Method and system for detecting corrosion of metal component in soil medium based on pulsed eddy current

Abstract

The invention relates to the technical field of nondestructive testing, in particular to a method and a system for detecting corrosion of a metal component in a soil medium based on pulsed eddy current, which are characterized in that background eddy current loss parameters of soil are quantitatively calculated by establishing an eddy current field distribution model in the soil medium; then, physical mapping characteristics such as derivative zero-crossing time and eddy current loss characteristic values are extracted from a detection coil signal on the basis, and a quantitative mapping relation between the physical mapping characteristics and the metal conductivity is established; and finally, fusing the physical characteristics and waveform transformation characteristics extracted from the signals, and inverting through a neural network model to obtain geometric parameters such as depth, length and the like of the corrosion defect. According to the method, soil interference is quantified through modeling, and targeted compensation and fusion are performed on the soil interference in feature extraction and inversion, so that the defects that a traditional pulsed eddy current technology is low in signal-to-noise ratio and cannot perform quantitative evaluation in a soil medium are fundamentally overcome, and rapid, non-excavation and high-precision quantitative detection on the corrosion state of the buried metal component is realized.

Description

Technical Field

[0001] This invention relates to the field of nondestructive testing technology, and in particular to a method and system for detecting corrosion of metal components in soil media based on pulsed eddy currents. Background Technology

[0002] Pulsed eddy current testing technology is widely used for non-destructive testing of metal structures due to its advantages such as sensitivity to defects in metallic materials and non-contact measurement. Current techniques primarily focus on exposed metal surfaces or structures with thin coatings, assessing defects by extracting and analyzing specific characteristics of the response signal.

[0003] However, the testing environment changes fundamentally when the metal component being inspected is buried in soil. Soil is a non-homogeneous, porous medium composed of solid, liquid, and gas phases. Its electromagnetic properties fluctuate significantly due to variations in water content, composition, and density, which strongly interferes with the propagation of the excitation electromagnetic field and the distribution of the eddy current field in the inspected object. This renders traditional detection models and feature extraction methods based on the assumption of a homogeneous medium ineffective, resulting in extremely low signal-to-noise ratios and difficulty in accurately retrieving corrosion defect information.

[0004] Therefore, there is an urgent need in this field for a pulsed eddy current corrosion detection method that specifically models and interprets signals for soil media characteristics. Summary of the Invention

[0005] In view of this, the purpose of this invention is to propose a method and system for detecting corrosion of metal components in soil media based on pulsed eddy currents, so as to solve the problems of failure of traditional detection models and feature extraction methods based on the assumption of homogeneous media, extremely low signal-to-noise ratio, and difficulty in accurately retrieving corrosion defect information.

[0006] To achieve the above objectives, this invention provides a method for detecting corrosion of metal components in soil media based on pulsed eddy currents. The method employs a detection system including an excitation coil and a detection coil, and includes the following steps:

[0007] Step S1: Perform electromagnetic modeling on the soil medium in the detection area, calculate the equivalent electromagnetic parameters of the soil, and quantitatively analyze the attenuation effect of the soil on the excitation electromagnetic field and the background eddy current loss generated by the attenuation operation.

[0008] Step S2: Based on the calculation results of step S1, extract at least one physical mapping feature quantity related to the corrosion state of the metal component from the induced voltage signal obtained from the detection coil, and establish a quantitative mapping relationship between the feature quantity and the local equivalent conductivity of the metal component.

[0009] Step S3: Combine the physical mapping features extracted in step S2 with the waveform transformation features further extracted from the induced voltage signal to construct a multidimensional feature vector. Then, use a machine learning model to invert and obtain the geometric parameters of corrosion defects in the metal component.

[0010] Preferably, the fundamental frequency and main spectral components of the excitation current pulse in the excitation coil are limited to a frequency band below 10MHz.

[0011] Preferably, step S1 specifically includes:

[0012] S101. Treat the non-uniform soil medium as an equivalent medium with uniform electromagnetic parameters, and calculate its equivalent conductivity σ. eff With equivalent permeability μ eff ;

[0013] S102, based on the equivalent conductivity σ eff With equivalent permeability μ eff Calculate the total eddy current loss power induced by the excitation electromagnetic field in the soil. and its equivalent loss resistance in the equivalent circuit of the detection system .

[0014] Preferably, the physical mapping feature quantities extracted in step S2 include the zero-crossing time T0 of the derivative of the induced voltage signal and the eddy current loss feature value P constructed based on the eddy current loss power density formula. e .

[0015] Preferably, the quantitative mapping relationship in step S2 includes calculating the local equivalent conductivity σ of the metal component using the time T0 when the derivative crosses zero. m Relationship: σ m =k1T0+b1, where the coefficients k1 and b1 are obtained through a calibration procedure, which includes measuring and performing regression analysis on standard samples in various simulated soil environments with different equivalent loss resistances R_ec to determine the coefficients k1 and b1.

[0016] Preferably, the waveform transformation features include time-frequency domain transformation features and signal statistical transformation features. The time-frequency domain transformation features include amplitude, derivative, or harmonic component features extracted from the signal waveform or spectrum. The signal statistical transformation features include principal component features obtained by performing multivariate statistical analysis on the signal sequence.

[0017] Preferably, step S3 specifically includes:

[0018] S301: Perform anti-interference preprocessing on the induced voltage signal;

[0019] S302: Construct the multidimensional feature vector and train a machine learning model using a data-augmented training set to establish a mapping relationship between the multidimensional feature vector and the geometric parameters of the corrosion defect.

[0020] Preferably, the machine learning model is a feedforward neural network, comprising at least one input layer, one output layer, and two hidden layers; the hidden layers employ non-linear activation functions, and the output layers employ linear activation functions.

[0021] Preferably, the geometric parameters of the corrosion defect include at least: the depth d of the corrosion location from the surface of the soil medium and the length l of the corrosion extending along the component.

[0022] This invention also provides a pulsed eddy current detection system for detecting corrosion of metal components in soil media, used to perform the above-described method. The system includes:

[0023] The excitation coil is configured to receive an excitation current pulse to generate an excitation electromagnetic field in the soil medium;

[0024] The detection coil is configured to receive the induced voltage signal generated by the eddy current field induced by the metal component in the soil;

[0025] A processor, electrically connected to the excitation coil and the detection coil, is configured to execute the following modules:

[0026] The parameter modeling module is configured to perform step S1 in the method, establish a eddy field distribution model in the soil medium, and quantitatively calculate the background eddy loss parameters of the soil.

[0027] The feature extraction and mapping module is configured to perform step S2 in the method, extract physical mapping feature quantities from the induced voltage signal and establish a mapping relationship between them and the conductivity of the metal component;

[0028] The fusion inversion module is configured to perform step S3 in the method, fusing multiple types of feature quantities and inverting the geometric parameters of corrosion defects through a machine learning model.

[0029] The beneficial effects of this invention are:

[0030] 1. This invention breaks through the limitation of traditional pulsed eddy current technology being only applicable to air medium, and constructs an eddy current field distribution model in soil medium. It incorporates the quantitative calculation of soil eddy current loss and electric field distribution into the detection and analysis system, making the detection method fundamentally adaptable to buried scenarios and promoting the upgrade of this technology from open space to concealed engineering applications.

[0031] 2. This invention establishes a mapping model between metal conductivity and eddy current attenuation characteristics, integrates time-frequency domain features and data-driven features to extract multi-dimensional information, and constructs a clear inversion link from microscopic changes in corrosion to macroscopic signal response. This overcomes the shortcomings of traditional methods in soil, such as low signal-to-noise ratio and reliance on experience, and provides a reliable basis for high-precision quantitative assessment.

[0032] 3. This invention provides an intelligent solution for condition-based maintenance of buried metal components, enabling maintenance personnel to quickly and intuitively obtain quantitative results such as corrosion depth and extent under non-excavation and non-contact conditions, providing solid technical support for intelligent supervision and life management of critical infrastructure. Attached Figure Description

[0033] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only for this invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0034] Figure 1 This is a schematic diagram of the main process of the corrosion detection method for metal components in soil medium based on pulsed eddy current according to an embodiment of the present invention. Detailed Implementation

[0035] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.

[0036] It should be noted that, unless otherwise defined, the technical or scientific terms used in this invention should have the ordinary meaning understood by one of ordinary skill in the art to which this invention pertains. The terms "first," "second," and similar terms used in this invention do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" mean that the element or object preceding the word encompasses the elements or objects listed following the word and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are used only to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0037] Example 1:

[0038] like Figure 1As shown in Embodiment 1, this method provides a corrosion detection method for metal components in soil media based on pulsed eddy currents. It employs a detection system including an excitation coil and a detection coil. The excitation coil generates a specific time-varying magnetic field capable of penetrating the soil medium and effectively inducing eddy currents. The generated magnetic field exhibits minimal attenuation in conductive soil media, allowing it to penetrate to a sufficient depth to reach the buried metal component and ensuring sufficient energy to induce detectable eddy currents within the metal. The generated magnetic field not only acts on the target metal but also induces eddy currents and parasitic eddy currents in the soil medium itself. The detection coil's function is to highly sensitively capture the weak voltage signal induced by the eddy current field in the metal component, superimposed with soil background interference. According to the law of electromagnetic induction, when the eddy current field in the metal component (induced by the excitation magnetic field) changes over time, a voltage signal is induced in the detection coil. The amplitude, waveform, and attenuation characteristics of this signal directly carry information about the conductivity distribution (i.e., corrosion state) of the metal component.

[0039] This method specifically includes the following steps:

[0040] Step 1: Modeling the eddy field distribution in the soil medium.

[0041] Specifically, it includes:

[0042] Step 1.1 Electromagnetic propagation attenuation model in soil:

[0043] First, based on Maxwell's equations, the propagation characteristics of electromagnetic fields in soil are analyzed. The wave equations for the electric and magnetic fields in the soil medium under the high-frequency electromagnetic field generated by the coil can be expressed as:

[0044] (1);

[0045] In the formula, E is the electric field strength in the soil medium; H is the magnetic field strength in the soil medium; and the parameter γ is the propagation constant, which can be expressed as:

[0046] γ=α+jβ(2;

[0047] In the formula, α is the attenuation constant of the electromagnetic wave; β is the constant of the phase change of the electromagnetic wave; and j is the imaginary unit.

[0048] For a typical soil medium that satisfies the condition σ / ωε>>1 (σ is electrical conductivity, ω is angular frequency, ε is dielectric constant), the electromagnetic wave attenuation constant α can be simplified as:

[0049] (3);

[0050] In the formula, f is the frequency and μ is the permeability. This formula shows that the attenuation of electromagnetic waves in the soil medium is proportional to the square root of the frequency. To ensure the effective penetration depth and signal strength of the excitation magnetic field in the soil, the fundamental frequency and main spectral components of the excitation current pulse supplied to the excitation coil in the entire detection system must be strictly limited to below 10MHz, and preferably operate in the low-frequency range of 1Hz to 10kHz. This is a prerequisite and system design constraint for the effective execution of all subsequent analyses and detections.

[0051] Step 1.2 Calculation of equivalent electromagnetic parameters for non-uniform soil:

[0052] To address the heterogeneous nature of soil, a mixed-media model is used for equivalence. For a soil composed of n components, the volume fraction v of the i-th component is known. i Conductivity σ i and permeability μ i (Assuming σ2>σ1, μ2>μ1), its equivalent conductivity σ eff The formulas for calculating the upper and lower boundaries are:

[0053] (4);

[0054] (5);

[0055] Its equivalent permeability μ eff The formulas for calculating the upper and lower boundaries are:

[0056] (6);

[0057] (7);

[0058] This is then equivalent to a homogeneous isotropic medium. This step transforms the complex real soil into model parameters that can be used for analytical calculations.

[0059] Step 1.3 Electric Field Distribution and Eddy Current Loss Model in Soil Medium

[0060] When a pulsed current is applied to the excitation coil, the time-varying magnetic field it generates induces a vortex electric field E in the soil medium. In cylindrical coordinates, the circumferential component of this electric field can be solved according to steps 1.1 and 1.2 above.

[0061] The electric field E at any point (ρ, Φ, z) in the soil can be expressed as:

[0062] (8);

[0063] In the formula, I is the coil current; R is the coil radius; J1 is the first-order Bessel function; d is the soil medium thickness; ρ, Φ, z are the radial, angular, and axial coordinates of the observation point, respectively; λ is the radial spatial frequency; and u is the axial propagation constant, which can be expressed as:

[0064] (9);

[0065] In the formula, k s The complex wave number of the soil can be expressed as:

[0066] (10);

[0067] In the formula, σ and μ are the electrical conductivity and magnetic permeability of the soil, respectively, which are determined by constraints in step 1.2.

[0068] Driven by the aforementioned induced electric field E, closed parasitic eddies are formed in the soil, causing energy to dissipate in the form of Joule heat, resulting in eddy current losses in the soil. The power density per unit volume of loss is J. s E=σE 2 J s =σE is the induced current density in the soil. The total eddy current loss power in the soil of the entire detection area is:

[0069] (11);

[0070] In the formula, V represents the total eddy current loss power in the soil. soil To detect the soil volume in the area.

[0071] This loss directly consumes the excitation energy, weakening the magnetic field effectively acting on the underground metal structure and becoming a significant background interference in the detection signal. At the circuit level, this loss can be equivalent to a frequency-dependent loss resistor connected in series with the excitation coil, and can be expressed as:

[0072] (12);

[0073] In the formula, The equivalent eddy current loss resistance is the core factor that causes the quality factor of the detection system to decrease and the signal amplitude to drop.

[0074] Step 2: Modeling the mapping between electrical conductivity of metal components and eddy current field attenuation.

[0075] Specifically, it includes:

[0076] Metal corrosion leads to a decrease in its local equivalent conductivity σ mAccording to the principle of electromagnetic induction, under step excitation, the decay time constant of eddy currents in a conductor is related to the magnitude of the equivalent conductivity. This relationship is reflected in the induced voltage V of the detection coil. r During the transition of (t), the first derivative of the signal V is defined. r '(t) = dV r / dt. The derivative crosses zero at time T0 (i.e., V r (T0)=0) shows a strong linear correlation with the equivalent conductivity within a specific range:

[0077] σ m =k1T0+b1(13)

[0078] T0 provides a rapid, preliminary estimate of the electrical conductivity of the corroded region. In the formula, k1 and b1 are coefficients obtained through calibration using standard samples. These coefficients are obtained through a specific calibration procedure that requires calibration on a series of known σ values ​​in various simulated soil environments with different equivalent electromagnetic parameters (representing different soil types). m The standard metal sample was measured, and its T0 value was collected and determined by regression analysis. This process ensures that the established mapping relationship inherently includes the adaptability and robustness to different soil background losses calculated in step S1.

[0079] To further enhance the sensitivity and anti-interference ability of the feature to micro-corrosion, an eddy current loss characteristic value P is constructed. e Based on the physical principle of eddy current loss in step 1.3, the formula for eddy current loss power density is obtained, and it is formalized into a computable characteristic through time-domain integration:

[0080] (14);

[0081] In the formula, [t1, t2] is the integration interval, covering the main decay stage after the falling edge of the pulse; P e It comprehensively reflects the energy dissipation process of the eddy current field and is more sensitive to early or microscopic corrosion.

[0082] Step 3: Multidimensional feature fusion and inversion method for corrosion defects.

[0083] Specifically, it includes:

[0084] Step 3.1 Anti-interference preprocessing and multi-dimensional feature fusion

[0085] First, the original induced voltage signal V r(t) Denoising preprocessing is performed, including first using an adaptive filter to reference the environmental noise signal and suppress related interference components, then decomposing the signal into different scales through wavelet transform, performing soft thresholding on the high-frequency detail coefficients representing noise, and finally reconstructing the denoised clean signal V(t). This process can be modeled as follows:

[0086] V(t) = H Hyb (V r (t),N ref (t))(15;

[0087] In the formula, H Hyb (·) represents the hybrid noise reduction operator; N ref (t) represents the synchronously acquired reference noise signal.

[0088] Three types of features are extracted and fused from V(t) to form a feature vector F that comprehensively describes the signal state:

[0089] ① Physical mapping characteristics: namely, the zero-crossing time T0 of the derivative extracted in step 2 and the characteristic value P of eddy current loss. e This is directly related to changes in material conductivity and energy dissipation;

[0090] ② Traditional time-frequency domain characteristics: Classical characteristics based on signal waveform and spectrum, including the steady-state peak value P. m Peak value of the first derivative P g The fundamental (first harmonic) amplitude W1 and the third harmonic amplitude W3 are characteristics that are sensitive to changes in signal amplitude, attenuation rate and spectral structure.

[0091] ③ Data-driven statistical features: Principal component analysis is applied to a complete periodic sampling sequence of V(t). Let the number of sampling points be m, forming a vector v = [V(t1), V(t2), ..., V(t...]. m )] T By projecting the principal components onto a new orthogonal basis through linear transformation, the principal component scores are obtained. The scores of the top three principal components with the highest cumulative contribution rates are selected and denoted as S1, S2, and S3.

[0092] Finally, by integrating all the above features, a multi-dimensional feature vector is constructed:

[0093] F=[T0,P e ,P m ,P g [W1,W3,S1,S2,S3] T (16);

[0094] Step 3.2 Small Sample Data Augmentation

[0095] To address the issue of model overfitting and weak generalization due to the limited number of samples with corrosion pits in real-world applications, for minority class samples F... i Randomly select sample F from its k nearest neighbors nn Generate new samples:

[0096] F new =F i +ξ·(F nn -F i (17);

[0097] In the formula, ξ∈[0,1] is a random number, which is used to expand the training dataset, improve the generalization ability of the model, and prevent overfitting.

[0098] Step 3.3 Corrosion Defect Parameter Inversion

[0099] A dual-hidden-layer feedforward neural network is constructed as the core computational model for inversion, reflecting the precise mapping relationship from feature F to defect parameters D (including the depth d of the corrosion location from the soil surface and the length l of the corrosion):

[0100] D=f out (W (3) ·σ(W (2) ·σ(W (1) ·F+b (1) )+b (2) )+b (3) (18);

[0101] In the formula, F is the input feature vector, calculated in step 3.1; D is the output corrosion defect parameter vector; W (1) W (2) W (3) The weight matrices connecting the input layer to the first hidden layer, the first hidden layer to the second hidden layer, and the second hidden layer to the output layer determine the way feature information is transferred and transformed between the network layers; b (1) b (2) b (3) These are the bias vectors for each corresponding layer, used for addition with the weighted sum to increase the model's flexibility; σ(·) is the nonlinear activation function of the hidden layer, using the hyperbolic tangent function tanh(·), and its expression is σ(x) = (e x -e -x ) / (e x +e -x The result of the linear combination is subjected to a nonlinear transformation; f out (·) represents the activation function of the output layer, which is a linear activation function, i.e., f out (x) = x, directly output the weighted sum as the final inversion estimate.

[0102] The network parameters Θ= {W} are adjusted using the augmented training set obtained in step 3.2. (1) W (2) W (3) ,b (1) ,b (2) ,b (3) Supervised training is performed, with the training process aimed at minimizing the network's output prediction value D. pred With the true value D true The mean square error between them is the target:

[0103] (19);

[0104] In the formula, Δ is the mean square error between the predicted output value and the true value; N is the number of training samples.

[0105] After training, the network parameters Θ are fixed and integrated into the embedded processor of the pulsed eddy current detector. In actual detection, the system inputs the feature vector F, which is acquired and preprocessed in real time, into the network. After forward propagation, it directly outputs the quantitative defect parameter estimate D, realizing a real-time, objective, and quantitative assessment of corrosion defects in metal components in soil.

[0106] Example 2:

[0107] This embodiment provides a pulsed eddy current detection system for detecting corrosion of metal components in soil media, used to execute the method provided in Embodiment 1. The system includes:

[0108] The excitation coil is configured to receive an excitation current pulse to generate an excitation electromagnetic field in the soil medium;

[0109] The detection coil is configured to receive the induced voltage signal generated by the eddy current field induced by the metal component in the soil;

[0110] A processor, electrically connected to the excitation coil and the detection coil, is configured to execute the following modules:

[0111] The parameter modeling module is configured to perform step S1 in the method, establish a eddy field distribution model in the soil medium, and quantitatively calculate the background eddy loss parameters of the soil.

[0112] The feature extraction and mapping module is configured to perform step S2 in the method, extract physical mapping feature quantities from the induced voltage signal and establish a mapping relationship between them and the conductivity of the metal component;

[0113] The fusion inversion module is configured to perform step S3 in the method, fusing multiple types of feature quantities and inverting the geometric parameters of corrosion defects through a machine learning model.

[0114] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed in this application can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0115] In the embodiments provided in this application, it should be understood that the disclosed devices / terminal equipment and methods can be implemented in other ways. For example, the device / terminal equipment embodiments described above are merely illustrative. For instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual coupling or direct coupling or communication connection may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0116] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0117] If the integrated module / unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying the computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc. It should be noted that the content included in the computer-readable medium can be appropriately added or removed according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electrical carrier signals and telecommunication signals.

[0118] The implementation of all or part of the processes in the methods of the above embodiments can also be accomplished by a computer program product. When the computer program product is run on a terminal device, the terminal device can implement the steps in the various method embodiments described above.

[0119] The embodiments described above are only used to illustrate the technical solutions of this application, and are not intended to limit it. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A method for detecting corrosion of metal components in soil media based on pulsed eddy currents, characterized in that, The method employs a detection system comprising an excitation coil and a detection coil, and includes the following steps: Step S1: Perform electromagnetic modeling on the soil medium in the detection area, calculate the equivalent electromagnetic parameters of the soil, and quantitatively analyze the attenuation effect of the soil on the excitation electromagnetic field and the background eddy current loss generated by the attenuation operation. Step S2: Based on the calculation results of step S1, extract at least one physical mapping feature quantity related to the corrosion state of the metal component from the induced voltage signal obtained from the detection coil, and establish a quantitative mapping relationship between the feature quantity and the local equivalent conductivity of the metal component. Step S3: Combine the physical mapping features extracted in step S2 with the waveform transformation features further extracted from the induced voltage signal to construct a multidimensional feature vector. Then, use a machine learning model to invert and obtain the geometric parameters of corrosion defects in the metal component.

2. The method for detecting corrosion of metal components in soil media based on pulsed eddy currents according to claim 1, characterized in that, The fundamental frequency and main spectral components of the excitation current pulse in the excitation coil are restricted to a frequency band below 10MHz.

3. The method for detecting corrosion of metal components in soil media based on pulsed eddy currents according to claim 1, characterized in that, Step S1 specifically includes: S101. Treat the non-uniform soil medium as an equivalent medium with uniform electromagnetic parameters, and calculate its equivalent conductivity σ. eff With equivalent permeability μ eff ; S102, based on the equivalent conductivity σ eff With equivalent permeability μ eff Calculate the total eddy current loss power induced by the excitation electromagnetic field in the soil. and its equivalent loss resistance in the equivalent circuit of the detection system .

4. The method for detecting corrosion of metal components in soil media based on pulsed eddy currents according to claim 3, characterized in that, The physical mapping features extracted in step S2 include the zero-crossing time T0 of the derivative of the induced voltage signal and the eddy current loss feature value P constructed based on the eddy current loss power density formula. e .

5. The method for detecting corrosion of metal components in soil media based on pulsed eddy currents according to claim 4, characterized in that, The quantitative mapping relationship in step S2 includes calculating the local equivalent conductivity σ of the metal component using the time T0 when the derivative crosses zero. m Relationship: σ m =k1T0+b1, where the coefficients k1 and b1 are obtained through a calibration procedure, which includes measuring and performing regression analysis on standard samples in various simulated soil environments with different equivalent loss resistances R_ec to determine the coefficients k1 and b1.

6. The method for detecting corrosion of metal components in soil media based on pulsed eddy currents according to claim 1, characterized in that, The waveform transformation features include time-frequency domain transformation features and signal statistical transformation features. The time-frequency domain transformation features include amplitude, derivative, or harmonic component features extracted from the signal waveform or spectrum. The signal statistical transformation features include principal component features obtained by performing multivariate statistical analysis on the signal sequence.

7. The method for detecting corrosion of metal components in soil media based on pulsed eddy currents according to claim 1, characterized in that, Step S3 specifically includes: S301: Perform anti-interference preprocessing on the induced voltage signal; S302: Construct the multidimensional feature vector and train a machine learning model using a data-augmented training set to establish a mapping relationship between the multidimensional feature vector and the geometric parameters of the corrosion defect.

8. The method for detecting corrosion of metal components in soil media based on pulsed eddy currents according to claim 7, characterized in that, The machine learning model is a feedforward neural network, which includes at least one input layer, one output layer and two hidden layers; the hidden layers use non-linear activation functions and the output layers use linear activation functions.

9. The method for detecting corrosion of metal components in soil medium based on pulsed eddy current according to claim 7, characterized in that, The geometric parameters of the corrosion defect include at least: the depth d of the corrosion location from the surface of the soil medium and the length l of the corrosion extending along the component.

10. A pulsed eddy current detection system for detecting corrosion of metal components in soil media, characterized in that, The system for performing the method according to any one of claims 1 to 9 comprises: The excitation coil is configured to receive an excitation current pulse to generate an excitation electromagnetic field in the soil medium; The detection coil is configured to receive the induced voltage signal generated by the eddy current field induced by the metal component in the soil; A processor, electrically connected to the excitation coil and the detection coil, is configured to execute the following modules: The parameter modeling module is configured to perform step S1 in the method, establish a eddy field distribution model in the soil medium, and quantitatively calculate the background eddy loss parameters of the soil. The feature extraction and mapping module is configured to perform step S2 in the method, extract physical mapping feature quantities from the induced voltage signal and establish a mapping relationship between them and the conductivity of the metal component; The fusion inversion module is configured to perform step S3 in the method, fusing multiple types of feature quantities and inverting the geometric parameters of corrosion defects through a machine learning model.

Images