Electromagnetic regulation and detection method and system for structural defects of offshore oil well control equipment

Abstract

The present application belongs to the technical field of nondestructive testing and defect evaluation, and particularly relates to a marine oil well control equipment structural defect electromagnetic regulation and control detection method and a detection system. The marine oil well control equipment structural defect electromagnetic regulation and control detection method and the detection system obtain crack multi-feature signals through electromagnetic excitation signal amplitude-frequency-phase regulation and control, obtain crack angle information with the aid of a residual network, and realize accurate identification of crack length and depth information. A service life prediction network is used to realize effective prediction of residual fatigue life. The detection system comprises a monitoring probe, an excitation signal generation unit and a monitoring signal processing and acquisition unit. The monitoring probe is composed of an excitation coil group, an array sensor and a front-end signal processing circuit. The excitation signal generation unit comprises a signal generator and a power amplifier. The monitoring signal processing and acquisition unit comprises a multiplexing and amplification filtering circuit, a band-pass filtering circuit, a phase-locked amplification circuit and a signal acquisition board.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of nondestructive testing and defect evaluation, and particularly relates to a method and system for electromagnetic regulation and detection of structural defects of offshore well control equipment. BACKGROUND

[0002] Under the influence of complex loads caused by wind, waves, and currents, offshore well control equipment is prone to fatigue damage. Defect cracks caused by this fatigue damage often expand in an irregular manner, which brings many challenges to defect detection and evaluation. For example, existing nondestructive testing techniques usually require large-scale removal of surface attachments and thorough cleaning of corrosion protection coatings when detecting structural defects of well control equipment, resulting in a cumbersome defect detection process, low efficiency, and high surface cleaning and coating repair costs.

[0003] Alternating current field measurement (ACFM) is a new nondestructive testing technique that is widely used in defect detection of surface cracks in conductive materials. This detection technique induces a uniform alternating current on the surface of a conductive test piece through a detection probe. When a defect exists, the path of the alternating current is disturbed, resulting in a distortion of the spatial magnetic field. By measuring the distorted magnetic field, the crack can be identified and quantitatively evaluated. Due to its non-contact detection and quantitative analysis capabilities, ACFM has been widely used in defect detection of various marine engineering structures.

[0004] However, further research has found that the detection object of existing ACFM detection techniques is mostly transverse cracks (i.e., cracks perpendicular to the direction of induced current). The crack depth is evaluated by the extreme value amplitude of the characteristic signal Bz, and the crack length is estimated by the extreme value interval. However, for longitudinal cracks (i.e., cracks parallel to the direction of induced current) and inclined cracks, it is difficult to achieve effective evaluation due to the significant reduction in detection sensitivity. In addition, existing ACFM detection techniques also have the disadvantages of inconsistent current field strength and large differences in defect response signal regulation, which seriously restrict the evaluation accuracy of structural defects. SUMMARY

[0005] The application provides a marine oil well control equipment structural defect electromagnetic regulation and control detection method and a detection system.

[0006] To solve the above technical problems, the application adopts the following technical solutions:

[0007] The marine oil well control equipment structural defect electromagnetic regulation and control detection method comprises the following steps:

[0008] Step S101: regulating and exciting induced current to obtain spatial magnetic field distortion of the marine oil well control equipment to be detected; and calculating an 8*8 magnetic field signal Bz matrix Z;

[0009] The magnetic field signal Bz matrix Z satisfies: ; wherein Bz0 ij Bz is the magnetic field signal value of the i-th row and the j-th column of the Bz matrix;

[0010] Step S102: performing absolute value operation on the magnetic field signal Bz matrix Z and drawing a gray scale image to obtain a Bz image;

[0011] Step S103: performing smoothing processing on the Bz image by using gray scale interpolation;

[0012] The crack images of different angles are fused by superposition to obtain complete crack contour information;

[0013] Step S104: adding a real angle label to all crack angle fusion images to form an angle evaluation database;

[0014] The angle evaluation database is trained by using a residual network; and the trained angle evaluation database can be used to obtain the angle prediction value of an unknown sample;

[0015] Step S105: regulating and controlling the amplitude and phase parameters to make the induced current perpendicular to the crack to obtain crack end point information;

[0016] Step S106: when the induced current is perpendicular to the crack, regulating and controlling the excitation signal frequency to change the skin depth of the induced current to obtain at least a first frequency image, a second frequency image and a third frequency image; wherein the first frequency is less than the second frequency, and the second frequency is less than the third frequency;

[0017] The first frequency image, the second frequency image and the third frequency image are respectively placed in R, G and B channels of a color image, and a crack size evaluation fusion image is obtained by stacking;

[0018] Step S107: Building a cross fusion network, extracting the Bz peak value in the Bz image and the peak distance ; using a random forest algorithm, a crack prediction system is established;

[0019] The prediction value of the cross fusion network is set as OUTn, the prediction value of the crack prediction system is set as OUTe, and the real value of the crack size is set as Label. The loss function obtained by constructing satisfies: ;

[0020] Among them, is the weight of the crack prediction system participating in the training;

[0021] Step S108: initializing the cross fusion network, embedding the crack prediction system, and training using training data; wherein the trained crack prediction system can be used to obtain the size prediction value of any angle crack in the structural defect of the offshore oil well control equipment.

[0022] More preferably, the following steps are also included:

[0023] Step S201: obtaining the control data, using the trained cross fusion network to extract the crack length information feature of the offshore oil well control equipment ;

[0024] and obtaining the stress data of the offshore oil well control equipment ;

[0025] Step S202: obtaining the remaining fatigue life by cyclic loading until the offshore oil well control equipment workpiece is completely broken ;

[0026] Step S203: constructing a multi-layer fully connected neural network topology structure;

[0027] Taking the crack length information feature of the offshore oil well control equipment , the stress data of the offshore oil well control equipment , the remaining fatigue life as the three nodes of the input layer, adopting a pyramid structure as the hidden layer, and taking the fatigue life prediction value as the single node of the output layer, a multi-dimensional physical constraint for fatigue life prediction is realized;

[0028] Step S204: training the life prediction network;

[0029] The regression evaluation index MSE is used to calculate the deviation between the predicted value and the measured value and build a crack length information feature containing offshore well control equipment , stress data of offshore well control equipment , residual fatigue life The physical consistency loss of the residual error of the three types of features ;

[0030] Wherein, the deviation between the predicted value and the measured value , satisfies: ; is the actual observation value, is the model prediction value, and n is the sample number;

[0031] The weighted loss function is designed, and the following is obtained: ; wherein, Determined by cross-validation;

[0032] , satisfies: ; wherein, , , is the weight, y is the model prediction value, B is the Basquin model output, P is the Paris model output, The non-physical value penalty term is used to ensure that y is non-negative, so as to avoid unreasonable results of model output;

[0033] Step S205: In the fatigue life prediction process, fix the time step, set the load as a fixed value, obtain the true crack length and simultaneously monitor the crack propagation image;

[0034] The multi-frequency fusion high-resolution magnetic field image of the induced current vertical crack is obtained through SRCNN, and the database is built;

[0035] Step S206: Construct the PredRNN++ spatiotemporal prediction neural network;

[0036] Wherein, the input of the PredRNN++ spatiotemporal prediction neural network is the known crack magnetic field image, and the output of the PredRNN++ spatiotemporal prediction neural network is the predicted crack magnetic field image;

[0037] Step S207: Train the PredRNN++ spatiotemporal prediction neural network;

[0038] The root mean square error RMSE and the structural similarity index SSIM are used as the indexes for evaluating the accuracy of the prediction results of the PredRNN++ spatiotemporal prediction neural network; wherein, the root mean square error RMSE satisfies: ; is the actual observation value, is the model prediction value, n is the sample number;

[0039] The structural similarity index SSIM satisfies: ; wherein, is the prediction value, is the actual true value, , is the mean value, , is the variance, is the covariance, , is a constant;

[0040] The PredRNN++ spatiotemporal prediction neural network trained can be used to obtain the crack propagation prediction image;

[0041] Step S208: adding a true length label to all the predicted crack propagation prediction images to form a length evaluation database; training the length evaluation database by using a cross-fusion network; wherein the trained length evaluation database can be used to obtain the crack propagation length prediction value of the crack propagation prediction image of the marine oil well control equipment structural defect.

[0042] On the other hand, the application also provides a marine oil well control equipment structural defect electromagnetic regulation and detection system, comprising a monitoring probe, an excitation signal generating unit, a monitoring signal processing and collecting unit;

[0043] The monitoring probe is composed of an excitation coil group, an array sensor and a front-end signal processing circuit; the excitation coil group comprises an upper excitation coil group and a lower excitation coil group in orthogonal distribution; wherein the upper excitation coil group and the lower excitation coil group are each composed of two semicircular coils, and the two semicircular coils of the same layer in the upper excitation coil group and the lower excitation coil group are connected with opposite sine alternating currents to generate a uniform induced current field; the array sensor is arranged above the uniform induced current field generated by the excitation coil group, and is used to collect the induced magnetic field signal perpendicular to the surface direction of the marine oil well control equipment to be detected; the front-end signal processing circuit is connected with the excitation coil group, and is used to adjust the resonant frequency of the monitoring probe according to the external signal;

[0044] The excitation signal generating unit comprises a signal generator and a power amplifier; wherein the signal generator is used to provide a sine alternating excitation signal for the excitation coil group; and the power amplifier is used to amplify and process the sine alternating excitation signal;

[0045] The monitoring signal processing and collecting unit comprises a multiplexing and amplification filtering circuit, a band-pass filtering circuit, a phase-locked amplification circuit and a signal collecting board card.

[0046] More preferably, the upper excitation coil group and the lower excitation coil group in the excitation coil group adopt the same frequency parameters, so as to induce X-direction induced current Jx and Y-direction induced current Jy on the surface of the marine oil well control equipment to be detected;

[0047] Wherein, the X-direction induced current Jx and the Y-direction induced current Jy respectively satisfy:

[0048] ;

[0049] ; 、 is the amplitude, is the angular frequency, 、 is the initial phase;

[0050] The X-direction included angle of the superimposed current Js obtained by vector addition of the X-direction induced current Jx and the Y-direction induced current Jy , satisfies:

[0051] and ;

[0052] The amplitude of the superimposed current Js obtained by vector addition of the X-direction induced current Jx and the Y-direction induced current Jy , satisfies: ;

[0053] The skin depth of the regulated induced current , satisfies: ; wherein, is the angular frequency, is the magnetic permeability, is the electrical conductivity.

[0054] More preferably, the array sensor adopts a rectangular coil in differential arrangement.

[0055] This invention provides a method and system for detecting electromagnetic control of structural defects in offshore oil well control equipment. The method includes the following steps: Step S101: Adjusting the excitation induced current to obtain the spatial magnetic field distortion of the offshore oil well control equipment under inspection; Step S102: Performing absolute value calculation on the magnetic field signal Bz matrix Z and plotting a grayscale image to obtain a Bz image; Step S103: Smoothing the Bz image using grayscale interpolation; fusing crack images from different angles using overlay to obtain complete crack contour information; Step S104: Adding real angle labels to all fused crack angle images to form an angle evaluation database; training the angle evaluation database using a residual network; and training the angle evaluation database to complete the angle evaluation... The database can be used to obtain the angle prediction value of unknown samples; Step S105: Adjust the amplitude and phase parameters to make the induced current perpendicular to the crack and obtain the crack endpoint information; Step S106: When the induced current is perpendicular to the crack, adjust the excitation signal frequency to change the skin depth of the induced current and obtain at least a first frequency image, a second frequency image, and a third frequency image; wherein, the first frequency is less than the second frequency, and the second frequency is less than the third frequency; place the first frequency image, the second frequency image, and the third frequency image in the R, G, and B channels of the color image respectively, and stack them to obtain a crack size evaluation fusion image; Step S107: Build a cross-fusion network and extract the Bz peak value in the Bz image. and peak spacing Using the random forest algorithm, a crack prediction system is established; Step S108: Initialize the cross-fusion network, embed the crack prediction system, and train it using training data; The trained crack prediction system can then be used to obtain the size prediction value of cracks at any angle in the structural defects of offshore oil well control equipment.

[0056] The electromagnetic control detection method and system for structural defects in offshore oil well control equipment with the above-mentioned characteristics have at least the following technical advantages compared with existing technologies:

[0057] (1) The electromagnetic control detection system for structural defects in marine oil well control equipment provided by this invention uses a circular structure composed of semi-circular coils for both the upper and lower excitation coil groups. Compared with existing rectangular coils, this effectively avoids the local concentration of induced current at corners, making the induced current distribution more uniform and concentrated at the center of the array sensor position, thereby significantly enhancing the sensitivity of the monitoring probe and the crack detection rate. The front-end signal processing circuit connected to the excitation coil group is used to adjust the resonant frequency of the monitoring probe according to the external signal. When the excitation coil group is in series resonance, the current driven by the voltage source is maximized, thereby enhancing the alternating magnetic field to increase the detection eddy current density. When the detection coils are in parallel resonance, the impedance is maximized, resulting in the highest amplitude of the induced voltage, thereby improving the sensitivity of capturing the eddy current disturbance signal caused by defects. In addition, the array sensor is composed of differentially arranged rectangular coils, which improves the area utilization rate while effectively reducing missed detections, and can effectively reduce coil mutual inductance and signal noise.

[0058] (2) The electromagnetic control and detection system for structural defects in marine oil well control equipment provided by this invention can control the direction of the induced current field by adjusting the sinusoidal signal parameters in the upper and lower excitation coil groups, changing the amplitude and phase at the same frequency, and obtaining a uniform current field with uniform amplitude. This solves the problem of the inability to separate the leakage magnetic signal and the AC electromagnetic field signal in time in the prior art, and eliminates the adverse effects of uneven current amplitude on the detection results. In addition, when the frequency parameters are changed, the skin depth of the induced current changes accordingly, thereby obtaining a variety of depth feature information.

[0059] (3) The electromagnetic control detection method for structural defects in offshore oil well control equipment provided by this invention obtains more information about cracks using a high-resolution network; by constructing a cross-fusion network, the single forward channel of the neural network is divided into two channels, which are used to extract the length and depth features of the crack, respectively. The feature fusion is achieved through the cross-fusion of the two channels. Then, the physical features in the crack Bz image are extracted, and a crack prediction system is established using the random forest algorithm; and the parameters are used to detect the crack. By adjusting the weights involved in the training of the crack prediction system, the training direction of the residual network was corrected. Finally, while enhancing the interpretability of the method, the prediction accuracy of crack length and depth was increased, avoiding misjudgments caused by single-dimensional feature analysis compared to the single detection mode of existing technologies.

[0060] (4), the marine oil well control equipment structural defect electromagnetic regulation and control detection method provided by the application, through the life prediction network embedded with physical information, three major fatigue core laws of Paris equation, damage rule and SN curve are converted into calculable constraint terms; through the synergistic effect of the three, multidimensional physical constraints of fatigue life prediction are realized, the generalization ability of the model is improved, and prediction distortion caused by data deviation is avoided. Then, the space-time prediction neural network captures the space-time dependence, especially the time sequence correlation that the stress concentration intensifies due to the increase of crack depth and then accelerates the expansion; in addition, local details such as crack edges can also be captured, so as to integrate global distribution information, and realize the technical purpose of more accurate prediction of crack propagation length. BRIEF DESCRIPTION OF DRAWINGS

[0061] The accompanying drawings are included to provide a further understanding of the application, and constitute a part of the specification, illustrate the application together with the embodiments of the application, and do not constitute a limitation on the application. In the following drawings:

[0062] Figure 1 The structural diagram of the marine oil well control equipment structural defect electromagnetic regulation and control detection system provided by the application is shown in the figure;

[0063] Figure 2 The structural diagram of the excitation coil group and array sensor in the monitoring probe is shown in the figure;

[0064] Figure 3 The structural diagram of the array sensor in the monitoring probe is shown in the figure;

[0065] Figure 4a The induced current density diagram of the rectangular coil is shown in the figure;

[0066] Figure 4b The induced current density diagram of the circular coil is shown in the figure;

[0067] Figure 5a The induced current direction diagram with an angle of 0° is shown in the figure;

[0068] Figure 5b The induced current direction diagram with an angle of 45° is shown in the figure;

[0069] Figure 5c The induced current direction diagram with an angle of 60° is shown in the figure;

[0070] Figure 5d The induced current direction diagram with an angle of 90° is shown in the figure;

[0071] Figure 5e The induced current direction diagram with an angle of 150° is shown in the figure;

[0072] Figure 6aDiagram of skin depth for induced current at 100 Hz frequency;

[0073] Figure 6b Diagram of skin depth for induced current at 1 kHz frequency;

[0074] Figure 6c Diagram of skin depth for induced current at 10 kHz frequency;

[0075] Figure 7a Diagram of regulated induced current amplitude;

[0076] Figure 7b Diagram of regulated induced current perpendicular to horizontal crack;

[0077] Figure 7c Diagram of regulated induced current perpendicular to vertical crack;

[0078] Figure 8 Diagram of equivalent resonant circuit of monitoring probe;

[0079] Figure 9a Diagram of excitation coil group series resonance (without resonance);

[0080] Figure 9b Diagram of detection coil parallel resonance (with resonance);

[0081] Figure 10 Diagram of flow of the method for detecting structural defects of offshore oil well control equipment provided by the application;

[0082] Figure 11 Diagram of structure of SRCNN network (super-resolution network);

[0083] Figure 12 Diagram of crack contour complete information obtained by fusing and superimposing crack images at different angles;

[0084] Figure 13 Diagram of crack size evaluation fusion image obtained by stacking 100 Hz image, 1 kHz image and 10 kHz image;

[0085] Figure 14 Diagram of structure of cross fusion network constructed;

[0086] Figure 15 Diagram of result of detecting crack propagation and complete fracture of test piece;

[0087] Figure 16 Diagram of structure of multi-layer fully connected neural network topology constructed;

[0088] Figure 17aA schematic diagram of crack length extension experimental results is shown in FIG. 1.

[0089] Figure 17b A schematic diagram of crack length extension prediction results is shown in FIG. 2.

[0090] Figure 18 A schematic diagram of the structure of the constructed PredRNN++ spatiotemporal prediction neural network is shown in FIG. 3. DETAILED DESCRIPTION

[0091] The application provides a marine oil well control equipment structural defect electromagnetic regulation and detection method and a detection system. Crack multi-feature signals are obtained through electromagnetic excitation signal amplitude-frequency-phase regulation, the angle information of the crack is obtained with the aid of a residual network, the amplitude-frequency-phase regulation and a physical heuristic neural network are used to realize accurate identification of crack length and depth information, a life prediction network is used to realize effective prediction of residual fatigue life, and a spatiotemporal prediction neural network is used to realize effective evaluation of crack propagation, thereby providing accurate data support for health evaluation of the underwater structure of the marine oil well control equipment.

[0092] In order to enable those skilled in the art to better understand the application, a marine oil well control equipment structural defect electromagnetic regulation and detection system provided by the application is first explained as follows. As shown in FIG. 4, the marine oil well control equipment structural defect electromagnetic regulation and detection system provided by the application comprises a monitoring probe, an excitation signal generation unit, and a monitoring signal processing and collecting unit. Figure 1

[0093] The monitoring probe is further composed of an excitation coil group, an array sensor, and a front-end signal processing circuit. Specifically, as shown in FIG. 5, the excitation coil group comprises an upper excitation coil group and a lower excitation coil group in orthogonal distribution. The upper excitation coil group and the lower excitation coil group are each composed of two semicircular coils. The two semicircular coils of the same layer in the upper excitation coil group and the lower excitation coil group are connected to opposite sine alternating currents, so as to generate a uniform induced current field. Figure 2 As shown in FIG. 6, the array sensor is arranged above the uniform induced current field generated by the excitation coil group (which is preferably composed of differentially arranged rectangular coils), and is used to collect an induced magnetic field signal perpendicular to the surface direction of the marine oil well control equipment to be detected. It should be noted that the array sensor adopts differential rectangular coil arrangement because the geometric shape of the rectangular coil is more in line with the regularity requirement of the array layout. Adjacent coils can be closely arranged, and the edge gap is smaller. For example, the filling factor (coil effective area / total area) of the rectangular coil is about 20%-30% higher than that of the circular coil, the area utilization rate is increased, and crack detection is not easy to occur. For reference,

[0094] Figure 3 Figure 4a ,​​​Figure 4b schematic diagram of induced current density of a rectangular coil, Figure 4a schematic diagram of induced current density of a rectangular coil, Figure 4b schematic diagram of induced current density of a rectangular coil). In addition, the differential arrangement reduces the mutual inductance between the coils, reduces the noise in the signal; the right-angle structure of the rectangular coil can make the magnetic field distribution more concentrated in the coil, and the edge magnetic field decays faster, further reducing the magnetic field coupling strength between adjacent coils. And the front-end signal processing circuit is connected with the excitation coil group, which is used to adjust the resonance frequency of the monitoring probe according to the external signal.

[0095] The excitation signal generating unit includes a signal generator and a power amplifier. The signal generator is used to provide a sinusoidal alternating excitation signal for the excitation coil group; the power amplifier is used to amplify the sinusoidal alternating excitation signal. And the monitoring signal processing and collecting unit includes a multiplexer and an amplification and filtering circuit, a band-pass filter circuit, a phase-locked amplification circuit and a signal acquisition board.

[0096] In addition, as a more preferred embodiment of the present application, the upper excitation coil group and the lower excitation coil group in the excitation coil group use the same frequency parameters, so as to induce X-direction induced current Jx and Y-direction induced current Jy on the surface of the marine oil well control equipment to be detected.

[0097] Wherein, the X-direction induced current Jx and the Y-direction induced current Jy respectively satisfy:

[0098] Formula (1);

[0099] Formula (2);

[0100] The X-direction included angle of the superimposed current Js obtained by vector addition of the X-direction induced current Jx and the Y-direction induced current Jy satisfies:

[0101] and Formula (3);

[0102] The amplitude of the superimposed current Js obtained by vector addition of the X-direction induced current Jx and the Y-direction induced current Jy satisfies: Formula (4);

[0103] The skin depth of the regulated induced current satisfies: Formula (5).

[0104] Specifically, as shown in Figure 5a- Figure 5e , respectively select , ; 、 ; 、 ; 、 ; 、 , so as to correspond to the angle of 0°, 45°, 60°, 90°, 150° induced current direction. By controlling the direction of the induced current, the information of the crack at any angle can be obtained. Compared with the existing rotating alternating electromagnetic field (the current is more concentrated in a single direction), the problem that the leakage magnetic signal and the alternating electromagnetic field signal cannot be separated in time is effectively solved.

[0105] Further, let 、 , the frequency is selected as 100Hz, 1kHz, 10kHz respectively, the current skin depth diagram under different frequencies is obtained, as shown in Figure 6a 、 Figure 6b 、 Figure 6c . It can be found that the induced current skin depth is different by multi-frequency excitation, so as to obtain a variety of depth feature information.

[0106] It should be pointed out that the amplitude-frequency phase control can make the current direction concentrated and more uniform, as shown in Figure 7a . Among them, Figure 7a is the superimposed current with the same size of the amplitude information of the induced current in each direction under the control of 1kHz frequency. It can be found from the figure that the above control method can highlight the influence of the current direction on the magnetic field distortion at the crack end point, and eliminate the adverse effects of the uneven current amplitude on the detection results. Further, as shown in Figure 7b 、 Figure 7c , it can be found that the regulated induced current is perpendicular to the horizontal crack (reference Figure 7b ) and the regulated induced current is perpendicular to the vertical crack (reference Figure 7c ), and the obtained magnetic field image amplitude is basically consistent.

[0107] As an optional embodiment of the present application, the impedance analyzer is used to measure that the inductance of the excitation coil group and the array sensor is 170.9μH and 27.7μH respectively. In this process, the front-end signal processing circuit is externally connected with an adjustable capacitor (which can automatically adjust the capacitance value), so that the monitoring probe resonant frequency is the excitation frequency, at this time the equivalent resonant circuit of the monitoring probe can be referred to as Figure 8 . When the excitation coil group is in series resonance, the current is maximum under the driving of the voltage source, as shown in Figure 9aAs shown, the alternating magnetic field can be significantly enhanced to improve the eddy current density of the test piece. When the detection coil is in parallel resonance, the impedance of the detection coil is maximum, and the amplitude of the induced voltage is highest. The small eddy current disturbance caused by defects can significantly change the voltage signal of the detection coil, such as Figure 9b As shown; at this time, the sensitivity is enhanced by about 30% under the self-adjusting resonance of 1kHz excitation.

[0108] The specific implementation process of the marine oil well control equipment structural defect electromagnetic regulation and detection system can be described as follows: adjust the required signal generator parameters to output a double-channel sinusoidal alternating signal; amplify the excitation signal amplitude through a power amplifier to reach the position of the excitation coil group of the monitoring probe, and induce an induced current on the surface of the structure 、 At this time, the marine oil well control equipment structural defects (especially cracks) cause the induced current disturbance to induce spatial magnetic field distortion, which is received by the array sensor. For example, the 64-way differential signal emitted by the 8×8 array sensor is simplified to 4-way differential signal through multiplexing and amplification filtering circuit. After processing by the band-pass filtering circuit, the high-frequency noise and coupled interference signals are removed. The AC signal is amplified by 2 times by the phase-locked amplification circuit, and is converted into DC amplitude and DC phase signals, and finally the signal acquisition board card collects the signals and sends them to the PC end.

[0109] On the other hand, the present application also provides a marine oil well control equipment structural defect electromagnetic regulation and detection method, as shown in Figure 10 The method comprises the following steps:

[0110] Step S101: Regulate the excitation induced current to obtain the spatial magnetic field distortion of the marine oil well control equipment to be detected. The 8×8 size magnetic field signal Bz matrix Z is calculated.

[0111] The magnetic field signal Bz matrix Z satisfies: .

[0112] In order to facilitate those skilled in the art to understand the detection method, the following size parameters of the test piece (the surface is provided with a crack with a length of 20mm, a depth of 5mm and an angle of 60°) are provided, and the detection process of the test piece is described as an example.

[0113] Specifically, by changing the amplitude and phase parameters of the orthogonal excitation coil group, the purpose of regulating the excitation induced current is achieved, so as to obtain the spatial magnetic field distortion caused by the induced current disturbance; then, the background magnetic field obtained by calibration is subtracted, and the 8×8 size magnetic field signal Bz matrix Z is obtained.

[0114] Step S102: Perform absolute value operation on the magnetic field signal Bz matrix Z and draw a gray scale image to obtain the Bz image.

[0115] Step S103: using gray scale interpolation, the Bz image is smoothed;

[0116] The crack profile complete information is obtained by fusing the crack images of different angles by superposition.

[0117] On the basis of completing step S101, the absolute value operation is further performed on the magnetic field signal Bz matrix Z, and a gray scale image is drawn to obtain a Bz image; and the Bz image is smoothed. Specifically, as shown in Figure 11 , first, the low-resolution Bz image can be processed by using the trained SRCNN network; for example, by using a three-layer convolution structure, feature extraction, nonlinear mapping and high-resolution image reconstruction are sequentially completed, so as to output a high-resolution image. The purpose of this is to obtain more crack information by using limited sensor data. Then, the crack images of different angles are fused by superposition, and the crack profile complete information is obtained. The crack profile complete information obtained after superposition can be referred to as shown in Figure 12 .

[0118] Step S104: a real angle label is added to all crack angle fusion images to form an angle evaluation database.

[0119] The residual network is used to train the angle evaluation database; and the trained angle evaluation database can be used to obtain the angle prediction value of an unknown sample.

[0120] On the basis of completing steps S102 and S103, the angle evaluation database is further constructed. Specifically, the residual network (preferably, a Resnet-18 residual network) is used to train the angle evaluation database to obtain the angle prediction value of the crack, which is 59.78°.

[0121] Step S105: the amplitude and phase parameters are regulated so that the induced current is perpendicular to the crack to obtain crack end point information.

[0122] Specifically, the regulation of the amplitude and phase parameters in step S105 can be referred to as , . The purpose is to make the induced current perpendicular to the crack to obtain the crack end point information.

[0123] Step S106: when the induced current is perpendicular to the crack, the frequency of the excitation signal is regulated to change the skin depth of the induced current, and at least a first frequency image, a second frequency image and a third frequency image are obtained; wherein the first frequency is less than the second frequency, and the second frequency is less than the third frequency.

[0124] The first frequency image, the second frequency image and the third frequency image are respectively placed in the R, G and B channels of a color image to stack a crack size evaluation fusion image.

[0125] On the basis of completing step S105, further through stacking, the crack size evaluation fusion image is obtained. As a more preferred embodiment of the present application, the frequency of the excitation signal is regulated herein to obtain 100Hz image, 1kHz image and 10kHz image. Among them, the images of different frequencies correspond to different induced current skin depths, which can reflect the characteristics of cracks at different depth levels.

[0126] Further, the three images are respectively placed in the R, G and B channels of the color image, and a three-dimensional sample of RGB type is constructed by using stacking defense to generate a crack size evaluation fusion image. For details, please refer to the following figure: Figure 13 The fusion method can integrate the information of different induced current skin depths in the same image; compared with the image of a single frequency, it can more comprehensively and intuitively present the crack depth information, and effectively reduce the depth quantization error caused by single frequency detection.

[0127] Step S107: Building a cross-fusion network to extract the Bz peak value in the Bz image and the peak distance . A crack prediction system is established by using a random forest algorithm.

[0128] The prediction value of the cross-fusion network is set as OUTn, the prediction value of the crack prediction system is set as OUTe, and the true value of the crack size is set as Label. The loss function constructed satisfies: ;

[0129] Among them, is the weight of the crack prediction system participating in the training.

[0130] On the basis of completing step S106, a crack prediction system is further constructed. It is worth noting that, as shown in the following figure, a cross-fusion network is first constructed. Among them, the cross-fusion network extracts crack length and depth features respectively through a double-channel structure, reduces the mutual quantization interference and strengthens the feature correlation through cross-fusion between channels. Then, through the weight parameter, the weight of the crack prediction system participating in the training is adjusted to realize the correction of the training direction of the crack prediction system to the residual network. Figure 14

[0131] Step S108: Initialize the cross-fusion network, embed the crack prediction system, and use the training data for training; wherein the crack prediction system trained can be used to obtain the size prediction value of any angle crack in the structure defect of the offshore oil well control equipment.

[0132] ​Specifically, the crack size evaluation fusion image is loaded into the cross fusion network trained, and finally the length of the crack is 20.32 mm and the depth of the crack is 5.09 mm. It can be found that, compared with the traditional single detection method, the marine oil well control equipment structure defect electromagnetic regulation and detection method provided by the application can simultaneously consider the multi-dimensional information of crack length and depth, and avoid size misjudgment caused by single feature analysis.

[0133] In addition, the marine oil well control equipment structure defect electromagnetic regulation and detection method provided by the application further comprises the following steps:

[0134] Step S201: obtaining regulation data, using the trained cross fusion network to extract crack length information features of the marine oil well control equipment .

[0135] and stress data of the marine oil well control equipment .

[0136] In order to facilitate those skilled in the art to understand the application, the test piece is further explained as follows. Specifically, a workpiece containing a 10 mm pre-crack is selected for cyclic loading, and the stress sensor obtains the workpiece stress 4e 7 Pa, and the above detection method is used to obtain a pre-crack prediction length of about 10.03 mm.

[0137] Step S202: obtaining the remaining fatigue life by cyclic loading until the marine oil well control equipment workpiece is completely broken .

[0138] On the basis of completing step S201, the remaining fatigue life is further obtained by cyclic loading . Wherein, the cyclic loading is until the crack of the test piece expands and completely breaks, and the remaining fatigue life thereof is about 5.2e 4 , which can be referred to as shown in Figure 15 .

[0139] Step S203: constructing a multi-layer fully connected neural network topology.

[0140] The crack length information features of the marine oil well control equipment , the stress data of the marine oil well control equipment , and the remaining fatigue life are used as three nodes of the input layer, a pyramid structure is used as the hidden layer, and a fatigue life prediction value is used as a single node of the output layer, so as to realize multi-dimensional physical constraints for fatigue life prediction.

[0141] Building upon step S202, a multi-layer fully connected neural network topology is further constructed. For details, please refer to... Figure 16 As shown, the input layer of this multi-layer fully connected neural network topology uses crack length information features from offshore oil well control equipment. Stress data of offshore oil well control equipment Remaining fatigue life Three types of features are used as input nodes; and a pyramid structure is used as the hidden layer, with the number of neurons in the hidden layer decreasing by 2 / 3 of the previous layer (preferably, the ReLU activation function is used to introduce nonlinear mapping capability, thereby enhancing the model's expressive energy for complex fatigue life relationships); the output layer outputs the fatigue life prediction value through a single node.

[0142] Step S204: Train the lifetime prediction network.

[0143] The regression evaluation index MSE is used to calculate the deviation between predicted and measured values. And construct features including crack length information of offshore oil well control equipment. Stress data of offshore oil well control equipment Remaining fatigue life Physical consistency loss of three types of characteristic residuals .

[0144] Among them, the deviation between the predicted value and the measured value ,satisfy: ;

[0145] Designing the weighted loss function, we obtain: ;

[0146] in, This was determined through cross-validation.

[0147] Building upon step S203, the life prediction network is further trained. Specifically, the workpiece stress 4e... 7 Pa and a crack length of 10.03 mm were applied to the trained PINN network; after inverse normalization, the predicted remaining fatigue life of the workpiece was approximately 5.18 e. 4 It is worth noting that the prediction results embedded with physical information not only fit the experimental data but also satisfy the theoretical relationship between stress intensity factor and crack propagation rate, reducing errors while being suitable for small datasets.

[0148] Step S205: In the fatigue life prediction process, a fixed time step is set, the load is set to a fixed value, the actual crack length is obtained, and the crack propagation image is monitored simultaneously.

[0149] The multi-frequency fusion high-resolution magnetic field image of the induced current vertical crack is obtained by the SRCNN, and a database is built.

[0150] On the basis of completing step S204, the real crack length is further obtained and the crack propagation image is monitored synchronously. Specifically, during the crack propagation process, after every 4000 times of loading, a static load of 3kN is set, the real crack length is obtained by using a vernier caliper, and the crack image after propagation is obtained by using the method in the above example, as shown in Figure 17a 、 Figure 17b ; wherein, as shown in Figure 17a , it is a schematic diagram of crack length expansion experimental results, Figure 17b is a schematic diagram of crack length expansion prediction results. Then, the initial crack image obtained above is arranged in time sequence as the first 5 frames (historical crack image) as the known image, and arranged in time sequence as the last 5 frames (historical crack image) as the prediction image for comparison, forming an input sequence. The purpose of this is to improve the prediction accuracy of the crack propagation by multi-frequency fusion.

[0151] Step S206: constructing a PredRNN++ spatiotemporal prediction neural network.

[0152] Wherein, the input of the PredRNN++ spatiotemporal prediction neural network is the known crack magnetic field image, and the output of the PredRNN++ spatiotemporal prediction neural network is the predicted crack magnetic field image.

[0153] On the basis of completing step S205, the PredRNN++ spatiotemporal prediction neural network is further constructed. As shown in Figure 18 , first, the spatial features and temporal dependencies are synchronously extracted by 5 layers of Causal LSTM convolutional layers. Each layer of Causal LSTM contains 3x3 convolution and gating mechanism, and residual connection is used between layers to alleviate the degradation problem of deep network; then, the gradient highway unit (GHU) is embedded, which directly transmits long-distance gradient through the gating signal to capture the long-term dynamic law of crack propagation. Finally, the output layer generates a single-channel predicted crack magnetic field image through 1x1 convolution and upsampling.

[0154] Step S207: training the PredRNN++ spatiotemporal prediction neural network.

[0155] The root mean square error RMSE and the structural similarity index SSIM are used as the indicators for evaluating the accuracy of the prediction results of the PredRNN++ spatiotemporal prediction neural network. Wherein, the root mean square error RMSE satisfies: ; the structural similarity index SSIM satisfies: .

[0156] The trained PredRNN++ spatio-temporal prediction neural network can be used to obtain the crack propagation prediction image.

[0157] On the basis of completing step S206, the PredRNN++ spatio-temporal prediction neural network is further trained. The output layer of the PredRNN++ spatio-temporal prediction neural network outputs the predicted crack magnetic field image of the workpiece, which can be referred to as shown in FIG. 6. Figure 17b The order of magnitude of the RMSE of the predicted image is 10e -2 , which indicates that the prediction error of the model is at a low level. At the same time, the SSIM is greater than 0.94, which means that there is a high similarity between the predicted value and the true value. Therefore, the synchronous extraction of spatial features and time-dependent relationships can significantly enhance the accuracy of the prediction results.

[0158] Step S208: Adding a true length label to all the predicted crack propagation prediction images to form a length evaluation database; training the length evaluation database by using a cross-fusion network; wherein the trained length evaluation database can be used to obtain the crack propagation length prediction value of the crack propagation prediction image of the structural defect of the offshore oil well control equipment.

[0159] It is worth noting that loading the aforementioned crack propagation prediction image into the trained length evaluation database and cross-fusion network can obtain the crack propagation length prediction value: 18.66 mm, 20.98 mm, 23.34 mm, 27.48 mm, 32.30 mm; and the corresponding true crack length is 19.3 mm, 21.0 mm, 23.3 mm, 26.0 mm, and 30.5 mm. After comparison, it can be concluded that the average error is about 0.532 mm.

[0160] The application provides a marine oil well control equipment structural defect electromagnetic regulation and detection method and a detection system. The marine oil well control equipment structural defect electromagnetic regulation and detection method comprises the following steps: step S101: regulating and exciting induced current to obtain spatial magnetic field distortion of the marine oil well control equipment to be detected; step S102: performing absolute value operation on a magnetic field signal Bz matrix Z and drawing a gray scale image to obtain a Bz image; step S103: performing smoothing processing on the Bz image by using gray scale interpolation; performing fusion on crack images at different angles by using superposition to obtain complete crack contour information; step S104: adding a real angle label to all crack angle fusion images to form an angle evaluation database; training the angle evaluation database by using a residual network; the trained angle evaluation database can be used to obtain angle prediction values of unknown samples; step S105: regulating and controlling amplitude and phase parameters to make the induced current perpendicular to the crack to obtain crack end point information; step S106: when the induced current is perpendicular to the crack, regulating and controlling an excitation signal frequency to change the skin depth of the induced current to obtain at least a first frequency image, a second frequency image and a third frequency image; the first frequency is smaller than the second frequency, and the second frequency is smaller than the third frequency; the first frequency image, the second frequency image and the third frequency image are respectively placed in R, G and B channels of a color image to stack to obtain a crack size evaluation fusion image; step S107: building a cross fusion network to extract Bz peak values and peak spacings in the Bz image; using a random forest algorithm to establish a crack prediction system; step S108: initializing the cross fusion network, embedding the crack prediction system and training by using training data; the trained crack prediction system can be used to obtain size prediction values of cracks at any angle in the marine oil well control equipment structural defect.

[0161] Compared with the prior art, the marine oil well control equipment structural defect electromagnetic regulation and detection method and the detection system have at least the following technical advantages:

[0162] (1) The marine oil well control equipment structural defect electromagnetic regulation and detection system provided by the application, the upper and lower excitation coil groups of which are both circular structures composed of semicircular coils; compared with existing rectangular coils, the local concentration of induced current at the corner is effectively avoided, the induced current is more evenly distributed and concentrated at the central array sensor position, thereby significantly enhancing the sensitivity and crack detection rate of the monitoring probe. The front-end signal processing circuit connected with the excitation coil group is used to adjust the resonant frequency of the monitoring probe according to the external signal; when the excitation coil group is in series resonance, the current is maximized under the driving of the voltage source, thereby enhancing the alternating magnetic field to improve the detection eddy current density; when the detection coil is in parallel resonance, the impedance is maximized, so that the induced voltage amplitude is the highest, thereby improving the sensitivity of capturing the eddy current disturbance signal caused by defects. In addition, the array sensor is composed of differentially arranged rectangular coils, which effectively reduces the missed detection situation while improving the area utilization rate, and can effectively reduce the coil mutual inductance and signal noise;

[0163] (2) The marine oil well control equipment structural defect electromagnetic regulation and detection system provided by the application can realize the regulation of the direction of the induced current field by adjusting the sine signal parameters in the upper and lower excitation coil groups under the same frequency, that is, by changing the amplitude and phase, and obtain a uniform current field with uniform amplitude, thereby solving the problem that the leakage magnetic signal and the alternating electromagnetic field signal cannot be separated in time in the prior art, and eliminating the adverse effects of uneven current amplitude on the detection results. In addition, when the frequency parameter is changed, the skin depth of the induced current changes accordingly, thereby obtaining a variety of depth characteristic information;

[0164] (3) The marine oil well control equipment structural defect electromagnetic regulation and detection method provided by the application obtains more information of the crack by means of a high-resolution network; by building a cross-fusion network, the single forward channel of the neural network is divided into two channels, and the two channels are respectively used to extract the length and depth features of the crack. Through the cross of the two channels, the fusion of the features is realized. Then, the physical features in the crack Bz image are extracted, and a crack prediction system is established by means of a random forest algorithm. The weight of the crack prediction system participating in the training is adjusted, and the correction of the training direction of the crack prediction system to the residual network is realized. Finally, while enhancing the interpretability of the method, the prediction accuracy of the crack length and depth is increased, and compared with the single detection mode of the prior art, the misjudgment caused by single-dimensional feature analysis is avoided;

[0165] ​(4) The marine oil well control equipment structural defect electromagnetic regulation and control detection method provided by the application converts Paris equation, damage rule and SN curve three major fatigue core laws into calculable constraint terms through a life prediction network embedded with physical information; through the synergistic effect of the three, multidimensional physical constraints of fatigue life prediction are realized, the generalization ability of the model is improved, and prediction distortion caused by data deviation is avoided. Then, the spatiotemporal prediction neural network captures the spatiotemporal dependence relationship, especially the time sequence correlation that stress concentration intensifies due to crack depth increase and then accelerates expansion; in addition, it can also capture local details such as crack edges, thereby integrating global distribution information, and achieving the technical purpose of more accurate prediction of crack propagation length.

[0166] The above is only a specific embodiment of the application, but the protection scope of the application is not limited thereto, any person skilled in the art can easily think of changes or replacements within the technical range disclosed by the application, which should be covered within the protection scope of the application. Therefore, the protection scope of the application should be subject to the protection scope of the claims.

Claims

1. The method for detecting the structural defects of offshore oil well control equipment by electromagnetic regulation and control, characterized in that, Comprising the following steps: Step S101: regulating and exciting induced current, obtaining spatial magnetic field distortion of the marine oil well control equipment to be detected; The calculated 8x8 size magnetic field signal Bz matrix Z; wherein the magnetic field signal Bz matrix Z satisfies: ; Step S102: absolute value operation is performed on the magnetic field signal Bz matrix Z, and a gray scale image is drawn, obtaining a Bz image; Step S103: using gray scale interpolation, the Bz image is smoothed; Different angle crack images are fused by superposition to obtain complete crack profile information; Step S104: a true angle label is added to all crack angle fusion images to form an angle evaluation database; The residual network is used to train the angle evaluation database; the trained angle evaluation database can be used to obtain the angle prediction value of the unknown sample; Step S105: regulating the amplitude and phase parameters to make the induced current perpendicular to the crack, and obtaining the crack end point information; Step S106: when the induced current is perpendicular to the crack, the frequency of the excitation signal is regulated to change the skin depth of the induced current, and at least a first frequency image, a second frequency image and a third frequency image are obtained; wherein the first frequency is less than the second frequency, and the second frequency is less than the third frequency; The first frequency image, the second frequency image and the third frequency image are respectively placed in the R, G and B channels of the color image to stack and obtain a crack size evaluation fusion image; Step S107: build a cross-fusion network to extract the Bz peak value in the Bz image and peak distance ; and use the random forest algorithm to establish a crack prediction system; The prediction value of the cross fusion network is set as OUTn, the prediction value of the crack prediction system is set as OUTe, and the true value of the crack size is set as Label, and the loss function obtained by constructing satisfies: ； wherein, is the weight the crack prediction system participates in at training time; Step S108: initializing the cross fusion network, embedding the crack prediction system, and training using training data; wherein the trained crack prediction system can be used to obtain the size prediction value of any angle crack in the structural defect of the marine oil well control equipment.

2. The method according to claim 1, wherein, Further comprising the following steps: Step S201: Obtain the control data, use the trained cross fusion network to extract the crack length information features of the offshore oil well control equipment ; And, acquiring stress data of offshore oil well control equipment ; Step S202: Obtain the remaining fatigue life by cyclic loading until the marine oil well control equipment workpiece is completely broken ; Step S203: constructing a multi-layer fully connected neural network topology structure; Crack length information characteristics of offshore oil well control equipment Stress data of offshore oil well control equipment Remaining fatigue life As three nodes in the input layer, a pyramid-shaped structure is used as the hidden layer, and the fatigue life prediction value is used as a single node in the output layer to realize multi-dimensional physical constraints on fatigue life prediction. Step S204: training the life prediction network; Adopting regression evaluation index MSE, the deviation between the predicted value and the measured value is calculated And build crack length information features containing offshore oil well control equipment Stress data of offshore oil well control equipment Residual fatigue life Physical consistency loss of three types of feature residuals ; Wherein, the prediction value and the measured value deviation , satisfy: ; is the actual observation value, is the model prediction value, n is the sample number; The weighted loss function is designed, and the following is obtained: ; wherein, Determined by cross-validation , satisfying: ; wherein, , , is a weight, y is a model prediction, B is a Basquin model output, P is a Paris model output, represents a non-physical value penalty term to ensure y is non-negative, thus avoiding unreasonable results from the model output; Step S205: in the fatigue life prediction process, fixing the time step and setting the load as a fixed value, obtaining the true crack length and synchronously monitoring the crack propagation image; Through SRCNN, a multi-frequency fusion high-resolution magnetic field image of the induced current perpendicular to the crack is obtained, and a database is built; Step S206: constructing a PredRNN++ spatiotemporal prediction neural network; The input of the PredRNN++ spatiotemporal prediction neural network is the known crack magnetic field image, and the output of the PredRNN++ spatiotemporal prediction neural network is the predicted crack magnetic field image; Step S207: training the PredRNN++ spatiotemporal prediction neural network; The root mean square error RMSE and the structural similarity index SSIM are used as indexes for evaluating the accuracy of the prediction results of the PredRNN++ spatiotemporal prediction neural network; wherein the root mean square error RMSE satisfies: ; is an actual observation value, is a model prediction value, and n is a sample quantity. The structural similarity index SSIM satisfies: ; wherein x is a predicted value, y is an actual real value, μ x ,μ y is a mean value, , xy is a covariance, and C1 and C2 are constants; The trained PredRNN++ spatiotemporal prediction neural network can be used to obtain the crack propagation prediction image; Step S208: a true length label is added to all predicted crack propagation prediction images to form a length evaluation database; the cross fusion network is used to train the length evaluation database; wherein the trained length evaluation database can be used to obtain the crack propagation length prediction value of the crack propagation prediction image of the structural defect of the marine oil well control equipment.

Images