Crack quantification method and apparatus based on orthogonal twinning and used for oil and gas pipelines, and storage medium

Abstract

A crack quantification method and apparatus based on orthogonal twinning and used for oil and gas pipelines, and a storage medium. The method comprises: acquiring a triaxial magnetic flux leakage measurement signal modulus value of an inner wall crack or an outer wall crack under a condition of a unidirectional direct-current excitation, and acquiring a dynamic magnetic signal or an eddy current signal under a condition of a dynamic magnetic field excitation that is orthogonal to the direct-current excitation; inputting the triaxial magnetic flux leakage measurement signal modulus value into a crack signal orthogonal twin model, so as to obtain a magnetic flux leakage enhancement estimation signal, wherein the magnetic flux leakage enhancement estimation signal is a function of an orthogonal twin triaxial magnetic flux leakage signal modulus value, the orthogonal twin triaxial magnetic flux leakage signal modulus value is a magnetic flux leakage response signal modulus value under a condition of a virtual orthogonal twin direct-current excitation, and the virtual orthogonal twin direct-current excitation and the unidirectional direct-current excitation are in the same plane and have equal magnitudes and perpendicular directions; and extracting feature vectors from the magnetic flux leakage enhancement estimation signal and the dynamic magnetic signal or the eddy current signal, and inputting the feature vectors into a crack scale estimation model, so as to obtain the size and inclination angle of a crack.

Description

Oil and gas pipeline crack quantification method, device and storage medium based on orthogonal twinning

[0001] This application claims priority to the Chinese patent application filed with the China Patent Office on August 16, 2023, with application number 2023110347949 and invention name “Oil and gas pipeline crack quantification method, device and storage medium based on orthogonal twinning”, the content of which should be understood as incorporated into this application by reference. Technical Field

[0002] The embodiments of the present disclosure relate to, but are not limited to, the field of crack detection technology, and in particular to a method and device for quantifying cracks in oil and gas pipelines based on orthogonal twinning, and a storage medium. The applicable objects of twinning include, but are not limited to, pipe body crack signals, elbow crack signals, girth weld crack signals of oil and gas pipelines or hydrogen pipelines, high-speed rail crack signals, tank bottom plate cracks, and other ferromagnetic metal material crack signals. Background Art

[0003] Ferromagnetic metal materials such as oil and gas pipelines, hydrogen pipelines, high-speed rails, and oil storage tanks are subject to long-term operation in complex natural environments and pressure loads. Stress cracks, fatigue cracks, brittle cracks, hydrogen-induced cracks, and other microcracks can develop within the materials or on their inner and outer walls. These cracks can then develop into visible fissures or larger metal losses. Cracks are one of the primary sources of defects in metal materials during service.

[0004] Compared to cracks, metal loss is easier to detect due to its larger size. Currently, several mature technologies are available to accurately detect, identify, and quantify metal loss defects. However, cracks are typically small in size, making detection, identification, and quantification of these early-stage microcracks more challenging, especially at inspection speeds greater than 2 m / s. This is especially true when cracks are in their initial stages of development or are in the early stages of service.

[0005] Furthermore, the crack response signal during crack detection is closely related to the excitation signal, excitation direction, and sensing sensor. For example, in unidirectional magnetic flux leakage (MFL) inspection of oil and gas pipelines (axial or circumferential MFL), based on the excitation direction, for cracks of the same size, the larger the tilt angle modulus, the greater the amplitude of the MFL response signal; the smaller the tilt angle modulus, the smaller the amplitude of the MFL response signal. Therefore, the signal of an axial crack or fissure under axial excitation is extremely weak, and is even considered unattainable by the industry at industrial level. Changing from axial to circumferential excitation can increase the strength of the axial crack or fissure signal, but circumferential excitation detectors are less effective at detecting circumferential cracks or fissures. Consequently, a serial combination MFL detector solution has emerged in the industry, combining an axial excitation MFL detector with a circumferential excitation MFL detector. This combination MFL detector solution significantly increases equipment manufacturing and inspection engineering costs. Furthermore, due to the limitations of its principle, traditional MFL testing is clearly inadequate or impossible to detect, identify, and quantify tiny cracks.

[0006] Summary of the Invention

[0007] The disclosed embodiments provide a method, device, and storage medium for quantifying cracks in oil and gas pipelines based on orthogonal twinning. The concept of "orthogonal twinning" is defined as obtaining a response signal corresponding to a twin magnetic field that is coplanar, equal in magnitude, and perpendicular to the unidirectional DC excitation, given the original response signal of a known unidirectional DC excitation. This response signal is defined as a "twin response signal" (e.g., a twin magnetic flux leakage response signal or a twin dynamic magnetic / eddy current response signal). This response signal also forms a twin relationship with the response signal of the original unidirectional DC excitation, i.e., the "original response signal" (e.g., the original magnetic flux leakage response signal or the original dynamic magnetic or eddy current response signal). By jointly processing the "twin response signal" and the "original response signal," the quantization accuracy of "small-angle cracks" in oil and gas pipelines can be significantly improved, especially the quantization accuracy of cracks distributed parallel to the excitation direction, including the crack length accuracy, width accuracy, depth accuracy, and inclination angle accuracy. A "small-angle crack" here refers to a crack with a positive or negative deviation of no more than 20° from the excitation direction. Orthogonal twinning methods include, but are not limited to, machine learning model mapping, mathematical analysis, circuit (or electromagnetic) simulation, and real-signal generation from hardware devices. The process of transforming the "original response signal" into the "twin response signal" is defined as "orthogonal twin transformation."

[0008] The technical solutions adopted in the embodiments of the present disclosure are:

[0009] An embodiment of the present disclosure provides an oil and gas pipeline crack quantification method based on orthogonal twinning, including: obtaining the modulus of a three-axis leakage magnetic measurement signal of an inner wall crack or an outer wall crack under unidirectional DC excitation conditions, and obtaining a dynamic magnetic signal or eddy current signal under a dynamic magnetic field excitation condition orthogonal to the DC excitation, wherein the unidirectional DC excitation refers to a DC excitation method with a single direction; inputting the modulus of the three-axis leakage magnetic measurement signal into an orthogonal twin model of a crack signal to obtain a leakage magnetic enhancement estimation signal, wherein the leakage magnetic enhancement estimation signal is a function of the simulated orthogonal twin three-axis leakage magnetic signal modulus, wherein the orthogonal twin three-axis leakage magnetic signal modulus is the modulus of a three-axis leakage magnetic response signal under virtual orthogonal twin DC excitation conditions, wherein the virtual orthogonal twin DC excitation is in the same plane as the unidirectional DC excitation, has equal size, and is perpendicular in direction; extracting a characteristic vector from the leakage magnetic enhancement estimation signal and the dynamic magnetic signal or eddy current signal, and inputting the characteristic vector into a crack scale estimation model to obtain the size and inclination angle of the crack.

[0010] Optionally, extracting a eigenvector from the leakage magnetic enhancement estimation signal and the dynamic magnetic signal or eddy current signal includes: obtaining at least one first eigenvalue based on the leakage magnetic enhancement estimation signal; obtaining at least one second eigenvalue based on the dynamic magnetic signal or eddy current signal; and combining the first eigenvalue and the second eigenvalue into a eigenvector.

[0011] Optionally, the first eigenvalue includes: a major axis, a minor axis, a major axis inclination angle, and a peak value. Obtaining the first eigenvalue based on the leakage magnetic enhancement estimation signal includes: binarizing the leakage magnetic enhancement estimation signal according to a preset binarization threshold to obtain a binary leakage magnetic signal; calculating the major axis, minor axis, major axis inclination angle of the edge contour of the binary leakage magnetic signal and the peak value of the leakage magnetic enhancement estimation signal, and using the obtained major axis, minor axis, major axis inclination angle, and peak value as the first eigenvalue; the dynamic magnetic signal or eddy current signal includes a front and rear coil differential signal and a left and right coil differential signal, and the second eigenvalue includes: the peak value and peak-to-peak value spacing of the front and rear coil differential signal, and the peak value and peak value spacing of the left and right coil differential signal.

[0012] Optionally, the crack signal orthogonal twin model is an autoencoder model implemented by a convolutional neural network, and the autoencoder model includes an encoder part and a decoder part; the crack scale estimation model is a fully connected neural network.

[0013] Optionally, the leakage magnetic enhancement estimation signal is a simulated optimal three-axis leakage magnetic signal modulus S Ao Compared with the simulated orthogonal twin three-axis magnetic leakage signal modulus S ⊥ The function f(αS Ao ,αS ⊥ ), the simulated optimal three-axis magnetic leakage signal modulus S AoIs to make ||αS A -M A || F The minimum simulated three-axis magnetic leakage signal modulus, where M A is the modulus of the three-axis magnetic flux leakage measurement signal, S A Substituting the crack size into the magnetic dipole model to obtain the simulated triaxial magnetic flux leakage signal modulus, ‖·‖ F is the Frobenius norm of the matrix, and α is the adjustment factor; are respectively the X-axis component, Y-axis component and Z-axis component of the simulated three-axis magnetic leakage signal, the X-axis direction is the same as the excitation direction of the unidirectional DC excitation, the Y-axis direction is perpendicular to the X-axis direction in the plane of the oil and gas pipeline, and the Z-axis direction is perpendicular to the X-axis direction and the Y-axis direction respectively; They are the X-axis component, Y-axis component and Z-axis component of the orthogonal twin three-axis magnetic flux leakage signal; K≥1; the adjustment factor α is equal to M A The maximum element of the matrix is ​​divided by the maximum element of the initial matrix, where the initial matrix is ​​a matrix composed of the simulated triaxial magnetic flux leakage signal modulus values ​​calculated by substituting the measured crack dimensions into the magnetic dipole model.

[0014] Optionally, the method further comprises: for a plurality of crack samples, using the modulus of the three-axis magnetic flux leakage measurement signal under unidirectional DC excitation conditions and the corresponding magnetic flux leakage enhancement estimation signal to train the crack signal orthogonal twin model, wherein the magnetic flux leakage enhancement estimation signal corresponding to each crack sample is generated by the following method: taking the measured scale of the crack sample as the initial crack scale; using M A The maximum element of the matrix is ​​divided by the maximum element of the initial matrix to obtain the adjustment factor α, M A is the modulus of the three-axis magnetic flux leakage measurement signal; calculate ||αS A -M A || F , where S A Substituting the crack size into the magnetic dipole model to obtain the simulated triaxial magnetic flux leakage signal modulus, ‖·‖ F is the Frobenius norm of the matrix; repeatedly adjust the crack scale and input the crack scale into the crack magnetic dipole model until ||αS is obtained. A -M A || F The minimum optimal crack size, S corresponding to the optimal crack size A That is the optimal simulation three-axis magnetic leakage signal modulus S Ao ; Input the optimal crack size into the crack orthogonal magnetic dipole model to obtain the simulated orthogonal twin triaxial magnetic leakage signal modulus S ⊥; Substitute the said S Ao and S ⊥ into the formula K≥1, thereby generating the magnetic flux leakage enhancement estimation signal.

[0015] Optionally, for the multiple crack samples, training the crack signal orthogonal siamese model using the modulus value of the triaxial magnetic flux leakage measurement signal under the condition of unidirectional DC excitation and the corresponding magnetic flux leakage enhancement estimation signal, includes: Marking the modulus value of the triaxial magnetic flux leakage measurement signal of each crack sample under the condition of unidirectional DC excitation and the corresponding magnetic flux leakage enhancement estimation signal as a set of orthogonal siamese mapping pairs; Randomly splitting N sets of orthogonal siamese mapping pairs into a training set and a test set according to the ratio k:(1-k), where 0<k<1; Training the crack signal orthogonal siamese model using k×N sets of the orthogonal siamese mapping pairs, and testing the crack signal orthogonal siamese model using (1-k)×N sets of the orthogonal siamese mapping pairs.

[0016] Optionally, after training the crack signal orthogonal siamese model, training the crack size estimation model using the feature vector, includes: Marking the first eigenvalue and the second eigenvalue of each crack sample and the true value of the corresponding crack size and inclination angle as a set of crack size estimation mapping pairs; Training the crack size estimation model using the crack size estimation mapping pairs corresponding to k×N sets of the orthogonal siamese mapping pairs, and testing the crack size estimation model using the crack size estimation mapping pairs corresponding to (1-k)×N sets of the orthogonal siamese mapping pairs.

[0017] The embodiment of the present disclosure also provides an oil and gas pipeline crack quantification device based on orthogonal siamese, including: a magnetic flux leakage sensor probe under the condition of unidirectional DC excitation, a dynamic magnetic or eddy current sensor probe under the conditions of unidirectional DC excitation and dynamic magnetic field excitation orthogonal to the DC excitation, a memory storing instructions, algorithms, and models, a processor executing the oil and gas pipeline crack quantification method described above, and a bus system connecting each unit.

[0018] The embodiment of the present disclosure also provides a storage medium, on which a computer program is stored, and when the program is executed by a processor, it implements the oil and gas pipeline crack quantification method based on orthogonal siamese described in any embodiment of the present disclosure.

[0019] The oil and gas pipeline crack quantification method, device, and storage medium based on orthogonal siamese in the embodiment of the present disclosure determine the objective existence of cracks by relying on crack response signals based on multiple detection principles. When the signal-to-noise ratio of the crack signal is low, orthogonal siamese is performed on the crack signal to obtain an enhanced signal corresponding to the crack, and then based on the enhanced crack signal of orthogonal siamese, a machine learning model is applied to achieve high-precision size quantification of the weak crack signal of the original measurement.

[0020] The oil and gas pipeline crack quantification method of the disclosed embodiment, through the orthogonal twin model of crack signals and based on unidirectional DC excitation conditions, proposes a new method and technology for accurately detecting, identifying, and quantifying tiny cracks, especially cracks distributed parallel to the excitation direction. This method has important practical significance for the safe operation of major infrastructure such as oil and gas pipelines, hydrogen pipelines, high-speed rails, and oil storage tanks.

[0021] Other features and advantages of the present disclosure will be described in the following description, and in part will become apparent from the description, or will be understood by practicing the present disclosure. Other advantages of the present disclosure can be realized and obtained through the solutions described in the description and the drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] The accompanying drawings are used to provide an understanding of the technical solution of the present disclosure and constitute a part of the specification. Together with the embodiments of the present disclosure, they are used to explain the technical solution of the present disclosure and do not constitute a limitation to the technical solution of the present disclosure.

[0023] FIG1 is a schematic diagram of an ellipsoidal profile and orthogonal magnetic field of a crack in an oil and gas pipeline according to an exemplary embodiment of the present disclosure;

[0024] FIG2 is a flow chart of an oil and gas pipeline crack quantification method and device based on orthogonal twinning and a storage medium according to an exemplary embodiment of the present disclosure;

[0025] FIG3A is a schematic diagram of a framework structure of an oil and gas pipeline crack quantification method and apparatus based on orthogonal twinning, and a storage medium in a model training phase according to an exemplary embodiment of the present disclosure;

[0026] FIG3B is a schematic diagram of a framework structure of an oil and gas pipeline crack quantification method and apparatus based on orthogonal twinning, and a storage medium in a model use phase according to an exemplary embodiment of the present disclosure;

[0027] FIG4A is a diagram of an exemplary embodiment of the present disclosure showing a "crack signal orthogonal twin model" (ML OT )’s loss function iteration process;

[0028] FIG4B is a diagram of an exemplary embodiment of the present disclosure showing a "crack signal orthogonal twin model" (ML OT )'s coefficient of determination iterative process;

[0029] FIG5A is a diagram of a “crack size estimation model” (ML size )’s loss function iteration process;

[0030] FIG5B is a diagram of a “crack size estimation model” (ML size )'s coefficient of determination iterative process;

[0031] FIG6A is a 3-D diagram showing the signal enhancement effect of a crack with a 0° tilt angle according to an exemplary embodiment of the present disclosure;

[0032] FIG6B is a top view of the signal enhancement effect of a crack with a 0° tilt angle according to an exemplary embodiment of the present disclosure;

[0033] FIG7A is a 3-D diagram showing the signal enhancement effect of a crack with a 3° tilt angle according to an exemplary embodiment of the present disclosure;

[0034] FIG7B is a top view of the signal enhancement effect of a 3° tilt angle crack according to an exemplary embodiment of the present disclosure;

[0035] FIG8A is a 3-D diagram showing the signal enhancement effect of an 18° tilt angle crack according to an exemplary embodiment of the present disclosure;

[0036] FIG8B is a top view of the signal enhancement effect of a crack with an inclined angle of 18° according to an exemplary embodiment of the present disclosure;

[0037] FIG9 is a schematic structural diagram of an oil and gas pipeline crack quantification device based on orthogonal twinning according to an exemplary embodiment of the present disclosure. DETAILED DESCRIPTION

[0038] To make the objectives, technical solutions and advantages of the present disclosure more clearly understood, the embodiments of the present disclosure will be described in detail below with reference to the accompanying drawings. It should be noted that, unless there is a conflict, the embodiments and features in the embodiments of the present disclosure can be combined with each other in any manner.

[0039] Unless otherwise defined, the technical or scientific terms used in the embodiments of the present disclosure should have the ordinary meaning understood by people with ordinary skills in the field to which the present disclosure belongs. The words "first", "second" and similar words used in the embodiments of the present disclosure do not indicate any order, quantity or importance, but are only used to distinguish different components. The words "include" or "comprising" and similar words mean that the elements or objects preceding the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects.

[0040] The present disclosure proposes the concept of "orthogonal twinning", as shown in FIG1 , which aims to have only one direction of DC excitation H A Under the condition of , the virtual twin generates an excitation magnetic field H that is in the same plane, equal in size and perpendicular to the DC excitation magnetic field. ⊥ The leakage magnetic response signal is generated. Then, the leakage magnetic enhancement estimation signal is obtained by jointly processing the original response signal and the twin response signal. Then, the leakage magnetic enhancement estimation signal is used to estimate the crack length, width, depth, and inclination angle, and finally, high-precision quantification of cracks, especially cracks with small inclination angles, is achieved. As shown in Figure 1, the inclination angle of the crack is defined as the difference between the major axis direction of the crack ellipsoid contour and the original excitation direction H. AThe angle between them ranges from (-90°, 90°]. The crack quantification method disclosed in the present invention is applicable to cracks with small tilt angles but is not limited to cracks with small tilt angles. Here, "cracks with small tilt angles" refer to cracks with a positive or negative deviation of no more than 20° tilt angle based on the excitation direction. In the embodiment of the present invention, unidirectional DC excitation refers to a DC excitation method with a single direction, including but not limited to the existing axial DC excitation method, circumferential DC excitation method, and spiral DC excitation method.

[0041] As shown in Figure 1, a spatial rectangular coordinate system is established on the inner wall of the oil and gas pipeline. The geometric center of the crack ellipsoid outline is used as the coordinate origin, and the excitation direction (H A The direction perpendicular to the X-axis in the pipe wall plane is the Y-axis. The direction perpendicular to the pipe wall and pointing to the inside of the pipe is the Z-axis. The opposite direction of the Y-axis is the H-axis. ⊥ direction; in the plane of the pipe wall, the major axis direction of the crack ellipsoid outline is the X′ axis direction, and the inclination angle relative to the X axis direction is θ; the direction perpendicular to the X′ axis in the plane of the pipe wall is the Y′ axis direction, and its inclination angle relative to the Y axis direction is θ; the direction coinciding with the Z axis is the Z′ axis direction; the flow direction of oil and gas in the pipeline is the axial direction; the direction perpendicular to the axial direction and along the circumference of the pipe wall is the circumferential direction; the direction coinciding with the Z axis is the radial direction.

[0042] As shown in FIG2 , the embodiment of the present disclosure provides a method and device for quantifying cracks in oil and gas pipelines based on orthogonal twinning, and a storage medium, including the following steps:

[0043] 201. Establish (train) a quantitative model of oil and gas pipeline cracks based on orthogonal twins;

[0044] 202. Use (test) a crack quantification model for oil and gas pipelines based on orthogonal twinning.

[0045] The crack quantification method of the embodiment of the present disclosure is described using oil and gas pipelines as an example. The implementation steps in other ferromagnetic metal materials can refer to this method and will not be repeated in this disclosure.

[0046] In some exemplary embodiments, referring to FIG. 3A , given the crack size and its response signal, step 201 may include the following steps:

[0047] 1) According to different detection principles, a weak matrix signal of crack response with a scale of (L, W, D, θ) is obtained under unidirectional DC excitation conditions. According to different detection principles, a 3-axis leakage magnetic field and a 2-axis dynamic magnetic field or eddy current sensor ultra-high resolution integrated probe is used to obtain a weak matrix signal of crack response with a scale of (L, W, D, θ) under unidirectional DC excitation conditions. For example, a DC excitation field H in direction A is obtained. A Three-axis magnetic flux leakage measurement signal modulus under the conditions And the DC excitation field H in direction A A Dynamic magnetic or eddy current signal under conditions Where (L, W, D, θ) are the actual measured length, width, depth, and inclination angle of the crack, respectively; Respectively represent the X-axis component, Y-axis component and Z-axis component of the three-axis magnetic flux leakage measurement signal; They represent the dynamic magnetic or eddy current signals of the front and rear coil differential (hereinafter referred to as "front and rear differential") and the left and right coil differential (hereinafter referred to as "left and right differential") of the ultra-high resolution integrated probe circuit board of the 3-axis leakage magnetic field and the 2-axis dynamic magnetic or eddy current sensor respectively. A and D A All are matrix signals. At least include the indicator signal of crack existence, such as and The crack presence indication signal includes the boundary information of the crack. The signal can distinguish whether the crack is an inner wall crack (ID) or an outer wall crack (OD).

[0048] 2) In the same direction A, the DC excitation magnetic field H A Under these conditions, the inner wall crack magnetic dipole model or the outer wall crack magnetic dipole model is applied to generate the scale (L s ,W s ,D s ,θ s ) Simulated triaxial magnetic flux leakage signal modulus of virtual crack Among them, S A is the matrix signal, (L s ,W s ,D s ,θ s ) are the length, width, depth and inclination angle of the virtual crack, which are intermediate variables. Their initial values ​​can be the corresponding known crack measured dimensions (L, W, D, θ); They are the X-axis component, Y-axis component and Z-axis component of the simulated three-axis magnetic leakage signal. If it is an inner wall crack, based on the "inner wall crack magnetic dipole model", the scale (L s ,W s ,D s ,θ s )calculate The formulas are (1) to (2); if it is an external wall crack, based on the “external wall crack magnetic dipole model”, the scale (L s ,W s ,D s ,θ s )calculate The formulas are (3) to (5). Among them, f xX′(·), f yX′ (·), f zX′ (·) are respectively in H A The three-axis (X′ axis, Y′ axis, Z′ axis) integral operator of the magnetic dipole under the excitation of the X′ axis component; f xY′ (·), f yY′ (·), f zY′ (·) are respectively in H A The three-axis (X′, Y′, and Z′) integral operator of the magnetic dipole excited by the Y′-axis component of μ0 = 4π×10 -7 (H / m) is the vacuum permeability, μ r is the relative magnetic permeability of the pipe wall under saturation conditions; σ x′ is the magnetic charge surface density of the magnetic dipole in the X′ axis direction, σ y′ is the surface density of the magnetic charge of the magnetic dipole in the Y′-axis direction; if it is an inner wall crack, h is the lift-off value of the sensor integrated probe; if it is an outer wall crack, h is the sum of the remaining wall thickness of the pipe at the crack and the lift-off value of the sensor integrated probe.

[0049] 3) In order to make the effect of crack simulation consistent with the effect of actual crack measurement, the objective function and constraint conditions are established: min||αS A -M A || F (6)

[0050] in,‖·‖ F is the Frobenius norm of the matrix, O s (L s ,W s ,D s ,θ s )=1 is the ellipsoid equation for fitting the crack contour; adjustment factor: α=max(M A ) / max(S A_initial ), equal to M A The maximum element of the matrix is ​​divided by the maximum element of the initial matrix, and the initial matrix (S A_initial ) is the simulated triaxial magnetic flux leakage signal modulus obtained by substituting the actual crack size (L, W, D, θ) into the magnetic dipole model. By adjusting the virtual crack size (L s ,W s ,D s ), so that the above objective function (6) is satisfied. The optimal virtual crack size that satisfies (6) is recorded as The corresponding simulated three-axis magnetic leakage signal modulus is recorded as In this paper, crack scale includes crack size and inclination angle.

[0051] Complex crack shapes can be decomposed into a combination of simple strip contours, and the strip contours can be abstracted into simple geometric shapes such as cuboids, ellipsoids, and elliptical cylinders.

[0052] In the crack quantization method of the embodiment of the present disclosure, the crack contour is approximated as an ellipsoid. In other exemplary embodiments, the crack contour can also be approximated as a rectangular parallelepiped, an elliptical cylinder, etc., which is not limited in the embodiment of the present disclosure.

[0053] 4) Apply the inner wall crack orthogonal magnetic dipole model or the outer wall crack orthogonal magnetic dipole model to calculate the orthogonal twin triaxial magnetic leakage signal modulus. A Twin magnetic fields H in the same plane, equal in magnitude, and perpendicular in direction ⊥ Under the conditions of incentive (|H ⊥ |=|H A |), for scale The virtual crack is applied to the magnetic dipole model to generate the corresponding leakage magnetic response signal modulus, that is, the simulated orthogonal twin triaxial leakage magnetic signal modulus in They are the X-axis component, Y-axis component and Z-axis component of the simulated orthogonal twin three-axis magnetic leakage signal. If it is an inner wall crack, based on the "inner wall crack orthogonal magnetic dipole model", the scale calculate The formulas are shown in (8) to (9); if it is an external wall crack, based on the “external wall crack orthogonal magnetic dipole model”, the scale calculate The formulas are shown in (10) to (12). xOT′ is the surface density of the orthogonal magnetic charge of the magnetic dipole in the X′ axis direction, σ yOT′ is the surface density of the orthogonal magnetic charge of the magnetic dipole in the Y′ axis direction.

[0054] 5) For the inner wall crack and the outer wall crack, obtain N orthogonal twin mapping pairs of different sizes and different inclination angles and the measured size set. A ,αS Ao ,αS ⊥ , and the measured scale (L, W, D, θ) form a one-to-one correspondence, and the i-th group of orthogonal twin mapping pairs is defined as The corresponding actual measurement scales are (L i , W i , D i , θ i ); then change the values of (L, W, D, θ), that is, replace with new measured cracks, and repeat steps 1) to 4) until N orthogonal twin mapping pairs with different sizes and different inclination angles are obtained for inner wall cracks and outer wall cracks respectively, that is The set of corresponding measured scales is expressed as {(L i , W i , D i , θ i ), i = 1, 2,..., N}.

[0055] 6) For inner wall cracks and outer wall cracks, respectively, randomly split the N groups of orthogonal twin mapping pairs and the set of measured scales into training sets and test sets according to the ratio k:(1 - k) (0 < k < 1), and their numbers are k×N and (1 - k)×N respectively. In addition, for inner wall cracks and outer wall cracks, respectively, randomly select a part from the k×N training sets as the validation set during model training.

[0056] 7) For inner wall cracks and outer wall cracks, for the training set, with as the input and as the output, i = 1, 2,..., k×N, build and train the machine learning model ML OT , and name it the "orthogonal twin model of crack signals", including the "orthogonal twin model of inner wall crack signals" and the "orthogonal twin model of outer wall crack signals". Among them, represents performing a logical operation on and , for example K≥1; obviously, that is, the output signal after the orthogonal twin transformation is enhanced.

[0057] 8) For inner wall cracks and outer wall cracks, for the training set, based on step 7), when i = 1, 2,..., k×N, then input into the trained ML OT model to obtain k×N groups of estimation results

[0058] 9) For inner wall cracks and outer wall cracks, for the training set, with as the threshold (c is between 0 and 1,示例性的, c = 0.5), perform 0 - 1 binary processing on to obtain where i = 1, 2,..., k×N.

[0059] 10) For the inner wall cracks and outer wall cracks, for the training set, by and as well as Compute k×N sets of eigenvectors. The edge contour is usually an ellipse. The major axis of the edge profile is a i , the minor axis is b i , the inclination angle of the long axis is The peak value is p i , The peak-to-peak value is The peak-to-peak spacing is The peak value is The peak spacing is When i=1,2,…,k×N, a total of k×N groups of eigenvectors are formed

[0060] 11) For the inner wall cracks and outer wall cracks, for the training set, As input, (L i ,W i ,D i ,θ i ) is the output, i=1,2,…,k×N, build and train the machine learning model ML size , named it "crack scale estimation model", which includes "inner wall crack scale estimation model" and "outer wall crack scale estimation model".

[0061] At this point, the trained ML OT Models and ML size The models together constitute the "crack quantification model".

[0062] In some exemplary embodiments, in the second stage, the use or testing of the "crack quantification model" is mainly applicable to a test set that has not been trained with the model in the first stage or to new cracks on a real oil and gas pipeline. Referring to FIG3B , step 202 may include the following steps:

[0063] 1) According to different detection principles, the weak matrix signal of crack response under unidirectional DC excitation conditions is obtained. For example, the DC excitation field H in direction A is obtained. A Three-axis magnetic flux leakage measurement signal modulus under the conditions And the DC excitation field H in direction A A Dynamic magnetic or eddy current signal under conditions At least include indicators of crack presence, such as and The crack presence indication signal includes the boundary information of the crack. The signal can distinguish whether the crack is an inner wall crack (ID) or an outer wall crack (OD).

[0064] 2) For inner wall cracks or outer wall cracks, M A The signal is input to the ML trained in the first stage OT ("Inner wall crack signal orthogonal twin model" or "Outer wall crack signal orthogonal twin model"), output leakage magnetic enhancement estimation signal

[0065] 3) is the threshold value, Perform 0-1 binary processing and get c is between 0 and 1, and for example, c=0.5.

[0066] 4) Pass and as well as Calculate eigenvectors remember The major axis of the edge profile is a, the minor axis is b, and the inclination angle of the major axis is The peak value is p, The peak-to-peak value is The peak-to-peak spacing is The peak value is The peak spacing is This set of feature vectors

[0067] 5) The feature vector Input to the ML trained in stage 1 size ("Inner wall crack size estimation model" or "Outer wall crack size estimation model"), output the estimated value of the inner wall or outer wall crack size

[0068] 6) Repeat steps 1) to 5) until all new cracks in the test set of the first stage or the real oil and gas pipeline are quantified.

[0069] In order to better understand the oil and gas pipeline crack quantification method, device and storage medium based on orthogonal twinning provided by the present disclosure, the following is combined with (X80 steel, 15.30 mm wall thickness) This exemplary embodiment of oil and gas pipeline crack quantification further illustrates the technical solution of the present disclosure.

[0070] This embodiment uses an ultra-high-resolution integrated probe based on 3-axis magnetic flux leakage and 2-axis dynamic magnetic or eddy current sensors to collect weak crack signals. 3-axis magnetic flux leakage refers to the axial, radial, and circumferential components of the leakage magnetic field; 2-axis dynamic magnetic or eddy current refers to the probe circuit board's ability to simultaneously collect front-to-back differential dynamic magnetic or eddy current signals and left-to-right differential dynamic magnetic or eddy current signals. The front-to-back differential dynamic magnetic or eddy current signals can detect and identify cracks in all directions, but as the crack tilt angle modulus decreases, the amplitude of the front-to-back differential dynamic magnetic or eddy current signals gradually decreases. Meanwhile, the amplitude of the left-to-right differential dynamic magnetic or eddy current signals increases as the crack tilt angle modulus decreases. The two complement each other, making up for each other's shortcomings. Therefore, using an ultra-high-resolution integrated probe with 3-axis magnetic flux leakage and 2-axis dynamic magnetic or eddy current sensors can determine the presence of a crack, distinguish internal from external cracks, and identify suspected crack areas.

[0071] The entire implementation process is divided into two stages. The first stage is the establishment stage of the "crack quantification model" (model training stage); the second stage is the use stage of the "crack quantification model" (model testing stage).

[0072] In the first stage, the crack size and its response signal are known, and the steps of establishing the "crack quantification model" (model training) include:

[0073] According to step 1), use 3-axis leakage magnetic field and 2-axis dynamic magnetic field or eddy current detector to carry out the A pulling experiment was conducted on an artificially cracked oil and gas pipeline (X80 steel, 15.30mm wall thickness), thereby obtaining the pulling data of the artificial crack. The use of axial excitation means that the excitation direction (X-axis direction) coincides with the axial direction, and the established coordinate system is as shown in Figure 1. First, 2160 artificial cracks of different sizes and inclination angles were processed on the pulling pipeline as samples for model training, including 1080 inner wall cracks (ID) and 1080 outer wall cracks (OD); for the inner wall cracks and outer wall cracks, 75% of the samples (810) were randomly assigned as the training set (marked by train), and the other 25% of the samples (270) were assigned as the test set (marked by test). 25% of the samples (203) randomly selected from the training set were defined as the validation set (marked by val), and their scale distribution ranges are shown in Table 1. Import the above pulling data into the "Crack Original Sample Calibration Software", select and export the crack original sample data file in the known area of ​​the processed crack, and save it in txt format. Each data file is named with the measured crack "length (L) - width (W) - depth (D) - inclination angle (θ) - inside or outside", thereby obtaining the different scales (L) under axial excitation conditions. i ,W i ,D i ,θ i ) crack response weak signal as well as i=1,2,…,2160. Each crack original sample data file consists of 5 matrix data, namely, leakage magnetic axial data matrix Magnetic Flux Leakage Radial Data Matrix Magnetic Flux Leakage Toroidal Data Matrix Dynamic magnetic or eddy current before and after differential data matrix and dynamic magnetic or eddy current left and right differential data matrix and, For inner wall cracks or outer wall cracks, 810 samples in the training set are prepared for the first stage training of the "crack quantification model", and 270 samples in the test set are prepared for the second stage testing of the "crack quantification model".

[0074] Table 1 Sample distribution of model training

[0075] According to step 2), the axial excitation field H of the pulling experiment is measured in advance. A Then, under the same axial excitation field conditions, for the inner wall crack and the outer wall crack, 1080 artificial crack samples of different sizes and different inclination angles were measured with the measured scale recorded in their file names. As the initial value, the magnetic dipole model is applied to generate the simulated triaxial magnetic leakage signal modulus of the virtual crack i=1,2,…,1080; if it is an inner wall crack, based on the “inner wall crack magnetic dipole model”, the scale calculate The formulas are (13) to (14); if it is an external wall crack, based on the “external wall crack magnetic dipole model”, the scale calculate The formulas are (15) to (17). Among them, f xX′ (·), f yX′ (·), f zX′ (·) are respectively in H A The three-axis (X′ axis, Y′ axis, Z′ axis) integral operator of the magnetic dipole under the excitation of the X′ axis component; f xY′ (·), f yY′ (·), f zY′ (·) are respectively in H A The three-axis (X′, Y′, and Z′) integral operator of the magnetic dipole excited by the Y′-axis component of μ0 = 4π×10 -7 (H / m), is the vacuum permeability, μ r is the relative magnetic permeability of the pipe wall under saturation conditions is the magnetic charge surface density of the magnetic dipole in the X′ axis direction corresponding to the i-th artificial crack sample, is the magnetic charge surface density of the magnetic dipole in the Y′ axis direction corresponding to the i-th artificial crack sample; if it is an inner wall crack, h i is the lift-off value of the sensor integrated probe at the i-th artificial crack sample. If it is an outer wall crack, h i It is the sum of the remaining wall thickness of the pipe at the i-th artificial crack sample and the lift-off value of the sensor integrated probe.

[0076] According to step 3), in order to make the effect of crack simulation consistent with the actual crack measurement effect, for the inner wall crack and outer wall crack, 1080 artificial crack samples of different sizes and different inclination angles were respectively established, and the objective function and constraint conditions were established one by one:

[0077] Among them, when i=1,2,…,1080, is the ellipsoid equation of the fitting crack profile corresponding to the i-th artificial crack sample; adjustment factor: express The maximum element of the matrix is ​​divided by the maximum element of the i-th initial matrix, and the i-th initial matrix That is, the measured crack size (L i ,W i ,D i ,θ i ) is substituted into the simulated three-axis magnetic leakage signal modulus calculated by the magnetic dipole model. i ,W i ,D i ,θ i ) is the initial value and the virtual crack size is adjusted The above objective function (18) is satisfied. The optimal virtual crack size that satisfies (18) is denoted as The corresponding simulated three-axis magnetic flux leakage signal is recorded as

[0078] According to step 4), the orthogonal magnetic dipole model of the inner wall crack and the orthogonal magnetic dipole model of the outer wall crack are used to calculate the orthogonal twin triaxial magnetic leakage signal modulus of 2160 artificial crack samples with different sizes and different inclination angles. A Twin magnetic fields H in the same plane, equal in magnitude, and perpendicular in direction ⊥ Under the conditions of incentive (|H ⊥ |=|H A|), for inner wall cracks or outer wall cracks, when i=1,2,…,1080, for scale The virtual crack is applied to the magnetic dipole model to generate the corresponding leakage magnetic response signal modulus, that is, the simulated orthogonal twin triaxial leakage magnetic signal modulus If it is an inner wall crack, based on the "inner wall crack orthogonal magnetic dipole model", the scale calculate The formulas are shown in (20) to (21); if it is an external wall crack, based on the “external wall crack orthogonal magnetic dipole model”, the scale calculate The formulas are shown in (22) to (24). is the surface density of the orthogonal magnetic charge of the magnetic dipole in the X′ axis direction corresponding to the i-th artificial crack sample, is the surface density of the magnetic charge orthogonal to the Y′ axis of the magnetic dipole corresponding to the i-th artificial crack sample.

[0079] According to step 5), for the inner wall crack and the outer wall crack, 1080 orthogonal twin mapping pairs of different sizes and different inclination angles and the measured size set are obtained respectively. In the above steps 1) to 4), for the inner wall crack or the outer wall crack, There are 1080 orthogonal twin mapping pairs of different sizes and tilt angles, and the corresponding measured scale set is expressed as {(L i ,W i ,D i ,θ i ),i=1,2,…,1080}.

[0080] According to step 6), for inner wall cracks and outer wall cracks, the ratio coefficient of splitting the training set and the test set in this embodiment is set to k=0.75, that is, the number of training set samples accounts for 75% of the total number of samples; and the proportion of the validation set in the training set is 25%.

[0081] According to step 7), for the inner wall crack and outer wall crack, for the training set samples (810), As input, For output, i = 1, 2, ..., 810, build and train the machine learning model ML OT, namely the "orthogonal twin model of crack signal", including the "orthogonal twin model of inner wall crack signal" and the "orthogonal twin model of outer wall crack signal". This embodiment specifically uses the convolutional neural network autoencoder (CNN Autoencoder) algorithm to build the orthogonal twin model of crack signal. The model requires that when i = 1, 2, ..., 810, the input For a 200×200 matrix, output The matrix is ​​also 200×200, with a physical step of 0.5 mm between two points. If the physical step exceeds 0.5 mm, cubic spline interpolation can be used. If the matrix dimensions do not meet the requirements, the matrix center is used as the matrix center, and edges exceeding 200 dimensions are pruned, and edges less than 200 dimensions are padded with zeros.

[0082] The convolutional neural network autoencoder designed in this embodiment has a total of 37 layers, of which the encoding part contains 18 layers and the decoding part contains 19 layers. The network architecture built using Python language is as follows:

[0083] In the above Python code, Conv2D represents a two-dimensional convolutional layer, SpatialDropout2D represents a two-dimensional spatial random zeroing layer, BatchNormalization represents a batch normalization layer, ReLU represents an activation function layer, MaxPooling2D represents a two-dimensional pooling layer, UpSampling2D represents a two-dimensional upsampling layer, and Activation represents an activation layer. Each line of Python represents a layer, and the output of each layer serves as the input to the next layer, which in turn serves as the input to the next layer.

[0084] According to step 8), for the inner wall crack and outer wall crack, for the training set samples (810), based on step 7), when i = 1, 2, ..., 810, Input to the trained ML OT In the model, 810 sets of estimation results were obtained

[0085] According to step 9), for the inner wall crack and outer wall crack, for the training set samples (810), when i = 1, 2, ..., 810, is the threshold value, Perform 0-1 binary processing and get

[0086] According to step 10), for the inner wall crack and outer wall crack, for the training set samples (810), when i = 1, 2, ..., 810, by and as well as Calculate 810 sets of eigenvectors. The edge contour is usually an ellipse. The major axis of the edge profile is a i , the minor axis is b i , the inclination angle of the long axis is The peak value is p i , The peak-to-peak value is The peak-to-peak spacing is The peak value is The peak spacing is When i=1,2,…,810, a total of 810 sets of eigenvectors are formed

[0087] According to step 11), for the inner wall crack and the outer wall crack, for the training set samples (810), when i = 1, 2, ..., 810, As input, (L i ,W i ,D i ,θ i ) as output, build and train the machine learning model ML size , namely the "crack scale estimation model", which includes the "inner wall crack scale estimation model" and the "outer wall crack scale estimation model". This embodiment specifically uses the fully connected neural network (DNN) algorithm to build the crack scale estimation model. The model requires that when i = 1, 2, ..., 810, the input is an 8×1 vector, output (L i ,W i ,D i ,θ i ) is a 4×1 vector.

[0088] The fully connected neural network designed in this embodiment has a total of 5 layers, namely 1 input layer, 3 hidden layers, and 1 output layer. The network architecture built using Python language is as follows:

[0089] In the above Python code, Dense is used to implement the fully connected layer. The middle layer has 10 neurons and the output layer has 4 neurons. kernel_initializer is the keyword that specifies the initialization method, and activation is the keyword that specifies the activation function.

[0090] Taking the “Orthogonal Twin Model of Inner Wall Crack Signal” as an example, the number of iteration steps of model training is set to 500. The iteration process of model training is shown in Figures 4A, 4B, 5A, and 5B. Figure 4A is the “Orthogonal Twin Model of Crack Signal” (ML OT ) loss function iteration process, Figure 4B is the “crack signal orthogonal twin model” (ML OT The coefficient of determination (R 2 ) iterative process, the iteration ends "crack signal orthogonal twin model" (ML OT ) are shown in Table 2. OT After the model converged, the loss function of the training set and the validation set was finally reduced to 0.0002, indicating that ML OT The final mean square error (MSE) of the model is very small and there is no overfitting. On the other hand, ML OT The coefficient of determination of the model training set finally reached 0.9371, and the coefficient of determination of the validation set finally reached 0.9281, indicating that ML OT The final fitting effect of the model is good and basically meets the requirements of orthogonal twin transformation.

[0091] Table 2 The “crack signal orthogonal twin model” (ML OT )

[0092] Taking the “inner wall crack signal orthogonal twin model” as an example, Figure 5A is the “crack scale estimation model” (ML size ) loss function iteration process, Figure 5B is the “crack size estimation model” (ML size The coefficient of determination (R 2 ) iterative process, the iteration ends "crack size estimation model" (ML size ) are shown in Table 3. size After the model converged, the loss function of the training set was finally reduced to 0.0044, and the loss function of the validation set was finally reduced to 0.0027, indicating that ML size The final mean square error (MSE) of the model is also very small, and there is no overfitting. On the other hand, ML size The coefficient of determination of the model training set finally reached 0.9907, and the coefficient of determination of the validation set finally reached 0.9940, indicating that ML size The final fitting effect of the model is very good, and the estimation accuracy of crack length, width, depth and inclination angle is very high.

[0093] Table 3 “Crack size estimation model” (ML size )

[0094] In the second stage, the model was tested on the 270 test set samples that were not trained in the first stage. The steps for using the "crack quantization model" (model testing) are as follows:

[0095] According to step 1), the weak matrix signal of crack response under unidirectional DC excitation is obtained according to different detection principles. After completing step 1) of the first stage, the weak matrix signal of crack response of 270 test set samples has been obtained under axial excitation conditions for inner wall cracks or outer wall cracks. as well as i=1,2,…,270. Each crack original sample data file consists of 5 matrix data, namely, leakage magnetic axial data matrix Magnetic Flux Leakage Radial Data Matrix Magnetic Flux Leakage Toroidal Data Matrix Dynamic magnetic or eddy current before and after differential data matrix and dynamic magnetic or eddy current left and right differential data matrix At least include indicators of crack presence, such as and The crack presence indication signal includes the boundary information of the crack. The signal can distinguish whether the crack is an inner wall crack (ID) or an outer wall crack (OD).

[0096] According to step 2), for inner wall cracks or outer wall cracks, for the test set samples (270), when i = 1, 2, ..., 270, Signal input to ML OT ("Inner wall crack signal orthogonal twin model" or "Outer wall crack signal orthogonal twin model"), output leakage magnetic enhancement estimation signal

[0097] According to step 3), for inner wall cracks or outer wall cracks, for the test set samples (270), when i = 1, 2, ..., 270, is the threshold value, Perform 0-1 binary processing and get

[0098] According to step 4), for inner wall cracks or outer wall cracks, for the test set samples (270), when i = 1, 2, ..., 270, by and as well as Calculate 270 sets of eigenvectors. The edge contour is usually an ellipse. The major axis of the edge profile is a i , the minor axis is b i , the inclination angle of the long axis is The peak value is p i , The peak-to-peak value is The peak-to-peak spacing is The peak value is The peak spacing is When i=1,2,…,270, a total of 270 sets of eigenvectors are formed

[0099] According to step 5), for the inner wall crack or outer wall crack, for the test set samples (270), when i = 1, 2, ..., 270, the feature vector Input to ML size ("Inner wall crack size estimation model" or "Outer wall crack size estimation model"), output the estimated value of the inner wall or outer wall crack size

[0100] Follow step 6) to ensure that all 270 cracks in the first phase test set are quantified for inner wall cracks or outer wall cracks.

[0101] Figures 6A, 6B, 7A, 7B, 8A, and 8B are orthogonal twinning effect diagrams of three groups of inner wall crack signals of different scales in the test set samples. From left to right, the three figures are the original input leakage magnetic measurement signal modulus M A , the picture title is "Input Signal", after ML OT Output leakage magnetic enhancement estimation signal The image is titled "Predicting Output Signals"; and ML OT The ideal output signal f(αS Ao ,αS ⊥ ), the image title is “True Output Signal.”

[0102] Figures 6A and 6B show an inner wall crack with L = 30 mm, W = 0.3 mm, D = 1.4 mm, and θ = 0° in the test set. OT (“Inner wall crack signal orthogonal twin model”) Signal enhancement effect diagram. It can be seen that the original input signal M A The leakage magnetic field peak is only 10Gs. After orthogonal twin transformation, ML OTThe peak value of the leakage magnetic field of the output signal reaches 120Gs, and the signal is enhanced by 12 times. In addition, since the inclination angle θ of the crack sample is 0°, it is an axial crack. Figure 6B shows that the original input signal M A It shows a double peak feature, and it is even difficult to distinguish it as an axial crack. However, after orthogonal twin transformation, ML OT Output signal It clearly reflects its axial crack characteristics, so The signal provides favorable conditions for quantifying the size of the axial crack.

[0103] Figures 7A and 7B show an inner wall crack with L = 40 mm, W = 0.3 mm, D = 2.5 mm, and θ = 3° in the test set. OT (“Inner wall crack signal orthogonal twin model”) Signal enhancement effect diagram. It can be seen that the original input signal M A The leakage magnetic field peak is only 12Gs. After orthogonal twin transformation, ML OT The leakage magnetic field peak of the output signal reaches 150Gs, and the signal is enhanced by 12.5 times. Similarly, since the inclination angle of the crack sample is θ = 3°, it belongs to a small inclination angle crack. Figure 7B shows that after the orthogonal twin transformation, the ML OT Output signal The crack signal characteristics are enhanced, providing favorable conditions for high-precision quantification of its scale.

[0104] Figures 8A and 8B show an inner wall crack with L = 60 mm, W = 0.4 mm, D = 2.9 mm, and θ = 18° in the test set. OT (“Inner wall crack signal orthogonal twin model”) Signal enhancement effect diagram. It can be seen that the original input signal M A The leakage magnetic field peak is only 60Gs. After orthogonal twin transformation, ML OT The peak value of the leakage magnetic field of the output signal reaches 200Gs, and the signal is enhanced by 3.3 times, so ML OT The output signal is also conducive to the high-precision quantification of the crack.

[0105] By performing interval estimation on the estimation errors of the 270 inner wall cracks in the test set, the calculation results are shown in Table 4 at a 90% confidence level. It can be seen that the confidence interval of the inner wall crack length estimation error is (-1.046mm, 1.026mm), the confidence interval of the width estimation error is (-0.006mm, 0.005mm), the confidence interval of the depth estimation error is (-0.214mm, 0.193mm), and the confidence interval of the inclination angle estimation error is (-2.282°, 2.452°). In summary, the oil and gas pipeline crack quantification method, device, and storage medium based on orthogonal twinning disclosed in this disclosure have very high scale quantization accuracy.

[0106] Table 4 Confidence intervals of the estimation error of the crack quantification model

[0107] The disclosed embodiments also provide an orthogonal twin-based oil and gas pipeline crack quantification device, comprising: 1) an ultra-high-resolution integrated probe of 3-axis magnetic leakage and 2-axis dynamic magnetic / eddy current sensors, capable of collecting front and rear differential dynamic magnetic / eddy current signals as well as left and right differential dynamic magnetic / eddy current signals; 2) a memory for storing instructions, algorithms, and models; 3) a processor for executing an orthogonal twin-based oil and gas pipeline crack quantification method; 4) a display for displaying quantification results; and 5) a bus system for connecting various units.

[0108] In one example, as shown in FIG9 , an orthogonal twin-based oil and gas pipeline crack quantification device may include: a 3-axis magnetic flux leakage and 2-axis dynamic magnetic / eddy current sensor ultra-high resolution integrated probe 910, a memory 920, a processor 930, a display 940, and a bus system 950, wherein the 3-axis magnetic flux leakage and 2-axis dynamic magnetic / eddy current sensor ultra-high resolution integrated probe 910, the memory 920, the processor 930, and the display 940 are connected via the bus system 950; the 3-axis magnetic flux leakage and 2-axis dynamic magnetic / eddy current sensor ultra-high resolution integrated probe 910, the memory 920, the processor 930, and the display 940 are connected via the bus system 950; The ultra-high-resolution integrated probe 910 is used to detect and obtain weak crack response signals with a scale of (L, W, D, θ) under unidirectional DC excitation conditions, including three-axis leakage magnetic signals and front-to-back differential dynamic magnetic / eddy current signals and left-to-right differential dynamic magnetic / eddy current signals. The memory 920 is used to store instructions and "crack signal orthogonal twin model" and "crack scale estimation model", etc. The processor 930 is used to execute the instructions stored in the memory 920 to quantify cracks using the orthogonal twin-based oil and gas pipeline crack quantification method. Specifically, the processor 930 can train the "crack signal orthogonal twin model" and "crack scale estimation model" respectively; it can enhance weak crack signals at different inclination angles and perform scale estimation; and finally, the quantification results are displayed on the display 940.

[0109] It should be understood that the memory 920 may include read-only memory and random access memory, and provides instructions and data, including the aforementioned "crack signal orthogonal twin model" and "crack size estimation model," to the processor 930. A portion of the memory 920 may also include non-volatile random access memory. For example, the memory 920 may also store device type information.

[0110] The processor 930 may be a central processing unit (CPU), or may be another general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. A general-purpose processor may be a microprocessor, or the processor 930 may be any conventional processor.

[0111] In addition to displaying the quantified results of the cracks, the display 940 can also display the crack detection data in the memory 920 through the crack analysis software.

[0112] In addition to the data bus, the bus system 950 may also include a power bus, a control bus, a status signal bus, and the like.

[0113] During implementation, the processing performed by the orthogonal twin-based oil and gas pipeline crack quantification device can be completed by hardware integrated logic circuits or software instructions in the processor 930. That is, the steps of the orthogonal twin-based oil and gas pipeline crack quantification device in the embodiment of the present disclosure can be executed by a hardware processor, or by a combination of hardware and software modules in the processor 930. The software module can be located in a storage medium such as random access memory, flash memory, read-only memory, programmable read-only memory, or electrically erasable programmable memory, registers, etc. The storage medium is located in the memory 920, and the processor 930 reads the information in the memory 920 and completes the steps of the above method in combination with its hardware. To avoid repetition, it will not be described in detail here.

[0114] An embodiment of the present disclosure also provides a storage medium storing executable instructions, which, when executed by a processor, can implement the oil and gas pipeline crack quantification method based on orthogonal twinning provided in any of the above embodiments of the present disclosure; in addition, the storage medium can also store crack quantification models, including a "crack signal orthogonal twin model" and a "crack scale estimation model"; the small inclination angle cracks are quantified using the trained "crack signal orthogonal twin model" and "crack scale estimation model" respectively.

[0115] It will be appreciated by those skilled in the art that all or some of the steps, systems, and functional modules / units in the methods disclosed above may be implemented as software, firmware, hardware, and appropriate combinations thereof. In hardware implementations, the division between the functional modules / units mentioned in the above description does not necessarily correspond to the division of physical components; for example, a physical component may have multiple functions, or a function or step may be performed by several physical components in cooperation. Some or all components may be implemented as software executed by a processor, such as a digital signal processor or a microprocessor, or implemented as hardware, or implemented as an integrated circuit, such as an application-specific integrated circuit. Such software may be distributed on a computer-readable medium, which may include a computer storage medium (or non-transitory medium) and a communication medium (or temporary medium). As is well known to those skilled in the art, the term computer storage medium includes volatile and non-volatile, removable, and non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules, or other data). Computer storage media include, but are not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital versatile disks (DVD) or other optical disk storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to store the desired information and can be accessed by a computer. In addition, it is well known to those skilled in the art that communication media generally embodies computer-readable instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave or other transport mechanism, and may include any information delivery media.

[0116] Although the embodiments disclosed in this disclosure are as described above, the contents described are merely embodiments adopted to facilitate understanding of the disclosure and are not intended to limit the disclosure. Any person skilled in the art to which the disclosure belongs may make any modifications and changes in the form and details of the implementation without departing from the spirit and scope of the disclosure. However, the scope of protection of the disclosure shall still be based on the scope defined by the appended claims.

Claims

1. A method for quantifying cracks in oil and gas pipelines based on orthogonal twinning, comprising: Obtaining a modulus value of a triaxial magnetic flux leakage measurement signal of an inner wall crack or an outer wall crack under unidirectional DC excitation, and obtaining a dynamic magnetic signal or an eddy current signal under a dynamic magnetic field excitation orthogonal to the DC excitation, wherein the unidirectional DC excitation refers to a DC excitation mode with a single direction; Inputting the modulus of the three-axis magnetic flux leakage measurement signal into the orthogonal twin model of the crack signal to obtain a magnetic flux leakage enhancement estimation signal, wherein the magnetic flux leakage enhancement estimation signal is a function of the modulus of the simulated orthogonal twin three-axis magnetic flux leakage signal, and the modulus of the orthogonal twin three-axis magnetic flux leakage signal is the modulus of the three-axis magnetic flux leakage response signal under the condition of virtual orthogonal twin DC excitation, wherein the virtual orthogonal twin DC excitation is in the same plane as the unidirectional DC excitation, has equal size, and is perpendicular to the direction; Extracting a characteristic vector from the leakage magnetic enhancement estimation signal and the dynamic magnetic signal or eddy current signal, and inputting the characteristic vector into a crack scale estimation model to obtain the size and inclination angle of the crack; The orthogonal twin is defined as, under the condition that the original response signal of unidirectional DC excitation is known, a response signal corresponding to a twin magnetic field that is in the same plane, equal in size, and perpendicular to the unidirectional DC excitation is obtained. The obtained response signal is defined as a twin response signal. The twin response signal and the original response signal form a twin relationship. The process of transforming the original response signal into the twin response signal is defined as the orthogonal twin transformation.

2. The method according to claim 1, wherein Extracting a feature vector from the magnetic flux leakage enhancement estimation signal and the dynamic magnetic signal or the eddy current signal includes: Obtain at least one first eigenvalue according to the magnetic flux leakage enhancement estimation signal; Obtain at least one second characteristic value according to the dynamic magnetic signal or the eddy current signal; The first eigenvalue and the second eigenvalue are combined into an eigenvector.

3. The method according to claim 2, wherein: The first eigenvalue includes: a major axis, a minor axis, a major axis tilt angle, and a peak value. The first eigenvalue is obtained according to the magnetic flux leakage enhancement estimation signal, including: Binarizing the magnetic flux leakage enhancement estimation signal according to a preset binarization threshold to obtain a binarized magnetic flux leakage signal; Calculating the major axis, minor axis, major axis tilt angle of the edge profile of the binary magnetic flux leakage signal and the peak value of the magnetic flux leakage enhancement estimation signal, and using the obtained major axis, minor axis, major axis tilt angle and peak value as the first eigenvalue; The dynamic magnetic signal or eddy current signal includes a front and rear coil differential signal and a left and right coil differential signal, and the second characteristic value includes: the peak value and peak-to-peak distance of the front and rear coil differential signal, and the peak value and peak distance of the left and right coil differential signal.

4. The method according to claim 2, wherein: The crack signal orthogonal twin model is an autoencoder model implemented by a convolutional neural network, and the autoencoder model includes an encoder part and a decoder part; the crack scale estimation model is a fully connected neural network.

5. The method according to claim 1, wherein The leakage magnetic enhancement estimation signal is the simulated optimal three-axis leakage magnetic signal modulus S Ao Compared with the simulated orthogonal twin three-axis magnetic leakage signal modulus S ⊥ The function f(αS Ao ,αS ⊥ ), the simulated optimal three-axis magnetic flux leakage signal modulus S Ao Is to make ||αS A -M A || F The minimum simulated three-axis magnetic leakage signal modulus, where M A is the modulus of the three-axis magnetic flux leakage measurement signal, S A Substituting the crack size into the magnetic dipole model to obtain the simulated triaxial magnetic flux leakage signal modulus, ‖·‖ F is the Frobenius norm of the matrix, and α is the adjustment factor; are respectively the X-axis component, Y-axis component and Z-axis component of the simulated three-axis magnetic leakage signal, the X-axis direction is the same as the excitation direction of the unidirectional DC excitation, the Y-axis direction is perpendicular to the X-axis direction in the plane of the oil and gas pipeline, and the Z-axis direction is perpendicular to the X-axis direction and the Y-axis direction respectively; They are the X-axis component, Y-axis component and Z-axis component of the orthogonal twin three-axis magnetic flux leakage signal; The adjustment factor α is equal to M A The maximum element of the matrix is ​​divided by the maximum element of the initial matrix, where the initial matrix is ​​a matrix composed of the simulated triaxial magnetic flux leakage signal modulus values ​​calculated by substituting the measured crack dimensions into the magnetic dipole model.

6. The method according to claim 1, wherein The method further comprises: For multiple crack samples, the crack signal orthogonal twin model is trained using the modulus of the three-axis magnetic flux leakage measurement signal under unidirectional DC excitation conditions and the corresponding magnetic flux leakage enhancement estimation signal, wherein the magnetic flux leakage enhancement estimation signal corresponding to each crack sample is generated by the following method: Taking the measured scale of the crack sample as the initial crack scale; Use M A The maximum element of the matrix is ​​divided by the maximum element of the initial matrix to obtain the adjustment factor α, M A is the modulus of the three-axis magnetic flux leakage measurement signal; Calculate ||αS A -M A || F , where S A Substituting the crack size into the magnetic dipole model to obtain the simulated triaxial magnetic flux leakage signal modulus, ‖·‖ F is the Frobenius norm of the matrix; The crack size is repeatedly adjusted and input into the crack magnetic dipole model until the crack size is obtained. A -M A || F The minimum optimal crack size, S corresponding to the optimal crack size A That is the optimal simulation three-axis magnetic leakage signal modulus S Ao ; The optimal crack size is input into the crack orthogonal magnetic dipole model to obtain the simulated orthogonal twin triaxial magnetic leakage signal modulus S ⊥ ; The S Ao With S ⊥ Substitute into the formula K≥1, thereby generating the magnetic flux leakage enhancement estimation signal.

7. The method according to claim 6, wherein: For the plurality of crack samples, the crack signal orthogonal twin model is trained using the modulus of the three-axis magnetic flux leakage measurement signal under unidirectional DC excitation conditions and the corresponding magnetic flux leakage enhancement estimation signal, including: The modulus of the triaxial magnetic flux leakage measurement signal of each crack sample under unidirectional DC excitation and the corresponding magnetic flux leakage enhancement estimation signal are marked as a set of orthogonal twin mapping pairs; The N sets of orthogonal twin mapping pairs are randomly split into training sets and test sets according to the ratio k:(1-k), where 0 <k<1; The crack signal orthogonal twin model is trained using k×N groups of the orthogonal twin mapping pairs, and the crack signal orthogonal twin model is tested using (1-k)×N groups of the orthogonal twin mapping pairs.

8. The method according to claim 7, wherein: After training the crack signal orthogonal twin model, using the feature vector to train the crack scale estimation model includes: Marking the first eigenvalue and the second eigenvalue of each crack sample with the corresponding true values ​​of the crack size and tilt angle as a set of crack size estimation mapping pairs; The crack scale estimation model is trained using the crack scale estimation mapping pairs corresponding to the k×N groups of orthogonal twin mapping pairs, and the crack scale estimation model is tested using the crack scale estimation mapping pairs corresponding to the (1-k)×N groups of orthogonal twin mapping pairs.

9. An oil and gas pipeline crack quantification device based on orthogonal twinning, comprising: A magnetic flux leakage sensor probe under unidirectional DC excitation conditions, a dynamic magnetic field or eddy current sensor probe under unidirectional DC excitation and dynamic magnetic field excitation conditions orthogonal to DC excitation, a memory for storing instructions, algorithms, and models, a processor for executing the oil and gas pipeline crack quantification method according to any one of claims 1 to 8, and a bus system connecting the various units.

10. A storage medium storing a program for an oil and gas pipeline crack quantification method based on orthogonal twinning, wherein the program, when executed by a processor, implements the oil and gas pipeline crack quantification method according to any one of claims 1 to 8.

Images