Internal detection technology for defects using multi-source electromagnetic coupling method

Abstract

This paper proposes an internal inspection system based on electromagnetic coupling sensing and permanent magnet magnetization, providing a theoretical study on alternating magnetic field induction and magnetic flux leakage coupling technology. Simulation analysis and experimental validation were conducted to evaluate the feasibility and reliability of the proposed magnetic rope system (MRS). Through research it has been found that a novel magnetic ring structure has been demonstrated to generate an axial magnetic leakage field and a circumferential uniform alternating electromagnetic field. The excitation method combining electromagnetic coupling sensing with permanent magnets enhances wall penetration capability and magnetic flux intensity, enabling the identification of subsurface defects and shallow surface defects. The amplitude of magnetic field intensity fluctuations of surface defect detection can reach 0.011 T, which is only 0.008 T compared to the test results of subsurface defects. By comparison, the amplitude of fluctuation for the probe structure using permanent magnets alone is only 0.002 T. The combined configuration of MRS and magnets optimizes defect detection rate and sensitivity, making it suitable for complex internal inspection environments and enhancing the reliability and stability. The sensitivity range of detection technology without using permanent magnets is between 0.24 and 0.36. The use of permanent magnets can increase sensitivity to 0.41∼0.77. The defects have lower sensitivity caused by the complex metal environment, and the use of permanent magnet technology can increase sensitivity by more than 0.3. The MRS overcomes the limitations of single-technology approaches in detecting multiple types of defects, thereby improving the detection capability for both near-surface and deeply buried defects.

1 Introduction

Driven by global energy transition goals, the energy industry is witnessing unprecedented development opportunities (Martinez et al., 2023; Patel et al., 2022; Zhao et al., 2022). However, alongside this rapid growth in installed capacity, operational safety issues are becoming increasingly prominent (Kaya et al., 2018; Rezaeiha et al., 2017; Yan and Archer, 2018). The service conditions of core components become increasingly severe (Otero and Ponta, 2018; Chu and Lam, 2020; Wang et al., 2023). These critical components are subjected to multiple impacts over long periods, including alternating wind loads, temperature cycling, salt spray corrosion, and lightning strikes, posing serious challenges to their structural integrity (Das et al., 2019; Leon, 2022; Mishnaevsky, 2019; Herring et al., 2019).

The tower, as the vertical support structure, is made of multi-section steel plates rolled and welded together (Klein and Marx, 2023). The weld regions are the mechanical weak points, where fatigue cracks are easily initiated under alternating loads. More troublesome, the outer surface of the tower is coated with an anti-corrosion layer, so even if cracks of millimeter or even several millimeters in width have developed at the weld, they cannot be detected by external visual inspection (Cheng et al., 2024; Chun-Lung and Tsung-Ying, 2017). In the field of energy equipment damage inspection, internal inspection refers to a technical method that, without destroying the equipment structure or interrupting its normal operation, uses specialized sensors, crawling carriers, or hand-held devices to directly probe and quantitatively characterize the blade cavity structure, tower weld seams, and other concealed parts. Compared with indirect inspection methods that infer internal conditions from external features, internal inspection offers unique advantages such as direct information acquisition, accurate defect localization, and nondestructive testing.

In recent years, significant progress has been made in internal inspection technologies for metal components of equipment. The enterprises and scholars both domestically and internationally have developed many new methods in the fields of structural health inspection technologies and instrument applications. Internationally, the magnetic flux leakage (MFL) inspection system developed by Germany’s Rosen GmbH can provide MFL internal inspection services for metal parts of various diameters (Ilueng et al., 1997; Shah and Mehtre, 2015; Wang et al., 2014). Canada’s BJ Services has conducted research on pipeline dent identification based on triaxial MFL signals. Companies such as Tuboscope (United States), TDW (United States), GE and PII (United States), British Gas (United Kingdom), and NGKH (Russia) have carried out extensive work in the application of pipeline non-destructive testing (NDT) technologies and instrument development (Gloria et al., 2009; Zhukova et al., 2019). NDT technologies have become indispensable tools for structural health inspection and monitoring in modern high-end industrial equipment, providing guidance and basis for the inspection, assessment, and maintenance of major equipment and critical components (Nestleroth and Davis, 2007; Jiles and Melikhov, 2007). Long-term corrosion can lead to pipeline wall thinning, resulting in structural aging; localized corrosion on pipelines can also develop into perforations, causing energy leaks; the interaction of localized dents may lead to pipeline fractures under increased stress or stress pulses. Common NDT techniques such as magnetic flux leakage, ultrasonic testing, eddy current testing, and electromagnetic acoustic transducer (EMAT) technologies can identify metal parts defects (Jiles and Atherton, 1986; Zatsepin, 1966). Apart from differences in the physical detection mechanisms of sensing probes, internal inspection tools also vary in mechanical structures, including spherical and snake-shaped acoustic emission sensors, which not only overcome the problem of limited internal equipment space but are also suitable for inspecting urban composite equipment. Currently, multi-technology integrated metal parts inspection methods have become a research focus among scholars. Recent studies involve the deployment of robots equipped with visual and ultrasonic testing for metal parts assessment, while in-line inspection (ILI) technologies utilize various sensors for data collection, providing rich information sources for metal components of equipment inspection (Bunge, 1989; Charap, 1974). Due to limitations in detection principles, single inspection technologies struggle to achieve comprehensive coverage of multiple defect types such as surface flaws, internal cracks, corrosion, and pinholes on metal parts. By implementing multi-modal sensing NDT through the integration of multiple detection technologies at the physical end or module level, the detection capability for various defect types can be enhanced, compensating for the shortcomings of single technologies in detecting multiple defect types (Hwang and Lord, 1975; Dutta et al., 2009; Chen and Wu, 2023).

Recent advances in electromagnetic coupling sensing detection technology have been achieved in areas such as sensor array integration, electromagnetic-acoustic emission fusion, and magnetoelectric composite materials. For example, array sensors based on the giant magneto-impedance (GMI) effect can expand the detection range, while the electromagnetic-acoustic emission fusion technique operates without a couplant, enabling non-contact online inspection. The advantages of this technology include non-contact detection, the ability to penetrate non-conductive coatings, and simultaneous acquisition of multiple parameter information such as resistance and inductance. Its main limitations include limited detection capability for deeply buried defects due to the skin effect, susceptibility to signal interference in complex electromagnetic environments, and relatively bulky probe structures that reduce adaptability in confined spaces.

Compared with other NDT methods, electromagnetic coupling sensing offers unique advantages in detecting various defects on both the inner and outer walls of metal components (Teng et al., 2019). In contrast to MFL, which relies on single-direction magnetization and exhibits significant sensitivity differences between circumferential and axial defects as well as a bulky excitation system, electromagnetic coupling sensing employs multi-directional excitation design to simultaneously generate axial and circumferential magnetic fields, maintaining high sensitivity to defects of arbitrary orientations on both inner and outer walls (Nakamura, 2012). Compared with conventional eddy current NDT, which is primarily effective for near-surface defects and struggles to detect inner-wall defects in thick-walled components, electromagnetic coupling sensing utilizes medium-to-low frequency excitation to effectively penetrate stainless steel plates, detecting both embedded defects on the top surface and micro-defects on the bottom surface (Hu et al., 2014). In comparison with multimodal sensing NDT, which often requires time-division or frequency-division excitation and suffers from signal crosstalk and high system complexity, electromagnetic coupling sensing achieves high-resolution simultaneous detection of inner- and outer-wall defects with a single, simplified probe configuration, offering minimal scanning blind zones and higher integration. In summary, electromagnetic coupling sensing inspection technology holds certain advantages in detecting various defects on both the internal and external walls of metal parts. It possesses potential industrial application value in assessing the integrity and reliability of metal structures (Li et al., 2006; Feng et al., 2016; Wu et al., 2017; Khan et al., 2021).

The energy equipment play a vital role in global energy infrastructure, and internal inspection technology serves as a critical tool for assessing metal parts integrity (Feng et al., 2018; Matthews et al., 2018; Ege and Coramik, 2018). However, it faces numerous detection challenges (Tianhan and Shizhong, 2024; Yang et al., 2022; Long et al., 2022). This paper proposes an internal inspection system based on electromagnetic coupling sensing and a permanent magnet composite excitation architecture. First, the proposed excitation structure generates an axial magnetic leakage field and a circumferential uniform alternating current field inside the metal parts. Second, by superimposing a permanent magnet magnetization structure, the excitation configuration is optimized to enhance wall penetration capability and magnetic flux intensity. Meanwhile, while ensuring the safe passage of the internal inspection tool through metal parts, it overcomes the limitations of single-technology approaches in detecting multiple types of defects, thereby improving the detection capability for both near-surface and deeply buried defects.

2 Inspection theory based on physical-level electromagnetic fusion mechanisms

2.1 Magnetic ring-based metal inspection systems

Currently, metal inspection faces challenges such as limitations in sensor detection capabilities, energy consumption, and pass-through performance. This paper proposes an internal inspection system based on an electromagnetic coupling mechanism. The probe excitation unit consists of a coil, high-permeability silicon steel material, and permanent magnets, as shown in Figure 1. To address the issue of mutual interference between channels, an ADC module is employed to convert the analog signals output by the magnetic sensors into digital signals. Simultaneously, an FPGA is used to control signal generation and data storage, while a battery unit powers the normal operation of the detector. A support system is utilized to balance the operational state of the detector, ensuring that each inspection unit maintains an equal distance from the metal wall and guaranteeing consistency in inspection results. The proposed sensing structure can simultaneously generate a uniform alternating current field and a magnetic leakage field on the metal wall, enabling a single excitation structure to achieve both physical effects. This enhances the detection and discrimination capabilities for different types of defects.

FIGURE 1

A schematic diagram of the detection principle for the in-line inspection probe proposed in this paper is shown in Figure 2. An FPGA is used to generate a low-frequency sinusoidal signal with adjustable amplitude and frequency. This signal is amplified by a power amplifier to drive the probe. The excitation structure simultaneously generates a strong axial magnetization field and a circumferential uniform eddy current field. The axial magnetization field produced at both ends of the magnetic yoke magnetizes the metal wall. Due to the nonlinear relationship between the magnetization field and magnetic permeability, the penetration depth of the electromagnetic field can be increased by reducing the magnetic permeability of the metal wall, in accordance with the principle of skin depth. Concurrently, the coil acts directly on the metal wall surface, inducing a circumferential uniform eddy current field, which enables the detection of axial defects and enhances the ability to detect multiple types of defects. The configuration of permanent magnets with the low-frequency excitation structure further increases the magnetic field strength and the magnetization capability of the metal without additional power consumption (Tianhan and Shizhong, 2024). Moreover, the magnetization intensity generated by the low-frequency excitation is adjustable. It can be modified according to inspection needs, such as adjusting the magnetization intensity for different metal wall thicknesses. In this paper, the superposition of permanent magnets and the excitation structure enhances the magnetization intensity, thereby improving the detection capability for subsurface defects.

FIGURE 2

2.2 Electromagnetic coupling detection theory under multi-magnetic excitation architecture

2.2.1 Magnetic circuit analysis of electromagnetic coupling

Considering the complexity of analyzing the mechanism of the composite excitation structure, a magnetic circuit model under the multi-magnetic excitation structure is established to qualitatively analyze the distribution patterns of electromagnetic and magnetic flux leakage information in defect areas (Yang et al., 2022; Long et al., 2022; Sun et al., 2021). This model simplifies the analysis process of the detection mechanism. Generally, the electrical conductivity of metal specimens is approximately 10⁵ to 10⁶ times higher than that of air and is not easily affected by external magnetic fields. However, due to the nonlinear characteristics of the magnetic permeability of ferromagnetic materials, external magnetic fields magnetizing the specimen can cause changes in its magnetic permeability, affecting the physical properties of the specimen, which is further reflected in the defect detection signals. Typically, the magnetic permeability of carbon steel metal materials is about 10² times higher than that of air. When an external magnetic field is applied to the metal wall, the concentration of a large amount of magnetic flux in the metal wall region leads to magnetic saturation. According to B-H curve analysis, once the specimen enters saturation, further increases in the external magnetic field do not increase the magnetic saturation intensity but cause some magnetic flux to leak into the air. Additionally, the magnetic circuit contains not only dynamic induced magnetic flux but also static DC magnetic flux generated by the permanent magnets. The specific magnetic circuit model is shown in Figure 3.

FIGURE 3

F₀ is the magnetomotive force generated by the low-frequency excitation coil, serving as the excitation source for both the magnetization field and the induced field. F₁ is the magnetomotive force produced by the permanent magnet, which is typically considered constant, and its direction aligns with that of F₀. The reason is that the excitation coil generates an alternating excitation output based on forward bias, ensuring the output magnetic field remains a forward-alternating field (Cacciola et al., 2010). Its characteristic lies in compensating for the attenuation of wall magnetization intensity caused by the addition of steel brushes through increased magnetic flux intensity (Tian and Sophian, 2005). The magnetic flux φ is calculated by Formula 1, where R_y represents the magnetic reluctance of the yoke (indicated by the blue region). R_m is the magnetic reluctance of the permanent magnet (pink region). R_n denotes the magnetic reluctance of the defect-free region in the test specimen. R_b is the magnetic reluctance of the steel brush (light gray region), generally considered constant and corresponding to the probe lift-off value, calculated using Formula 2.

Here, μ_b is the magnetic permeability of the steel brush, S_b is the cross-sectional area of the steel brush, and l is the length of the steel brush. R_mfl represents the magnetic reluctance in the defect region of the specimen, which varies depending on the type and size of the defect. R_dis refers to the magnetic leakage reluctance, primarily influenced by magnetization intensity. As shown in Formula 3, R_d denotes the change in magnetic reluctance caused by distortions in the electromagnetic flow due to defects, with maximum disturbance occurring when the current direction is perpendicular to the defect, leading to an increase in magnetic reluctance; conversely, magnetic reluctance is smaller under other conditions (Sophian et al., 2017). R_c represents the compressive magnetic reluctance in the specimen, mainly affected by F₀ and F₁. The magnetic reluctance in the defect region can be expressed by Formula 4.

The magnetic flux φ can be expanded and calculated using Formula 5. F_ec is the electromagnetic field disturbance potential caused by defects, representing a secondary induced electromagnetic field that acts as a demagnetizing force, with its direction opposite to that of F₀. R_y, R_m, and R_n are typically treated as constants.

2.2.2 Electromagnetic model of defects in the semi-magnetized state

An important factor affecting detection performance is the magnetizing capability of the probe and the depth of magnetic field penetration. Both over-saturation and under-magnetization generally fail to meet detection requirements. On one hand, over-saturation causes the magnetic field leaking from inside the metal to be compressed back into the metal by the original magnetic field, simultaneously affecting the detection range of the receiving sensor, thereby impacting the probe’s detection sensitivity. On the other hand, under-magnetization results in a lower external magnetization intensity on the metal wall, leading to minimal magnetic flux leakage, and the weak leakage field is difficult for the sensor to capture.

As shown in Figure 4, μ_max represents the maximum magnetic permeability of the test specimen. μ₁ and μ₂ are the magnetic permeabilities of the defective area and the non-defective area, respectively. B₁ and B₂ are the magnetic flux densities of the defective and non-defective areas calculated by Formula 6, respectively. B_sat is the initial magnetic flux density of the specimen in a saturated magnetization state.

FIGURE 4

When the applied magnetic field H exceeds the field value corresponding to μ_max, the sensing area of the specimen enters a semi-saturated magnetization state. In this stage, the gradients of magnetic permeability and magnetization intensity are relatively large. When encountering a defective area, both change rapidly, making it easier to generate perturbations in the magnetic field. This further enhances the receiving sensor’s ability to capture defect information. The vicinity of the maximum gradient is suitable for setting the optimal external magnetic field range. Analysis reveals that the point of maximum slope can be designated as a pseudo-cross point, serving as a criterion for determining the magnetization intensity range for the optimal semi-saturated magnetization state. Therefore, assuming the magnetic permeability of the non-defective area is μ₁, when a defect occurs in the specimen, the magnetic reluctance R_d decreases, and the corresponding magnetic flux increases. Simultaneously, as the magnetic permeability in this area drops to μ₂, the magnetic flux density increases, forcing more magnetic flux to leak into the air, thereby enhancing the probe’s detection capability.

Furthermore, based on the optimal configuration state of semi-saturated magnetization, considering the spatial constraints of small-diameter metals, passability can be improved by adding steel brushes. However, adding steel brushes effectively increases the probe’s lift-off, which reduces the magnetization intensity of the excitation field applied to the metal, making it more difficult to detect deeper subsurface defects. As shown in Figure 5, the traditional method for increasing the penetration depth rate is based on Equation 7. If the excitation frequency is reduced to enhance penetration capability, it leads to increased power requirements for the excitation source of the detection system, raising power consumption, and consequently reducing the operational runtime of the in-line inspection system. Therefore, considering that the electrical conductivity of the specimen cannot be changed, the loss of magnetization intensity in the specimen caused by the increased lift-off is compensated by adding permanent magnets. Using the pseudo-cross point as a reference, the appropriate magnetic field strength of the permanent magnets is determined, converting the original standard penetration depth δ₁ to a new depth δ₂.

FIGURE 5

The magnetic circuit model clearly expresses the magnetic flux distribution in the defect area. This paper establishes a magnetic dipole model to analyze the MFL information at the defect. When the magnetization direction is perpendicular to the defect, according to the magnetic dipole model, assuming the defect length is infinite, the magnetic charge is mainly uniformly distributed on both sides of the defect wall, and the linear magnetic charge density is set as ρ. The magnetic polarities on the two walls of the defect are opposite. The magnetic charge wall can be divided into countless segments, each part having a length of dγ. Therefore, according to Equation 8, the magnetic density dH at point p (x, y) is generated by a pair of magnetic charges dγ.

Furthermore, by integrating the magnetic charges on the defect wall and superposing the integrated magnetic charges at point p (x, y), the magnetic field components H_x and H_z are shown in Equation 9 and Equation 10, respectively.

When a crack occurs, and d ≥ ω, H_z can be simply described as Equation 11. Particularly, when neglecting the influence of the magnetic field on the specimen, ρ = 1, and only considering the magnetic field distribution, it is as shown in Equation 12.

H_a represents the internal magnetic field strength of the specimen, without considering the influence of magnetic field density on material properties. On the other hand, when the defect is parallel to the magnetization direction, the coil directly induces an alternating electromagnetic field inside the metal. The circumferential eddy current flow direction is obstructed by the defect. Considering the skin effect of the alternating magnetic field, eddy currents gather on the surface of the specimen and concentrate at both ends of the defect, while part of the eddy current flows along the bottom of the defect or takes a detour. According to the Biot-Savart law, the surface eddy current can be regarded as several current lines with width dx, which can be considered as a long straight line carrying current I, where I=F_ec/R_dis. As shown in Equation 13.l is the height of the eddy current starting from the test point x. m is the width of the area covered by the eddy current. Meanwhile, as the eddy current flows around both ends of the defect, a curvature of the eddy current bending is generated on both sides of the defect. If the radian of the micro-eddy current is Idr and the radius is r, then at any point in space, the superposition law of electromagnetic fields holds true. The generated magnetic density B_z is:

Therefore, combining Equations 9, 10, 13, 14, the superposition of the electromagnetic field vectors at any point in space is given by Equation 15 and Equation 16.

3 Finite element modeling and simulation analysis of electromagnetic coupling detection

To investigate the vector distribution characteristics of electromagnetic fields for surface and subsurface defects under different excitation conditions, a finite element method (FEM) was employed to establish a dynamic-static magnetic field simulation model for electromagnetic coupling detection. By analyzing the relationship between electromagnetic components and defect signals under different excitation configurations, the distribution patterns of electromagnetic field vectors around defects were studied to further validate the detection capability and reliability of the proposed method. Due to the complexity of the MRS (Magnetic Rope System) and metal structural models, only one unit of the MRS was selected for simulation verification to simplify computation time. Considering the geometric symmetry of the model, a two-dimensional model was adopted to meet the requirements of simulation analysis.

Table 1 lists the relevant geometric parameters and physical properties of the simulation model. The simulation model corresponding to Table 1 is shown in Figure 2, and the specific materials and positions of each component are presented in Figure 2. The basic components of this simulation model are the magnetic yoke, excitation coil, and permanent magnet. The proposed sensing structure simulation model includes both internal and external surface defects of the test specimen. In the physical field settings, both magnetic field and magnetic field without current coupling were established. Steady-state and frequency-domain coupling analyses were conducted to study the distribution of electromagnetic fields. The excitation coil having 520 turns was set to a current of 0.5 A with the frequency of 100 Hz. Under different excitation frequencies, the proposed probe and permanent magnet excitation configuration significantly affect the distribution of magnetization intensity and eddy current density. The parametric sweep method is adopted to determine the excitation coil frequency and number of turns. The optimization objective is to balance the detection sensitivity for both inner-wall and outer-wall defects while ensuring sufficient penetration depth to cover the inner wall. For each combination, the change in magnetic flux density and the signal-to-noise ratio in the defect region are calculated simultaneously. Finally, the actual detection performance under the optimal parameters is verified through simulation. By studying the electromagnetic field distribution inside the specimen under different excitation frequencies, simulations were conducted for three configurations: MRS probe excitation alone, permanent magnet excitation alone, and combined excitation of both. The trends of magnetic leakage field and alternating field intensity in surface and subsurface defect regions were analyzed to determine the optimal excitation frequency for optimizing the excitation structure configuration and enhancing defect detection capability (Shin et al., 2009).

When detecting surface defects, the magnetic field intensity exhibits a declining trend as the frequency increases. According to Faraday’s law, alternating excitation electric fields generate alternating magnetic fields acting on the specimen surface. The magnetic field produced by eddy currents opposes the direction of the primary magnetic field. Therefore, the displayed eddy current vector direction is negative, with positive and negative values mainly representing vector direction. The trend shows an initial rapid increase, followed by a slower increase, gradually stabilizing. Simultaneously, based on the magnetization and penetration principles of ferromagnetic materials, magnetic field intensity decreases as frequency increases. During specimen magnetization, changes in magnetic permeability with increasing magnetic field intensity lead to an initial decrease and subsequent increase in eddy current density. Higher frequencies are unfavorable for detecting subsurface defects.

To enhance the detection capability of the proposed probe configuration for both surface and subsurface defects and based on experimental validation, an excitation frequency of 100 Hz was selected to meet detection requirements. The optimal excitation configuration is characterized by high magnetic field intensity and eddy current density, effectively meeting the detection needs for various defect types. Furthermore, to optimize the probe’s detection performance, analyzing the optimal position of the receiver sensor is crucial for improving detection sensitivity. By evaluating the distribution of magnetic field intensity and eddy current density at different detection points, the region with minimal signal fluctuation was identified as the ideal sensor detection position. This area is less susceptible to background magnetic field interference, significantly enhancing defect detection sensitivity and reliability.

4 Simulation results analysis and discussion

The study found that using a configuration of permanent magnets and the MRS structure enhances the capture capability of magnetic leakage and electromagnetic disturbance information at defect locations under composite excitation.

4.1 Magnetic field distribution for different types of defects

Figures 6–8 present the magnetic flux cloud distribution under different sensing configurations. The color intensity in the cloud maps corresponds to magnetic field intensity. By comparing subsurface defects #1 and #2 with different burial depths, Figure 6 shows that when only the MRS acts on the specimen, more magnetic flux and disturbance fields concentrate on the specimen’s surface for defects with shallow burial depths. Larger penetration depths lead to more divergent magnetic field distributions, making subsurface defects harder to detect. Additionally, the disturbance field direction opposes the primary magnetic field, consistent with theoretical analysis. As shown in Figure 7, when only permanent magnets are applied, the magnetic field exhibits higher penetration capability, reaching subsurface defects but with weaker intensity. Figure 8 demonstrates that under combined MRS and permanent magnet excitation, substantial magnetic fields concentrate in regions above subsurface defects, and magnetic flux lines bend around defect areas. This configuration enhances specimen magnetization intensity, improves magnetic field penetration capability, increases the intensity of disturbance electromagnetic fields at defect locations, and enhances the probe’s detection capability and sensitivity.

FIGURE 6

FIGURE 7

FIGURE 8

4.2 Simulation results for different types of defects

Following the analysis of the magnetic field distribution patterns of the probe structure in surface and subsurface defect areas, to more intuitively verify the detection capability of the probe, different defect types on the specimen were scanned, and signal variations and characteristics were utilized to analyze the defect conditions. In the simulation analysis, based on the defect detection results under different configurations shown in Figure 9, when only the MRS configuration is used, it exhibits high sensitivity to surface defect #1 but struggles to identify subsurface defects. When only permanent magnet excitation is applied, it demonstrates better detection capability for subsurface defects, but its sensitivity is limited for detecting shallow surface defects. According to the detection results in Figure 10, it is evident that when the proposed MRS and magnet configuration act jointly on the specimen, it exhibits good sensitivity for both surface and subsurface defects. The amplitude of magnetic field intensity fluctuations for surface defect detection is greater than that of subsurface defects, and the amplitude of fluctuation can reach 0.011 T, which is only 0.008 T compared to the test results of subsurface defects. By comparison, the amplitude of fluctuation for the probe structure using permanent magnets alone is only 0.002 T. Moreover, as the burial depth of subsurface defects increases, the signal variation shows a decaying trend. This is due to the increased penetration depth of the magnetic field, which leads to a divergent distribution and reduced intensity of the electromagnetic field, thereby affecting the detection signal strength of the defects.

FIGURE 9

FIGURE 10

4.3 Experimental validation of internal detection for metal defects

To validate the defect detection capability of the proposed MRS probe structure in practical environments, pull-through tests were conducted on actual metals. In this experiment, four repeated pull-through tests were performed on the metal to verify the reliability of the internal detector for metal defect detection. Experiments were carried out under two configurations of the MRS probe: with and without the effect of permanent magnets, corresponding to the simulation results for different types of defects. The detection results are shown in Figure 11. A comparison of the detection results in Figures 11a,b reveals that when permanent magnets are added to the MRS, subsurface defects can be effectively detected. On the other hand, without the use of permanent magnets, only defects with shallow burial depths can be detected, while deeper subsurface defects cannot be effectively identified. The results indicate that, in the absence of permanent magnets, the magnetization intensity of the excitation structure on the specimen does not reach the optimal magnetization state, preventing the magnetic field from penetrating deeper defect regions and making it difficult to detect deeper defects. Considering the need for higher penetration depth, a structural configuration combining MRS and permanent magnets was employed. By superimposing the static magnetic field generated by the permanent magnets, the penetration depth and intensity of the magnetic field were increased, and the amplitude of the magnetic field strength of the deeper defect can display a fluctuation of 0.2 T.

FIGURE 11

4.4 Analysis of detection capabilities for different types of defects

To further verify the detection capability of the electromagnetic coupling detection system for different types of defects as shown in Figure 12, the defect information shown in Table 2 was extracted. Metal specimens 1 and 2 include surface and subsurface defects, respectively; among them, there are four surface defects of different shapes and subsurface defects with varying burial depths. Constructing grayscale images can effectively represent all defect information within the metal, enabling intuitive detection and analysis of these defect features. The test results are shown in Figures 13, 14. When only MRS was functioning, subsurface defects D1-4 and D1-5 could not be detected, indicating that the defect information was susceptible to background noise. Meanwhile, experimental data for circumferential defect D1-2 and axial defect D1-6 demonstrate that the electromagnetic coupling detection method can effectively optimize detection results. When MRS was applied without the permanent magnet, using grayscale imaging effectively avoided missed detections of D2-2, D2-3, D2-4, and D2-5 caused by the complex metal environment, as shown in Figure 14. With the addition of the permanent magnet, the grayscale imaging effect for defects was significantly enhanced. Furthermore, both circumferential and axial defects could be detected, verifying the distinct advantage of electromagnetic coupling detection that it is unaffected by defect orientation, and effectively identifying three different types of defects: D2-1, D2-4, and D2-5.

FIGURE 12

FIGURE 13

FIGURE 14

In the configuration without the permanent magnet and MRS, the signal fluctuations at these defect locations were relatively weak and easily influenced by background noise as shown in Table 3, which could lead to inaccurate judgments regarding defect detection and result in misinterpretation. The sensitivity results in Table 3 are determined based on the maximum amplitude of changes in the test magnetic field strength. For the detection of D1 type defects, the sensitivity range of detection technology without using permanent magnets is between 0.24 and 0.36. The use of permanent magnets can increase sensitivity to 0.41∼0.77. D2 type defects have lower sensitivity caused by the complex metal environment, and the use of permanent magnet technology can increase sensitivity by more than 0.3. When both the permanent magnet and MRS were applied simultaneously, the static magnetic field generated by the permanent magnet superimposed with the axial alternating magnetization field produced by MRS, jointly acting on the metal to enhance its magnetization intensity. Based on the detection results, it was found that the sensors of the electromagnetic coupling detection system were able to detect all defects in metal specimens 1 and 2.

5 Conclusion

This study proposes a new architecture based on electromagnetic coupling sensing combined with permanent magnet composite excitation. Its core innovations and substantive contributions are as follows: It simultaneously generates an axial leakage magnetic field and a circumferentially uniform alternating current field, achieving synergistic excitation of static magnetic field and alternating electromagnetic field, thereby overcoming the limitations of single-technology approaches for detecting multiple defect types. By superimposing a permanent magnet magnetisation structure and optimising its configuration, the magnetic flux density and penetration depth into the pipe wall are significantly increased, particularly improving detection sensitivity for shallow-surface and buried defects. This study provides a new multi-physics synergistic excitation solution for the in-line inspection of metallic components, effectively extending the detectable depth range and achieving a comprehensive improvement in detection capability—from a single defect type to surface, subsurface, and buried defects.

6 Future work

References

• 1

  BungeH. (1989). J. Texture and magnetic properties. Textures Microstructures11, 75–91. 10.1155/tsm.11.75

  • CrossRef
  • Google Scholar
  • View reference in article

• 2

  CacciolaM.MegaliG.PellicanoD.CalcagnoS.VersaciM.MorabitoF. C. (2010). Modelling and validating Ferrite-core probes for GMR-eddy current testing in metallic plates. Piers Online6 (3), 237–241. 10.2529/piers090910070900

  • CrossRef
  • Google Scholar
  • View reference in article

• 3

  CharapS. (1974). Magnetic hysteresis model. IEEE Trans. Magnetics10 (4), 1091–1096. 10.1109/tmag.1974.1058525

  • CrossRef
  • Google Scholar
  • View reference in article

• 4

  ChenW.WuD. (2023). Resistance-frequency eddy current method for electrical conductivity measurement. Measurement209, 112501. 10.1016/j.measurement.2023.112501

  • CrossRef
  • Google Scholar
  • View reference in article

• 5

  ChengY.ZhaoY. N.QiH. T.ZhouX. (2024). Intelligent optimal design of steel-concrete hybrid wind turbine tower based on evolutionary algorithm. J. Constr. Steel Res.218, 108729. 10.1016/j.jcsr.2024.108729

  • CrossRef
  • Google Scholar
  • View reference in article

• 6

  ChuY.-J.LamH.-F. (2020). Comparative study of the performances of a bio-inspired flexible-bladed wind turbine and a rigid-bladed wind turbine in centimeter-scale. Energy213, 118835. 10.1016/j.energy.2020.118835

  • CrossRef
  • Google Scholar
  • View reference in article

• 7

  Chun-LungC.Tsung-YingL. (2017). “Impact analysis of transmission capacity constraints on wind power penetration and production cost in generation dispatch,” in IEEE International Conference on Intelligent Systems Applications to Power Systems, Taiwan.

  • Google Scholar
  • View reference in article

• 8

  DasS.Mohamed SajeerM.ChakrabortyA. (2019). Vibration control of horizontal axis offshore wind turbine blade using SMA stiffener. Smart Mater. Struct.28 (9), 095025. 10.1088/1361-665x/ab1174

  • CrossRef
  • Google Scholar
  • View reference in article

• 9

  DuttaS. M.GhorbelF. H.StanleyR. K. (2009). Simulation and analysis of 3-D magnetic flux leakage. IEEE Trans. Magnetics45 (4), 1966–1972. 10.1109/tmag.2008.2011896

  • CrossRef
  • Google Scholar
  • View reference in article

• 10

  EgeY.CoramikM. (2018). A new measurement system using magnetic flux leakage method in pipeline inspection. Measurement123, 163–174. 10.1016/j.measurement.2018.03.064

  • CrossRef
  • Google Scholar
  • View reference in article

• 11

  FengB.KangY.SunY.YangY.YanX. (2016). Influence of motion induced eddy current on the magnetization of steel pipe and MFL signal. Int. J. Appl. Electromagn. Mech.52 (1-2), 357–362. 10.3233/jae-162076

  • CrossRef
  • Google Scholar
  • View reference in article

• 12

  FengB.RibeiroA. L.RochaT. J.RamosH. G. (2018). Comparison of inspecting non-ferromagnetic and ferromagnetic metals using velocity induced eddy current probe. Sensors18 (10), 3199. 10.3390/s18103199

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 13

  GloriaN. B. S.AreizaM. C. L.MirandaI. V. J.RebelloJ. (2009). Development of a magnetic sensor for detection and sizing of internal pipeline corrosion defects. NDT and E Int.42 (8), 669–677. 10.1016/j.ndteint.2009.06.009

  • CrossRef
  • Google Scholar
  • View reference in article

• 14

  HerringR.DyerK.MartinF.WardC. (2019). The increasing importance of leading edge erosion and a review of existing protection solutions. Renew. Sustain. Energy Rev.115, 109382. 10.1016/j.rser.2019.109382

  • CrossRef
  • Google Scholar
  • View reference in article

• 15

  HuY.ZhangW.LiuQ. H. (2014). Multiphysics Coupling of Dynamic Fluid Flow and Electromagnetic Fields for Subsurface Sensing. IEEE.

  • Google Scholar
  • View reference in article

• 16

  HwangJ.LordW. (1975). Finite element modeling of magnetic field/defect interactions. ASTM J. Test. Eval.3 (1), 21–25. 10.1520/jte10129j

  • CrossRef
  • Google Scholar
  • View reference in article

• 17

  IluengU.ShanS.FoutchD. A.ShanS.FoutchD. A. (1997). Magnetic flux leakage (MFL) technology for natural gas pipeline inspection. Ndt and E Int.30 (1), 36.

  • Google Scholar
  • View reference in article

• 18

  JilesD. C.AthertonD. L. (1986). Theory of ferromagnetic hysteresis. J. Mag. Mag. Mater61, 48–60. 10.1016/0304-8853(86)90066-1

  • CrossRef
  • Google Scholar
  • View reference in article

• 19

  JilesD. C.MelikhovY. (2007). “Modelling of nonlinear behaviour and hysteresis in magnetic materials,” in Handbook of Magnetism and Advanced Magnetic Materials.

  • Google Scholar
  • View reference in article

• 20

  KayaM. N.KoseF.InghamD.MaL.PourkashanianM. (2018). Aerodynamic performance of a horizontal axis wind turbine with forward and backward swept blades. J. Wind Eng. Industrial Aerodynamics176, 166–173. 10.1016/j.jweia.2018.03.023

  • CrossRef
  • Google Scholar
  • View reference in article

• 21

  KhanM. A.SunJ.LiB.PrzybyszA.KoselJ. (2021). Magnetic sensors-A review and recent technologies. Eng. Res. Express3 (2), 022005. 10.1088/2631-8695/ac0838

  • CrossRef
  • Google Scholar
  • View reference in article

• 22

  KleinF.MarxS. (2023). Influence of adjacent segments on the torsional load-bearing behaviour of assembled half-shell towers. Eng. Struct.280, 115624. 10.1016/j.engstruct.2023.115624

  • CrossRef
  • Google Scholar
  • View reference in article

• 23

  LeonM.Jr (2022). Root causes and mechanisms of failure of wind turbine blades: overview. Mater.15 (9), 2959. 10.3390/ma15092959

  • CrossRef
  • Google Scholar
  • View reference in article

• 24

  LiY.TianG. Y.WardS. (2006). Numerical simulation on magnetic flux leakage evaluation at high speed. NDT and E Int.39 (5), 367–373. 10.1016/j.ndteint.2005.10.006

  • CrossRef
  • Google Scholar
  • View reference in article

• 25

  LongY.ZhangJ.HuangS.PengL.WangW.WangS.et al (2022). A novel crack quantification method for ultra-high-definition magnetic flux leakage detection in pipeline inspection. IEEE Sensors J.22 (16), 16402–16413. 10.1109/jsen.2022.3190684

  • CrossRef
  • Google Scholar
  • View reference in article

• 26

  MartinezA.MurphyL.IglesiasG. (2023). Evolution of offshore wind resources in northern Europe under climate change. Energy269, 126655. 10.1016/j.energy.2023.126655

  • CrossRef
  • Google Scholar
  • View reference in article

• 27

  MatthewsJ. C.NestlerothJ. B.ConditW. (2018). Evaluating emerging structural inspection technologies for high-risk cast iron water mains. Tunn. Undergr. Space Technol.77, 288–294. 10.1016/j.tust.2018.04.014

  • CrossRef
  • Google Scholar
  • View reference in article

• 28

  MishnaevskyL.Jr. (2019). Repair of wind turbine blades: review of methods and related computational mechanics problems. Renew. Energy140, 828–839. 10.1016/j.renene.2019.03.113

  • CrossRef
  • Google Scholar
  • View reference in article

• 29

  NakamuraS. (2012). Study on sensing based on electromagnetic resonant coupling and its applications. J. Soc. Instrum. Control Eng.51 (5), 486. 10.11499/sicejl.51.486

  • CrossRef
  • Google Scholar
  • View reference in article

• 30

  NestlerothJ. B.DavisR. J. (2007). Application of eddy currents induced by permanent magnets for pipeline inspection. NDT and E Int.40 (1), 77–84. 10.1016/j.ndteint.2006.07.002

  • CrossRef
  • Google Scholar
  • View reference in article

• 31

  OteroA. D.PontaF. L. (2018). On the sources of cyclic loads in horizontal-axis wind turbines: the role of blade-section misalignment. Renew. Energy117, 275–286. 10.1016/j.renene.2017.10.049

  • CrossRef
  • Google Scholar
  • View reference in article

• 32

  PatelR. P.KachhwahaS. S.SurisettyV. A. K.Singh KachhwahaS. (2022). A revised offshore wind resource assessment and site selection along the Indian coast using ERA5 near-hub-height wind products. Ocean. Eng.254, 111341. 10.1016/j.oceaneng.2022.111341

  • CrossRef
  • Google Scholar
  • View reference in article

• 33

  RezaeihaA.KalkmanI.BlockenB. (2017). Effect of pitch angle on power performance and aerodynamics of a vertical axis wind turbine. Appl. Energy197, 132–150. 10.1016/j.apenergy.2017.03.128

  • CrossRef
  • Google Scholar
  • View reference in article

• 34

  ShahS.MehtreB. M. (2015). An overview of vulnerability assessment and penetration testing techniques. J. Comput. Virology and Hacking Tech.11 (1), 27–49. 10.1007/s11416-014-0231-x

  • CrossRef
  • Google Scholar
  • View reference in article

• 35

  ShinY. K.ChoiD. M.KimY. J.LeeS. S. (2009). Signal characteristics of differential-pulsed eddy current sensors in the evaluation of plate thickness. NDT and E Int.42 (3), 215–221. 10.1016/j.ndteint.2008.09.006

  • CrossRef
  • Google Scholar
  • View reference in article

• 36

  SophianA.TianG.FanM. (2017). Pulsed eddy current non-destructive testing and evaluation: a review. Chin. J. Mech. Eng.30 (3), 500–514. 10.1007/s10033-017-0122-4

  • CrossRef
  • Google Scholar
  • View reference in article

• 37

  SunH.PengL.HuangS.LiS.LongY.WangS.et al (2021). Development of a physics-informed doubly fed cross-residual deep neural network for high-precision magnetic flux leakage defect size estimation. IEEE Trans. Industrial Inf.18 (3), 1629–1640. 10.1109/tii.2021.3089333

  • CrossRef
  • Google Scholar
  • View reference in article

• 38

  TengC.ZhangW. H.TamK. W.ChoiW. W.HoS. K (2019). Wireless Angular Displacement Sensing System Based on Electromagnetic Coupling of UHF RFID Tag Array. IEEE8892018. 10.1109/RFID-TA.2019.8892018

  • CrossRef
  • Google Scholar
  • View reference in article

• 39

  TianG. Y.SophianA. (2005). Reduction of lift-off effects for pulsed eddy current NDT. NDT and E Int.38 (4), 319–324. 10.1016/j.ndteint.2004.09.007

  • CrossRef
  • Google Scholar
  • View reference in article

• 40

  TianhanY.ShizhongL. (2024). The impact of magnet structures on target magnetic anomaly detection. J. Phys. Conf. Ser.2891 (12), 122009. 10.1088/1742-6596/2891/12/122009

  • CrossRef
  • Google Scholar
  • View reference in article

• 41

  WangY.LiJ.ViehlandD. (2014). Magnetoelectrics for magnetic sensor applications: status, challenges and perspectives. Mater. Today17 (6), 269–275. 10.1016/j.mattod.2014.05.004

  • CrossRef
  • Google Scholar
  • View reference in article

• 42

  WangZ.LuZ.YiW.HaoJ.ChenY. (2023). A study of nonlinear aeroelastic response of a long flexible blade for the horizontal axis wind turbine. Ocean. Eng.279, 113660. 10.1016/j.oceaneng.2023.113660

  • CrossRef
  • Google Scholar
  • View reference in article

• 43

  WuJ.SunY.FengB.KangY. (2017). The effect of motion-induced eddy current on circumferential magnetization in MFL testing for a steel pipe. IEEE Trans. Magnetics53 (7), 1–6. 10.1109/tmag.2017.2655483

  • CrossRef
  • Google Scholar
  • View reference in article

• 44

  YanC.ArcherC. L. (2018). Assessing compressibility effects on the performance of large horizontal-axis wind turbines. Appl. Energy212 (15), 33–45. 10.1016/j.apenergy.2017.12.020

  • CrossRef
  • Google Scholar

• 45

  YangL.LiJ.ZhengW.LiuB. (2022). Simulation and design of a balanced-field electromagnetic technique sensor for crack detection in long-distance oil and gas pipelines. Energies15 (14), 5274. 10.3390/en15145274

  • CrossRef
  • Google Scholar

• 46

  ZatsepinN. N. (1966). Calculation of the magnetostatic field of surface defects. Defektoskopiya2, 385–399.

  • Google Scholar

• 47

  ZhaoZ.JiangR.FengJ.LiuH.WangT.ShenW.et al (2022). Researches on vortex generators applied to wind turbines: a review. Ocean. Eng.253, 111266. 10.1016/j.oceaneng.2022.111266

  • CrossRef
  • Google Scholar

• 48

  ZhukovaV.Corte-LeonP.IpatovM.BlancoJ. M.Gonzalez-LegarretaL. (2019). Development of magnetic microwires for magnetic sensor applications. Sensors19 (21), 4767. 10.3390/s19214767

  • Pubmed Abstract
  • CrossRef
  • Google Scholar