1,2]. The pipeline network
safety is a priority in the gas and oil industries, which require low-cost NDT techniques for real-time
monitoring of defects and flaws in the pipeline. Thus, the pipeline network requires periodic inspections
and suitable maintenances that avoid future accidents. Generally, in these inspections conventional
NDT techniques are used, such as infrared thermography testing, radiographic testing, visual testing,
ultrasonic testing, acoustic emission, liquid penetrant testing and Eddy current testing [3–6]. Liquid

Appl. Sci. 2019, 9, 2695; doi:10.3390/app9132695 www.mdpi.com/journal/applsci

Appl. Sci. 2019, 9, 2695 2 of 10

penetrant testing is useful for inspection of surface defects in ferromagnetic samples, although these
defects should be contaminants free [7]. On the other hand, ultrasonic testing has high sensitivity
for detection of small flaws, but it needs reference standards for calibrating the equipment as well as
operators with extensive training and experience [8,9]. Another NDT technique is the acoustic emission,
which is used for monitoring discontinuities on a large area using multiple sensors whose signals
must be interpreted by inspectors with extensive technical knowledge [10]. Infrared thermography
is another NDT technique that has high sensitivity to temperature variations and it needs expensive
equipment [5]. Also, radiographic testing has a high sensitivity for detecting flaws. Nevertheless,
this technique presents limitations such as radiation hazard and highly expensive equipment, and it
requires operators with high skill level [10–12]. Finally, Eddy current testing is employed for inspection
of defects based on electromagnetic induction [12,13]. However, this technique needs an external
magnetic field source to induce Eddy currents.
Recently, several studies about the defects inspection of ferromagnetic structures have been
reported using the metal magnetic memory (MMM) method [14–21]. The MMM method is a passive
magnetic testing technique, which is different to the conventional magnetic flux leakage method. This
MMM method may be used for monitoring defects or degrees of stress concentration in ferromagnetic
materials measuring their spontaneous surface micro-magnetic signals [20,21]. These signals can
be affected by several factors such as different manufacture processes applied to the ferromagnetic
materials, considering types of heat treatment, welding, casting and forging [22]. This method could
be employed for monitoring defects of ferromagnetic structures based on the analysis of self-magnetic
leakage fields distribution around each defect. The MMM method does not require additional
equipment, external magnetic field sources and special treatment on the structures surface. Thus,
this magnetic method has a simple operation principle and it could be used for defects inspection
of pipeline network using magnetic sensors with high resolution. However, more studies related
to this method are necessary to predict the relation between the defect size (depth and width) and
variations of the amplitude and shape of the MMM signals. In order to study this relation, we measure
the tangential and normal MMM response around three different rectangular defects on the external
surface of a ferromagnetic pipe (ASTM A-36 steel) using a simple measurement system. This system
uses a low-cost magnetoresistive sensor, an Arduino nano and a virtual instrumentation. In addition,
it can be used without special equipment, additional treatment on the surface of ferromagnetic pipes
or operators with high skill level. The proposed system could be employed for real-time monitoring of
the size and location of rectangular defects on the external surface of ferromagnetic pipes.
This paper consists of the following sections. Section 2 includes the analytical response of tangential
and normal MMM signals of rectangular defects. Section 3 presents the measurement system and
experimental results of the tangential and normal MMM signals of three different rectangular detects
of a ferromagnetic pipe. Section 4 summarizes the conclusions and outlines future research works.

2. Analytical Model of the MMM Signals

This section describes the MMM signals of a rectangular defect of a ferromagnetic material
considering the 2D magnetic dipole model. Figure 1 shows a rectangular defect that contains 2a width,
b depth and a magnetic charge density ± Sm around the defect edges. A 2D magnetic dipole model is
employed to predict MMM signals around the rectangular defect. This model can determine the MMM
signals of defects with simple geometries [23,24]. Forster [25–27] developed a theoretical modeling and
experimentation of the MMM signals of discontinuities on test specimens. In addition, Zatsepin and
Schcerbinin [28,29] proposed analytical models for the tangential and normal MMM signals around a
2D rectangular defect, considering a constant pole density on its surfaces. These models were modified
by Edwards and Palmer [30] to evaluate the magnetic field (Sm ) inside the gap of a rectangular defect.
     
Ho n    1  1 
b  y  a   tan 1  b  y  a 
Hy    
tan (1)
2 tan 1 n  n       
  y  a  z  z  b    y  a   z  z  b  
2 2
 
Appl. Sci. 2018, 8, x FOR PEER REVIEW 3 of 10

  2 2 
  y  a    z  b    y  a   z  
Appl. Sci. 2019, 9, 2695 2 2 3 of 10
H0 n    1 
Hoz n    11  1 ln  b  y  a    b  y2 a  (2)
H 
Hy  4 tan
nn 
n 
 n 
tan
   (Hz)
  


 MMM
2
 a  signals
2   tan 1 
  y 2 a    z    y  a    z  b2 
 2 
  (1)
    y  a defect
z zon a ferromagnetic
1
Thus, the tangential2 tan
(Hy) and normal 

y   z z  for
b  a
 rectangular   b 
    
pipe can be calculated by [27,28]:
where n = b/a, H0 is the uniform magnetic  field along the y-axis (i.e.,axial axis) of the ferromagnetic

πn µ  2 b(2y−a)

 2 
H ( − 1 ) b (
y +
a   z 
a )  b− defect,
 
 2
y rectangular y  arespectively,
z

 
pipe, b and o
H y 2a
= are the depth and width of the y and
 −1 
  −1
 z are the
 
tan    tan
    (1)
    
−1 H
µ)
n
n(n +respectively,
2 tanH axes,
1    2   2   
z 
0
horizontal and vertical and
( yln+ a)µ is+ the
z(z relative
+ b) permeability ( y − a)of+the z(zmaterial.
+ b) 
   
(2)

The magnetic field density n  n  the
4 tan 1inside  defect
 h y (S a m) was
2

2
z  determined
  a   by
y ih
2
b  i  and Palmer [30]
2
z Edwards

  2 
 2 2 2 
as: H0 πn(µ − 1)  ( y + a) + (z + b) ( y − a) + z 


Hz = ln (2)

 1
−1 n(n + µ)  ( y + a)2 + (z)2 ( y − a)2 + (z + b)2 
h ih i
4 tan
where n = b/a, H0 is the uniform magnetic field

0 along
H 0 nthe  y-axis (i.e., axial axis) of the ferromagnetic

Sm  1rectangular defect, respectively, y and z are (3)
pipe, bn and
where = b/a,2aH0are theuniform
is the depth and width
magnetic of tan
field the
along n the  (i.e., axial axis) of the ferromagnetic
n  y-axis the
horizontal
pipe, b and 2aand vertical
are the depthaxes,andrespectively,
width of the and µ is the relative
rectangular defect, permeability
respectively, yofand thezmaterial.
are the horizontal
where The
µ magnetic
0 is the field
magnetic density
permeabilityinside of thefreedefect
space. (S m) was determined by Edwards and Palmer [30]
and vertical axes, respectively, and µ is the relative permeability of the material.
as:

 0  H 0  n    1
Sm  (3)
tan 1 n  n   

where µ0 is the magnetic permeability of free space.

Figure1.1. Schematic
Figure Schematic of
of aa rectangular
rectangular defect
defectwith
withmagnetic
magneticcharge density±± SSmm.
chargedensity

The magnetic
Equations field consider
(1)–(3) density inside
a smallthe defect
size for (Smeach
) wasrectangular
determineddefect
by Edwards and Palmer
in comparison [30]the
with as:
diameter of the pipe (i.e., the curvature of the pipe wall is neglected). Figure 2 depicts the behaviour
µ0 µH0 πn(µ − 1)
of the tangential (Hz) and normal (Hy) SmMMM
= signals of a rectangular defect on a ferromagnetic (3)
sample using Equations (1) and (2). For this case, tan−1wenuse µ) following data: µ = 180, H0 = 0.7958 A/m,
(n +the
a = 2.25µmm
where andFigure
b = 3.51.mm.
Schematic
For Hz ofcomponent,
a rectangular the defect with magnetic
maximum valuecharge
occursdensity ± Sm. center. With
at the defect
0 is the magnetic permeability of free space.
respect to the Hy component,
Equations a it presents fora each
zerorectangular
value at the defect center andwithtwo the
peaks with
Equations(1)–(3)
(1)–(3)consider
consider small
a smallsize size defect
for each rectangular in comparison
defect in comparison diameter
with the
maximum
of and minimum
the pipeof(i.e., the curvaturevalues on both
of the pipe sides of
wallpipe the rectangular
is neglected). defect.
Figure 2Figure
depicts the behaviour of the
diameter the pipe (i.e., the curvature of the wall is neglected). 2 depicts the behaviour
tangential (Hz) and normal (Hy) MMM signals of a rectangular defect on
of the tangential (Hz) and normal (Hy) MMM signals of a rectangular defect on a ferromagnetica ferromagnetic sample using
Equations (1) and (2). For this case, we use the following data:
sample using Equations (1) and (2). For this case, we use the following data: µ = 180, H 0 = 0.7958 A/m, a = 2.25 mm
µ = 180, H0 = 0.7958 A/m,
and b = mm
a = 2.25 3.5 mm.
and For Hzmm.
b = 3.5 component, the maximum
For Hz component, the value
maximumoccurs at the
value defect
occurs atcenter. With
the defect respect
center. to
With
the Hy component, it presents a zero value at the defect center and two
respect to the Hy component, it presents a zero value at the defect center and two peaks with peaks with maximum and
minimum
maximumvalues on both sides
and minimum valuesofontheboth
rectangular
sides of defect.
the rectangular defect.

Figure 2. Variation of the normal (Hy) and tangential (Hz) metal magnetic memory (MMM) signals
for a rectangular defect of a ferromagnetic sample with magnetic charge density ± Sm.

Figure2.2. Variation
Figure Variationofofthe
thenormal
normal(Hy)
(Hy) and
and tangential
tangential (Hz)
(Hz) metal
metal magnetic
magnetic memory
memory (MMM)
(MMM) signals
signals for
for a rectangular defect of a ferromagnetic sample with magnetic charge density
a rectangular defect of a ferromagnetic sample with magnetic charge density ± Sm . ± Sm .
Appl. Sci.
Appl.Sci.
Appl. Sci.2018,
2018, 9,
8,x2695
2019,8, xFOR
FORPEER
PEERREVIEW
REVIEW 44 of
of 10
10

3. Results
3. Results and and Discussion
Discussion
3. Results and Discussion
In this
In this section,
section, thethe MMM
MMM signals
signals ofof three
three different
different rectangular
rectangular defects
defects of of an
an ASTM
ASTM A-36 A-36 steel
steel
In this section, the MMM signals of three different rectangular defects of an ASTM A-36 steel pipe
pipe are
pipe are presented.
presented. This This pipe
pipe hashas the
the following
following dimensions:
dimensions: 247.0
247.0 mm mm length,
length, 48.06
48.06 mmmm outerouter
are presented. This pipe has the following dimensions: 247.0 mm length, 48.06 mm outer diameter and
diameter and
diameter and 3.65
3.65mm mmthickness.
thickness.
3.65 mm thickness.
Figure 33 shows
Figure shows the the measurement
measurement system system of of MMM
MMM signals
signals around
around three
three different
different rectangular
rectangular
Figure 3 shows the measurement system of MMM signals around three different rectangular
defects of
defects of anan ASTM
ASTM A-36 A-36 steel
steel pipe.
pipe. This
This system
system includes
includes aarotatory
rotatory mechanism,
mechanism, aa magnetoresistive
magnetoresistive
defects of an ASTM A-36 steel pipe. This system includes a rotatory mechanism, a magnetoresistive
sensor (MAG3110),
sensor (MAG3110), an an Arduino
Arduino nano nano (ATmega328)
(ATmega328) and and aa virtual
virtual instrumentation
instrumentation developeddeveloped in in
sensor (MAG3110), an Arduino nano (ATmega328) and a virtual instrumentation developed in Delphi
Delphi Borland
Delphi Borland code code [31].
[31]. The
The rotatory
rotatory mechanism
mechanism uses uses aa motor
motor (Bühler
(Bühler GmbH,GmbH, Braunschweig,
Braunschweig,
Borland code [31]. The rotatory mechanism uses a motor (Bühler GmbH, Braunschweig, Germany)
Germany) that
Germany) that is
is supplied
supplied with with 1.5
1.5 VVdcdc to
to generate
generate aarotational
rotationalmotion
motion of of 22 rpm.
rpm. TheThemeasurement
measurement
that is
system is supplied
is fabricated with
fabricated of 1.5 V dc
ofnon-magnetic to generate
non-magnetic materials a rotational
materialssuchsuch as motion
asnilamide
nilamideandof 2 rpm. The
and aluminium, measurement
aluminium, which whichdo donot
not system
affect
system affect
is fabricated of non-magnetic materials such as nilamide and aluminium, which do not affect the
theself-magnetic
the self-magnetic leakage leakagefluxfluxofofthe
theferromagnetic
ferromagneticpipe. pipe.The
Thepipe
pipesample
sampleis iscollocated
collocatedin inthe
therotatory
rotatory
self-magnetic leakage flux of the ferromagnetic pipe. The pipe sample is collocated in the rotatory
supports, keeping
supports, keeping aa constant
constant distance
distance of of 22 mm
mm between
between thethe external
external surface
surface of of the
the pipe
pipe andand thethe
supports, keeping a constant distance of 2 mm between the external surface of the pipe and the
magnetoresistivesensor.
magnetoresistive sensor.TheThepipe
pipehashasthree
threerectangular
rectangulardefects
defectswith
withdifferent
differentdepthdepthand andwidth
widthalong
along
magnetoresistive
its length.
length. These sensor. The
These defects
defects pipe
(see Figure
Figure has4)three
haverectangular
the following
followingdefects with different
dimensions: S1 (4.36 depth
(4.36 mmand width
width andalong
3.65
its (see 4) have the dimensions: S1 mm width and 3.65
its length.
mm depth),
depth), S2These defects
S2 (3.80
(3.80mm (see
mm width Figure
width and 4)
and 2.0 have
2.0mm the
mmdepth) following
depth) and dimensions:
and S3
S3 (2.80
(2.80mm S1
mm width (4.36
width and mm
and 0.50 width
0.50mm and
mm depth). 3.65
depth). Themm
The
mm
depth), S2 (3.80 mm width and 2.0 mm depth) and S3 (2.80 mm width and 0.50 mm depth). The
tangential and
tangential and normal
normal MMM MMM signalssignals around
around the the three
three defects
defects are are measured
measured using using the the
tangential and normal MMM signals around the three defects are measured using the magnetoresistive
magnetoresistive sensor. The sensor has a resolution of 30 nT, a sensibility
magnetoresistive sensor. The sensor has a resolution of 30 nT, a sensibility of 100 nT, 2and it uses an of 100 nT, and it uses an
sensor. The sensor has a resolution of 30 nT, a sensibility of 100 nT, and it uses an I C interface to
II C
22 C interface
interface to to communicate
communicate with with Arduino
Arduino nano nano (ATmega328).
(ATmega328). The The measured
measured MMM MMM signals signals of of the
the
communicate
three defects
defects are withprocessed
are Arduino nanousing(ATmega328). The measured MMM is signals of the inthree defects are
three processed using aa virtual
virtual instrumentation,
instrumentation, which
which is developed
developed in Delphi
Delphi Borland
Borland
processed
code[31].[31]. Forusing
Forthe a virtual
theanalytical instrumentation,
analyticalmodel,
model,values
valuesof which
ofHH00for is developed
for each
eachdefect
defectare in Delphi
aredetermined Borland
determinedto code
toestimate [31].
estimatethe For
thebest the
bestfit
fit
code
analytical
of Equations model,
Equations (1)–(3) values
(1)–(3) for of
for the H 0 for each
thecorresponding defect are
corresponding experimental determined
experimental data. to estimate
data. For
For these the
thesecases,best
cases, we fit of Equations
we assumed (1)–(3)
assumed aa relative
relative
of
for the corresponding
permeability 180. experimental data. For these cases, we assumed a relative permeability of 180.
of 180.
permeability of

Figure 3.
Figure
Figure 3.Measurement
3. Measurement system
Measurement systemof
system ofthe
of thetangential
the tangentialand
tangential and normal
and normalMMM
normal MMM signals
MMM signals around
signals around three
around threerectangular
three rectangular
rectangular
defects
defects of
of the
the ASTM
ASTM A-36
A-36 steel
steel pipe.
pipe.
defects of the ASTM A-36 steel pipe.

Figure 4.Dimensions
Figure Dimensions ofthree
three rectangulardefects
defects ofthe
the ASTMA-36
A-36 steelpipe.
pipe.
Figure 4.
4. Dimensions of
of three rectangular
rectangular defects of
of the ASTM
ASTM A-36 steel
steel pipe.

Figure555depicts
Figure
Figure depictsthe
depicts thetangential
the tangentialMMM
tangential MMMsignal
MMM signalof
signal ofthe
of thefirst
the firstrectangular
first rectangulardefect
rectangular defect(S1),
defect (S1),which
(S1), whichis
which isobtained
is obtained
obtained
using
using the analytical model and proposed measurement system. The measured
using the analytical model and proposed measurement system. The measured tangential MMMMMM
the analytical model and proposed measurement system. The measured tangential
tangential MMM
signal
signal exhibits
signal exhibits aa peak
peak (173.2
(173.2 µµT)
T) at
at the
the defect
defect center
center (y
(y == 00 mm).
mm). This
This peak
peak corresponds
corresponds with
with the
the
Appl. Sci. 2019, 9, 2695 5 of 10

Appl. Sci. 2018, 8, x FOR PEER REVIEW 5 of 10

Appl. Sci. 2018, 8, x FOR PEER REVIEW 5 of 10
exhibits a peak (173.2 µT) at the defect center (y = 0 mm). This peak corresponds with the maximum
maximum
depth
maximum (3.65 depth
mm) of
depth (3.65
the mm)
(3.65 mm) of
defect the
ofS1. defect
theThis peak
defect S1. This
This peak
S1.width width
is related
peak width tois
istherelated
defectto
related the
the defect
towidth (4.36 width
defect mm). The
width (4.36 mm).
(4.36width
mm).
The
of
The width
thewidth of
measured the
of the measured
tangential
measured MMMtangential MMM
signal MMM
tangential signal
is less than
signal is less
theissignal than
of the
less than the
the signal
analytical of the
signal ofmodel. analytical model.
This difference
the analytical model.
This
can difference
Thisoccur because
difference can
cantheoccur because
2D magnetic
occur becausedipolethe
the 2D magnetic
2Dmethod
magnetic dipole
dipole amethod
considers constant
method considers
magneticaapermeability
considers constant
constant magnetic
of the
magnetic
permeability
ferromagnetic of the
sample. ferromagnetic
In the positions sample.
from In
−5.0 the
mm positions
to 5.0 from
mm, the
permeability of the ferromagnetic sample. In the positions from −5.0 mm to 5.0 mm, the analytical−5.0 mm
analytical to 5.0 mm,
tangential the
MMM analytical
signal
tangential
has
tangential MMM
MMM signal
a high absolute has
has aa(18.52
difference
signal high
high µT absolute
and 47.18
absolute difference
µT) with
difference (18.52
respect
(18.52 µµTT toand
andthe47.18 µT)
µT) with
measured
47.18 MMM
with respect
signal.
respect to the
to On
the
measured
the other MMM
hand, thesignal.
normal On the
MMM other hand,
signal of the
defectnormal
S1 hasMMMtwo signal
peaks
measured MMM signal. On the other hand, the normal MMM signal of defect S1 has two peaks (112.2 of
(112.2defect
µT S1
and has two
−143.6 peaks
µT) (112.2
close to
µ T
both and
µ T and −143.6
edges of the
−143.6 µ T) close
defect
µ T) to both
closeS1to(see edges
bothFigure
edges6).of the
of These defect
the defect S1
two peaks (see Figure
S1 (seerepresent
Figure 6). 6). These
a change two
These two peaks
in the represent
polarity
peaks of theaa
represent
change
normal
changeMMM in the polarity
signal due
in the polarity of the
of to normal
thedifferent MMM
normal polarities
MMM signalsignal due
of the
due to different
magnetic
to differentchargespolarities
on theof
polarities of the
defect magnetic charges
edges, incharges
the magnetic which
on
the the defect
normal MMMedges, in
signal which
reachedthe anormal
value MMM
zero at signal
the defectreached
center a(yvalue
=
on the defect edges, in which the normal MMM signal reached a value zero at the defect centery-axis 0 zero
mm). at
Thethe defect
distance center
along (y
(y == 00
mm).
between
mm). TheThe distance
the two peaks
distance along
along y-axis
values between
y-axisisbetween
related to thethe
the two
two peaks
defect
peaks values
width
values is
(4.36 related
mm). The
is related to the defect
maximum
to the width
defect width (4.36
and minimum mm).
(4.36 mm).
The
The maximum and minimum magnitudes of the normal MMM signal are achieved along
maximum
magnitudes of theand minimum
normal MMM magnitudes
signal are of
achieved the normal
along the MMM
positions signal
close are
to achieved
±2.0 mm. The normal
along the
the
positions
MMM close
signal to ±2.0
registers mm.
a The
magnetic normal
flux MMM
offset signal
about registers
±80.0 µT. Fora magnetic
this
positions close to ±2.0 mm. The normal MMM signal registers a magnetic flux offset about ±80.0 µT. case, flux
the offset
normal about
MMM ±80.0 µT.
signal
For
For this
this case,
obtained case, the
through normal
normal MMM
the analytical MMMmodel signal
has aobtained
signal through
through analytical
good approximation
obtained in comparison
analytical model
model has with
has aa good
the
good approximation
measured MMM
approximation
in
in comparison with the measured MMM signal. In the positions from −5.0 mm to 5.0 mm, these
comparison
signal. In the with
positions the measured
from −5.0 mm MMMto 5.0 signal.
mm, In
these the positions
signals have from
an −5.0
absolute mm to 5.0
difference mm,
of 8.22 µT
these
signals
and 24.52
signals have
have an absolute
respectively.
µT, an difference of 8.22 µ T and 24.52 µ
absolute difference of 8.22 µ T and 24.52 µ T, respectively. T, respectively.

Figure
Figure 5. Tangential
5. Tangential
Figure5. MMM
TangentialMMM signal
MMMsignal of
signalof the
ofthe rectangular
therectangular defect
rectangulardefect S1
defectS1 of
S1of the
ofthe ASTM
theASTM A-36
ASTMA-36 steel
A-36steel pipe.
steelpipe.
pipe.

Figure
Figure 6. Normal
6.Normal
Figure6. MMM
NormalMMM signal
MMMsignal of
signalof the
ofthe rectangular
therectangular defect
rectangulardefect S1
defectS1 of
S1of the
ofthe ASTM
theASTM A-36
ASTMA-36 steel
A-36steel pipe.
steelpipe.
pipe.

Figure
Figure 7 shows the measured tangential MMM signal of the second rectangular defect (S2). This
Figure 77 shows
shows the
the measured
measured tangential
tangential MMMMMM signal
signal ofof the
the second
second rectangular
rectangular defect
defect (S2).
(S2). This
This
MMM
MMM signal has a maximum value (137.1 µT) at the defect center. This value is 36.1 µT less than the
MMM signal
signal has
has aa maximum
maximum value value (137.1
(137.1 µµT)
T) at
at the
the defect
defect center.
center. This
This value
value is
is 36.1
36.1 µµTT less
less than
than the
the
maximum
maximum field
field of
of the
the defect
defect S1,
S1, which
which represents
represents aareduction
reduction of
of20.8%.
20.8%. This
This decrease
decrease is
is related
related to
to the
the
maximum field of the defect S1, which represents a reduction of 20.8%. This decrease is related to the
less
less depth (2.0 mm) of the defect S2 in comparison to the depth (3.65 mm) of defect S1. The width of
less depth
depth (2.0
(2.0 mm)
mm) of of the
the defect
defect S2S2 in
in comparison
comparison to to the
the depth
depth (3.65
(3.65 mm)
mm) of of defect
defect S1.
S1. The
The width
width ofof
the
the tangential MMM signal is related to the defect width (3.80 mm). The measured tangential MMM
tangential MMM signal is related to the defect width (3.80 mm). The measured tangential MMM
response
response hashas aa magnetic
magnetic fluxflux offset
offset close
close to
to 35
35 µT
µT atat the
the right
right edge
edge ofof the
the defect
defect S2.
S2. The
The measured
measured
Appl. Sci. 2019, 9, 2695 6 of 10

the tangential
Appl.
MMMPEER signal is related to the defect width (3.80 mm). The measured tangential MMM
Appl. Sci.
Sci. 2018,
2018, 8,
8, xx FOR
FOR PEER REVIEW
REVIEW 66 of
of 10
10
response has a magnetic flux offset close to 35 µT at the right edge of the defect S2. The measured
tangential
tangentialMMM MMM signal signalhashasaa behaviour
behavioursimilar
similar totothat
that obtained
obtainedby by the
the analytical
analyticalmodel.
model. On On the
the other
other
hand, the measured normal MMM signal (see Figure 8) of the defect
hand, the measured normal MMM signal (see Figure 8) of the defect S2 has maximum and minimumS2 has maximum and minimum
values
values (25.8
(25.8µT µTand and−27.2
−27.2µT)µT)around
aroundboth
bothdefect
defectedges.
edges.TheThepolarity
polarityofof
this MMM
this MMM signal
signalis altered
is altered at
the defect
at the center
defect S2 (yS2
center = 0(ymm).
= 0 In addition,
mm). the normal
In addition, the MMM
normalsignal
MMM obtained
signal through
obtainedthe analytical
through the
model agree well with the experimental data in the positions between
analytical model agree well with the experimental data in the positions between −2.0 and 2.0 mm.−2.0 and 2.0 mm. However,
the analytical
However, the response
analyticalof the tangential
response MMM signal
of the tangential MMM hassignal
a high
hasabsolute difference
a high absolute (18.67 µT(18.67
difference and
20.17 µT) with respect to the measured MMM signal in the positions from −5.0
µ T and 20.17 µ T) with respect to the measured MMM signal in the positions from −5.0 m to 5.0 mm. m to 5.0 mm. The offset
of
Theboth tangential
offset of bothand normal MMM
tangential signalsMMM
and normal is duesignals
to the magnetic
is due toflux
the along the pipe
magnetic flux surface
along the close
pipe to
the defects.
surface closeIntoaddition,
the defects.the results of both
In addition, theMMM
resultssignals
of both determined
MMM signals with determined
the 2D magnetic with dipole
the 2D
method
magnetic dipole method have a discrepancy with respect to measured MMM responses. This is that
have a discrepancy with respect to measured MMM responses. This is due to the fact due
the analytical
to the fact thatmodel considers
the analytical a constant
model considersmagnetic permeability
a constant magnetic of the ferromagnetic
permeability material. In
of the ferromagnetic
addition,
material. the defect S2 has
In addition, the edges
defectwith small
S2 has inclinations
edges with small thatinclinations
can affect the measured
that can affect response of both
the measured
MMM signals.
response of both MMM signals.

Figure 7.
Figure 7.
Figure Tangential
7. Tangential MMM
TangentialMMM signal
MMMsignal of
signalof the
ofthe rectangular
therectangular defect
rectangulardefect S2
defectS2 of
S2of the
ofthe ASTM
theASTM A-36
ASTMA-36 steel
A-36steel pipe.
steelpipe.
pipe.

Figure 8.
8. Normal
Figure8.
Figure Normal MMM
MMM signal
signal of
of the rectangular
the rectangular defect
rectangular defect S2
defectS2 of
S2of the
ofthe ASTM
theASTM A-36
ASTMA-36 steel
A-36steel pipe.
steelpipe.
pipe.

For
For the
the third
third rectangular
rectangular defect
defect (S3),
(S3), the
the measured
measured tangential
tangential MMM
MMM signal
signal has
has aa maximum
maximum valuevalue
(53.6 µ T)atatthe
(53.6µT) thedefect
defectcenter
center(see
(seeFigure
Figure9).9).This
Thisvalue
valueis is
119.6 µTµT
119.6 less than
less that
than obtained
that obtainedof the defect
of the S1,
defect
which represents
S1, which a reduction
represents of 69.1%.
a reduction ThisThis
of 69.1%. largest variation
largest is related
variation to thetosmallest
is related depthdepth
the smallest (0.50 mm)
(0.50
of
mm)the defect
of the S3. In addition,
defect the width the
S3. In addition, of this tangential
width of thisMMM response
tangential MMM decreases in comparison
response decreases to in
the MMM signal
comparison of the
to the MMMdefects S1 and
signal of S2,
therespectively. For the
defects S1 and S2,defect S3, both tangential
respectively. MMMS3,
For the defect signals
both
tangential MMM signals have a similar behaviour between the positions from −1.4 mm to 1.4 mm.
The distance between these positions corresponds to the width of the defect S3. Nevertheless, the
absolute difference between both MMM signals increases for positions out of this width. For y = ±5.0
mm, the absolute difference between both tangential MMM signals are 8.12 µT and 20.68 µT,
reason for the differences between measured signals values and the analytical model-predicted ones.
For instance, the analytical model considers a constant magnetic permeability of the ferromagnetic
material. In addition, this analytical model is suitable for defects with a completely rectangular shape.
In the real defect, the shape is approximately rectangular with small inclinations along its edges. Also,
the analytical model does consider a uniform distribution of both MMM signals around each defect.
Appl. Sci. 2019, 9, 2695 7 of 10
The defect edges generate two peaks in the shape of the normal MMM signals, which indicate a
polarity shift of the magnetic flux at the defect center. The defect with the largest depth achieves the
have
highesta similar
value of behaviour between
the tangential MMMthe positions
response,from while −1.4
themm to 1.4
defect mm.
with The distance
smallest depth has between these
the lowest
positions corresponds
value of tangential MMM to the width
signal. of the
Thus, these defect S3. Nevertheless,
parameters the absolute
could be employed difference
to predict the size between
(depth
both MMM signals
and width) increases
and location for positions
of rectangular out ofofthis
defects ferromagnetic = ±5.0
width. For ypipes mm,the
using theMMM
absolute difference
method. The
between
measurement both tangential
system could MMM signals
be used forare 8.12 µTmonitoring
real-time and 20.68 µT, respectively.defects
of rectangular On the onother hand, the
ferromagnetic
normal MMMsystem
pipes. This response of the
does defect
not S3 has
require two peaks
special values (48.3
treatment of the and −59.5 µT) close
µT ferromagnetic sample,to their edges,
complex
achieving
equipmenta and polarity shift at
operators theextensive
with defect center (see Figure
training. 10). Thewith
In comparison defect S3 has
other NDT less width (2.80
techniques suchmm) as
in
thecomparison
Eddy current to testing
the defects S2 (3.80 mm)
and magnetic flux and S3 (4.36
leakage mm),
testing, therespectively,
MMM method which
doescauses
not needa reduction
to apply
in the width
external of thefield
magnetic tangential MMM signal. pipe.
on the ferromagnetic In addition,
It allowsthethemaximum
decrease of and minimum
additional magnitudes
equipment and
(48.3
special and −59.5on
µTtreatment of pipe
µT)the the measured
surface. On normal MMMhand,
the other signalthereach close
X-ray to theis edges
testing an NDT of the
thatdefect
uses
S3. For the equipment
expensive positions between −1.4 mm
and operators andextensive
with 1.4 mm, the analytical
technique normal MMM
experience. Another signal
NDT has a good
testing is
approximation
liquid penetrant with respect towhich
inspection, the measured MMM
has low cost, signal.
but it canHowever,
only detect thesurface
absolute difference
flaws. between
In addition, it
these signals
requires pipeincreases
surfaces for
freepositions out of theThe
of contaminants. range
MMM y = ±method
1.4 mm.could
For the
be integrated = ± 5.0
positions y with mm,
other NDTthe
absolute
testing such difference between
as electrical both normal
resistance MMM signals
measurement are 40.69
and acoustic µT andtechnique.
emission 8.12 µT, respectively.
For instance,The the
ideal approximation
comparison of analysis of the analytical
related model is resistance
to the electrical the main reason for theand
monitoring differences between measured
acoustic emission technique
signals
could be values
used and theassessment
in the analytical model-predicted
of damage and defects ones. For instance, the
in structures analytical
[32–35]. For model
structuresconsiders
formed a
constant magnetic permeability
with ferromagnetic materials, the of the ferromagnetic
results material. could
of both techniques In addition,
be usedthiswith
analytical
those model
obtained is
suitable
throughfor MMMdefects with afor
method completely
monitoring rectangular shape.
the location, In the real and
dimensions defect, the shape
shape of their is approximately
defects. Thus,
rectangular
these NDT tests with could
small inclinations
be complementaryalong itsforedges. Also, the
inspection analyticalinmodel
of damages does with
structures consider a uniform
ferromagnetic
distribution
materials. of both MMM signals around each defect.

Appl. Sci. 2018, 8, x FOR PEER REVIEW 8 of 10

Figure 9. Tangential MMM signal of
of the
the rectangular
rectangular defect
defect S3
S3 of
of the
the ASTM
ASTM A-36
A-36 steel
steel pipe.
pipe.

Figure 10.
Figure 10. Normal MMM signal of the rectangular defect S3 of the ASTM A-36
A-36 steel
steel pipe.
pipe.

4. Conclusions
A measurement system for real-time monitoring of the rectangular surface defects of
ferromagnetic pipes is reported. This system detects the variations of the tangential and normal
MMM signals around the defects, which could be used to determine their location and size. The
proposed system uses a low-cost magnetoresistive sensor, an Arduino nano and a virtual
instrumentation. This system could allow the inspection of rectangular defects and it does not need
Appl. Sci. 2019, 9, 2695 8 of 10

The defect edges generate two peaks in the shape of the normal MMM signals, which indicate
a polarity shift of the magnetic flux at the defect center. The defect with the largest depth achieves
the highest value of the tangential MMM response, while the defect with smallest depth has the
lowest value of tangential MMM signal. Thus, these parameters could be employed to predict the
size (depth and width) and location of rectangular defects of ferromagnetic pipes using the MMM
method. The measurement system could be used for real-time monitoring of rectangular defects on
ferromagnetic pipes. This system does not require special treatment of the ferromagnetic sample,
complex equipment and operators with extensive training. In comparison with other NDT techniques
such as the Eddy current testing and magnetic flux leakage testing, the MMM method does not need to
apply external magnetic field on the ferromagnetic pipe. It allows the decrease of additional equipment
and special treatment on the pipe surface. On the other hand, the X-ray testing is an NDT that uses
expensive equipment and operators with extensive technique experience. Another NDT testing is
liquid penetrant inspection, which has low cost, but it can only detect surface flaws. In addition, it
requires pipe surfaces free of contaminants. The MMM method could be integrated with other NDT
testing such as electrical resistance measurement and acoustic emission technique. For instance, the
comparison of analysis related to the electrical resistance monitoring and acoustic emission technique
could be used in the assessment of damage and defects in structures [32–35]. For structures formed
with ferromagnetic materials, the results of both techniques could be used with those obtained through
MMM method for monitoring the location, dimensions and shape of their defects. Thus, these NDT
tests could be complementary for inspection of damages in structures with ferromagnetic materials.

4. Conclusions
A measurement system for real-time monitoring of the rectangular surface defects of ferromagnetic
pipes is reported. This system detects the variations of the tangential and normal MMM signals around
the defects, which could be used to determine their location and size. The proposed system uses a
low-cost magnetoresistive sensor, an Arduino nano and a virtual instrumentation. This system could
allow the inspection of rectangular defects and it does not need expensive equipment and operators
with extensive experience. The size (depth and width) of the defects is related to the amplitude
and shape of these magnetic signals. The measured MMM signals have non-uniform distributions,
which registered differences with respect those calculated through a 2D magnetic dipole model. These
differences were caused by small inclinations along the edges of the three rectangular defects. In
addition, the analytical model considered a constant magnetic permeability of the ferromagnetic pipe.
Future studies will include the analysis of MMM signals around small defects with different
shapes (e.g., triangular, cylindrical and spherical shape) on the surface of ferromagnetic pipes. In
addition, we will study the relations between size and shape of the defects with respect to the shifts in
their measured MMM signals.

Author Contributions: J.J.V.-S. and J.J.D.-C. developed the measurement system; C.A.C.-Á. measured the MMM
signals; R.J.-A. and S.M.D.-N. designed the data processing system; F.L.-H. and A.L.H.-M. contributed to the
writing, review and editing of the paper.
Funding: This research was funded by project PRODEP “Estudio de Dispositivos Electrónicos y Electromecánicos
con Potencial Aplicación en Fisiología y Optoelectrónica” and PFCE 2018 “DES Técnica Veracruz 30MSU0940B-21”.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Tong, S.J.; Wu, Z.Z.; Wang, R.J.; Wu, H. Fire risk study of long-distance oil and gas pipeline based on QRA.
Procedia Eng. 2016, 135, 369–375. [CrossRef]
2. Iqbal, H.; Tesfamariam, S.; Haider, H.; Sadiq, R. Inspection and maintenance of oil & gas pipelines: A review
of policies. Struct. Infrastruct. Eng. 2016, 13, 1–23. [CrossRef]
3. Dehui, W.; Zhitian, L.; Xiaohong, W.; Lingxin, S. Composite magnetic flux leakage detection method for
pipelines using alternating magnetic field excitation. NDT E Int. 2017, 91, 148–155. [CrossRef]
Appl. Sci. 2019, 9, 2695 9 of 10

4. Gholizadeh, S. A review of non-destructive testing methods of composite materials. Procedia Struct. Integr.
2016, 1, 50–57. [CrossRef]
5. Hellier, C.J. Handbook of Nondestructive Evaluation, 2nd ed.; McGraw-Hill Education: New York, NY, USA, 2013.
6. Liu, B.; He, L.-Y.; Zhang, H.; Cao, Y.; Fernandes, H. The axial crack testing model for long distance oil-gas
pipeline based on magnetic flux leakage internal inspection method. Measurement 2017, 103, 275–282.
[CrossRef]
7. Lu, Q.Y.; Wong, C.H. Applications of non-destructive testing techniques for post-process control of additively
manufactured parts. Virtual Phys. Prototyp. 2017, 12, 301–321. [CrossRef]
8. Purna-Chandra-Rao, B. Non-destructive Testing and Damage Detection. In Aerospace Materials and Material
Technologies; Prasad, N., Wanhill, R., Eds.; Indian Institute of Metals Series; Springer: Singapore, 2017;
Volume 2, pp. 209–218, ISBN 978-981-10-2143-5.
9. Liu, J.; Xu, G.; Gu, X.; Zhou, G. Ultrasonic test of resistance spot welds based on wavelet package analysis.
Ultrasonics 2015, 56, 557–565. [CrossRef]
10. Nazarchuk, Z.; Skalskyi, V.; Serhiyenko, O. Acoustic Emission: Methodology and Application; Springer
International Publishing: Cham, Switzerland, 2017; ISBN 978-3-319-49348-0.
11. Yang, R.; He, Y.; Zhang, H. Progress and trends in nondestructive testing and evaluation for wind turbine
composite blade. Renew. Sustain. Energy Rev. 2016, 60, 1225–1250. [CrossRef]
12. Cheng, L.; Kogia, M.; Mohimi, A.; Kappatos, V.; Selcuk, C.; Gan, T.-H. Crack characterisation using
invariable feature extraction in stainless steel specimen used for absorber tubes of CSP applications via
EMAT. Renew. Energy 2017, 101, 771–781. [CrossRef]
13. García-Martín, J.; Gómez-Gil, J.; Vázquez-Sánchez, E. Non-Destructive Techniques Based on Eddy Current
Testing. Sensors 2011, 11, 2525–2565. [CrossRef]
14. Xu, K.; Qiu, X.; Tian, X. Theoretical investigation of metal magnetic memory testing technique for detection
of magnetic flux leakage signals from buried defect. Nondestruct. Test. Eval. 2018, 33, 45–55. [CrossRef]
15. Xu, K.; Qiu, X.; Tian, X. Investigation of metal magnetic memory signals of welding cracks. J. Nondestruct. Eval.
2017, 20. [CrossRef]
16. Li, Y.; Zeng, X.; Wei, L.; Wan, Q. Characterizations of damage induced magnetization for X80 pipeline steel
by metal magnetic memory testing. Int. J. Appl. Electromagn. Mech. 2017, 1, 23–35. [CrossRef]
17. Dubov, A.; Kolokolnikov, S. Assessment of the material state of oil and gas pipeline based on the metal
magnetic memory method. Weld. World 2012, 56, 11–19. [CrossRef]
18. Peng, Y.; Feng, P.S.; Chao, L. Numerical studies to signal characteristics with the metal magnetic memory-effect
based of cast iron. Appl. Mech. Mater. 2014, 633–634, 277–283. [CrossRef]
19. Zhang, Y.; Zhou, D.; Jiang, P.; Zhang, H. The state of the art surveys for application of metal magnetic
memory testing in remanufacturing. Adv. Mater. Res. 2011, 301–303, 366–372. [CrossRef]
20. Doubov, A.A. Development of a metal magnetic memory method. Chem. Pet. Eng. 2012, 47, 837–839.
[CrossRef]
21. Pengpeng, S.; Ke, J.; Xiaojing, Z. A magnetomechanical model for the magnetic memory method. Int. J.
Mech. Sci. 2017, 124–125, 229–241. [CrossRef]
22. Leng, J.; Xu, M.; Zhou, G.; Wu, Z. Effect of initial remanent states on the variation of magnetic memory
signals. NDT E Int. 2012, 52, 23–27. [CrossRef]
23. Shcherbinin, V.; Pashagin, A. Influence of the extension of a defect on the magnitude of its magnitude field.
Defektoskopiya 1972, 4, 74–82.
24. Shcherbinin, V.; Pashagin, A. On the volumen polarization of cracks. Defektoskopiya 1974, 4, 106–110.
25. Forster, F. Nondestructive inspection by the method of magnetic leakage fields. Defektoskopiya 1982, 11, 3–25.
26. Forster, F. On the way from the know-how to the know why in the magnetic leakage field method of
nondestructive testing. Mater. Eval. 1985, 43, 1154–1661.
27. Forster, F. New findings in the field of nondestructive magnetic leakage field inspection. NDT Int. 1986,
19, 3–14. [CrossRef]
28. Zatsepin, N.N.; Shcherbinin, V.E. Calculation of the magneto static field of surface defects. I. Field topography
of defect models. Defektoskopiya 1966, 5, 50–59.
29. Shcherbinin, V.E.; Zatsepin, N.N. Calculation of the magneto static field of surface defects. II. Experimental
verification of the principal theoretical relationships. II. Defektoskopiya 1966, 5, 59–65.
Appl. Sci. 2019, 9, 2695 10 of 10

30. Edwards, C.; Palmer, S.B. The magnetic leakage field of surface-breaking cracks. J. Phys. D Appl. Phys. 1986,
19, 657. [CrossRef]
31. Juárez-Aguirre, R.; Domínguez-Nicolás, S.M.; Manjarrez, E.; Tapia, J.A.; Figueras, E.; Vázquez-Leal, H.;
Aguilera-Cortés, L.A.; Herrera-May, A.L. Digital signal processing by virtual instrumentation of a MEMS
magnetic field sensor for biomedical applications. Sensors 2013, 13, 15068–15084. [CrossRef] [PubMed]
32. Lemaitre, J.; Dufailly, J. Damage measurements. Eng. Fract. Mech. 1987, 28, 643–661. [CrossRef]
33. Todoroki, A.; Ueda, M.; Hirano, Y. Strain and damage monitoring of CFRP laminates by means of electrical
resistance measurement. J. Solid Mech. Mater. Eng. 2007, 1, 947–977. [CrossRef]
34. Chen, B.; Liu, J. Damage in carbon fiber-reinforced concrete, monitored by both electrical resistance
measurement and acoustic emission analysis. Constr. Build. Mater. 2008, 22, 2196–2201. [CrossRef]
35. Niccolini, G.; Borla, O.; Accornero, F.; Lacidogna, G.; Carpinteri, A. Scaling in damage by electrical resistance
measurement: And application to the terracotta statues of the Sacred Mountain of Varallo Renaissance
Complex (Italy). Rend. Lincei 2015, 26, 203–209. [CrossRef]

© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).