Article
Weak Magnetic Internal Signal Characteristics of Pipe Welds
under Internal Pressure
Bin Liu, Yanduo Fu *, Luyao He, Hao Geng and Lijian Yang
Citation: Liu, B.; Fu, Y.; He, L.;
Geng, H.; Yang, L. Weak Magnetic
Internal Signal Characteristics of
Pipe Welds under Internal Pressure.
Sensors 2023, 23, 1147. https://
doi.org/10.3390/s23031147
Academic Editor: Songling Huang
Received: 25 December 2022
Revised: 14 January 2023
Accepted: 16 January 2023
Published: 19 January 2023
Copyright: © 2023 by the authors. Li-
censee MDPI, Basel, Switzerland.
This article is an open access article
distributed under the terms and con-
ditions of the Creative Commons At-
tribution (CC BY) license (https://cre-
ativecommons.org/licenses/by/4.0/).
Sensors 2023, 23, 1147. https://doi.org/10.3390/s23031147
School of Information Science and Engineering, Shenyang University of Technology, Shenyang 110870, China
* Correspondence: fyd2917@163.com; Tel.: +86-183-0417-1453
Abstract: Weak magnetic detection technology is an effective method to identify stress-induced
damage to ferromagnetic materials, and it especially possesses great application potential in long-
distance oil and gas pipeline weld crack detection. In the process of pipeline operation, due to in-
ternal pressure and external loads, local stress concentration may be generated, and partial stress
concentration may lead to local cracks and expansion of the pipe. In order to improve the accuracy
of magnetic signal analysis for ferromagnetic materials under internal pressure, the causes of mag-
netic signal generation at pipeline welds were analyzed from a microscopic perspective. The distri-
butions of magnetic signals at pipeline welds, weld cracks, and base metal cracks under different
internal pressures were numerically analyzed. The variation trends of magnetic signal characteris-
tics, such as peak values of axial and radial components, gradient K, maximum gradient Kmax, and
gradient energy factor S(K), were analyzed. In addition, experiments were carried out to verify the
numerical data. It was revealed that with the elevation of internal pressure, the peak values of the
axial and radial components, gradient K, maximum gradient Kmax, and gradient energy factor S(K)
linearly increased. However, the magnitude and average change of S(K) were larger, which can
more directly indicate variations of magnetic signals. The radial growth rate νy of S(K) was 3.24%
higher than the axial growth rate νx, demonstrating that the radial component of the magnetic signal
was more sensitive to variations of stress. This study provided a theoretical and experimental basis
for detection of stress-induced damage to long-distance oil and gas pipelines.
Keywords: pipeline weak magnetic stress; simulation analysis; energy factor; experimental test
1. Introduction
Pipelines are the most economic and efficient means of oil and natural gas transpor-
tation over long distances in different environments, owing to their advantages of contin-
uous transportation, low cost, high efficiency, climate resilience, and high reliability. Oil
and gas pipeline leakages may not only cause serious air, water, and soil pollution, but
also result in huge economic losses. According to the pipeline incident statistics published
in recent years, weld cracks are one of the main causes of oil and natural gas transporta-
tion failure over long distances [1,2], and it is urgent to realize the online detection of small
weld cracks (where the opening width is less than 1 mm). At present, in situ pipeline
detection technology is the most effective method for testing pipeline safety that is glob-
ally recognized by the pipeline industries [3–6]. Under the normal operation of the pipe-
line, the internal detection equipment is driven by oil and natural gas (speed, 1–5 m/s) in
order to realize the noncontact and dynamic detection of corrosion, cracks, metal loss,
pinholes, stress, and other defects in pipelines.
Traditional nondestructive detection techniques, such as eddy currents, magnetic
powder, ultrasound, radiation, etc., have shown some deficiencies in pipeline assessment
[7–10]. The internal detection of magnetic flux leakage has become the mainstream tech-
nology in the field of pipeline internal detection due to its advantages of being contactless,
www.mdpi.com/journal/sensors
Sensors 2023, 23, 1147 2 of 28

anti-interference, fast signal acquisition, etc., and the technology can identify a certain
opening width of crack defects (where the opening width is greater than 1 mm). However,
it is infeasible to effectively and accurately identify microcracks in welded pipes [11–15].
Therefore, research on the internal inspection of microcracks in long-distance welded oil
and gas pipelines has noticeably attracted scholars’ attention.
The weak magnetic detection method can realize contactless and dynamic online de-
tection of the stressed areas of ferromagnetic components using the magnetomechanical
effect of ferromagnetic materials. It has great application prospects in internal stress de-
tection of pipelines [16]. This technology has been widely used to detect pipeline defects
[17–21]. Under internal pressure and external loading of oil and gas pipelines, abnormal
stress distribution may occur at the microcracks of welded pipelines. Stress distribution is
closely associated with crack size. Using the weak magnetic detection method is advanta-
geous to identify abnormal changes of stress and microcracks in welded pipelines [22,23].
In this study, the causes of magnetic signals generated at welded pipelines were first
analyzed from a microscopic perspective. Second, a magnetomechanical model of the
composite stressed area was established based on the finite element method, the magne-
tomechanical effects at the microcracks of a welded pipeline were calculated quantita-
tively, and the characteristics of weak magnetic signals in pipeline welds, weld cracks,
and base metal cracks were studied. In addition, the variation patterns of characteristic
parameters of weak magnetic signals, such as gradient value K, maximum gradient Kmax,
and gradient energy factor S(K) in different stressed areas were analyzed and verified ex-
perimentally. This research facilitated detection of microcracks in long-distance oil and
gas pipelines.

2. Mechanism of Weak Magnetic Internal Detection in Long-Distance Oil and Gas

Pipelines
Based on the force magnetic coupling model and the idea of calculus and integration,
the spatial distribution model of magnetic signals in pipeline welds, weld cracks, and base
metal cracks was established.

2.1. Magnetic Detection Mechanism

Based on micromagnetism theory and Weis molecular field theory, the force-mag-
netic coupling model was derived from the ferromagnetism theory proposed by Jiles and
Atherton (J–A theory).
It is assumed that the angle between the magnetic moment of the atom μJ and the
external magnetic field H is θi, according to the statistics, and the partition function Z(H)
of the system is formulated as follows [24]:
N
JH cos j / KBT  4KBT  JH 
N

Z ( H ) =  sin id  = 
2 

 
0
d e
0   JH
sh
 KBT
 ,

(1)

where KB is the Boltzmann constant, T denotes the temperature, N represents the number
of atoms per unit volume, and θi is the angle in the range of 0~π.
According to:
 = − KBT ln Z , (2)
and

= −M (3)
H ,

the magnetization strength M can be expressed as:

 
M = KBT

ln Z ( H ) = NJ cth JH − KBT  , (4)
H   KBT JH 
Sensors 2023, 23, 1147 3 of 28

According to:
JH
= ,
(5)
K BT
L(α) can be expressed as:
1
L( ) = cth - (6)
 ,

where L(α) is the Langevin function.

According to the J–A theory, a modified Langevin function can be used to fit the
magnetization curve of the material:
  H + Man  a 
Man = Ms coth − , (7)
  a  H + Man 
The stress action is equivalent to an additional magnetic field and combines the ap-
proach principle to form the stress magnetization model.
The effective field He is expressed as:
3  d 
H e = H + M + H = H + M +   (8)
2 0  dM  ,
The relationship between the hysteresis expansion coefficient and the magnetization
strength of the material can be formulated as follows:
 = ( 1(0) +  1' (0) )M 2 + ( 2(0) +  2 ' (0) )M 4 , (9)
Then, Equation (8) is simplified as:
3
H e = H + M + ((2 1(0) + 2 1' (0) )M + (4 2(0) + 4 2 ' (0) )M 3 ) , (10)
2 0
It is supposed that the irreversible magnetization follows the law of proximity:
dMirr 1
= ( Man − Mirr ) , (11)
dW 
where ξ is a constant dependent on the energy per unit volume, and Mirr denotes the irre-
versible component of the magnetization. The derivative of magnetization strength to
stress is expressed as:
dM 1 dMan
= 2  (1 − c)( Man − Mirr ) + c
d  d ,
(12)

Taking Equation (11) into Equation (12), the relationship between stress σ and mag-
netization M is obtained as follows:
  a  3 
( M an − M ) + cM s csch 2  e  − 2   ( 1 + 11 ) M +2 ( 2 + 21 ) M 3  
H
dM E   a  H e   0 
= , (13)
d   H  a   3 
1 − cM s csch 2  e  − 2   (1 2 ) 
 + 6  M 2
+ 
  a  H e   0 
where α is the coupling parameter, H is the external magnetic field, σ denotes stress, λ is
the magnetostriction coefficient, M represents magnetization, and c is a reversible coeffi-
cient.
According to Equation (13), material parameters were taken as c = 0.25, µ0 = 4π × 10−7
NA−2, γ1 =7 × 10−18 A−2∙m2, γ11 = −1 × 10−25 A−2∙m2∙Pa−1, γ2 = −3.3 × 10−30 A−4∙m4, and γ12 = 2.1
× 10−38 A−4∙m4∙Pa−1 [25]. The magnetic curve output was plotted as shown in Figure 1. It
Sensors 2023, 23, 1147 4 of 28

was revealed that stress corresponded to magnetization one-to-one, and that magnetiza-
tion increased with the increase of stress (Figure 1).

Figure 1. Stress-magnetization curve.

According to magnetization and the relative permeability μr:

M = ( r − 1)  H , (14)
Taking Equation (14) into Equation (13), the relationship between relative permeabil-
ity μr and stress σ is formulated as follows [26]:
  H  a  3
 M an − ( r − 1)  H  + cM s csch 2  e  − 2   ( 1 +  11 )  ( r − 1)  H + 2 ( 2 + 2 21 ) (  r − 1)  H 3 
3

E 
d r   a  H e  0
= , (15)
d   a   3  
 1 + cM s 
csch 2  He 
− 
 a  H2    1
 
 + 6(
 2( r
 − 1)
2
 H 2
+   ) H
  e   0  
According to Equation (15), magnetic permeability, as an intermediate value, could
effectively establish the coupling relationship between weak magnetic signals and stress,
thus directly reflecting the influences of stress on ferromagnetic materials and providing
a numerical basis for the following magnetic finite element analysis.

2.2. Research on the Weak Magnetic Internal Detection Mechanism of the Pipeline
Weak magnetic internal detection was applied for nondestructive testing of ferro-
magnetic pipelines using the natural magnetization of the ferromagnetic field. The pipe
weld is different from the pipe base metal in metallophase, organization, and stress, thus,
the magnetic distribution was obviously different.
The ferromagnetism of the material was linearly correlated with the martensite con-
tent [27].
M sa = 99 fM − 2.9 , (16)
where Msa is the saturation magnetization (emu/g) and fM is the martensitic volume frac-
tion (%).
In the process of pipe welding, the weld metal goes through three stages: heating,
melting, and crystallization and solid phase transformation from the beginning of for-
mation to cooling to room temperature. During welding, the microstructure of the weld
material changes, and a large amount of martensite is generated. When the pipeline is
running, the loading also induces martensite transformation, and the greater the stress,
the more martensite transformation occurs [28–30].
According to Equation (16), the pipe is therefore ferromagnetic in the weld area, and
the weak magnetic internal detection technique can identify magnetic signals. The macro-
scopic manifestation is the sudden change of the self-leakage magnetic field around the
weld or weld crack, as shown in Figure 2.
Sensors 2023, 23, 1147 5 of 28

Figure 2. Schematic diagram of in-pipeline detection.

2.3. Characteristic Analysis of the Weak Magnetic Signal

Using ANSYS finite element simulation software, the mechanical analysis model of
the weld was established in Cartesian coordinate system (x, y, z). Solid70 was used to
solve the temperature field. When the grid was divided, the mesh at the weld of the pipe-
line was finer, 2 mm, and the mesh at both sides of the matrix was thicker, 5 mm. The
initial room temperature was 25 °C, and the welding temperature was 1500 °C. The pipe
material used in this study was X70, the pipe length was 1000 mm, the outer diameter was
1219 mm, the thickness was 16 mm, the weld width was 20 mm, the welding line speed V
was 1 mm/s, the welding voltage U was 36 V, the welding current I was 32 A, and the
welding thermal efficiency η was 0.75. The expression of heat generation intensity Q in
unit time was formulated as follows, and Q can be calculated by taking corresponding
parameters into account [31]:
I  U 
Q= (17)
V
The result of temperature field simulation can be read into the stress field simulation
to get the result of the stress distribution of the weld.
The von Mises stress σv can be expressed as [32]:

v =
1
2

( − z )2 + (z − r )2 + (r −  )2  (18)

where σz is axial stress (MPa), σθ denotes circumferential stress (MPa), and σr is radial
stress (MPa).

2.3.1. Weak Magnetic Signals in the Weld Pipe

The welding stress diagram is shown in Figure 3a, and the von Mises stress distribu-
tion diagram on the extracted y = 0 path is illustrated in Figure 3b. It was revealed that the
stress distribution at the pipe weld was uneven, and there was stress concentration that
gradually weakened both sides. The maximum stress was tensile stress, and it was located
near the weld, which could justify why the center of the weld would be prone to cracks
and other defects.
Sensors 2023, 23, 1147 6 of 28

3.5x108
Stress(Pa) 3.0x108
2.5x108
2.0x108
1.5x108
1.0x108
5.0x107

100

80

60

40

20
(a) (b)
Figure 3. Results of welding stress simulation. (a) Welding stress diagram; (b) von Mises stress dis-
tribution diagram on the extracted y = 0 path.

As found with Equation (15), the magnetic signal distribution results are shown in
Figure 4.
100
Radial component(A/M) 80
200
60
40
150
20
100 0
-20
Axial component(A/M) 50 -40
-60
0
100
100

80

60

40

20
80

60

40

20

2.018

(a) (b)
Figure 4. Magnetic signals of the pipe weld. (a) Axial component of weak magnetic signal; (b) radial
component of weak magnetic signal.

There was a peak in the axial component of the weak magnetic signal at the pipe
weld and a sinusoidal fluctuation in the radial component, which is the typical feature of
stress concentration, thus confirming the feasibility of detecting stress in the pipe weld
using the weak magnetic inner detection method.

2.3.2. Weak Magnetic Signals in Weld Cracks

According to the above-described weld simulation model, cracks were created with
a size of 2 mm × 0.95 mm × 1 mm (length × width × depth), and the magnetic simulation
results are shown in Figure 5.
Sensors 2023, 23, 1147 7 of 28

(a) (b)
Figure 5. Magnetic signal of weld cracks in the pipe. (a) Axial component of weak magnetic signal;
(b) radial component of weak magnetic signal.

There was a peak for the axial component of the weld crack and two sinusoidal fluc-
tuations in the radial component. The magnetic signal’s characteristics were compared
with the weld magnetic signal (Figure 4), and the weld crack could be identified.

2.3.3. Weak Magnetic Signal of Base Material Crack

A crack was created with the size of 2 mm × 0.95 mm × 1 mm (length × width × depth)
on the pipe base material (consistent with the weld base material parameters). The results
of magnetic simulation are shown in Figure 6.

50
Axial Radial 100

component(A/M) 40
component(A/M) 50
30
0
20

10 -50
160

140

120

100
140

120

100

80

60

40

20
80

60

40

20

1.516

2.018

(a) (b)
Figure 6. Crack magnetic signal of pipe base material. (a) Axial component of weak magnetic signal;
(b) radial component of weak magnetic signal.

It was revealed that there was a peak in the axial component of weak magnetic signal
at the crack of the pipe base material and a sinusoidal fluctuation in the radial component,
which is the typical feature of stress concentration, thus confirming the feasibility of de-
tecting cracks using the weak magnetic inner detection method.

3. Analysis of Influences of Internal Pressure on Weak Magnetic Signal Characteris-

tics
3.1. Influences of Internal Pressure on Weak Magnetic Signal Characteristics of Weld
According to the pipe weld model established in Section 2.3.1, an internal pressure
of 0–3 MPa was applied to the pipe, and simulation was conducted with an interval pres-
sure of 0.5 MPa. The results of testing the weak magnetic signal at the weld are shown in
Figure 7.
Sensors 2023, 23, 1147 8 of 28

400
0MPa
350 0.5MPa
1MPa
1.5MPa
300 2MPa
2.5MPa

Hx(A/M)
3MPa
250

200

150

100
0 20 40 60 80 100
Position(mm)
(a) (b)
Figure 7. Magnetic signal of weld under different internal pressure. (a) Axial component of weak
magnetic signal; (b) radial component of weak magnetic signal.

According to the above-mentioned results of magnetic simulation, the peak values of

the axial components and the peak-to-peak values of the radial components, gradient K,
maximum gradient Kmax, and gradient energy factor S(K) of magnetic signal characteris-
tic parameters were analyzed.

3.1.1. Peak Values in Axial and Radial Components at Weld Varies

As there were peak values in axial and radial components, the peak values of the
axial components and the peak-to-peak values of the radial components changed with the
internal pressure as follows (Figure 8):

380 Peak of the axial componment 160 Peak to peak of the

Linear fitting curve radial componment
360 Linear fitting curve
140
340
Hy（A/M）
Hx（A/M）

320 120
300
100
280

260 80

240
60
0.0 0.5 1.0 1.5 2.0 2.5 3.0 0.0 0.5 1.0 1.5 2.0 2.5 3.0
Stress（MPa） Stress（MPa）
(a) (b)
Figure 8. Peak values of weak magnetic signal at weld varies with internal pressures. (a) Peak values
of the axial component; (b) peak-to-peak values of the radial component.

It was revealed that peak values in axial and radial components increased linearly
with the increase of internal pressure. For an increase of pressure with an interval pressure
of 0.5 MPa, the average change of peak values in the axial component was 20 A/M, and
the average change of peak-to-peak values in the radial component was 13.33 A/M.

3.1.2. Magnetic Field Strength Gradient K at Weld Varies

Due to the stress concentration at the weld, the magnetic field strength at the weld
significantly changed, and, accordingly, the magnetic field strength gradient remarkably
varied. The magnetic field intensity gradient was calculated as follows:
Hp ( y )
K = (19)
lk
where K is the gradient value of the magnetic field strength, |ΔHp(y)| is the difference in
the magnetic field intensity between the two adjacent detection points, and Δlk is the space
between two adjacent detection points.
Sensors 2023, 23, 1147 9 of 28

The variations of magnetic signal gradient K with internal pressure in the axial and
radial components are shown in Figure 9.

(a) (b)
Figure 9. Magnetic signal gradient K variation with internal pressure. (a) The axial gradient compo-
nent Kx; (b) the radial gradient component Ky.

It was revealed that the axial gradient component Kx and the radial gradient compo-
nent Ky gradually increased with the increase of internal pressure.

3.1.3. Maximum Magnetic Field Strength Gradient Kmax at Weld Varies

The variation of Kmax with internal pressure for axial and radial components is illus-
trated in Figure 10.

18 Extreme value of radial gradient

Extreme value of axial gradient 12
Linear fitting curve Linear fitting curve
16 11
10
Kymax(A/M/mm)
Kxmax(A/M/mm)

14
9
12 8
7
10
6
8 5
0.0 0.5 1.0 1.5 2.0 2.5 3.0 0.0 0.5 1.0 1.5 2.0 2.5 3.0
Stress（MPa） Stress（MPa）
(a) (b)
Figure 10. Kmax variation with internal pressure. (a) Axial component of Kmax; (b) radial compo-
nent of Kmax.

It was revealed that in axial and radial components, Kmax linearly increased with the
elevation of internal pressure. For an increase of pressure with an interval pressure of 0.5
MPa, the average change of Kmax for the axial component was 2.83 A/M/mm, and the
average change of Kmax for the radial component was 1.95 A/M/mm.

3.1.4. Gradient Energy Factor S(K) at Weld Varies

A new parameter, namely the gradient energy factor S(K), is presented in this study.
S(K) is the area surrounded by the gradient curve of magnetic field strength and the ab-
scissa axis, and it can more directly reflect the degree of stress damage.
The variations of the S(K) parameter with internal pressure in the axial and radial
components are shown in Figure 11.
Sensors 2023, 23, 1147 10 of 28

500 240
Axial gradient area 220 Radial gradient area
450 Linear fitting curve Linear fitting curve
200
400
180
Sx（K）

Sy（K）
350
160
300 140

250 120

100
200
0.0 0.5 1.0 1.5 2.0 2.5 3.0 0.0 0.5 1.0 1.5 2.0 2.5 3.0
Stress（MPa） Stress（MPa）
(a) (b)
Figure 11. S(K) variation with internal pressure. (a) Axial component of S(K); (b) radial component
of S(K).

It was found that S(K) increased with the elevation of internal pressure. For an in-
crease of pressure with an interval pressure of 0.5 MPa, the average change of S(K) for the
axial component was 77.5, and the average change of S(K) for the radial component was
36.7.
The year-on-year growth rate is expressed as follows:
A A2 − A1
= = (20)
A1 A1
where ν is the year-on-year growth rate, A2 denotes the secondary current value, A1 is the
primary current value in the same period, and ΔA is the current increment.
Therefore, the year-on-year growth rate of energy factor S(K) for axial component νx
was 106.67%, and the growth rate of energy factor S(K) for radial component νy was
109.43%. The radial growth rate νy of S(K) was 3.24% higher than the axial growth rate νx,
indicating that the radial component of the magnetic signal at the weld of the ferromag-
netic material was more sensitive to variations of stress.
It was revealed that the change of energy factor S(K) was larger and more obvious
than that of K and Kmax in terms of order of magnitude and average change, which could
more intuitively reflect the change of the magnetic signal. Therefore, S(K) can be used as
a new parameter to comprehensively reflect the damage state of the weld.

3.2. Influences of Internal Pressure on the Weak Magnetic Signal of Weld Crack
The oil and gas pipelines must tolerate internal pressure during normal operation,
and their combination with stress concentration may seriously affect the operational
safety of the pipeline. Therefore, it is extremely necessary to analyze the influences of in-
ternal pressure on a pipeline’s weak magnetic signal.
The initial residual stress obtained from the simulation in Section 2.3.2 was set as
prestress, and an internal pressure of 0.5–3 MPa was applied to the pipe. The simulation
was executed with a pressure interval of 0.5 MPa, and the influences of internal pressure
on the weak magnetic field at the weld crack of the pipe were analyzed. The results of this
test of the weak magnetic signal at the weld crack are shown in Figure 12:
Sensors 2023, 23, 1147 11 of 28

(a) (b)
Figure 12. Magnetic signal of weld crack under different internal pressures. (a) Axial component of
weak magnetic signal; (b) radial component of weak magnetic signal.

According to the above-mentioned results of magnetic simulation, the peak values of

the axial components and the peak-to-peak values of the radial components, gradient K,
maximum gradient Kmax, and gradient energy factor S(K) of magnetic signal characteris-
tic parameters were analyzed.

3.2.1. Peak Values in Axial and Radial Components at Weld Crack

The variations of the peak values of the axial components and the peak-to-peak val-
ues of the radial components with internal pressure are shown in Figure 13.

850
520 Peak of the axial componment
Linear fitting curve Peak to peak of the
500 800 radial componment
Linear fitting curve
480
750
Hx（A/M）

Hy（A/M）

460

440 700
420
650
400

380 600
360 0.5 1.0 1.5 2.0 2.5 3.0
0.5 1.0 1.5 2.0 2.5 3.0
Stress（MPa） Stress(MPa)

(a) (b)
Figure 13. The peak values of weak magnetic signal at weld crack changes with internal pressures.
(a) Peak values of the axial component; (b) peak-to-peak values of the radial component.

It was revealed that the peak values in axial and radial components increased linearly
with internal pressure. For an increase of pressure with an interval pressure of 0.5 MPa,
the average change of peak value in the axial component was 83.33 A/M, and the average
change of peak-to-peak value in the radial component was 137.5 A/M.

3.2.2. Magnetic Field Strength Gradient K at Weld Crack

The variations of magnetic field strength gradient K for the axial and radial compo-
nents under different internal pressures are shown in Figure 14.
Sensors 2023, 23, 1147 12 of 28

(a) (b)
Figure 14. Magnetic signal gradient K at weld crack changes with internal pressures. (a) The axial
gradient component Kx; (b) the radial gradient component Ky.

It was found that the magnetic field strength gradient K for the axial and radial com-
ponents gradually increased with elevation of internal pressure.

3.2.3. Maximum Magnetic Field Strength Gradient Kmax at Weld Crack

The changes of maximum magnetic field strength gradient Kmax for axial and radial
components are illustrated in Figure 15.

75 Extreme value of axial gradient Extreme value of radial gradient

140
Linear fitting curve Linear fitting curve
70
130
65
Kxmax(A/M/mm)

Kymax(A/M/mm)

120
60

55 110

50 100

45
90
40
80
0.5 1.0 1.5 2.0 2.5 3.0 0.5 1.0 1.5 2.0 2.5 3.0
Stress（MPa） Stress（MPa）
(a) (b)
Figure 15. Kmax variations at weld crack with internal pressures. (a) Axial component of Kmax; (b)
radial component of Kmax.

It was revealed that in axial and radial components, Kmax linearly increased with the
elevation of internal pressure. For an increase of pressure with an interval pressure of 0.5
MPa, the average change of Kmax for the axial component was 12.5 A/M/mm, and the
average change of Kmax for the radial component was 23.3 A/M/mm.

3.2.4. Gradient Energy Factor S(K) at Weld Crack

The variations of gradient energy factor S(K) in axial and radial components with
internal pressure are shown in Figure 16.
Sensors 2023, 23, 1147 13 of 28

2100
750 Axial gradient area
Linear fitting curve 2000 Radial gradient area
700 Linear fitting curve
1900
650
1800
Sx（K）

Sy（K）
600
1700
550
1600
500 1500
450 1400

400 1300
0.5 1.0 1.5 2.0 2.5 3.0 0.5 1.0 1.5 2.0 2.5 3.0
Stress（MPa） Stress（MPa）
(a) (b)
Figure 16. S(K) varies with the internal pressure. (a) Axial component of S(K); (b) radial component
of S(K).

It was found that the gradient energy factor S(K) in axial and radial components grad-
ually increased with the elevation of internal pressure. For an increase of pressure with
an interval pressure of 0.5 MPa, the average change of the gradient energy factor S(K) in
the axial component was 125, and it was 340 in the radial component.
The year-on-year growth rate in the axial component νx of energy factor S(K) was
59.57%; the year-on-year growth rate in the radial component νy of energy factor S(K) was
61.54%. The year-on-year growth rate in the axial component was 1.97% higher than that
in the radial component, indicating that the radial component of the magnetic signal at
the weld crack of ferromagnetic materials was more sensitive to the variations of stress.
After Figure 16 was compared with Figures 14 and 15, it was found that the gradient
energy factor S(K), gradient K, and the maximum gradient Kmax followed the same pat-
tern of variations. Moreover, the variation of S(K) was more significant than that of K and
Kmax, indicating that S(K) can be replaced with the gradient factor to analyze the stress
state at the weld.

3.3. Effects of Internal Pressure on Weak Magnetic Signal of Pipeline Base Metal Crack
A crack was created with the size of 2 mm × 0.95 mm × 1 mm (length × width × depth)
on the pipe base metal (consistent with the weld base metal parameters). The results of
the weak magnetic field simulation are shown in Figure 17.

(a) (b)
Figure 17. Magnetic signal of parent material crack under different internal pressure. (a) Axial com-
ponent of weak magnetic signal; (b) radial component of weak magnetic signal.

According to the above-mentioned results of magnetic simulation, the peak values of

the axial components and the peak-to-peak values of the radial components, gradient K,
maximum gradient Kmax, and gradient energy factor S(K) of magnetic signal characteris-
tic parameters were analyzed.
Sensors 2023, 23, 1147 14 of 28

3.3.1. Peak Values in Axial and Radial Components at the Crack of Base Metal
The variations of the peak values of the axial components and the peak-to-peak val-
ues of the radial components with internal pressure are shown in Figure 18.

110 360
Peak of the axial componment
Linear fitting curve Peak to peak of the
100 340 radial componment
Linear fitting curve
320
90
Hx（A/M）

Hy（A/M）
300
80
280
70
260

60 240

50 220
0.5 1.0 1.5 2.0 2.5 3.0 0.5 1.0 1.5 2.0 2.5 3.0
Stress（MPa） Stress（MPa）
Peak of the axial componment
(a) Linear fitting curve
(b)
Figure 18. The peak values of weak magnetic signal at the crack of base metal changes with internal
pressures. (a) Peak values of the axial component; (b) Peak-to-peak values of the radial component.

It was revealed that peak values in axial and radial components increased linearly
with the internal pressure. For an increase of pressure with an interval pressure of 0.5
MPa, the average change of peak value in the axial component was 17.67 A/M, and the
average change of peak-to-peak value in the radial component was 52.17 A/M.

3.3.2. Magnetic Field Strength Gradient K at the Crack of Base Metal

The variations of magnetic field strength gradient K for the axial and radial compo-
nents under different internal pressures are shown in Figure 19.

(a) (b)
Figure 19. Magnetic signal gradient K at base metal crack changes with internal pressures. (a) The
axial gradient component Kx; (b) the radial gradient component Ky.

It was found that magnetic field strength gradient K for the axial and radial compo-
nents gradually increased with the elevation of internal pressure.

3.3.3. Maximum Magnetic Field Strength Gradient Kmax at the Crack of Base Metal
The changes of maximum magnetic field strength gradient Kmax for axial and radial
components are illustrated in Figure 20.
Sensors 2023, 23, 1147 15 of 28

Extreme value of axial gradient 21 Extreme value of radial gradient

Linear fitting curve Linear fitting curve
3.5 20

19

Kxmax(A/M/mm)

Kymax(A/M/mm)
18
3.0
17

16

2.5 15

14
0 1 2 3 0.5 1.0 1.5 2.0 2.5 3.0
Stress（MPa） Stress（MPa）
(a) (b)
Figure 20. Kmax variations at base metal crack with internal pressures. (a) Axial component of Kmax;
(b) radial component of Kmax.

It was revealed that in axial and radial components, Kmax linearly increased with the
elevation of internal pressure. For an increase of pressure with an interval pressure of 0.5
MPa, the average change of Kmax for axial component was 0.58 A/M/mm, and the average
change of Kmax for radial component was 3.5 A/M/mm.

3.3.4. Gradient Energy Factor S(K) at the Crack of Base Metal

The variations of gradient energy factor S(K) in axial and radial components with
internal pressure are shown in Figure 21.

150
90
Axial gradient area Radial gradient area
140
85 Linear fitting curve Linear fitting curve

80 130

75 120
Sx(k)

Sy(k)

70 110
65
100
60
90
55
80
50
0.5 1.0 1.5 2.0 2.5 3.0 0.5 1.0 1.5 2.0 2.5 3.0
Stress（MPa） Stress（MPa）
(a) (b)
Figure 21. S(K) varies with the internal pressure. (a) Axial component of S(k); (b) radial component
of S(k).

It was revealed that S(K) in both axial and radial components increased linearly with
the elevation of internal pressure. For an increase of pressure with an interval pressure of
0.5 MPa, the average change of S(K) in the axial component was 14.5, and it was 23.67 in
the radial component. The year-on-year growth rate in the axial component νx of energy
factor S(K) was 64.15%; the year-on-year growth rate in the radial component νy of energy
factor S(K) was 77.5%. The year-on-year growth rate in the axial component was 13.35%
higher than that in the radial component, indicating that the radial component of the mag-
netic signal at the weld crack of ferromagnetic materials was more sensitive to the varia-
tions of stress. From the perspective of magnitude and average change, S(K) had a greater
variation than K and Kmax, and the variation may be more obvious.
According to the above-mentioned parametric analysis of the weld, weld cracks, and
base metal cracks, it was noted that the magnitude and average change of the proposed
new parameter S(K) were larger than those of K and Kmax, thus, S(K) can be used as a new
parameter to comprehensively assess the stress-induced damage of pipelines.
Sensors 2023, 23, 1147 16 of 28

3.4. Comprehensive Analysis of Characteristic Parameters of Weak Magnetic Signals

The variations of Kmax and S(K) parameters in the weld, weld crack, pipe base metal
crack, and pipe damage were comprehensively analyzed (Figure 22).
80
140 weld
weld
70 Weld crack Weld crack
Base metal crack 120 Base metal crack
60

Kymax(A/M/mm)
100
Kxmax(A/M/mm)

50
80
40
60
30
40
20
20
10
0
0.5 1.0 1.5 2.0 2.5 3.0 0.5 1.0 1.5 2.0 2.5 3.0
Stress（MPa） Stress（MPa）
(a) (b)
800
2000 weld
weld Weld crack
700 Weld crack 1800 Base metal crack
Base metal crack
1600
600
1400
500
Sy(K)

1200
Sx(K)

1000
400
800
300
600
200 400
200
100
0.5 1.0 1.5 2.0 2.5 3.0 0.5 1.0 1.5 2.0 2.5 3.0
Stress（MPa） Stress（MPa）
(c) (d)
Figure 22. Variations of Kmax and S(K) parameters calculated by simulation under three conditions.
(a) Axial component of Kmax; (b) radial component of Kmax. (c) Axial component of S(k); (d) radial
component of S(k).

It can be observed from the above-illustrated figure that during weld crack detection,
the two parameters had generally great changes, and the order of magnitude of S(K) was
greater than that of Kmax, accompanied by more significant changes.
The variations of Kmax and S(K) parameters under the same internal pressure of 3
MPa are shown in Figure 23.

80 weld
Weld crack
160 weld
Weld crack
70 140
Kxmax(A/M/mm)

Kymax(A/M/mm)

Base metal crack Base metal crack

60 120
50 100
40 80
30 60
20 40
10 20
0 0
3MPa 3MPa
(a) (b)
Sensors 2023, 23, 1147 17 of 28

800
weld
weld
Weld crack
Base metal crack
2000 Weld crack
Base metal crack

700
600 1500
Sx(K)

Sy(K)
500
400 1000
300
200 500
100
0 0
3MPa 3MPa
(c) (d)
Figure 23. Variations of Kmax and S(K) parameters at 3 MPa. (a) Axial component of Kmax; (b) radial
component of Kmax. (c) Axial component of S(k); (d) radial component of S(k).

It can be observed from the above-illustrated figure that during weld crack detection,
Kmax and S(K) significantly changed, the stress concentration was large, and the risk of
failure was high. Therefore, the pipe weld crack should be detected in time with this
method.

4. Experiment and Analysis

In order to study the weak magnetic signal characteristics of weld cracks in long oil
and gas pipelines and verify the reliability of the numerical model, a pipeline pressuriza-
tion experiment was designed to analyze the changes of the weak magnetic signal charac-
teristic parameters of pipeline welds, weld cracks, and base metal cracks. The experi-
mental results verified the reliability of the numerical model.

4.1. Experimental Materials

The experimental material was a section of X70 welded pipe with artificial cracks.
The overall length of the pipe was 6000 mm, the diameter was 1012 mm, and the wall
thickness was 14.5 mm.

4.2. Experimental Equipment

The test equipment was TSC-2M-8 made by the Russian Power Diagnosis Company
(Moscow, Russia). The lift off value was set to 1 mm. The weak magnetic signal under
different internal pressures was measured perpendicular to the weld and crack. A dia-
gram of the test measurement is shown in Figure 24.

(a) (b) (c)

Figure 24. Experimental setup. (a) Experimental arrangement; (b) weld, weld crack, and base metal
crack; (c) pressure suppression equipment.
Sensors 2023, 23, 1147 18 of 28

4.3. Experimental Methods

First, both ends of the pipe and weld were blocked with a water nozzle with a diam-
eter of φ 50 mm. One end was considered the water inlet, and the other end was the water
outlet. During the experiment, the pressure pump was used to inject water into the water
inlet to simulate the working state of the pipeline during normal operation, and the water
pressure sensor was installed to monitor the change of water pressure in the pipeline to
prevent cracking in pipelines. A strain gauge was attached to the crack tip, and the stress
at the crack tip was detected with the stress and strain measurement devices. Once the
strain at the crack tip exceeded the preset threshold, the strain gauge would trigger an
alarm and terminate water injection into the pipes to reduce the internal pressure of the
pipes and ensure safety.
When the pipeline was pressurized, the pressure was held for 30 min with pressure
increments of 0.5 MPa. When stress distribution was stable, the weak magnetic signals of
the weld, weld crack, and base metal crack were measured.

4.4. Analysis of the Experimental Results

4.4.1. Weak Magnetic Signal of Weld
The variation pattern of weak magnetic signal characteristics at the pipe weld is
shown in Figure 25.

(a) (b)
Figure 25. Magnetic signal changes at the weld with internal pressures. (a) Axial component of weak
magnetic signal; (b) radial component of weak magnetic signal.

According to the experimental results, the magnetic signal characteristic parameters,

such as axial peak value and radial peak-to-peak value, gradient K, maximum gradient
Kmax, and gradient energy factor S(K), were analyzed.
1. Peak values in axial and radial components
The fitting results of axial peak value and radial peak-to-peak value versus internal
pressure are illustrated in Figure 26.

-55
Peak of the axial componment 70
Peak to peak of the
Linear fitting curve radial componment
-60
65 Linear fitting curve
Hy（A/M）
Hx（A/M）

-65
60

-70
55

-75 50

-80 45
0.5 1.0 1.5 2.0 2.5 3.0 0.5 1.0 1.5 2.0 2.5 3.0
Stress（MPa） Stress（MPa）
(a) (b)
Sensors 2023, 23, 1147 19 of 28

Figure 26. Peak value variations with internal pressure. (a) Peak values of the axial component; (b)
peak-to-peak values of the radial component.

It can be observed from the above-illustrated figure that the peak value of the axial
component and the peak-to-peak value of the radial component increased linearly with
the elevation of internal pressure.
2. Magnetic field strength gradient K
The variations of the magnetic field strength gradient K under different internal pres-
sures are shown in Figure 27.

(a) (b)
Figure 27. Variations of the magnetic field strength gradient K at the weld under different internal
pressures. (a) The axial gradient component Kx; (b) the radial gradient component Ky.

It can be observed that the magnetic field intensity gradient K increased with the
elevation of internal pressure, which is consistent with the numerical data.
3. Maximum magnetic field intensity gradient Kmax
The variations of Kmax under different internal pressures are shown in Figure 28.

1.8
Extreme value of axial gradient 3.2 Extreme value of radial gradient
Linear fitting curve Linear fitting curve
1.6 3.0
1.4 2.8
Kymax(A/M/mm)
Kxmax(A/M/mm)

1.2 2.6
2.4
1.0
2.2
0.8
2.0
0.6
1.8
0.4 1.6
0.5 1.0 1.5 2.0 2.5 3.0 0.5 1.0 1.5 2.0 2.5 3.0
Stress（MPa） Stress（MPa）
(a) (b)
Figure 28. Variations of Kmax at the weld under different internal pressures. (a) Axial component of
Kmax; (b) radial component of Kmax.

It can be observed that with the increase of internal pressure, Kmax became larger,
which is consistent with the numerical data.
4. Gradient energy factor S(K)
Figure 29 shows the variations of S(K) enclosed by the magnetic field intensity gra-
dient curve and the abscissa axis under different internal pressures.
Sensors 2023, 23, 1147 20 of 28

95
100 Radial gradient area
Axial gradient area
90 Linear fitting curve Linear fitting curve

95
85

Sy(K）
Sx（K) 90
80
85
75
80
70
0.5 1.0 1.5 2.0 2.5 3.0 0.5 1.0 1.5 2.0 2.5 3.0
Stress（MPa） Stress（MPa）
(a) (b)
Figure 29. Variations of S(K) at the weld under different internal pressures. (a) Axial component of
S(k); (b) radial component of S(k).

It can be observed from the above-illustrated fitting curve that S(K) increased with
the elevation of internal pressure, which is consistent with the numerical data. The year-
on-year growth rate of axial component νx of energy factor S(K) was 26.39%; the year-on-
year growth rate of radial component νy of energy factor S(K) was 30.77%. The year-on-
year growth rate of the radial component was larger than that of the axial component, and
the radial component of the magnetic signal at the weld crack was more sensitive to the
variations of stress, thus verifying the correctness of the numerical data.
In terms of order of magnitude, S(K) was greater than K and Kmax, and the degree of
variation was more obvious. The experimental results were consistent with the numerical
data, and it was confirmed that the new parameter S(K) can more intuitively and compre-
hensively indicate weld crack-induced failures in pipelines.

4.4.2. Weld Crack

The variation pattern of weak magnetic signal characteristics at the pipe weld crack
is shown in Figure 30.

(a) (b)
Figure 30. Magnetic signal changes at weld crack with internal pressures. (a) Axial component of
weak magnetic signal; (b) radial component of weak magnetic signal.

According to the experimental results, the magnetic signal characteristic parameters,

such as axial peak value and radial peak-to-peak value, gradient K, maximum gradient
Kmax, and gradient energy factor S(K) were analyzed.
1. Peak values in axial and radial components
The fitting results of axial peak value and radial peak-to-peak value versus internal
pressure are illustrated in Figure 31.
Sensors 2023, 23, 1147 21 of 28

54
Peak to peak of the
52 radial componment
Linear fitting curve
50
48

Hy（A/M）
46
44
42
40
38
36
0.5 1.0 1.5 2.0 2.5 3.0
Stress（MPa）
(a) (b)
Figure 31. Peak value varies at weld crack with internal pressures. (a) Peak values of the axial com-
ponent; (b) peak-to-peak values of the radial component.

It can be observed from the above-illustrated figure that the peak value of the axial
component and the peak value of the radial component increased linearly with the eleva-
tion of internal pressure.
2. Magnetic field strength gradient K
The variations of the magnetic field strength gradient K under different internal pres-
sures are shown in Figure 32.

(a) (b)
Figure 32. Variations of the magnetic field strength gradient K at weld crack under different internal
pressures. (a) The axial gradient component Kx; (b) the radial gradient component Ky.

It can be observed that the magnetic field intensity gradient K increased with the
elevation of internal pressure, which is consistent with numerical data.
3. Maximum magnetic field intensity gradient Kmax
The variations of Kmax under different internal pressures are shown in Figure 33.

8
Extreme value of axial gradient Extreme value of radial gradient
8 Linear fitting curve Linear fitting curve
7
7
Kxmax(A/M/mm)

Kymax(A/M/mm)

6
6

5 5

4 4

3 3
0.5 1.0 1.5 2.0 2.5 3.0 0.5 1.0 1.5 2.0 2.5 3.0
Stress（MPa） Stress（MPa）
(a) (b)
Sensors 2023, 23, 1147 22 of 28

Figure 33. Variations of Kmax at weld crack under different internal pressures. (a) Axial component
of Kmax; (b) radial component of Kmax.

It can be observed that with the increase of internal pressure, Kmax became larger,
which is consistent with the numerical data.
4. Gradient energy factor S(K)
Figure 34 shows the variations of the S(K) enclosed by the magnetic field intensity
gradient curve and the abscissa axis under different internal pressures.

175
180
Axial gradient area Radial gradient area
170
Linear fitting curve Linear fitting curve
165 170
160
Sx(K)

Sy(K)
155 160

150
150
145
140
140
135
130 130
0.5 1.0 1.5 2.0 2.5 3.0 0.5 1.0 1.5 2.0 2.5 3.0
Stress（MPa） Stress（MPa）
(a) (b)
Figure 34. Variations of S(K) at weld crack under different internal pressures. (a) Axial component
of S(k); (b) radial component of S(k).

It was revealed that S(K) increased with the elevation of internal pressure, which is
consistent with the numerical data. The year-on-year growth rate of axial component νx of
S(K) was 29.63%; the year-on-year growth rate of radial component νy of S(K) was 30.88%.
The year-on-year growth rate of the radial component was higher than that of the axial
component, and the radial component of the magnetic signal at the weld crack was more
sensitive to the variations of stress, thus verifying the correctness of the numerical data.

4.4.3. Crack of Pipe Base Metal

The variation pattern of weak magnetic signal characteristics at the crack in the pipe
base metal is illustrated in Figure 35.

(a) (b)
Figure 35. Magnetic signal changes at the crack of base metal with internal pressures. (a) Axial com-
ponent of weak magnetic signal; (b) radial component of weak magnetic signal.

According to the experimental results, the magnetic signal characteristic parameters,

such as axial peak value and radial peak-to-peak value, gradient K, maximum gradient
Kmax, and gradient energy factor S(K) were analyzed.
1. Peak values in axial and radial components
Sensors 2023, 23, 1147 23 of 28

The fitting results of axial peak value and radial peak-to-peak value versus internal
pressure are illustrated in Figure 36.

32
Peak to peak of the
radial componment
30 Linear fitting curve

Hy（A/M）
28

26

24

22
0.5 1.0 1.5 2.0 2.5 3.0
Stress（MPa）
(a) (b)
Figure 36. Peak value varies at the crack of base metal with internal pressures. (a) Peak values of the
axial component; (b) peak-to-peak values of the radial component.

It can be observed from the above-illustrated figure that the peak value of the axial
component and the peak-to-peak value of the radial component increased linearly with
the elevation of internal pressure.
2. Magnetic field strength gradient K
The variations of the magnetic field strength gradient K under different internal pres-
sures are shown in Figure 37.

(a) (b)
Figure 37. Variations of the magnetic field strength gradient K at the crack of base metal under
different internal pressures. (a) The axial gradient component Kx; (b) the radial gradient component
Ky.

It can be observed that the magnetic field intensity gradient K increased with the
elevation of internal pressure, which is consistent with numerical data.
3. Maximum magnetic field intensity gradient Kmax
The variations of Kmax under different internal pressures are shown in Figure 38.
Sensors 2023, 23, 1147 24 of 28

6.0
Extreme value of axial gradient
4.0 Extreme value of radial gradient
Linear fitting curve 5.5 Linear fitting curve
5.0
3.5

Kxmax(A/M/mm)

Kymax(A/M/mm)
4.5
4.0
3.0
3.5

2.5 3.0
2.5

2.0 2.0
1.5
0.5 1.0 1.5 2.0 2.5 3.0 0.5 1.0 1.5 2.0 2.5 3.0
Stress（MPa） Stress（MPa）
(a) (b)
Figure 38. Variations of Kmax at the crack of base metal under different internal pressures. (a) Axial
component of Kmax; (b) radial component of Kmax.

It can be observed that with the increase of internal pressure, Kmax became larger,
which is consistent with the numerical data.
4. Gradient energy factor S(K)
Figure 39 shows the variations of the S(K) enclosed by the magnetic field intensity
gradient curve and the abscissa axis under different internal pressures.

112
Axial gradient area 120 Radial gradient area
110 Linear fitting curve Linear fitting curve
108
110
106
Sy(K)
Sx(K)

104
100
102
100
90
98
96
80
94
0.5 1.0 1.5 2.0 2.5 3.0 0.5 1.0 1.5 2.0 2.5 3.0
Stress（MPa） Stress（MPa）
(a) (b)
Figure 39. Variations of S(K) at the crack of base metal under different internal pressures. (a) Axial
component of S(k); (b) radial component of S(k).

It was revealed that S(K) increased with the elevation of internal pressure, which is
consistent with the numerical data. From the perspective of magnitude, S(K) was greater
than K and Kmax, and the degree of variation was more obvious. The experimental results
were consistent with the numerical data. The year-on-year growth rate of axial component
νx of energy factor S(K) was 18.09%; the year-on-year growth rate of radial component νy
of energy factor S(K) was 50%. The year-on-year growth rate of the radial component was
larger than that of the axial component, and the radial component of the magnetic signal
at the base metal crack was more sensitive to the variations of stress, thus verifying the
correctness of the numerical data.
According to the parametric analysis of the weld, weld crack, and base metal crack,
it was noted that the order of magnitude and average change of the new parameter S(K)
were larger than K and Kmax, thus, S(K) can be used as a new parameter to comprehen-
sively indicate weld crack-induced failures in pipelines.

4.5. Comprehensive Analysis of Characteristic Parameters

The changes of Kmax and energy factor S(K) as related to the parameters of the weld,
weld crack, and base metal crack were comprehensively analyzed, as displayed in Figure
40.
Sensors 2023, 23, 1147 25 of 28

8
8
Weld
7 Base metal crack 7 Weld
Weld crack Base metal crack
Weld crack
6 6

Kxmax(A/M/mm)

Kymax(A/M/mm)
5
5
4

3 4

2 3
1
2
0
0.5 1.0 1.5 2.0 2.5 3.0 0.5 1.0 1.5 2.0 2.5 3.0
Stress（MPa） Stress（MPa）
(a) (b)
180
180 Weld
Base metal crack
160 Weld crack

Weld 160
Base metal crack
140 Weld crack
140

Sy(K)
Sx(K)

120
120
100
100
80
80
60
0.5 1.0 1.5 2.0 2.5 3.0 0.5 1.0 1.5 2.0 2.5 3.0
Stress（MPa） Stress（MPa）
(c) (d)
Figure 40. Variations of Kmax and S(K) parameters calculated by experiment under three conditions.
(a) Axial component of Kmax; (b) radial component of Kmax. (c) Axial component of S(k); (d) radial
component of S(k).

It can be observed from the above-illustrated figure that during weld crack detection,
the variations of the two parameters were large on the whole, and the order of magnitude
and degree of variation of S(K) were greater than those of Kmax. Therefore, the energy
factor S(K) can be used as a new parameter to comprehensively indicate weld crack-in-
duced failure in a pipeline, verifying the correctness of the numerical data.
The variations of Kmax and S(K) parameters under the same internal pressure of 3
MPa are shown in Figure 41.
9
8
8 Weld
Weld
Base metal crack
Base metal crack 7 Weld crack
7 Weld crack

6
6
Kymax(A/M/mm)
Kxmax(A/M/mm)

5
5
4
4
3
3

2
2

1 1

0 0
3MPa 3MPa
(a) (b)
Sensors 2023, 23, 1147 26 of 28

200
180
Weld
Base metal crack 180 Weld
160 Weld crack Base metal crack
Weld crack
160
140
140
120
120
Sx(K)

Sy(K)
100
100
80
80
60
60
40
40
20 20

0 0
3MPa 3MPa
(c) (d)
Figure 41. Variations of Kmax and S(K) parameters at an internal pressure of 3 MPa. (a) Axial com-
ponent of Kmax; (b) radial component of Kmax. (c) Axial component of S(k); (d) radial component of
S(k).

It can be observed from the above-illustrated figure that during weld crack detection,
the Kmax and S(K) parameters significantly varied, indicating that the stress concentration
was noticeable; thus, the weld crack was the most dangerous and should be detected in
time.

5. Conclusions
Through numerical simulation and experimental verification, the following conclu-
sions could be drawn.
When the pipeline is welded, the microstructure of the material at the weld changes,
thus resulting in a large amount of martensite. The linear relationship between the ferro-
magnetism of the material and the content of martensite is the cause of the magnetic signal
at the weld.
Under the action of internal pressure, the magnetic signal parameters (K, Kmax, and
energy factor S(K)) of the weld, weld crack, and base metal crack increased with the ele-
vation of internal pressure. The average variation of the newly proposed parameter, en-
ergy factor S(K), was larger, which could more intuitively indicate weld crack-induced
failure in a pipeline.
Under internal pressure, the radial year-on-year growth rate νy of the energy factor
S(K) of the weld, weld crack and base metal crack were greater than the axial year-on-year
growth rate νx (νy at the weld was 3.24% higher than νx, νy at the weld crack was 1.97%
higher than νx, and νy at the base metal crack was 13.35% higher than νx in this study),
indicating that the radial component of the magnetic signal is more sensitive to variations
of stress.
Subsequently, more in-depth research will be carried out in the field of micro-cracks
in pipeline welds to realize real-time warnings of micro-cracks and avoid accidents.

Author Contributions: Conceptualization, L.H.; Methodology, H.G.; Resources, L.Y.; Writing—

original draft, Y.F.; Writing—review & editing, B.L. All authors have read and agreed to the pub-
lished version of the manuscript.
Funding: This research was funded by the National Natural Science Foundation of China (Grant
No. 62101356).
Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.
Data Availability Statement: Not applicable.
Conflicts of Interest: The authors declare no conflict of interest.
Sensors 2023, 23, 1147 27 of 28

References
1. Wang, H.; Li, S.; Chen, S.; Luo, Y.; Zhang, X.; Zhou, S. Discussion on Fracture of Circumferential Weld of High Grade Natural
Gas Pipeline. Pet. Tubul. Goods Instruments. 2020, 6, 49–52.
2. Feng, G. Study on the Characteristics of Weld Crack Signal Based on Magnetic Memory Testing Technology. Master’s Thesis,
Shenyang University of Technology, Shenyang, China, 2021.
3. He, L. Based on the Stress Internal Detection Mechanism of Magnetic Memory Oil and Gas Pipeline. Ph.D. Thesis, Shenyang
University of Technology, Shenyang, China, 2020.
4. Li, G.; Shen, G.; Li, H. Nondestructive testing technologys for industrial pipelines. J. Papers 2006, 28, 89–93.
5. Li., Q.; Zhao, M.; Ren, X.; Wang, L.; Feng, X.; Niu, Y. Construction status and development trend of Chinese oil & gas pipeline.
Oil-Gas Field Surf. Eng. 2019, 38, 14–17.
6. Bo, Z.; Xiao-Jia, X.U.; Hua-Dong, S. Review of in-pipe inspection technology and systemdevelopment in pipeline. Pipeline Tech-
nol. Equip. 2018, 3, 22–25.
7. Yang, L.; Geng, H.; Gao, S. Internal magnetic leakage detection technology of long oil and gas pipeline. Instrument 2016, 37,
1736–1746.
8. Xie, S.; Chen, Z.; Takagi, T. Development of A Very Fast Simulator for Pulsed Eddy Current Testing Signals of Local Wall
Thinning. NDT E Int. 2012, 51, 45–50.
9. Xie, S.; Chen, Z.; Takagi, T. Quantitative Non-destructive Evaluation of Wall Thinning Defect in Double-layer Pipe of Nuclear
Power Plants Using Pulsed ECT Method. NDT E Int. 2016, 75, 87–95.
10. 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.
11. Xu, K.; Qiu, X.; Tian, X. Investigation of Metal Magnetic Memory Signals of Welding Cracks. J. Nondestruct. Eval. 2017, 36, 20.
12. Chen, H.; Wang, C.; Zuo, X. Research on methods of defect classification based on metal magneticmemory. NDT E Int. 2017, 92,
82–87.
13. Ni, C.; Hua, L.; Wang, X. Crack propagation analysis and fatigue life prediction for structural alloy steel based on metal mag-
netic memory testing. J. Magn. Magn. Mater. 2018, 462, 144–152.
14. Huang, H.; Qian, Z.; Yang, C.; Han, G.; Liu, Z. Magnetic memory signals of ferromagnetic weldment induced by dynamic
bending load. Nondestruct. Test. Eval. 2016, 32, 166–184.
15. Liu, B.; Fu, Y.; Ren, J. Modelling and analysis of magnetic memory testing method based on thedensity functional theory. Non-
destr. Test. Eval. 2015, 30, 13–25.
16. Yao, K.; Wang, Z.D.; Deng, B.; Shen, K. Experimental Research on Metal Magnetic Memory Method. Exp. Mech. 2012, 52, 305–
314.
17. Zeng, W.G.; Dang, X.F.; Li, S.H.; Wang, H.M.; Wang, H.; Wang, B. Application of non-contact magnetic corresponding on the
detection for natural gas pipeline. E3S Web Conf. 2020, 185, 1090.
18. He, G.; He, T.; Liao, K.; Zhu, H.; Zhao, S. A Novel Three-Dimensional Non-Contact Magnetic Stress Inspection Technology and
Its Application on LNG Pipeline. 2020 13th International Pipeline Conference, 2020. Available online:
https://xueshu.baidu.com/usercenter/paper/show?paperid=100c0tx0bx2106b0y9160ad089783964&site=xueshu_se (accessed on
24 December 2022).
19. Konnov, V.V. Processing of magnetometric data during remote diagnostics of underground steel pipelines. Oil Gas Bus. 2012,
6, 147–162.
20. Narkhov, E.D.; Sapunov, V.; Denisov, A.U.; Savelyev, D.V. Novel quantum NMR magnetometer non-contact defectoscopy and
monitoring technique for the safe exploitation of gas pipelines. WIT Trans. Ecolog. Environ. 2015, 186, 649–658.
21. Bhadran, V.; Shukla, A.; Karki, H. Non-contact flaw detection and condition monitoring of subsurface metallic pipelines using
magnetometric method. Mater. Today 2020, 28, 860–864.
22. Su, S.; Zhao, X.; Wang, W.; Zhang, X. Metal magnetic memory inspection of Q345 steel specimens with butt weld in tensile and
bending test. J. Nondestruct. Eval. 2019, 38, 64.
23. Lu, W.; Li, J.; Huang, Z.; Chen, Q.; Chen, T. Studies on the magnetic memory testing for the welding cold crack of low alloy
high strength steel. Value Eng. 2019, 38, 123–125, 2019.
24. Birss, R.R.; Faunce, C.A.; Isaac, E.D. Magnetomechanical effects in iron and iron-carbon alloys. J. Phys. D Appl. Phys. 1971, 4,
1040–1048.
25. Liu, B.; Zheng, S.; He, L.; Zhang, H.; Ren, J. Study on internal detection in oil–gas pipelines based on complex stress magneto-
mechanical modeling. IEEE Trans. Instrum. Meas. 2019, 69, 5027–5036.
26. Liu, Q.Y.; Luo, X.; Zhu, H.Y.; Han, Y.W.; Liu, J.X. Modeling plastic deformation effect on the hysteresis loops of ferromagnetic
materials based on modified Jiles-Atherton model. Acta Phys. Sinica. 2017, 66, 107501.
27. Peng, K. Research on the Mechanism and Performance Evolution of Martensite Phase Transition during the Cold Drawing of
304H Stainless Steel Wire. Master’s Thesis, Wuhan University of Science and Technology, Wuhan, China, 2021.
28. Li, Y.; Yu, D.; Li, B.; Chen, X. Martensitic transformation of anaustenitic stainless steel under non-proportional cyclic loading.
Int. J. Fatigue 2019, 124, 338–347.
29. Hu, B.; Liu, Y.; Yu, R. Numerical simulation on magnetic–mechanical behaviors of 304 austenite stainless steel. Measurement
2020, 151, 107185.
Sensors 2023, 23, 1147 28 of 28

30. Paul, S.K.; Stanford, N.; Hilditch, T. IJF-2017-austenite plasticitymechanisms and their behavior during cyclic loading. Int. J.
Fatigue 2017, 106, 185–195.
31. Yang, X. Residual Stress Analysis of Pipe Welding and Its Fatigue Life Assessment. Master’s Thesis, Harbin Engineering Uni-
versity, Harbin, China, 2019.
32. Wang, C. Numerical Simulation of the Welding Temperature Field and the Stress Field. Master’s Thesis, Shenyang University
of Technology, Shenyang, China, 2005.

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual au-
thor(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to
people or property resulting from any ideas, methods, instructions or products referred to in the content.