An alternating current magnetic flux leakage based thick coating metal crack detection method and system

Abstract

The application discloses a kind of based on alternating current magnetic flux leakage thick coating metal crack detection method and system, belong to nondestructive testing technical field.Method includes obtaining the detection working condition parameter of metal component to be detected;According to the detection working condition parameter and the preset alternating current magnetic flux leakage response relationship, determine target detection parameter combination, the alternating current magnetic flux leakage response relationship is used to characterize the corresponding relationship between the magnetic flux leakage signal characteristics under different detection working condition parameters and different detection parameter combinations;According to the target detection parameter combination, the alternating current magnetic field is applied to the metal component to be detected, and the magnetic flux leakage detection signal generated outside thick coating of the metal component to be detected is collected;From the magnetic flux leakage detection signal, extract crack feature data related to the magnetic field distortion caused by crack;According to the crack feature data, determine the crack detection result of the metal component to be detected.The application can improve the adaptability of alternating current magnetic flux leakage detection parameter under thick coating condition and the stability of crack feature extraction.

Description

Technical Field

[0001] This invention relates to the field of nondestructive testing technology, and in particular to a method and system for detecting cracks in thick-coated metals based on AC leakage magnetic flux. Background Technology

[0002] In marine engineering equipment, petrochemical pipelines, pressure vessels, aerospace structural components, and other metal components that are exposed to corrosive, humid, high-temperature, or salt spray environments for extended periods, the surface of the metal substrate is typically coated with anti-corrosion coatings, thermal insulation layers, heat insulation layers, or other protective layers. While these coatings improve the corrosion resistance and environmental adaptability of the metal components, they can also create obstacles during non-destructive testing, making it difficult to detect cracks and defects on or near the surface of the metal substrate in a timely manner.

[0003] Existing non-destructive testing methods for metal cracks mainly include ultrasonic testing, eddy current testing, magnetic particle testing, and magnetic flux leakage testing. For metal components without a thick coating, these methods can identify cracks to a certain extent. However, when the metal component is covered with a thick coating, the ultrasonic signal is easily attenuated and scattered at the coating interface, resulting in a weaker echo signal. Eddy current testing is affected by the skin effect, limiting its detection depth and penetration capability. Magnetic particle testing usually requires good accessibility of the surface being tested and good magnetic particle adhesion conditions, making it difficult to adapt to non-magnetic thick coatings or complex field environments. Therefore, detecting cracks in the metal substrate without removing the thick coating still presents significant engineering challenges.

[0004] AC magnetic flux leakage testing utilizes an alternating magnetic field applied to metal components. When cracks, pits, or other discontinuous defects exist in the metal substrate, the magnetic field distribution near the defect area is distorted, forming a magnetic flux leakage signal on the outer side of the coating that can be collected by a magnetic sensor. Compared with some traditional testing methods, AC magnetic flux leakage testing has a certain coating penetration capability, making it suitable for in-service testing of coated metal components.

[0005] However, under the testing conditions of thick coatings, AC magnetic flux leakage detection still has the following problems: First, as the coating thickness and detection lift-off height increase, the amplitude of the magnetic flux leakage signal attenuates significantly, and the magnetic field distortion characteristics caused by cracks are easily drowned out by background noise; Second, different metal materials, different coating materials, different coating thicknesses, and different detection gaps will cause changes in the optimal AC excitation parameters and the spatial parameters of the magnetic sensor, making it difficult to balance detection sensitivity and stability when fixing detection parameters; Third, during the on-site testing process, factors such as environmental magnetic field interference, coating thickness fluctuations, and sensor posture changes can easily lead to misjudgments or missed judgments; Fourth, existing solutions usually rely on empirical parameters or experimental parameter tuning, lacking a parameter determination mechanism that corresponds the detection condition parameters with the characteristics of the AC magnetic flux leakage signal.

[0006] Therefore, how to adaptively determine the AC leakage magnetic flux detection parameters based on the actual testing conditions of the metal component under thick coating conditions, and extract stable crack features from the weak leakage magnetic flux signal, is a technical problem that needs to be solved in the nondestructive testing of cracks in thick coated metals. Summary of the Invention

[0007] The purpose of this invention is to provide a method for detecting cracks in thick-coated metals based on AC magnetic flux leakage, so as to at least solve the problems of significant attenuation of magnetic flux leakage signal, difficulty in adapting detection parameters to different detection conditions, and insufficient stability of crack feature recognition in existing AC magnetic flux leakage detection under thick coating conditions.

[0008] To achieve the above objectives, the present invention provides a method for detecting cracks in thick-coated metal based on AC leakage magnetic flux, comprising: acquiring detection condition parameters of a metal component to be tested, wherein the detection condition parameters include at least coating parameters of the metal component to be tested; determining a target detection parameter combination that matches the detection condition parameters according to the detection condition parameters and a preset AC leakage magnetic flux response relationship, wherein the AC leakage magnetic flux response relationship is used to characterize the correspondence between leakage magnetic flux signal characteristics under different detection condition parameters and different detection parameter combinations, and the target detection parameter combination includes AC excitation parameters and magnetic sensor spatial parameters; applying an AC magnetic field to the metal component to be tested according to the target detection parameter combination, and acquiring leakage magnetic flux detection signals generated by the metal component to be tested on the outer side of the thick coating; extracting crack feature data related to magnetic field distortion caused by cracks from the leakage magnetic flux detection signals; and determining the crack detection result of the metal component to be tested based on the crack feature data.

[0009] Furthermore, the testing parameters also include at least one of the material parameters of the metal component to be tested and the structural parameters of the area to be tested; the coating parameters include at least one of the coating thickness, coating material type, and coating medium properties.

[0010] Furthermore, the AC leakage magnetic response relationship is obtained in the following way: an AC leakage magnetic response model is established through multiphysics coupling theory; multiple candidate detection parameter combinations are set, and leakage magnetic response data corresponding to each candidate detection parameter combination is obtained based on the AC leakage magnetic response model; and a correspondence between detection operating parameters, detection parameter combinations and leakage magnetic signal characteristics is established based on the leakage magnetic response data.

[0011] Furthermore, the AC leakage magnetic field response model includes a metal substrate model, a coating model, an AC magnetic field excitation model, a magnetic sensor acquisition model, and a crack defect response model. The AC leakage magnetic field response model is used to characterize the distribution changes of the leakage magnetic field in the metal substrate, thick coating, and crack defects.

[0012] Furthermore, establishing a correspondence between the detection operating parameters, the combination of detection parameters, and the characteristics of the magnetic flux leakage signal based on the magnetic flux leakage response data includes: determining the magnetic flux leakage penetration evaluation value and the crack distortion evaluation value corresponding to each candidate combination of detection parameters based on the magnetic flux leakage response data; determining the comprehensive evaluation index of each candidate combination of detection parameters based on the magnetic flux leakage penetration evaluation value and the crack distortion evaluation value; and establishing a mapping relationship between the detection operating parameters, the candidate combination of detection parameters, and the comprehensive evaluation index, as at least a part of the AC magnetic flux leakage response relationship.

[0013] Furthermore, the leakage magnetic penetration evaluation value is determined based on at least two of the leakage magnetic response amplitude on the outer side of the thick coating, the reference response amplitude in the crack-free region, and the background noise amplitude; the crack distortion evaluation value is determined based on at least one of the amplitude difference, phase difference, gradient difference, and harmonic energy difference between the cracked region and the crack-free region.

[0014] Furthermore, the AC excitation parameters include at least one of excitation frequency, excitation amplitude, and excitation waveform; the magnetic sensor spatial parameters include at least one of the spacing between adjacent magnetic sensors, magnetic sensor array arrangement, and detection lift-off height.

[0015] Further, the magnetic flux leakage detection signal generated on the outside of the thick coating of the metal component to be tested is collected, and crack feature data related to the magnetic field distortion caused by the crack is extracted from the magnetic flux leakage detection signal, including: collecting multi-channel magnetic flux leakage detection signals at different spatial positions on the outside of the thick coating through multiple magnetic sensor channels; performing inter-channel differential processing on the multi-channel magnetic flux leakage detection signal to obtain differential magnetic flux leakage signal; and extracting at least one of peak feature, gradient feature, phase feature and energy feature from the differential magnetic flux leakage signal to obtain the crack feature data.

[0016] Furthermore, after acquiring the magnetic flux leakage detection signal, the method further includes: determining the real-time signal-to-noise ratio or channel consistency evaluation value based on the magnetic flux leakage detection signal; when the real-time signal-to-noise ratio is lower than a preset signal-to-noise ratio threshold, or the channel consistency evaluation value does not meet a preset consistency condition, correcting at least one of the AC excitation parameter, magnetic sensor spatial parameter, sampling gain, and filtering parameter in the target detection parameter combination.

[0017] This invention also provides a thick-coated metal crack detection system based on AC leakage magnetic flux, comprising: a parameter acquisition module for acquiring detection condition parameters of the metal component to be inspected, the detection condition parameters including at least the coating parameters of the metal component to be inspected; a parameter determination module for determining a target detection parameter combination matching the detection condition parameters according to the detection condition parameters and a preset AC leakage magnetic flux response relationship, the AC leakage magnetic flux response relationship being used to characterize the correspondence between leakage magnetic flux signal characteristics under different detection condition parameters and different detection parameter combinations, the target detection parameter combination including AC excitation parameters and magnetic sensor spatial parameters; a leakage magnetic flux acquisition module for applying an AC magnetic field to the metal component to be inspected according to the target detection parameter combination and acquiring the leakage magnetic flux detection signal generated by the metal component to be inspected on the outer side of the thick coating; a feature extraction module for extracting crack feature data related to magnetic field distortion caused by cracks from the leakage magnetic flux detection signal; and a crack detection module for determining the crack detection result of the metal component to be inspected based on the crack feature data.

[0018] Compared with the prior art, the present invention has at least the following beneficial effects: First, the present invention determines the target detection parameter combination based on the detection condition parameters such as the coating parameters of the metal component to be tested, and in combination with the preset AC leakage magnetic response relationship, so that the AC excitation parameters and the spatial parameters of the magnetic sensor can be matched with the detection conditions of thick coatings, avoiding the reliance on fixed empirical parameters for detection, and improving the adaptability of detection parameters under different coating conditions.

[0019] Secondly, this invention characterizes the correspondence between detection operating parameters, detection parameter combinations and leakage magnetic signal characteristics through AC leakage magnetic response relationship, which can provide a basis for parameter selection before or during detection, so that the target detection parameter combination can take into account both leakage magnetic penetration capability and crack distortion characteristic intensity, thereby improving the detectability of weak leakage magnetic signals on the outside of thick coating.

[0020] Third, this invention extracts crack feature data related to magnetic field distortion caused by cracks from the magnetic flux leakage detection signal, rather than relying solely on a single signal amplitude to determine cracks. This helps to enhance the stability of crack identification results and reduce the impact of coating thickness fluctuations, detection lift-off changes, and background noise on the detection results.

[0021] Fourth, when using a multi-channel magnetic sensor to collect leakage magnetic detection signals, this invention can highlight the spatial response difference between cracked and non-cracked regions through inter-channel differential processing, suppress background interference from multiple channels, and improve the reliability of crack feature extraction.

[0022] Fifth, the present invention can correct the target detection parameter combination based on the real-time signal-to-noise ratio or channel consistency evaluation value, so that the detection process has a certain closed-loop adjustment capability, which is beneficial to improving the stability and repeatability of thick coating metal crack detection in complex field environments. Attached Figure Description

[0023] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention, and the embodiments in the accompanying drawings do not constitute any limitation on the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0024] Figure 1 A schematic flowchart of a method for detecting cracks in thick-coated metal based on AC leakage magnetic flux provided in an embodiment of the present invention; Figure 2 A flowchart illustrating a method for establishing an AC leakage magnetic response relationship according to an embodiment of the present invention; Figure 3 This is a schematic diagram of the ACLM detection principle provided in an embodiment of the present invention; Figure 4 A sectional view of a finite element simulation model provided in an embodiment of the present invention; Figure 5 A typical leakage magnetic field signal feature map provided in an embodiment of the present invention; Figure 6 An optimized detection parameter curve is provided for an embodiment of the present invention; Figure 7 This is a schematic diagram of a multi-channel magnetic flux leakage detection signal differential processing and crack feature extraction process provided in an embodiment of the present invention; Figure 8 This is a flowchart illustrating a target detection parameter combination correction method provided in an embodiment of the present invention; Figure 9 A system performance comparison chart provided for an embodiment of the present invention; Figure 10 This is yet another system performance comparison chart provided in an embodiment of the present invention; Figure 11 This is a schematic diagram of a thick-coated metal crack detection system based on AC leakage magnetic flux, provided as an embodiment of the present invention.

[0025] The realization of the purpose, functional features and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0026] The specific embodiments of the present invention will now be described with reference to the accompanying drawings. It should be understood that the following embodiments are only for illustrating the technical solutions of the present invention and are not intended to limit the scope of protection of the present invention. Where there is no conflict, the technical features in the following embodiments can be combined with each other.

[0027] Example 1 This embodiment provides a method for detecting cracks in thick-coated metals based on AC leakage magnetic flux. This method is applicable to metal components with surfaces covered by anti-corrosion coatings, heat insulation layers, thermal insulation layers, or other non-metallic protective layers, such as offshore platform steel structures, oil and gas pipelines, pressure vessels, spacecraft metal skins, and rocket launch pad support structures. These metal components may develop surface cracks, near-surface cracks, or fatigue cracks during long-term service. Due to the obscuring effect of thick coatings, traditional detection methods struggle to reliably identify crack defects without removing the coating.

[0028] like Figure 1 As shown, the method in this embodiment includes the following steps.

[0029] S101. Obtain the testing condition parameters of the metal component to be tested, wherein the testing condition parameters include at least the coating parameters of the metal component to be tested.

[0030] Specifically, coating parameters may include at least one of the following: coating thickness, coating material type, coating medium properties, coating surface roughness, and coating thickness fluctuation range. Coating thickness can be obtained using a coating thickness gauge, ultrasonic thickness gauge, design drawings, historical test data, or through on-site input. Coating material type may be an epoxy resin coating, anti-corrosion paint layer, thermal insulation layer, heat insulation layer, or other protective layer. Coating medium properties may include relative permeability, dielectric constant, conductivity, or other parameters that can affect the penetration of AC magnetic fields and the propagation of leakage magnetic signals.

[0031] In some embodiments, the detection parameters also include at least one of the material parameters of the metal component to be detected and the structural parameters of the area to be detected. The material parameters may include the type of metal matrix material, magnetic permeability, electrical conductivity, saturation magnetic induction intensity, thickness, and surface condition; the structural parameters of the area to be detected may include the curvature of the detection area, edge location, weld location, stiffener location, or other structural information that may affect the leakage magnetic field distribution.

[0032] S102. Based on the detection operating condition parameters and the preset AC leakage magnetic response relationship, determine the target detection parameter combination that matches the detection operating condition parameters.

[0033] The AC leakage magnetic response relationship is used to characterize the correspondence between leakage magnetic signal characteristics under different detection operating parameters and different combinations of detection parameters. The target detection parameter combination includes AC excitation parameters and magnetic sensor spatial parameters. The AC excitation parameters may include at least one of excitation frequency, excitation amplitude, excitation waveform, and excitation duration; the magnetic sensor spatial parameters may include at least one of the spacing between adjacent magnetic sensors, magnetic sensor array arrangement, detection lift-off height, and scanning step distance.

[0034] In this embodiment, the preset AC leakage magnetic response relationship can be established in advance through simulation modeling, standard specimen experiments, or joint calibration of simulation and experiment, and stored in the testing equipment or host computer. The AC leakage magnetic response relationship can be represented as a parameter mapping table, response surface, regression model, lookup table, neural network model, or other data relationship that can output the combination of target detection parameters according to the detection operating conditions.

[0035] For example, for different testing conditions with coating thicknesses of 5mm, 10mm, 15mm, and 20mm, the correspondence between coating thickness, excitation frequency, sensor spacing, detection lift-off height, and leakage magnetic signal amplitude, crack distortion intensity, and signal-to-noise ratio can be established in advance. During testing, the system determines the target testing parameter combination suitable for the current condition by searching, interpolating, or calculating the AC leakage magnetic response relationship based on the coating thickness and detection lift-off height of the metal component to be tested.

[0036] In a preferred embodiment, when determining the target detection parameter combination based on the detection operating conditions, not only is the maximum amplitude of the leakage magnetic signal considered, but also the penetration capability of the thick coating, the distinguishability of crack distortion, the background noise level, and the spatial resolution. This avoids the increased power consumption, sensor saturation, or decreased spatial resolution caused by simply increasing the excitation amplitude or simply pursuing the signal amplitude.

[0037] S103. Apply an alternating magnetic field to the metal component to be tested according to the target detection parameter combination, and collect the leakage magnetic field detection signal generated by the metal component to be tested on the outside of the thick coating.

[0038] Specifically, such as Figure 3 As shown, under AC excitation, an induced current and an AC magnetic field distributed along a preset direction are formed in the metal matrix of the metal component to be tested. When there are no cracks or defects in the metal matrix, the distribution of the induced current in the detection area is relatively continuous, and the change of the leakage magnetic flux detection signal collected on the outer side of the thick coating is relatively gradual; when there are cracks or defects in the metal matrix, the cracks hinder the continuous passage of the induced current along the original path, causing the induced current to circulate around the crack ends or around the crack, forming current convergence areas and current sparse areas.

[0039] Due to the change in the induced current distribution, the magnetic field components near the crack are distorted accordingly. Specifically, the magnetic field component Bx along the scanning direction or along the workpiece surface exhibits amplitude variations at current convergence and current sparseness points; the magnetic field component Bz perpendicular to the metal component surface forms positive and negative peaks depending on the different current flow directions around the crack. The processing unit can control the magnetic sensor array to acquire the Bx and / or Bz components at different spatial locations outside the thick coating along a preset scanning direction, and use these Bx and / or Bz components as leakage magnetic flux detection signals.

[0040] Furthermore, the processing unit can use the amplitude increase and decrease regions in the Bx component, as well as the positive and negative peak values ​​in the Bz component, as the magnetic field distortion response caused by the crack. By subsequently extracting the peak values, valley values, peak-to-peak values, peak-to-valley spacing, gradient changes, and phase changes in the Bx and / or Bz components, crack characteristic data for determining the crack location and crack size grade can be obtained.

[0041] S104. Extract crack feature data related to magnetic field distortion caused by cracks from the leakage magnetic field detection signal.

[0042] Specifically, the magnetic flux leakage detection signal can be a single-channel signal or a multi-channel magnetic flux leakage detection signal. For multi-channel magnetic flux leakage detection signals, each channel signal can be preprocessed. Preprocessing includes at least one of the following: zero-point correction, baseline drift removal, power frequency interference suppression, bandpass filtering, synchronization demodulation, FFT analysis, or wavelet denoising.

[0043] In one implementation, such as Figure 5 and Figure 7 As shown, multi-channel magnetic flux leakage detection signals are acquired at different spatial locations on the outer side of the thick coating through multiple magnetic sensor channels. The magnetic flux leakage detection signals of adjacent or spaced channels are then subjected to inter-channel differential processing to obtain differential magnetic flux leakage signals. Inter-channel differential processing can be expressed as: subtracting the magnetic flux leakage signal of the (i+1)th magnetic sensor channel from the magnetic flux leakage signal of the ith magnetic sensor channel to obtain the corresponding differential signal. Since environmental magnetic field interference, overall coating thickness variations, and overall changes in detection lift-off typically generate similar common-mode effects in multiple channels, inter-channel differential processing can suppress common-mode background signals and highlight the spatial response differences between cracked and crack-free regions.

[0044] After obtaining the differential leakage magnetic field signal, at least one of the following characteristics—peak features, gradient features, phase features, and energy features—can be extracted from the differential leakage magnetic field signal as crack feature data. Specifically, peak features may include peak value, valley value, peak-to-peak value, peak-valley spacing, and half-maximum width at half maximum (FWHM); gradient features may include maximum gradient value, average gradient value, and gradient abrupt change location; phase features may include the phase abrupt change or phase difference between the cracked region and the non-cracked region; and energy features may include fundamental frequency energy, harmonic energy, harmonic energy ratio, and wavelet decomposition energy. These multiple features can collectively constitute a crack feature vector.

[0045] In a specific example, the crack feature vector can be represented as a data set consisting of the peak-to-peak value, peak-to-valley spacing, maximum gradient value, phase abrupt change, and harmonic energy ratio of the leakage magnetic signal. The system uses this crack feature vector to determine whether a crack exists in the area to be detected, and further estimates the crack location and crack size class.

[0046] S105. Determine the crack detection result of the metal component to be tested based on the crack feature data.

[0047] Specifically, crack detection results may include at least one of the following: crack presence result, crack location result, crack length grade result, crack width grade result, crack depth grade result, and alarm prompt result.

[0048] In one implementation, the system compares crack feature data with a preset crack detection threshold. When at least one of the peak-to-peak value, gradient abrupt change value, phase abrupt change amount, or harmonic energy ratio exceeds the corresponding threshold, and multiple adjacent channels exhibit a peak-valley distribution spatially matching the crack, a crack is determined to exist in the detection area. Further, the crack center location is determined based on the location of the leakage magnetic flux characteristic peak, the crack length direction range is estimated based on the peak-valley spacing or characteristic response width, and the crack size level is determined based on the peak-to-peak value, gradient value, or energy characteristics.

[0049] In another implementation, crack feature data can be input into a pre-trained defect identification model, which outputs the probability of crack presence, crack location, and crack size level. The defect identification model can be a support vector machine, random forest, neural network, convolutional neural network, or other classification and regression model. It should be noted that the defect identification model is only one way to utilize crack feature data. The core of this embodiment lies in matching AC magnetic flux leakage detection parameters according to the thick coating detection conditions and extracting the magnetic field distortion features caused by the crack from the magnetic flux leakage detection signal.

[0050] Example 2 This embodiment describes the process of establishing the AC leakage magnetic flux response relationship. The AC leakage magnetic flux response relationship is used to characterize the correspondence between leakage magnetic flux signal characteristics under different detection operating parameters and different combinations of detection parameters, and is used to determine the target detection parameter combination based on the detection operating parameters of the metal component to be tested during actual testing. This process can be completed before the testing equipment leaves the factory, or it can be completed by a host computer or simulation platform before testing, and it can also be updated during the testing process based on newly added experimental data.

[0051] like Figure 2 As shown, the establishment of the AC leakage magnetic response relationship includes the following steps.

[0052] S201. An AC leakage magnetic response model is established using multi-physics coupling theory.

[0053] Specifically, such as Figure 4 As shown, a three-dimensional AC leakage magnetic response model can be established using finite element simulation software. This model can include a metal substrate model, a coating model, an AC magnetic field excitation model, a magnetic sensor acquisition model, and a crack defect response model.

[0054] The metal matrix model is used to represent the geometry, material type, permeability, conductivity, magnetization properties, and boundary conditions of the sample metal component. For ferromagnetic metal components, a nonlinear magnetization curve can be introduced into the metal matrix model to characterize the change in permeability of the metal matrix under the action of an alternating magnetic field.

[0055] A coating model is used to represent a thick coating covering the surface of a metal substrate. This thick coating can be an anti-corrosion coating, epoxy resin coating, heat insulation layer, thermal insulation layer, or other non-metallic protective layer. Parameters of the coating model may include coating thickness, coating material type, coating medium properties, and coating thickness variation range.

[0056] An AC magnetic field excitation model is used to represent a magnetic yoke, an excitation coil, and an AC excitation signal. The AC excitation signal can be a sine wave, square wave, triangular wave, or modulated wave. The magnetic yoke can be made of a high-permeability material to form the AC magnetic circuit and introduce the AC magnetic field into the metal matrix.

[0057] The magnetic sensor acquisition model represents the spatial arrangement of the magnetic sensor array relative to the thick coating surface, including the number of magnetic sensors, the spacing between adjacent magnetic sensors, the detection lift-off height, the scanning direction, and the sampling position. The crack defect response model represents the crack length, crack width, crack depth, crack direction, and the positional relationship of the crack relative to the magnetic field direction.

[0058] In a specific example, a three-dimensional AC magnetic flux leakage detection model is built using COMSOL Multiphysics. The model includes a U-shaped high-permeability ferrite yoke, a multilayer structure, a rectangular excitation coil, and a crack defect model. Specifically, the relative permeability of the U-shaped high-permeability ferrite yoke can be set to 2000; the multilayer structure includes a carbon steel metal substrate, a 10mm thick epoxy resin coating, and an air domain; the rectangular excitation coil can have 200 turns and a wire diameter of 0.5mm; and the crack defect model can be set to 10mm × 1mm × 5mm.

[0059] To improve the simulation accuracy of the crack region, local mesh refinement can be applied at the crack edge, the coating-air interface, the metal substrate-coating interface, and near the magnetic sensor acquisition location. The minimum element size for the crack region can be set to 0.1 mm. Boundary conditions can be set as magnetically insulating boundaries, infinitely far-field boundaries, or other boundary conditions suitable for AC electromagnetic field calculations. During simulation, leakage magnetic field components at different spatial locations outside the thick coating can be obtained, such as the magnetic field component along the scanning direction, the magnetic field component perpendicular to the component surface, and the corresponding amplitude, phase, and gradient changes.

[0060] S202. Set multiple candidate detection parameter combinations and obtain the leakage magnetic response data corresponding to each candidate detection parameter combination based on the AC leakage magnetic response model.

[0061] Candidate detection parameter combinations can include different excitation frequencies, excitation amplitudes, excitation waveforms, sensor spacing, sensor array arrangement, detection lift-off height, and scanning step distance. To improve parameter search efficiency, candidate detection parameter combinations can be set using methods such as equal-interval parameter scanning, orthogonal experimental design, response surface analysis, or genetic algorithm optimization.

[0062] In a specific example, the excitation frequency is set to 0.1 kHz to 10 kHz with a step size of 0.1 kHz; the sensor spacing is set to 5 mm to 30 mm with a step size of 1 mm; and the detection lift-off height is set to 5 mm to 15 mm with a step size of 1 mm. For each set of candidate detection parameters, the magnetic flux leakage response distribution on the outside of the thick coating is calculated based on the AC magnetic flux leakage response model, and the corresponding magnetic flux leakage response data is extracted.

[0063] The magnetic flux leakage response data may include at least one of the following: magnetic flux leakage response amplitude on the outer side of the thick coating, reference response amplitude in the crack-free region, response amplitude in the cracked region, amplitude difference between the cracked and crack-free regions, phase difference between the cracked and crack-free regions, gradient difference at the crack edge, harmonic energy difference, background noise amplitude, and spatial resolution index.

[0064] In one implementation, magnetic flux leakage response data can be extracted directly above the crack, at the edges of the crack, in a crack-free background area, and at each sampling channel of the sensor array. For example, the peak value of the vertical magnetic field component directly above the crack, the peak-to-valley difference between the crack and its sides, the average response amplitude in the crack-free area, and the noise amplitude in the background area can be extracted. This method allows the magnetic flux leakage response data to reflect not only signal strength but also the spatial distortion characteristics caused by the crack.

[0065] In another implementation, a response record can be generated for each candidate combination of detection parameters. The response record includes at least the detection operating parameters, candidate combination of detection parameters, leakage magnetic response amplitude, phase characteristics, gradient characteristics, energy characteristics, and spatial resolution indicators. Multiple response records constitute an AC leakage magnetic response sample library, providing a data foundation for subsequently establishing AC leakage magnetic response relationships.

[0066] S203. Establish the correspondence between the detection operating condition parameters, the combination of detection parameters and the characteristics of the leakage magnetic signal based on the leakage magnetic response data.

[0067] Specifically, based on the magnetic flux leakage response data, the magnetic flux leakage penetration evaluation value and crack distortion evaluation value corresponding to each candidate detection parameter combination are determined. The magnetic flux leakage penetration evaluation value is used to evaluate the ability of an AC magnetic field to form a measurable magnetic flux leakage signal on the outside of the thick coating after penetrating it; the crack distortion evaluation value is used to evaluate the intensity and identifiability of the disturbance to the magnetic flux leakage distribution caused by crack defects.

[0068] In one embodiment, the magnetic flux leakage penetration evaluation value can be determined based on at least two of the following: the magnetic flux leakage response amplitude outside the thick coating, the reference response amplitude in the crack-free region, and the background noise amplitude. Specifically, the ratio between the magnetic flux leakage response amplitude outside the thick coating and the background noise amplitude can be used as a component of the magnetic flux leakage penetration evaluation value, or the relative change between the magnetic flux leakage response amplitude outside the thick coating and the reference response amplitude in the crack-free region can be used as a component of the magnetic flux leakage penetration evaluation value. When the magnetic flux leakage response amplitude outside the thick coating is large and has high distinguishability relative to the background noise, the magnetic flux leakage penetration evaluation value is high.

[0069] In one embodiment, the crack distortion evaluation value can be determined based on at least one of the amplitude difference, phase difference, gradient difference, and harmonic energy difference between the cracked region and the crack-free region. Specifically, the crack distortion evaluation value can be determined based on the peak-to-peak value of the cracked region, the maximum gradient value at the crack edge, the phase abrupt change of the cracked region relative to the background region, and the harmonic energy change of the cracked region. When the cracked region has obvious peak-to-valley response, gradient abrupt change, phase abrupt change, or frequency domain energy change relative to the crack-free region, the crack distortion evaluation value is higher.

[0070] After determining the magnetic flux leakage penetration evaluation value and the crack distortion evaluation value, a comprehensive evaluation index can be further determined. The comprehensive evaluation index can be a weighted result of the magnetic flux leakage penetration evaluation value, crack distortion evaluation value, background noise evaluation value, and spatial resolution evaluation value. For example, the comprehensive evaluation index can be determined as follows: J = αP + βD + γR - δN; Where J represents the comprehensive evaluation index, P represents the magnetic flux leakage penetration evaluation value, D represents the crack distortion evaluation value, R represents the spatial resolution evaluation value, N represents the background noise evaluation value, and α, β, γ, and δ represent the corresponding weighting coefficients. The above formula is only one optional calculation method. In other embodiments, normalized weighting, threshold screening, ranking optimization, or multi-objective optimization methods can also be used to determine the comprehensive evaluation index.

[0071] The weighting coefficients can be set according to the object being inspected, the coating thickness, the range of crack size of interest, and the required inspection accuracy. For example, when the coating thickness is large, the weight corresponding to the magnetic flux leakage penetration evaluation value can be increased; when it is necessary to identify micro-cracks, the weight corresponding to the crack distortion evaluation value and the spatial resolution evaluation value can be increased; in inspection scenarios with strong environmental electromagnetic interference, the penalty weight corresponding to the background noise evaluation value can be increased.

[0072] When the comprehensive evaluation index meets the preset optimization conditions, the corresponding candidate detection parameter combination is used as the target detection parameter combination or the candidate target detection parameter combination. The preset optimization conditions can be: the comprehensive evaluation index is maximized, the comprehensive evaluation index exceeds a preset threshold, the signal-to-noise ratio exceeds a preset threshold and the spatial resolution meets the requirements, or the comprehensive evaluation index is optimal under the limitations of excitation power consumption, sensor saturation, and detection speed.

[0073] Furthermore, an AC leakage magnetic response relationship can be established based on multiple testing condition parameters, candidate testing parameter combinations, and comprehensive evaluation indicators. This AC leakage magnetic response relationship can be represented as a mapping table, response surface, regression model, lookup table, parameter database, or machine learning model. During actual testing, if the current testing condition parameters match existing testing condition parameters in the sample library, the corresponding target testing parameter combination can be directly called; if the current testing condition parameters are between two or more known testing conditions, the target testing parameter combination can be determined through interpolation; if the current testing condition parameters exceed the range of existing samples, simulation calculations can be performed again, or the AC leakage magnetic response relationship can be supplemented through standard specimen experiments.

[0074] In one implementation, experimental data can also be used to correct the AC magnetic flux leakage response relationship. Specifically, after testing a standard specimen or a known defect area in the field, the actual collected magnetic flux leakage detection signal is compared with the simulated magnetic flux leakage response data. When the deviation between the two exceeds a preset deviation threshold, at least one of the following evaluation indicators—magnetic flux leakage penetration evaluation value, crack distortion evaluation value, background noise evaluation value, or comprehensive evaluation index—is corrected, and the AC magnetic flux leakage response relationship is updated. This method improves the adaptability of the AC magnetic flux leakage response relationship to the actual testing environment.

[0075] In a specific example, parameter scanning analysis revealed that for a carbon steel specimen covered with a 10mm thick epoxy resin coating, a good balance between magnetic flux leakage penetration and spatial resolution could be achieved under conditions of an excitation frequency of 1kHz, a sensor spacing of 20mm, and a detection lift-off height of 10mm. The specific values ​​described above are merely a preferred embodiment and should not be construed as limiting the scope of protection of this invention. Under other coating thicknesses, other materials, or other detection lift-off conditions, different combinations of target detection parameters can be determined based on the AC magnetic flux leakage response relationship.

[0076] Example 3 This embodiment describes the AC magnetic flux leakage detection device and the detection signal acquisition process. The AC magnetic flux leakage detection device in this embodiment is used to apply an AC magnetic field to a metal component covered with a thick coating, according to the target detection parameter combination determined in Embodiment 1 or Embodiment 2, and to acquire magnetic flux leakage detection signals on the outside of the thick coating.

[0077] An AC magnetic flux leakage detection device may include an AC excitation unit, a magnetic yoke, an excitation coil, a magnetic sensor array, a signal conditioning unit, a data acquisition unit, and a processing unit. The AC excitation unit, signal conditioning unit, and data acquisition unit are electrically connected to the processing unit. The processing unit receives the target detection parameter combination and controls the AC excitation unit, magnetic sensor array, signal conditioning unit, and data acquisition unit to work together according to the target detection parameter combination.

[0078] The AC excitation unit may include a signal generator, a power amplifier, and an excitation control circuit. The signal generator generates an AC excitation signal according to the target detection parameters. The power amplifier amplifies the AC excitation signal. The excitation control circuit adjusts the excitation frequency, amplitude, and waveform of the AC excitation signal according to control commands issued by the processing unit. The AC excitation signal can be a sine wave, square wave, triangle wave, swept frequency signal, or modulated wave. In a specific example, the AC excitation unit uses a DDS signal generator or digital-to-analog converter to output a 1kHz sine wave, which is then amplified and applied to the excitation coil. The aforementioned 1kHz is merely a preferred parameter under a 10mm thick epoxy resin coating condition and is not intended to limit the scope of protection of this invention.

[0079] The magnetic yoke can be made of a high-permeability material, such as ferrite, silicon steel sheet, permalloy, or other soft magnetic materials. The yoke guides the alternating magnetic field into the metal component being tested and improves the concentration of the magnetic field distribution within the testing area. The yoke can be a U-shaped, E-shaped, C-shaped, or other structure capable of forming a local magnetic circuit. An excitation coil is wound on the yoke; after an alternating current is applied to the coil, an alternating magnetic circuit is formed between the yoke and the metal component being tested. When the surface of the metal component is covered with a thick coating, the alternating magnetic field penetrates the thick coating and acts on the metal substrate. When cracks or defects exist in the metal substrate, the magnetic field distribution near the crack area is distorted, forming a leakage magnetic field outside the thick coating that can be collected by the magnetic sensor array.

[0080] A magnetic sensor array is disposed on the outer side of the thick coating layer to acquire leakage magnetic field detection signals. The magnetic sensor array can be arranged along the detection scanning direction, perpendicular to the detection scanning direction, or in a two-dimensional array to obtain the spatial leakage magnetic field response distribution of the area to be detected. The magnetic sensors in the magnetic sensor array can be tunnel magnetoresistive sensors, giant magnetoresistive sensors, Hall sensors, fluxgate sensors, or other magnetic field sensitive elements. Preferably, the magnetic sensor array employs a TMR sensor array to improve the acquisition sensitivity of weak leakage magnetic field signals under thick coating conditions.

[0081] In a specific example, the magnetic sensor array employs a 16-channel TMR sensor array, with each channel arranged according to a preset sensor spacing. Each magnetic sensor channel can be used to acquire magnetic field components perpendicular to the surface of the metal component under test, magnetic field components along the scanning direction, or magnetic field components in multiple directions simultaneously. By acquiring leakage magnetic field detection signals at different spatial locations outside the thick coating through multiple magnetic sensor channels, the spatial response differences between the cracked area and the surrounding crack-free area can be obtained, providing a data foundation for subsequent inter-channel differential processing, gradient feature extraction, and crack location determination.

[0082] The signal conditioning unit amplifies, filters, and performs anti-interference processing on the electrical signal output from the magnetic sensor array. The signal conditioning unit may include at least one of an instrumentation amplifier, a low-noise amplifier, a programmable gain amplifier, a bandpass filter, a low-pass filter, a notch filter, and an anti-aliasing filter. The instrumentation amplifier is used to increase the amplitude of weak leakage magnetic signals; the programmable gain amplifier is used to adjust the amplification factor according to the strength of the leakage magnetic signal to avoid the signal being too small or the acquisition saturation; the filter is used to suppress low-frequency drift, high-frequency noise, power frequency interference, and environmental electromagnetic interference. In a specific example, the signal conditioning unit includes an instrumentation amplifier and a fourth-order Butterworth filter to improve the stability of signal acquisition.

[0083] The data acquisition unit is used to acquire the signal-conditioned magnetic flux leakage detection signal and send it to the processing unit. The data acquisition unit may include a multi-channel analog-to-digital converter, a sample-and-hold circuit, a synchronous sampling control circuit, and a communication interface. To ensure the phase and spatial correspondence between the multi-channel magnetic flux leakage detection signals, the data acquisition unit can use synchronous sampling to acquire the magnetic flux leakage detection signals from multiple magnetic sensor channels. In a specific example, the data acquisition unit uses an embedded processor in conjunction with a multi-channel analog-to-digital converter for sampling; the sampling rate can be set to 100 kS / s, and the sampling resolution can be set to 16 bits or higher.

[0084] The processing unit can be an embedded controller, microcontroller, digital signal processor, industrial computer, host computer, or cloud processing platform. It performs tasks such as receiving target detection parameter combinations, AC excitation control, sampling timing control, signal preprocessing, crack feature extraction, crack detection result output, and data storage. The processing unit can also be used to correct the target detection parameter combinations, sampling gain, or filtering parameters based on real-time signal-to-noise ratio or channel consistency evaluation values.

[0085] Before the detection begins, the processing unit can perform a background calibration process. Specifically, the AC magnetic flux leakage detection device is placed in a crack-free area or a standard reference area of ​​the metal component to be tested, an AC magnetic field is applied according to the target detection parameter combination, and background magnetic flux leakage signals are acquired. The processing unit determines the zero-point offset, reference amplitude, background noise amplitude, and channel consistency parameters of each magnetic sensor channel based on the background magnetic flux leakage signals. During subsequent detection, the processing unit can use the above background calibration data to perform zero-point correction, normalization, or channel compensation on the actually acquired magnetic flux leakage detection signals to reduce the impact of channel differences, overall coating thickness variations, and environmental magnetic field interference on crack identification results.

[0086] During the testing process, the AC magnetic flux leakage detection device can continuously scan along the surface of the metal component to be tested, or it can perform fixed-point testing at preset detection points. When using continuous scanning, the AC magnetic flux leakage detection device can be used in conjunction with a position encoder, displacement sensor, scanning drive mechanism, or manual scanning recording unit. The processing unit records the multi-channel magnetic flux leakage detection signal corresponding to each sampling moment or each scanning position, and establishes the correspondence between the magnetic flux leakage detection signal and the spatial position. Thus, the acquired magnetic flux leakage detection signal can be converted into a magnetic flux leakage response curve or a two-dimensional magnetic flux leakage response map distributed along the scanning path.

[0087] When using a fixed-point detection method, the AC magnetic flux leakage detection device can sequentially collect magnetic flux leakage detection signals at multiple preset detection points in the area to be detected. The processing unit generates local magnetic flux leakage response data for the area to be detected based on the spatial location of each detection point and the corresponding magnetic flux leakage detection signal. For welds, edges, areas near stiffeners, or other structurally complex areas, a smaller scanning step size or more detection points can be used to improve crack location accuracy.

[0088] Furthermore, when acquiring the leakage magnetic field detection signal, the processing unit can control the AC excitation unit to output the corresponding excitation frequency and amplitude according to the target detection parameter combination, while simultaneously controlling the data acquisition unit to sample according to the sampling sequence synchronized with the excitation signal. Through synchronous sampling, the amplitude and phase information of the leakage magnetic field detection signal relative to the AC excitation signal can be obtained, thus providing a basis for subsequent extraction of peak features, gradient features, phase features, and energy features.

[0089] In one embodiment, the AC magnetic flux leakage detection device may further include a display unit, a communication unit, and a storage unit. The display unit is used to display the current target detection parameter combination, real-time magnetic flux leakage waveform, crack detection results, and alarm information; the communication unit is used to upload the detection data to a host computer, a remote server, or a cloud platform; and the storage unit is used to store the detection operating parameters, target detection parameter combination, multi-channel magnetic flux leakage detection signals, crack characteristic data, and crack detection results.

[0090] The aforementioned AC magnetic flux leakage detection device enables the detection of thick-coated metal components according to a target detection parameter combination that matches the detection operating conditions, and collects multi-channel magnetic flux leakage detection signals with spatial positional correspondence. Compared with single fixed parameters and single-channel detection methods, this embodiment can provide more stable and complete magnetic flux leakage signal data for subsequent crack feature extraction and crack detection result generation.

[0091] Example 4 This embodiment describes the crack feature extraction process from multi-channel magnetic flux leakage detection signals. The crack feature extraction process in this embodiment can be executed by the processing unit in Embodiment 3, or by a host computer, industrial computer, or cloud processing platform that is communicatively connected to the AC magnetic flux leakage detection device.

[0092] like Figure 5 and Figure 7 As shown, after the magnetic sensor array acquires multi-channel magnetic flux leakage detection signals, the magnetic flux leakage detection signals of each channel are first preprocessed. The multi-channel magnetic flux leakage detection signals may include magnetic flux leakage signals acquired by different magnetic sensor channels at the same scanning position, or they may include a sequence of magnetic flux leakage signals acquired by the same magnetic sensor channel on a continuous scanning path.

[0093] Preprocessing may include at least one of the following: DC bias removal, baseline correction, bandpass filtering, synchronous demodulation, FFT analysis, wavelet denoising, and normalization. Specifically, DC bias removal is used to eliminate zero-point deviation of the magnetic sensor; baseline correction is used to eliminate low-frequency trend terms caused by slow changes in coating thickness, probe posture changes, or temperature drift during scanning; bandpass filtering is used to retain the effective magnetic flux leakage signal corresponding to the AC excitation frequency and its related frequency components; synchronous demodulation is used to obtain the amplitude and phase of the magnetic flux leakage detection signal relative to the AC excitation signal; FFT analysis is used to obtain the frequency domain energy distribution of the magnetic flux leakage detection signal; wavelet denoising is used to suppress transient interference and random noise; and normalization is used to reduce the impact of differences in gain between different sensor channels on subsequent feature extraction.

[0094] In one implementation, the processing unit first acquires the multi-channel magnetic flux leakage detection signal corresponding to the background region and establishes a background reference based on the mean, variance, and noise amplitude of the background region. Subsequently, the processing unit compares the multi-channel magnetic flux leakage detection signal of the region to be detected with the background reference to obtain a corrected magnetic flux leakage signal after eliminating background influence. This reduces the impact of overall coating thickness differences, environmental magnetic field interference, and channel zero-point drift on crack feature extraction.

[0095] After preprocessing, the multi-channel magnetic flux leakage detection signals undergo inter-channel differential processing. Differential processing may include at least one of the following: adjacent channel differential, symmetrical channel differential, and differential processing between the suspected crack region and the background region.

[0096] Adjacent channel differential refers to subtracting the leakage magnetic field signals acquired by two adjacent magnetic sensor channels at the same scanning position to obtain the adjacent channel differential signal. Symmetrical channel differential refers to subtracting the signals of symmetrical channels located on both sides of the reference position, using the center channel of the magnetic sensor array or the suspected crack center position as a reference, to obtain the symmetrical channel differential signal. Crack suspected area and background area differential refers to differentiating the leakage magnetic field signal of the suspected crack area with the leakage magnetic field signal of the crack-free background area to obtain the area differential signal.

[0097] The aforementioned differential processing can cancel or weaken common-mode interference present in multiple channels. This common-mode interference can include signal changes caused by factors such as ambient magnetic field interference, overall coating thickness variation, overall detection lift-off height variation, overall probe attitude deviation, and sensor temperature drift. Since crack defects typically cause leakage magnetic field distortion within a localized area, inter-channel differential processing can enhance the local response differences between cracked and crack-free regions.

[0098] In one specific implementation, let the preprocessed leakage magnetic signal of the i-th magnetic sensor channel be Si(x), and the preprocessed leakage magnetic signal of the (i+1)-th magnetic sensor channel be Si+1(x), where x represents the scanning position. Then, the differential signal between adjacent channels can be expressed as: ΔSi(x) = Si+1(x) - Si(x).

[0099] The processing unit can construct a differential leakage magnetic field signal sequence based on the differential signals from multiple adjacent channels, and further calculate the variation of the differential leakage magnetic field signal sequence in the scanning direction. When the differential leakage magnetic field signal shows obvious positive and negative peak changes near a certain spatial location, it indicates that there may be leakage magnetic field distortion caused by a crack in that location.

[0100] Subsequently, crack feature data is extracted from the differential leakage magnetic signal. The crack feature data may include at least one of peak features, gradient features, phase features, and energy features.

[0101] Peak characteristics are used to reflect abrupt changes in the amplitude of the leakage magnetic field signal in the crack region. Peak characteristics can include positive peak value, negative peak value, peak-to-peak value, peak-to-valley spacing, half-width at half-maximum (HWHM), and peak location. For typical crack defects, the differential leakage magnetic field signal usually forms positive and negative peaks with a certain spatial correspondence on both sides or above the crack. The processing unit can determine the crack response intensity based on the peak-to-peak values ​​of the positive and negative peaks, determine the crack location based on the center position of the positive and negative peaks, and estimate the crack influence range based on the peak-to-valley spacing or HWHM.

[0102] Gradient features are used to reflect the degree of spatial variation of the leakage magnetic field near the crack edge. Gradient features can include the maximum gradient value, average gradient value, gradient abrupt change location, and gradient sign change point. The magnetic field distribution changes abruptly at the crack edge, therefore there is usually a large gradient response near the crack edge. The processing unit can determine the maximum gradient value and the gradient abrupt change location by calculating the first-order difference or first-order derivative of the differential leakage magnetic signal relative to the scanning position.

[0103] Phase characteristics are used to reflect the phase shift of the AC leakage magnetic field signal in the crack region. Phase characteristics can include the phase difference between the crack region and the uncracked region, the amount of phase abrupt change, and the rate of phase change. Under synchronous sampling or demodulation conditions, the processing unit can obtain the phase of the leakage magnetic field detection signal of each channel relative to the AC excitation signal. When a crack alters the local electromagnetic field distribution, the phase response of the crack region will shift relative to the background region; this phase shift can serve as an auxiliary feature for crack identification.

[0104] Energy characteristics are used to reflect the redistribution of frequency domain energy caused by cracks. Energy characteristics can include fundamental frequency energy, second harmonic energy, higher harmonic energy, harmonic energy ratio, and wavelet decomposition energy. The processing unit can perform FFT analysis or wavelet decomposition on the differential magnetic flux leakage signal to obtain the energy distribution at different frequency bands or scales. When a crack causes local magnetic field distortion, the harmonic components or local-scale energy in the magnetic flux leakage detection signal may change; therefore, energy characteristics can be used as one of the bases for crack detection and crack size determination.

[0105] In one specific implementation, the processing unit performs peak-valley detection on the differential leakage magnetic field signal. When a set of positive and negative peaks with a preset spacing relationship appears on the continuous scanning path, and the peak-to-peak value of the set of positive and negative peaks exceeds a preset peak-to-peak value threshold, the region is identified as a suspected crack region. The preset spacing relationship can be determined based on the sensor spacing, scanning step size, crack direction, and standard specimen calibration results. The processing unit further calculates the maximum gradient value, phase abrupt change, and harmonic energy ratio of the suspected crack region. If at least one of the maximum gradient value, phase abrupt change, and harmonic energy ratio satisfies the corresponding judgment condition, the presence of a crack in the suspected crack region is confirmed.

[0106] In another specific implementation, the processing unit constructs a crack feature vector from the peak-to-peak value, peak-to-valley spacing, full width at half maximum (FWHM), maximum gradient value, phase abrupt change, and harmonic energy ratio. The crack feature vector can be expressed as: F=[Vpp, Wp, Gmax, Δφ, Er]; Where Vpp represents peak-to-peak value, Wp represents peak-to-valley spacing or feature width, Gmax represents maximum gradient value, Δφ represents phase abrupt change, and Er represents harmonic energy ratio. The above crack feature vector is only one example; in other implementations, feature terms can be added or removed according to the detection object and detection accuracy requirements.

[0107] The processing unit can determine the crack detection result based on the crack feature vector. Specifically, the crack feature vector can be compared with a preset judgment threshold, or the crack feature vector can be input into a pre-trained defect recognition model. The defect recognition model can be a support vector machine, random forest, neural network, convolutional neural network, or other classification and regression model. The defect recognition model can be trained based on standard specimen detection data. The training samples include magnetic flux leakage signals and their corresponding crack annotation results collected under different coating thicknesses, crack lengths, crack widths, crack depths, and detection lift-off heights.

[0108] Crack detection results can include crack presence, crack location, crack size class, and crack risk level. Crack location can be determined based on peak position, center of positive and negative peaks, gradient abrupt change location, or the location of maximum response across multiple channels. Crack size class can be determined by combining peak-to-peak value, peak-to-valley spacing, half-width at half-maximum, maximum gradient value, and energy characteristics. For example, crack size can be classified into micro-cracks, small cracks, medium cracks, and large cracks based on standard specimen calibration results. Crack risk level can be determined by combining crack size class, crack location, type of metal component under test, and service conditions.

[0109] In one implementation, to avoid misjudgment due to a single feature, the processing unit can employ a multi-condition joint judgment rule. For example, when the peak-to-peak value exceeds a preset peak-to-peak value threshold, the maximum gradient value exceeds a preset gradient threshold, and the phase abrupt change exceeds a preset phase threshold, a crack is confirmed to exist; when only the peak-to-peak value exceeds the threshold while the gradient and phase features do not meet the conditions, the area is marked as a suspected interference area or an area to be re-inspected. This improves the reliability of crack identification in complex detection environments with thick coatings.

[0110] Furthermore, the processing unit can generate crack characteristic curves or crack response maps based on crack characteristic data along the continuous scanning path. For two-dimensional magnetic sensor arrays, two-dimensional magnetic flux leakage characteristic maps can also be generated based on crack characteristic data corresponding to different channels and scanning positions. Inspectors can view suspected crack areas, crack center locations, and crack influence ranges through crack characteristic curves or two-dimensional magnetic flux leakage characteristic maps, thereby assisting in subsequent re-inspection and maintenance decisions.

[0111] Example 5 This embodiment describes the process of optimizing detection parameters and correcting the target detection parameter combination. The detection parameter optimization in this embodiment can be used to determine the initial target detection parameter combination before detection, or it can be used to perform closed-loop correction of the target detection parameter combination based on the real-time acquired magnetic flux leakage detection signal during the detection process.

[0112] In one specific embodiment, such as Figure 6 As shown, a detection parameter optimization curve or a detection parameter optimization surface can be generated based on the leakage magnetic response data corresponding to multiple candidate detection parameter combinations. Figure 6 One coordinate dimension represents the AC excitation frequency, the other represents the magnetic sensor spacing, and the vertical coordinate represents the signal-to-noise ratio or comprehensive evaluation index. For each set of candidate detection parameter combinations, the processing unit or simulation platform calculates the magnetic flux leakage response data on the outside of the thick coating based on the AC magnetic flux leakage response model, and determines the magnetic flux leakage penetration evaluation value, crack distortion evaluation value, background noise evaluation value, and spatial resolution evaluation value corresponding to the candidate detection parameter combination based on the magnetic flux leakage response data, thereby obtaining the comprehensive evaluation index.

[0113] Among them, the magnetic flux leakage penetration evaluation value is used to characterize the ability of an alternating magnetic field to form a measurable magnetic flux leakage response on the outside of the thick coating after penetrating the thick coating; the crack distortion evaluation value is used to characterize the amplitude difference, phase difference, gradient difference, or harmonic energy difference between the cracked area and the non-cracked area; the background noise evaluation value is used to characterize the influence of environmental electromagnetic interference, channel noise, or baseline drift on the magnetic flux leakage detection signal; and the spatial resolution evaluation value is used to characterize the influence of the spatial parameters of the magnetic sensor on the crack location accuracy and crack feature resolution capability.

[0114] like Figure 6 As shown, when the excitation frequency is too low, although the alternating magnetic field has good penetration ability, the spatial resolution of the crack region may decrease; when the excitation frequency is too high, the leakage magnetic signal is easily attenuated by the skin effect and coating lift-off, and the effective signal amplitude outside the thick coating decreases. Similarly, when the magnetic sensor spacing is too small, the difference response between adjacent channels is not obvious and is easily affected by channel noise; when the magnetic sensor spacing is too large, it may not accurately reflect the local distortion characteristics of the crack region. Therefore, the processing unit does not simply select the parameter combination with the largest signal amplitude, but determines the target detection parameter combination based on the peak region or the region with a high comprehensive evaluation index in the detection parameter optimization surface.

[0115] In a specific example, for a carbon steel specimen covered with a 10mm thick epoxy resin coating, the parameter optimization surface shows that when the AC excitation frequency is close to 1kHz and the magnetic sensor spacing is close to 20mm, the system signal-to-noise ratio or comprehensive evaluation index is high, achieving a better balance between leakage magnetic penetration capability, crack distortion response, and spatial resolution. Therefore, an excitation frequency around 1kHz and a sensor spacing around 20mm can be used as the initial target detection parameter combination for this detection condition. It should be noted that the above values ​​are only the optimal results for specific coating thickness, metal materials, and detection lift-off conditions. When the coating thickness, material parameters, or detection lift-off parameters change, adjustments can be made based on... Figure 6 The parameter optimization method shown redefines the combination of target detection parameters.

[0116] Furthermore, in actual testing, the coating thickness of the metal component under test may fluctuate locally, the lift-off height of the detection probe may change, and electromagnetic interference may exist in the environment. Therefore, relying solely on the initial target detection parameter combination determined before testing may not be sufficient to maintain stable testing results throughout the entire testing process. To address this, this embodiment also provides a closed-loop correction method for the target detection parameter combination.

[0117] like Figure 8As shown, after acquiring the magnetic flux leakage detection signal according to the target detection parameter combination, the processing unit first calculates the real-time signal-to-noise ratio (SNR) or channel consistency evaluation value. The real-time SNR can be determined based on the ratio between the effective response amplitude of the suspected crack region and the noise amplitude of the background region. For example, the peak-to-peak value, maximum gradient value, or specific frequency band energy of the suspected crack region can be used as the effective response value, and the root mean square noise, variance, or frequency band noise energy of the crack-free background region can be used as the noise value to determine the real-time SNR. The channel consistency evaluation value can be determined based on the amplitude differences, phase differences, baseline drift differences, or correlation coefficients of multiple magnetic sensor channels within the background region.

[0118] When the real-time signal-to-noise ratio meets the preset signal-to-noise ratio threshold and the channel consistency evaluation value meets the preset consistency condition, the processing unit maintains the current target detection parameter combination and outputs the crack detection result based on the currently acquired magnetic flux leakage detection signal. This indicates that the current target detection parameter combination is basically matched with the thick coating detection condition of the metal component to be detected, and the acquired magnetic flux leakage detection signal can meet the requirements for subsequent crack feature extraction.

[0119] When the real-time signal-to-noise ratio (SNR) is lower than the preset SNR threshold, or the channel consistency evaluation value does not meet the preset consistency conditions, the processing unit determines that the current detection state does not meet the preset conditions and corrects at least one of the AC excitation parameters, magnetic sensor spatial parameters, sampling gain, or filtering parameters. For example, when the real-time SNR is low and the acquired signal has not reached saturation, the excitation amplitude or sampling gain can be increased; when the acquired signal is close to saturation, the sampling gain or excitation amplitude can be reduced to avoid signal clipping distortion; when background noise is concentrated in the current detection frequency band, the excitation frequency or filtering parameters can be adjusted to make the detection frequency band avoid the main interference frequency band; when the crack spatial response is not obvious, the combination of sensor channels involved in differential processing, the scanning step distance, or the detection lift-off height can be adjusted; when the channel consistency is abnormal, the background reference signal can be re-acquired, channel compensation can be performed, abnormal channels can be eliminated, or the detection personnel can be prompted to correct the probe posture.

[0120] In one implementation, the processing unit can filter the corrected parameters based on the AC leakage magnetic response relationship. Specifically, the processing unit inputs the current detection operating parameters, real-time signal-to-noise ratio, channel consistency evaluation value, and current target detection parameter combination into the AC leakage magnetic response relationship, queries or calculates the comprehensive evaluation index corresponding to multiple candidate corrected parameter combinations, and uses the candidate corrected parameter combination whose comprehensive evaluation index meets the preset correction conditions as the corrected target detection parameter combination. The preset correction conditions can be the maximum comprehensive evaluation index, the expected improvement of real-time signal-to-noise ratio exceeding a preset improvement threshold, the expected channel consistency meeting a preset consistency condition, or the optimal comprehensive evaluation index under the constraints of excitation power consumption and sampling dynamic range.

[0121] After parameter correction, the processing unit re-acquires the magnetic flux leakage detection signal according to the corrected parameters and recalculates the real-time signal-to-noise ratio or channel consistency evaluation value. If the recalculated real-time signal-to-noise ratio and channel consistency evaluation value meet the preset conditions, the corrected parameters are used to continue detection; if the preset conditions are still not met after correction, the processing unit outputs a re-inspection prompt or marks the current detection area as an area to be manually reviewed.

[0122] In one implementation, to avoid instability in detection results due to frequent changes in the target detection parameter combination, the processing unit can set correction trigger conditions and correction hold conditions. For example, parameter correction is triggered only when the real-time signal-to-noise ratio is lower than a preset signal-to-noise ratio threshold for multiple consecutive sampling windows; channel calibration or probe posture prompts are triggered only when the channel consistency evaluation value does not meet the preset consistency condition for multiple consecutive sampling windows. After correction, the processing unit can maintain the corrected target detection parameter combination for a preset hold time or a preset scanning distance, avoiding repeated parameter switching in a short period of time.

[0123] Furthermore, the processing unit can update the AC leakage magnetic response relationship based on the parameter correction results. Specifically, when the corrected target detection parameter combination achieves a high real-time signal-to-noise ratio and good channel consistency in actual testing, the processing unit records the current testing parameters, the corrected target detection parameter combination, the actual leakage magnetic detection signal characteristics, and the crack detection results as new sample records and stores them in the AC leakage magnetic response relationship database. Subsequently, under the same or similar testing conditions, the processing unit can prioritize the use of this corrected target detection parameter combination, thereby improving parameter matching efficiency and field adaptability.

[0124] By combining the parameter optimization process before detection with the closed-loop parameter correction process during detection, this embodiment enables the detection system to adaptively adjust the AC magnetic flux leakage detection parameters based on coating thickness, material parameters, detection lift-off height, and on-site signal quality. This improves the detectability of magnetic flux leakage signals, the stability of crack feature extraction, and the repeatability of detection results during the detection of cracks in thick-coated metals.

[0125] Example 6 This embodiment provides an experimental verification method to illustrate the detection effect of the method of the present invention.

[0126] Standard specimens were prepared. The metal substrate material of the specimens was Q235 carbon steel, and the dimensions were 100mm×100mm×10mm. An epoxy resin coating was applied to the surface of the metal substrate, with a coating thickness of 10±0.2mm. Artificial cracks were machined on or near the surface of the metal substrate. The artificial cracks could be wire-cut cracks, with crack lengths ranging from 5mm to 20mm and crack widths ranging from 0.1mm to 1mm.

[0127] During the experiment, the target detection parameter combination was determined based on the aforementioned AC leakage magnetic flux response relationship. In a specific example, the target detection parameter combination included a 1kHz sine wave excitation, a 20mm sensor spacing, and a 10mm detection lift-off height. The AC leakage magnetic flux detection device tested the standard specimen according to this target detection parameter combination and acquired multi-channel leakage magnetic flux detection signals. The processing unit filtered, differentially processed, and extracted crack features from the multi-channel leakage magnetic flux detection signals to obtain the crack detection results.

[0128] like Figure 9 and Figure 10 As shown, the proposed ACLM method for detecting cracks in thick-coated metals based on AC leakage magnetic flux can be compared with ultrasonic testing (UT), eddy current testing (ECT), and AC field testing (ACFM) in terms of comprehensive performance.

[0129] Figure 9 This is used to compare the signal-to-noise ratio and detection efficiency of different detection methods. Figure 9 The bar chart on the left is a signal-to-noise ratio comparison chart, with the vertical axis representing the signal-to-noise ratio and the horizontal axis representing different detection methods. Figure 9 The bar chart on the right is a comparison of detection efficiency; the vertical axis represents scanning speed or detection efficiency, and the horizontal axis represents different detection methods. Figure 9 It can be seen that, under the same or similar conditions for detecting cracks in thick-coated metals, the signal-to-noise ratio of the method provided in this embodiment is higher than that of ultrasonic testing, eddy current testing, and AC field testing, indicating that this embodiment can improve the distinguishability of crack response signals relative to background noise. At the same time, the scanning speed or detection efficiency of this embodiment is relatively high, indicating that this embodiment can complete the acquisition of magnetic flux leakage signals and crack feature identification without removing the thick coating, thereby reducing detection preparation time and repeated detection time.

[0130] Figure 10 Each evaluation index can be represented by a normalized score, with a higher score indicating better detection performance for the corresponding index. These evaluation indexes may include signal-to-noise ratio (SNR), crack resolution, detection speed, thick coating adaptability, and cost control.

[0131] Among them, the signal-to-noise ratio is used to represent the degree of distinguishability of the crack response relative to the background noise in the detection signal; crack resolution capability is used to represent the ability of the detection system to identify small-sized cracks or crack edge features; detection speed is used to represent the detection length or detection area that can be completed per unit time; thick coating adaptability is used to represent the ability to adapt to different coating thickness conditions without removing the coating; cost control capability is used to represent the comprehensive level of the complexity of the detection device, detection preparation cost and on-site detection and maintenance cost.

[0132] Depend on Figure 10 It is known that traditional ultrasonic testing has a relatively mature application basis in the inspection of some metal components, but it is easily affected by coating interface attenuation and coupling state changes under thick coating conditions; eddy current testing has high surface detection sensitivity, but its ability to detect metal cracks under thick coatings is limited due to the skin effect; AC field testing can detect surface or near-surface cracks, but under thick coatings and complex field interference conditions, the adaptability of detection parameters and signal stability are still insufficient.

[0133] In contrast, this embodiment determines the target detection parameter combination that matches the detection operating conditions by using a preset AC magnetic flux leakage response relationship. Combined with differential processing of multi-channel magnetic flux leakage detection signals, crack feature extraction, and a target detection parameter combination correction mechanism, it can improve the signal-to-noise ratio of the magnetic flux leakage signal and the stability of crack feature recognition under thick coating conditions. Therefore, this embodiment achieves superior overall performance in terms of thick coating adaptability, crack resolution, and overall detection stability.

[0134] It should be noted that the above experimental parameters are only used to illustrate specific embodiments of the present invention and do not limit the present invention to be applicable only to 10mm thick epoxy resin coatings, nor do they limit the excitation frequency to 1kHz, the sensor spacing to 20mm, or the detection lift-off height to 10mm. For coatings of 5mm to 20mm or other thicknesses, as well as other metal materials and coating materials, the combination of target detection parameters can be re-determined according to the AC leakage magnetic response relationship provided by the present invention.

[0135] Example 7 This embodiment provides a thick-coated metal crack detection system based on AC leakage magnetic flux. For example... Figure 11 As shown, the system includes a parameter acquisition module, a parameter determination module, a magnetic flux leakage acquisition module, a feature extraction module, and a crack detection module.

[0136] The parameter acquisition module is used to acquire the testing condition parameters of the metal component to be tested, which include at least the coating parameters of the metal component. The parameter acquisition module can be connected to a coating thickness measurement device, a material database, a manual input interface, or a historical testing database to obtain the coating thickness, coating material type, metal material type, and structural parameters of the area to be tested.

[0137] The parameter determination module is used to determine the target detection parameter combination that matches the detection operating conditions based on the detection operating conditions and the preset AC leakage magnetic response relationship. The parameter determination module may include an AC leakage magnetic response relationship database, a parameter interpolation unit, a comprehensive evaluation index calculation unit, and a parameter output unit. The parameter determination module searches for or calculates the comprehensive evaluation index corresponding to different candidate detection parameter combinations based on the current detection operating conditions and outputs the target detection parameter combination that meets the preset optimization conditions.

[0138] The magnetic flux leakage acquisition module is used to apply an alternating magnetic field to the metal component under test according to the target detection parameter combination, and to acquire the magnetic flux leakage detection signal generated by the metal component on the outside of the thick coating. The magnetic flux leakage acquisition module may include an AC excitation unit, a magnetic yoke, an excitation coil, a magnetic sensor array, a signal conditioning unit, and a data acquisition unit.

[0139] The feature extraction module is used to extract crack feature data related to magnetic field distortion caused by cracks from the magnetic flux leakage detection signal. The feature extraction module can perform processing such as filtering, differencing, gradient calculation, phase analysis, frequency domain analysis, wavelet analysis, and eigenvector generation.

[0140] The crack detection module is used to determine the crack detection results of the metal component under inspection based on crack characteristic data. The module can output information such as the presence and location of cracks, crack size grade, crack risk level, and an inspection report. Crack detection results can be displayed locally or uploaded to a remote server or cloud database via the communication module.

[0141] The system in this embodiment is based on the same inventive concept as the aforementioned method embodiment. Therefore, for any content not detailed in this embodiment, please refer to the relevant descriptions in the aforementioned method embodiment.

[0142] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0143] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A method for detecting cracks in thick-coated metals based on AC leakage magnetic flux, characterized in that, include: Obtain the testing condition parameters of the metal component to be tested, wherein the testing condition parameters include at least the coating parameters of the metal component to be tested; Based on the detection operating parameters and the preset AC leakage magnetic response relationship, a target detection parameter combination matching the detection operating parameters is determined. The AC leakage magnetic response relationship is used to characterize the correspondence between leakage magnetic signal characteristics under different detection operating parameters and different detection parameter combinations. The target detection parameter combination includes AC excitation parameters and magnetic sensor spatial parameters. An alternating magnetic field is applied to the metal component to be tested according to the target detection parameter combination, and the leakage magnetic field detection signal generated by the metal component to be tested on the outside of the thick coating is collected. Extract crack feature data related to magnetic field distortion caused by cracks from the magnetic flux leakage detection signal; The crack detection result of the metal component to be tested is determined based on the crack feature data.

2. The method for detecting cracks in thick-coated metals based on AC leakage magnetic flux according to claim 1, characterized in that, The testing parameters also include at least one of the material parameters of the metal component to be tested and the structural parameters of the area to be tested; the coating parameters include at least one of the coating thickness, coating material type and coating medium properties.

3. The method for detecting cracks in thick-coated metals based on AC leakage magnetic flux according to claim 1, characterized in that, The AC leakage magnetic response relationship was obtained in the following way: An AC leakage magnetic response model is established using multiphysics coupling theory. Multiple candidate detection parameter combinations are set, and the leakage magnetic response data corresponding to each candidate detection parameter combination is obtained based on the AC leakage magnetic response model. Based on the magnetic flux leakage response data, establish the correspondence between the detection operating condition parameters, the combination of detection parameters, and the characteristics of the magnetic flux leakage signal.

4. The method for detecting cracks in thick-coated metals based on AC leakage magnetic flux according to claim 3, characterized in that, The AC magnetic flux leakage response model includes a metal substrate model, a coating model, an AC magnetic field excitation model, a magnetic sensor acquisition model, and a crack defect response model. The AC magnetic flux leakage response model is used to characterize the distribution changes of the AC magnetic field leakage field in the metal substrate, thick coating, and crack defects.

5. The method for detecting cracks in thick-coated metals based on AC leakage magnetic flux according to claim 3, characterized in that, Based on the leakage magnetic response data, establish the correspondence between detection operating parameters, detection parameter combinations, and leakage magnetic signal characteristics, including: Based on the magnetic flux leakage response data, determine the magnetic flux leakage penetration evaluation value and crack distortion evaluation value corresponding to each candidate detection parameter combination; The comprehensive evaluation index for each candidate detection parameter combination is determined based on the magnetic flux leakage penetration evaluation value and the crack distortion evaluation value. Establish a mapping relationship between the detection operating condition parameters, the candidate detection parameter combinations, and the comprehensive evaluation index, as at least a part of the AC leakage magnetic response relationship.

6. The method for detecting cracks in thick-coated metals based on AC leakage magnetic flux according to claim 5, characterized in that, The leakage magnetic penetration evaluation value is determined based on at least two of the leakage magnetic response amplitude on the outside of the thick coating, the reference response amplitude in the crack-free region, and the background noise amplitude; the crack distortion evaluation value is determined based on at least one of the amplitude difference, phase difference, gradient difference, and harmonic energy difference between the cracked region and the crack-free region.

7. The method for detecting cracks in thick-coated metals based on AC leakage magnetic flux according to claim 1, characterized in that, The AC excitation parameters include at least one of excitation frequency, excitation amplitude, and excitation waveform; the magnetic sensor spatial parameters include at least one of the spacing between adjacent magnetic sensors, magnetic sensor array arrangement, and detection lift-off height.

8. The method for detecting cracks in thick-coated metals based on AC leakage magnetic flux according to claim 1, characterized in that, The magnetic flux leakage detection signal generated on the outer side of the thick coating of the metal component under test is collected, and crack feature data related to the magnetic field distortion caused by the crack is extracted from the magnetic flux leakage detection signal, including: Multi-channel magnetic flux leakage detection signals are acquired at different spatial locations on the outside of the thick coating through multiple magnetic sensor channels; The multi-channel magnetic flux leakage detection signal is subjected to inter-channel differential processing to obtain a differential magnetic flux leakage signal; At least one of peak features, gradient features, phase features, and energy features is extracted from the differential leakage magnetic signal to obtain the crack feature data.

9. The method for detecting cracks in thick-coated metals based on AC leakage magnetic flux according to claim 1, characterized in that, After acquiring the leakage magnetic field detection signal, the process also includes: The real-time signal-to-noise ratio or channel consistency evaluation value is determined based on the magnetic flux leakage detection signal. When the real-time signal-to-noise ratio is lower than the preset signal-to-noise ratio threshold, or when the channel consistency evaluation value does not meet the preset consistency condition, at least one of the AC excitation parameter, magnetic sensor spatial parameter, sampling gain, and filtering parameter in the target detection parameter combination is corrected.

10. A thick-coated metal crack detection system based on AC leakage magnetic flux, characterized in that, include: The parameter acquisition module is used to acquire the detection condition parameters of the metal component to be tested, and the detection condition parameters include at least the coating parameters of the metal component to be tested. The parameter determination module is used to determine a target detection parameter combination that matches the detection operating conditions based on the detection operating conditions and the preset AC leakage magnetic response relationship. The AC leakage magnetic response relationship is used to characterize the correspondence between leakage magnetic signal characteristics under different detection operating conditions and different detection parameter combinations. The target detection parameter combination includes AC excitation parameters and magnetic sensor spatial parameters. The magnetic flux leakage acquisition module is used to apply an alternating magnetic field to the metal component to be tested according to the target detection parameter combination, and to acquire the magnetic flux leakage detection signal generated by the metal component to be tested on the outside of the thick coating. The feature extraction module is used to extract crack feature data related to magnetic field distortion caused by cracks from the magnetic flux leakage detection signal; A crack detection module is used to determine the crack detection result of the metal component to be inspected based on the crack feature data.

Images