Method for identifying damage position of underwater production system pipeline, electronic device, storage medium and program product

Abstract

Embodiments of the present application provide a method for identifying a damage position of a pipeline of an underwater production system, an electronic device, a storage medium and a program product. A preset noise reduction algorithm is used to perform noise reduction processing on the collected first reflected ultrasonic guided wave signal, which can remove noise interference in the signal, improve the quality and reliability of the signal, thereby providing more accurate data basis for subsequent damage feature extraction and analysis, and helping to improve the accuracy of damage identification. The one-dimensional convolutional neural network model can extract at least one of the time domain damage feature and the frequency domain damage feature from the second reflected ultrasonic guided wave signal, comprehensively utilize multiple features to judge the damage position of the pipeline, and can more comprehensively reflect the damage position of the pipeline. By dividing the pipeline into multiple pipeline position segments and using the one-dimensional convolutional neural network model to calculate the damage probability of each pipeline position segment, the damage position in the pipeline of the underwater production system can be more accurately located.

Description

Technical Field

[0001] This application relates to the field of marine engineering, and more particularly to a method for identifying the location of damage in underwater production system pipelines, electronic equipment, storage media, and program products. Background Technology

[0002] In offshore oil and gas extraction, subsea production system pipelines serve functions such as transporting oil and gas, connecting and supporting equipment, and play a crucial role in ensuring the normal operation of deepwater production systems and oil and gas production. However, in the complex marine environment, pipelines are susceptible to corrosion, fatigue, and third-party activities, leading to damage and increasing the risk of leaks.

[0003] Currently, ultrasonic guided wave detection technology is widely used in intelligent identification and monitoring of pipeline damage in underwater production systems. Its principle is to utilize the changes in propagation characteristics caused by reflection, refraction, and mode conversion when ultrasonic guided waves encounter damage in the pipeline. These changes are analyzed by signal processing algorithms to identify damage. Signal processing algorithms, such as Fourier transform, analyze the preprocessed signal to extract feature parameters related to pipeline damage, thereby realizing the identification and location of pipeline damage.

[0004] However, the propagation characteristics of ultrasonic guided waves in pipelines are affected by a variety of factors. For pipelines with complex structures, ultrasonic guided waves will undergo complex reflections, refractions and mode conversions at these locations, which can easily generate interference noise. Traditional signal processing methods are difficult to handle this interference noise, resulting in low accuracy in identifying and locating pipeline damage. Summary of the Invention

[0005] This application provides a method, electronic device, storage medium, and program product for identifying the damage location of pipelines in an underwater production system, in order to accurately identify the damage location of pipelines in an underwater production system.

[0006] In a first aspect, embodiments of this application provide a method for identifying the location of damage in an underwater production system pipeline, including:

[0007] Receive the first reflected ultrasonic guided wave signal from any underwater production system pipeline acquired by the ultrasonic guided wave detection system;

[0008] A preset noise reduction algorithm is used to denoise the first reflected ultrasonic guided wave signal to obtain the second reflected ultrasonic guided wave signal.

[0009] The underwater production system pipeline is divided into multiple pipeline location segments;

[0010] The second reflected ultrasonic guided wave signal is input into a trained one-dimensional convolutional neural network model for the following processing: damage features related to damage are extracted from the second reflected ultrasonic guided wave signal, wherein the damage features include at least one of time-domain damage features and frequency-domain damage features; based on at least one of the time-domain damage features and frequency-domain damage features, the damage probability of each pipeline location segment is calculated and the damage probability of each pipeline location segment is output.

[0011] The pipeline segment whose damage probability meets the preset conditions is identified as the damage location of the underwater production system pipeline.

[0012] In one possible implementation, the ultrasonic guided wave detection system includes an excitation signal source device and a piezoelectric transducer;

[0013] Accordingly, before receiving the first reflected ultrasonic guided wave signal from any underwater production system pipeline acquired by the ultrasonic guided wave detection system, the following steps are included:

[0014] Establish a cylindrical coordinate system for the cross-section of the underwater production system pipeline;

[0015] Based on the cylindrical coordinate system of the cross-section of the underwater production system pipeline, the wave equation and dispersion equation of the underwater production system pipeline are established.

[0016] The excitation signal is determined based on the wave equation and the dispersion equation;

[0017] An excitation signal is sent to the piezoelectric transducer through an excitation signal source device;

[0018] In response to the excitation signal, the piezoelectric transducer is triggered to transmit ultrasonic guided wave signals to the underwater production system pipeline in order to collect the reflected ultrasonic guided wave signals from the underwater production system pipeline.

[0019] In one possible implementation, a preset noise reduction algorithm is used to denoise the reflected ultrasonic guided wave signal to obtain the denoised reflected ultrasonic guided wave signal, including:

[0020] The reflected ultrasonic guided wave signal is input into the first preset algorithm to decompose the reflected ultrasonic guided wave signal, and the output obtained is n signal components, where n is an integer greater than 1;

[0021] Calculate the permutation entropy of n signal components;

[0022] Select signal components whose permutation entropy value is less than a preset threshold to obtain effective signal components;

[0023] Select signal components whose permutation entropy value is greater than or equal to a preset threshold to obtain noise signal components;

[0024] The noise signal components are decomposed and reconstructed using wavelet decomposition, and then subjected to wavelet threshold denoising to obtain the denoised signal components.

[0025] The effective signal component and the denoised signal component are combined to obtain the denoised reflected ultrasonic guided wave signal.

[0026] In one possible implementation, before inputting the reflected ultrasonic guided wave signal to the first preset algorithm, the following steps are included:

[0027] Based on the second preset algorithm, within a preset numerical range, a first preset number of first parameters and second parameters are generated;

[0028] Combine the first and second parameters in pairs and calculate the combined envelope entropy value;

[0029] Select combinations whose envelope entropy values ​​are less than a first preset threshold to establish a first preset algorithm.

[0030] In one possible implementation, it also includes:

[0031] Establish j one-dimensional convolutional neural network models with different structures, where j is an integer greater than 1;

[0032] Obtain the damage dataset;

[0033] The damage dataset is input into j one-dimensional convolutional neural network models with different structures, and the j one-dimensional convolutional neural network models with different structures are trained to obtain j trained one-dimensional convolutional neural network models with different structures.

[0034] Obtain the classification accuracy, precision, recall, and F1 score of j trained one-dimensional convolutional neural network models with different structures;

[0035] Based on classification accuracy, precision, recall, and F1 score, a one-dimensional convolutional neural network model that meets the preset requirements is selected and determined as a trained one-dimensional convolutional neural network model.

[0036] In one possible implementation, obtaining the damage dataset includes:

[0037] A simulation model of the underwater production system pipeline is established based on the preset pipeline parameters and preset damage types.

[0038] Damage datasets were obtained based on a simulation model of the underwater production system pipeline.

[0039] Secondly, embodiments of this application provide a damage location identification device for an underwater production system pipeline, comprising:

[0040] The receiving module is used to receive the first reflected ultrasonic guided wave signal from any underwater production system pipeline collected by the ultrasonic guided wave detection system;

[0041] The noise reduction module is used to perform noise reduction processing on the first reflected ultrasonic guided wave signal using a preset noise reduction algorithm to obtain the second reflected ultrasonic guided wave signal.

[0042] The processing module is used to divide the underwater production system pipeline into multiple pipeline location segments;

[0043] The processing module is also used to input the second reflected ultrasonic guided wave signal into a trained one-dimensional convolutional neural network model for the following processing: extracting damage-related damage features from the second reflected ultrasonic guided wave signal, wherein the damage features include at least one of time-domain damage features and frequency-domain damage features; calculating the damage probability of each pipeline location segment based on at least one of the time-domain damage features and frequency-domain damage features, and outputting the damage probability of each pipeline location segment;

[0044] The determination module is used to identify the pipeline location segment whose damage probability meets preset conditions as the damage location of the underwater production system pipeline.

[0045] Thirdly, embodiments of this application provide an electronic device, including: a memory and a processor;

[0046] The memory stores the instructions that the computer executes;

[0047] The processor executes computer execution instructions stored in memory, causing the processor to perform the first aspect and / or various possible implementations of the first aspect as described above.

[0048] Fourthly, embodiments of this application provide a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, are used to implement the first aspect and / or various possible implementations of the first aspect.

[0049] Fifthly, embodiments of this application provide a computer program product, including a computer program that, when executed by a processor, implements the first aspect and / or various possible implementations of the first aspect.

[0050] The underwater production system pipeline damage location identification method, electronic device, storage medium, and program product provided in this application employ a preset noise reduction algorithm to denoise the acquired first reflected ultrasonic guided wave signal, which can remove noise interference from the signal, improve signal quality and reliability, and thus provide a more accurate data foundation for subsequent damage feature extraction and analysis, helping to improve the accuracy of damage identification. A one-dimensional convolutional neural network model can extract at least one of time-domain and frequency-domain damage features from the second reflected ultrasonic guided wave signal, comprehensively utilizing multiple features to determine the pipeline damage location, which can more comprehensively reflect the pipeline damage location. By dividing the pipeline into multiple pipeline location segments and using a one-dimensional convolutional neural network model to calculate the damage probability of each pipeline location segment, the damage location in the underwater production system pipeline can be located more accurately. Attached Figure Description

[0051] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0052] Figure 1 A schematic diagram of an underwater production system pipeline provided in an embodiment of this application;

[0053] Figure 2a A flowchart illustrating the method for identifying the damage location of underwater production system pipelines provided in this application embodiment. Figure 1 ;

[0054] Figure 2b A schematic diagram of the architecture of a one-dimensional convolutional neural network model provided in an embodiment of this application;

[0055] Figure 3a A schematic flowchart (2) of the method for identifying the damage location of pipelines in an underwater production system provided in this application embodiment;

[0056] Figure 3b A schematic diagram of the group velocity dispersion curve of ultrasonic guided waves in the pipeline of an underwater production system provided in this application embodiment;

[0057] Figure 3c A schematic diagram of the phase velocity dispersion curve of an ultrasonic guided wave in an underwater production system pipeline provided in this embodiment of the application;

[0058] Figure 3d This is a schematic diagram of the propagation of ultrasonic guided waves in a pipeline provided in an embodiment of this application;

[0059] Figure 4 A schematic diagram of the structure of the damage location identification device for the underwater production system pipeline provided in this application embodiment;

[0060] Figure 5This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application.

[0061] The accompanying drawings illustrate specific embodiments of this application, which will be described in more detail below. These drawings and descriptions are not intended to limit the scope of the concept in any way, but rather to illustrate the concept of this application to those skilled in the art through reference to particular embodiments. Detailed Implementation

[0062] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0063] Figure 1 This is a schematic diagram of a scenario for an underwater production system pipeline provided in an embodiment of this application, such as... Figure 1 As shown, the specific application scenarios of this application include underwater production systems and their pipelines.

[0064] Subsea production systems are key facilities for deep-sea oil and gas development, undertaking functions such as transporting oil and gas, connecting and supporting equipment, and playing a vital role in ensuring the normal operation of deep-sea production systems and oil and gas production. In the deep-sea environment, subsea production system pipelines are subjected to high pressure, low temperature, strong corrosion, and complex ocean currents and waves for extended periods, making them highly susceptible to damage. Furthermore, the pipelines in subsea production systems are intricate and complex, with different inner and outer diameters and materials used for different pipelines.

[0065] Currently, ultrasonic guided wave detection technology is widely used in intelligent identification and monitoring of pipeline damage in underwater production systems. Its principle is to utilize the changes in propagation characteristics caused by reflection, refraction, and mode conversion when ultrasonic guided waves encounter damage in the pipeline. These changes are analyzed by signal processing algorithms to identify damage. Signal processing algorithms, such as Fourier transform, analyze the preprocessed signal to extract feature parameters related to pipeline damage, thereby realizing the identification and location of pipeline damage.

[0066] However, the propagation characteristics of ultrasonic guided waves in pipelines are affected by a variety of factors. For pipelines with complex structures, ultrasonic guided waves will undergo complex reflections, refractions and mode conversions at these locations, which can easily generate interference noise. Traditional signal processing methods are difficult to handle this interference noise, resulting in low accuracy in identifying and locating pipeline damage.

[0067] To address the aforementioned technical problems, the following technical concept is proposed: The inventors propose establishing a preset noise reduction algorithm applicable to reflected ultrasonic guided wave signals from underwater production system pipelines. This algorithm aims to suppress interference noise in the reflected ultrasonic guided wave signals and increase the proportion of effective signal. Furthermore, the inventors envision using a deep neural network model to more comprehensively and deeply mine damage characteristics in the noise-reduced reflected ultrasonic guided wave signals, thereby more accurately identifying damage locations. Further, to more precisely locate damage, the inventors propose refining the management of the underwater production system to more precisely detect and analyze damage in different parts of the pipeline.

[0068] The technical solution of this application and how the technical solution of this application solves the above-mentioned technical problems are described in detail below with specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments. The embodiments of this application will now be described with reference to the accompanying drawings.

[0069] Figure 2a A flowchart illustrating the method for identifying the damage location of underwater production system pipelines provided in this application embodiment. Figure 1 ,like Figure 2a As shown, this method can be applied to any electronic device, as detailed below:

[0070] S201. Receive the first reflected ultrasonic guided wave signal from any underwater production system pipeline collected by the ultrasonic guided wave detection system.

[0071] Specifically, the ultrasonic guided wave detection system emits ultrasonic guided waves into the underwater production system pipeline according to a predetermined detection strategy and parameter settings. The ultrasonic guided waves propagate within the pipeline, and when they encounter a point of pipeline damage, reflection occurs. Subsequently, the ultrasonic guided wave detection system, using its own signal receiving device, accurately acquires the first reflected ultrasonic guided wave signal from the pipeline, thus receiving the first reflected ultrasonic guided wave signal.

[0072] S202. A preset noise reduction algorithm is used to perform noise reduction processing on the first reflected ultrasonic guided wave signal to obtain the second reflected ultrasonic guided wave signal.

[0073] Specifically, a preset denoising algorithm is used to extract features from the first reflected ultrasonic guided wave signal, identifying noise components and effective signal features. Then, according to the denoising rules set within the preset algorithm, noise components are suppressed or eliminated. During processing, the algorithm continuously monitors and adjusts the signal in real time to ensure that effective information in the first reflected ultrasonic guided wave signal is preserved to the greatest extent possible while removing noise. After a series of calculations and processing steps by the denoising algorithm, a relatively pure second reflected ultrasonic guided wave signal is finally output.

[0074] S203. Divide the underwater production system pipeline into multiple pipeline location segments.

[0075] Specifically, based on the actual layout drawings of the underwater production system pipelines, key information such as pipeline routing, connection methods, and surrounding environmental characteristics is carefully analyzed. Simultaneously, experience data from previous underwater pipeline inspections and maintenance are referenced to comprehensively consider the varying degrees of impact from factors such as water flow impact, seawater corrosion, and mechanical stress at different pipeline locations. Then, based on this information, the underwater production system pipelines are divided according to pre-defined rules. The division criteria can be based on the physical length of the pipelines, such as setting a segment node every specific length (e.g., 50 meters); or on the functional areas of the pipelines, such as dividing sections connecting different production equipment into independent locations; or even incorporating topographic factors to distinguish pipelines located at different seabed depths or with different geomorphic features. In this way, the originally continuous underwater production system pipelines are systematically divided into multiple pipeline locations with clearly defined boundaries and characteristics.

[0076] S204. Input the second reflected ultrasonic guided wave signal into the trained one-dimensional convolutional neural network model for the following processing: extract damage-related features from the second reflected ultrasonic guided wave signal, wherein the damage features include at least one of time-domain damage features and frequency-domain damage features; calculate the damage probability of each pipeline location segment based on at least one of the time-domain damage features and frequency-domain damage features, and output the damage probability of each pipeline location segment.

[0077] Among them, one-dimensional convolutional neural network models, such as Figure 2b As shown, Figure 2b This is a schematic diagram of the architecture of a one-dimensional convolutional neural network model provided in an embodiment of this application. The model includes: an input layer for inputting the second reflected ultrasonic guided wave signal; a convolutional layer for extracting damage-related features from the second reflected ultrasonic guided wave signal; a pooling layer for reducing the dimensionality of the damage features, reducing computational cost, and maintaining feature invariance; a fully connected layer for integrating the damage features and calculating the damage probability of each pipeline segment; a dropout layer for randomly discarding a portion of neurons to prevent overfitting; and an output layer for outputting the damage probability of each pipeline segment.

[0078] Specifically, when the second reflected ultrasonic guided wave signal is input into a one-dimensional convolutional neural network (CNN) model, the convolutional layers within the model, using their specific convolutional kernels, scan the signal point-by-point. During this process, the convolutional kernels perform convolution operations with local regions of the second reflected ultrasonic guided wave signal, capturing local features. Through layer-by-layer filtering and refinement by multiple convolutional layers, the one-dimensional CNN model can accurately extract damage-related features from the complex second reflected ultrasonic guided wave signal. These damage features encompass at least one of the time-domain and frequency-domain damage features. Regarding time-domain damage features, the CNN model can identify information such as the arrival time difference of the reflected wave, changes in pulse width, and waveform distortion. These features directly reflect signs of damage to the pipeline in the time dimension. For frequency-domain damage features, the CNN model uses mathematical methods such as Fourier transform to convert the time-domain signal to the frequency domain for analysis, extracting features such as amplitude changes of specific frequency components, frequency curve offsets, and abnormal frequency band energy distribution.

[0079] After damage feature extraction, the fully connected layers of the one-dimensional convolutional neural network model comprehensively process at least one of the extracted time-domain and frequency-domain damage features. Through a pre-defined algorithm and weight allocation, these damage features are compared and matched with patterns of different damage levels and corresponding probabilities learned during the training of the one-dimensional convolutional neural network model. Based on this, the model calculates the damage probability of each pipeline segment and outputs the calculated damage probability for each pipeline segment.

[0080] It is possible that a one-dimensional convolutional neural network model, based on damage features, can identify the damage type of each pipeline segment, including no damage, cracks, corrosion, and fracture.

[0081] S205. The pipeline segment whose damage probability meets the preset conditions is determined as the damage location of the underwater production system pipeline.

[0082] Specifically, the preset conditions are set based on a comprehensive consideration of multiple factors. On one hand, design standards and material properties of underwater production system pipelines, as well as past damage cases and statistical data of similar pipelines, are referenced. For example, a reasonable damage probability threshold is set based on parameters such as the fatigue limit and corrosion rate of the materials, combined with the severity of the actual operating environment. For some critical pipeline sections, because they are crucial to the normal operation of the entire underwater production system, the preset damage probability threshold may be relatively low, such as 0.3 (i.e., when the damage probability of a certain pipeline section reaches 30% or higher, that pipeline section is considered to be potentially damaged). Optionally, for some minor or non-critical pipeline sections, the threshold can be appropriately increased, for example, set to 0.5.

[0083] For example, critical pipelines may be those connecting the subsea central manifold to the wellhead, or the subsea central manifold to the offshore pipeline. Damage to these critical pipelines can affect oil and gas collection and transportation, leading to a decrease in system production capacity or even shutdown. Secondary or non-critical pipelines may be flushing or sewage pipelines. A failure in a single pipeline will not cause system shutdown and can be resolved through isolation, switching, or temporary repairs.

[0084] After the damage probability of each pipeline segment is output, the damage probability of each pipeline segment is compared with preset conditions one by one. For those pipeline segments whose damage probabilities meet the preset conditions, the damaged locations of the underwater production system pipelines are determined.

[0085] If possible, select the pipeline segment with the highest probability of loss to determine the location of damage in the underwater production system pipeline.

[0086] In summary, the damage location identification method for underwater production system pipelines provided in this application, by employing a preset noise reduction algorithm to denoise the acquired first reflected ultrasonic guided wave signal, can remove noise interference from the signal, improve signal quality and reliability, and thus provide a more accurate data foundation for subsequent damage feature extraction and analysis, contributing to improved accuracy in damage identification. A one-dimensional convolutional neural network model can extract at least one of time-domain and frequency-domain damage features from the second reflected ultrasonic guided wave signal, comprehensively utilizing multiple features to determine the pipeline damage location, thus providing a more comprehensive reflection of the pipeline damage location. By dividing the pipeline into multiple pipeline segment segments and using a one-dimensional convolutional neural network model to calculate the damage probability of each pipeline segment, the damage location in the underwater production system pipeline can be located more accurately.

[0087] Figure 3a A schematic flowchart of the method for identifying the damage location of underwater production system pipelines provided in this application embodiment is shown in Figure 2. Figure 3a As shown, this method can be applied to any electronic device, as detailed below:

[0088] S301. Establish j one-dimensional convolutional neural network models with different structures, where j is an integer greater than 1.

[0089] The one-dimensional convolutional neural network model includes an input layer, a convolutional layer, a pooling layer, a fully connected layer, and an output layer. By adjusting parameters such as the number of convolutional and pooling layers, the filter size of the convolutional layer, and the pooling window size, j one-dimensional convolutional neural network models with different structures can be established.

[0090] S302. Obtain the damage dataset.

[0091] In some possible implementations, a simulation model of the underwater production system pipeline is established based on preset underwater production system pipeline parameters and preset damage types. Damage datasets are then obtained based on this simulation model.

[0092] Specifically, the preset underwater production system pipeline parameters include geometric and material parameters, and the preset damage types include no damage, cracks, corrosion, and fracture. Based on the preset underwater production system pipeline parameters and preset damage types, an underwater production system pipeline simulation model is established. Based on the sinusoidal wave function modulated by the Hanning window, the preset period, and the preset center frequency, the simulated excitation signal is determined. Based on the simulated excitation signal, simulated ultrasonic guided waves are excited and sent to the underwater production system pipeline simulation model. The underwater production system pipeline simulation model is subjected to extremely refined predefined processing, and simulated reflected ultrasonic guided waves are collected to obtain the loss dataset.

[0093] S303. Input the damage dataset into j one-dimensional convolutional neural network models with different structures, and train the j one-dimensional convolutional neural network models with different structures to obtain j trained one-dimensional convolutional neural network models with different structures.

[0094] Specifically, the damage dataset is input into j one-dimensional convolutional neural network models with different structures. Based on the Rectified Linear Unit (ReLU) activation function and the Softmax output layer activation function, the j one-dimensional convolutional neural network models with different structures are trained to obtain j trained one-dimensional convolutional neural network models with different structures.

[0095] S304. Obtain the classification accuracy, precision, recall, and F1 score of j trained one-dimensional convolutional neural network models with different structures.

[0096] Specifically, the test dataset from the loss dataset is input into j pre-trained one-dimensional convolutional neural network models with different structures. The output results of the one-dimensional convolutional neural network models are compared with the test dataset to obtain the classification accuracy, precision, recall, and F1 score of the j pre-trained one-dimensional convolutional neural network models with different structures. The F1 score is the harmonic mean of precision and recall.

[0097] S305. Based on classification accuracy, precision, recall, and F1 score, select a one-dimensional convolutional neural network model that meets the preset requirements and determine it as a trained one-dimensional convolutional neural network model.

[0098] Specifically, based on the actual needs of underwater production system pipeline damage detection, predefined requirements are established. For example, thresholds for each indicator are set according to the detection focus. Then, the indicators of each one-dimensional convolutional neural network model are compared and screened against the predefined requirements. If multiple one-dimensional convolutional neural network models meet the requirements, the comprehensive score is calculated by weighting each indicator according to its importance, and the optimal model is selected. If no model meets the requirements, the one-dimensional convolutional neural network model needs to be retrained. Finally, the one-dimensional convolutional neural network model that meets the predefined requirements is selected and determined as the trained one-dimensional convolutional neural network model.

[0099] S306. Receive the first reflected ultrasonic guided wave signal from any underwater production system pipeline acquired by the ultrasonic guided wave detection system.

[0100] The ultrasonic guided wave detection system includes an excitation signal source device and a piezoelectric transducer.

[0101] In one possible implementation, before receiving the first reflected ultrasonic guided wave signal from any underwater production system pipeline acquired by the ultrasonic guided wave detection system, the process includes: establishing a cylindrical coordinate system of the underwater production system pipeline cross-section; establishing wave equations and dispersion equations for the underwater production system pipeline based on the cylindrical coordinate system of the underwater production system pipeline cross-section; determining an excitation signal based on the wave equations and dispersion equations; sending the excitation signal to a piezoelectric transducer via an excitation signal source device; and triggering the piezoelectric transducer in response to the excitation signal to transmit ultrasonic guided wave signals to the underwater production system pipeline to acquire the reflected ultrasonic guided wave signal from the underwater production system pipeline.

[0102] Specifically, a cylindrical coordinate system is established for the pipe cross-section to determine parameters such as the pipe's inner and outer diameters a and b, density ρ, elastic modulus E, and Poisson's ratio ν. By introducing displacement potentials φ and ψ, the equations of motion are solved to obtain the wave equation satisfied by the displacement potential variables, as shown in the following formula (1), which includes:

[0103]

[0104] Where φ is the scalar displacement potential; ψ is the vector displacement potential; c L c is the longitudinal wave velocity; T t represents the transverse wave velocity; t represents time.

[0105] Furthermore, substituting the displacement potential into the boundary conditions yields a homogeneous system of equations for constants A1, A2, B1, and B2. Solving this system reveals the relationship between the wave number k and the frequency ω, i.e., the dispersion equation, as shown in formula (2) below, which includes:

[0106] det[A1A2B1B2]=0 (2)

[0107] Where A1 is the first general solution coefficient of the scalar displacement potential; A2 is the second statistical coefficient of the scalar displacement potential; B1 is the first general solution coefficient of the vector displacement potential; and B2 is the second general solution coefficient of the vector displacement potential.

[0108] Possibly, the dispersion curve of ultrasonic guided waves in underwater production system pipelines is as follows: Figure 3b and Figure 3c As shown, Figure 3b This is a schematic diagram of the group velocity dispersion curve of ultrasonic guided waves in the pipeline of an underwater production system provided in this application embodiment. Figure 3c This is a schematic diagram of the phase velocity dispersion curve of the ultrasonic guided wave in the pipeline of the underwater production system provided in this application embodiment.

[0109] Based on the wave equation and the dispersion equation, the group velocity and phase velocity dispersion curves of the ultrasonic guided wave in the pipeline of the underwater production system are obtained. Based on the group velocity and phase velocity dispersion curves, the frequency of the excitation signal is calculated to determine the excitation signal.

[0110] An excitation signal is sent to the piezoelectric transducer via an excitation signal source device. Responding to the excitation signal, the piezoelectric transducer is triggered. The piezoelectric transducer is a key device based on the piezoelectric effect; upon receiving the excitation signal, its internal piezoelectric material undergoes electromechanical coupling, converting the electrical signal into mechanical vibration. This mechanical vibration is further emitted as ultrasonic guided waves into the underwater production system pipeline. The ultrasonic guided waves propagate along a specific path within the pipeline. During propagation, if they encounter structural changes, damage, or the end of the pipeline, the ultrasonic guided waves will be reflected, such as... Figure 3d As shown, Figure 3d This is a schematic diagram of the propagation of ultrasonic guided waves in a pipeline provided in an embodiment of this application. The ultrasonic guided waves emitted by the piezoelectric transducer propagate to both ends of the pipeline. When they encounter a loss in the pipeline, they will be reflected. At this time, the ultrasonic guided wave signals reflected back from the pipeline are collected using a corresponding signal acquisition device.

[0111] S307. A preset noise reduction algorithm is used to perform noise reduction processing on the first reflected ultrasonic guided wave signal to obtain the second reflected ultrasonic guided wave signal.

[0112] In one possible implementation, the reflected ultrasonic guided wave signal is input to a first preset algorithm to decompose the reflected ultrasonic guided wave signal, and the output obtained is n signal components, where n is an integer greater than 1. The permutation entropy value of the n signal components is calculated, and the signal components with permutation entropy values ​​less than a preset threshold are selected to obtain effective signal components. The signal components with permutation entropy values ​​greater than or equal to the preset threshold are selected to obtain noise signal components. The noise signal components are decomposed and reconstructed by wavelet decomposition, and wavelet threshold denoising is performed to obtain denoised signal components. The effective signal components and the denoised signal components are merged to obtain the denoised reflected ultrasonic guided wave signal.

[0113] Among them, wavelet threshold denoising is used to indicate the use of a soft threshold function to denoise the noise signal components.

[0114] For example, the first preset algorithm can be an improved complete ensemble empirical mode decomposition algorithm with adaptive noise (ICEEMDAN), where the signal components are intrinsic mode function (IMF) components.

[0115] In one possible implementation, before inputting the reflected ultrasonic guided wave signal into the first preset algorithm, the method includes: generating a first preset number of first parameters and second parameters within a preset numerical range based on a second preset algorithm; combining the first parameters and second parameters pairwise and calculating the envelope entropy value of the combinations; selecting combinations whose envelope entropy values ​​are less than a first preset threshold to establish the first preset algorithm.

[0116] For example, the second preset algorithm is the Northern Goshawk Optimization (NGO) algorithm.

[0117] If possible, select the combination with the smallest envelope entropy value to establish the first preset algorithm.

[0118] S308. Divide the underwater production system pipeline into multiple pipeline location segments.

[0119] S309. Input the second reflected ultrasonic guided wave signal into the trained one-dimensional convolutional neural network model for the following processing: extract damage-related features from the second reflected ultrasonic guided wave signal, wherein the damage features include at least one of time-domain damage features and frequency-domain damage features; calculate the damage probability of each pipeline location segment based on at least one of the time-domain damage features and frequency-domain damage features, and output the damage probability of each pipeline location segment.

[0120] S310. The pipeline segment whose damage probability meets the preset conditions is determined as the damage location of the underwater production system pipeline.

[0121] It should be noted that the specific implementation methods of steps S308 to S310 can refer to the specific implementation methods of steps S203 to S205, and will not be elaborated further here.

[0122] In summary, the damage location identification method for underwater production system pipelines provided in this application employs a preset noise reduction algorithm to denoise the acquired first reflected ultrasonic guided wave signal. This removes noise interference from the signal, improves signal quality and reliability, and provides a more accurate data foundation for subsequent damage feature extraction and analysis, thus contributing to improved accuracy in damage identification. A one-dimensional convolutional neural network model can extract at least one of time-domain and frequency-domain damage features from the second reflected ultrasonic guided wave signal. By comprehensively utilizing multiple features to determine the pipeline damage location, it can more comprehensively reflect the pipeline damage location. By dividing the pipeline into multiple pipeline segment segments and using a one-dimensional convolutional neural network model to calculate the damage probability of each segment, the damage location in the underwater production system pipeline can be located more accurately.

[0123] Furthermore, we construct j one-dimensional convolutional neural network models with different structures, each with a unique architecture. These models can complement each other, reducing the risk of overfitting. We obtain the classification accuracy, precision, recall, and F1 score for each model. These metrics reflect the model's performance from different perspectives. Based on these multi-dimensional metrics, we can more accurately select the model that meets the preset requirements. Different types of damage data may have different characteristics and distributions. By constructing and training multiple models with different structures, we can find the model best suited for a specific type of damage data.

[0124] By pre-setting the pipeline parameters and damage types of the underwater production system, the system can accurately simulate potential damage to the pipeline under various actual operating conditions. Using the established simulation model, damage datasets under different operating conditions can be systematically generated, significantly reducing experimental costs and improving research efficiency.

[0125] Furthermore, establishing a cylindrical coordinate system for the cross-section of the underwater production system pipeline allows for a more accurate description of its physical characteristics based on the pipeline's geometry. Establishing wave and dispersion equations for the underwater production system pipeline using this cylindrical coordinate system provides a theoretical understanding of the propagation patterns of ultrasonic guided waves within the pipeline. Determining the excitation signal based on the wave and dispersion equations ensures the signal's relevance and effectiveness. By sending the excitation signal to a piezoelectric transducer via an excitation signal source device, and then having the piezoelectric transducer emit ultrasonic guided wave signals to collect the reflected ultrasonic guided wave signals, a reliable signal excitation and acquisition process is achieved.

[0126] Furthermore, the reflected ultrasonic guided wave signal is input into a first preset algorithm for decomposition, resulting in n signal components. This decomposition method breaks down the complex reflected ultrasonic guided wave signal into several relatively simple components, facilitating a more detailed analysis of the signal characteristics. By calculating the permutation entropy values ​​of the n signal components and selecting effective and noise signal components based on preset thresholds, effective distinction between signal and noise is achieved. Wavelet decomposition and reconstruction of the noise signal components, followed by wavelet threshold denoising, effectively removes noise. The effective and denoised signal components are then combined to obtain the denoised reflected ultrasonic guided wave signal, significantly improving signal quality.

[0127] Furthermore, by generating a first preset number of first and second parameters based on a second preset algorithm within a preset numerical range, the possible parameter space can be systematically searched. This helps to find optimal or near-optimal parameter combinations to adapt to different underwater production system pipeline inspection scenarios and the characteristics of reflected ultrasonic guided wave signals, thereby improving the algorithm's adaptability and generalization ability. Calculating the envelope entropy values ​​of pairwise combinations of the first and second parameters, and selecting combinations with envelope entropy values ​​less than a first preset threshold to establish the first preset algorithm, provides an effective parameter selection criterion. Using the selected parameter combinations to establish the first preset algorithm can optimize the algorithm's performance.

[0128] Figure 4 This is a schematic diagram of the structure of the damage location identification device for underwater production system pipelines provided in the embodiments of this application, as shown below. Figure 4 As shown, the underwater production system pipeline damage location identification device 40 provided in this embodiment includes: a receiving module 401, a noise reduction module 402, a processing module 403, and a determination module 404.

[0129] The receiving module 401 is used to receive the first reflected ultrasonic guided wave signal of any underwater production system pipeline collected by the ultrasonic guided wave detection system.

[0130] The noise reduction module 402 is used to perform noise reduction processing on the first reflected ultrasonic guided wave signal using a preset noise reduction algorithm to obtain the second reflected ultrasonic guided wave signal.

[0131] Processing module 403 is used to divide the underwater production system pipeline into multiple pipeline location segments.

[0132] The processing module 403 is further configured to input the second reflected ultrasonic guided wave signal into a trained one-dimensional convolutional neural network model for the following processing: extracting damage-related features from the second reflected ultrasonic guided wave signal, wherein the damage features include at least one of time-domain damage features and frequency-domain damage features; calculating the damage probability of each pipeline location segment based on at least one of the time-domain damage features and frequency-domain damage features, and outputting the damage probability of each pipeline location segment.

[0133] The determination module 404 is used to determine the pipeline location segment whose damage probability meets the preset conditions as the damage location of the underwater production system pipeline.

[0134] In one possible implementation, the ultrasonic guided wave detection system includes an excitation signal source device and a piezoelectric transducer, and the receiving module 401 is specifically used for:

[0135] Establish a cylindrical coordinate system for the cross-section of the underwater production system pipeline;

[0136] Based on the cylindrical coordinate system of the cross-section of the underwater production system pipeline, the wave equation and dispersion equation of the underwater production system pipeline are established.

[0137] The excitation signal is determined based on the wave equation and the dispersion equation;

[0138] An excitation signal is sent to the piezoelectric transducer through an excitation signal source device;

[0139] In response to the excitation signal, the piezoelectric transducer is triggered to transmit ultrasonic guided wave signals to the underwater production system pipeline in order to collect the reflected ultrasonic guided wave signals from the underwater production system pipeline.

[0140] In one possible implementation, the noise reduction module 402 is specifically used for:

[0141] The reflected ultrasonic guided wave signal is input into the first preset algorithm to decompose the reflected ultrasonic guided wave signal, and the output obtained is n signal components, where n is an integer greater than 1;

[0142] Calculate the permutation entropy of n signal components;

[0143] Select signal components whose permutation entropy value is less than a preset threshold to obtain effective signal components;

[0144] Select signal components whose permutation entropy value is greater than or equal to a preset threshold to obtain noise signal components;

[0145] The noise signal components are decomposed and reconstructed using wavelet decomposition, and then subjected to wavelet threshold denoising to obtain the denoised signal components.

[0146] The effective signal component and the denoised signal component are combined to obtain the denoised reflected ultrasonic guided wave signal.

[0147] In one possible implementation, the noise reduction module 402 is specifically used for:

[0148] Based on the second preset algorithm, within a preset numerical range, a first preset number of first parameters and second parameters are generated;

[0149] Combine the first and second parameters in pairs and calculate the combined envelope entropy value;

[0150] Select combinations whose envelope entropy values ​​are less than a first preset threshold to establish a first preset algorithm.

[0151] In one possible implementation, the processing module 403 is specifically used for:

[0152] Establish j one-dimensional convolutional neural network models with different structures, where j is an integer greater than 1;

[0153] Obtain the damage dataset;

[0154] The damage dataset is input into j one-dimensional convolutional neural network models with different structures, and the j one-dimensional convolutional neural network models with different structures are trained to obtain j trained one-dimensional convolutional neural network models with different structures.

[0155] Obtain the classification accuracy, precision, recall, and F1 score of j trained one-dimensional convolutional neural network models with different structures;

[0156] Based on classification accuracy, precision, recall, and F1 score, a one-dimensional convolutional neural network model that meets the preset requirements is selected and determined as a trained one-dimensional convolutional neural network model.

[0157] In one possible implementation, the processing module 403 is specifically used for:

[0158] A simulation model of the underwater production system pipeline is established based on the preset pipeline parameters and preset damage types.

[0159] Damage datasets were obtained based on a simulation model of the underwater production system pipeline.

[0160] The underwater production system pipeline damage location identification device provided in this embodiment can execute the method provided in the above method embodiment. Its implementation principle and technical effect are similar, and will not be described in detail here.

[0161] Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Figure 5 As shown, the electronic device 50 provided in this embodiment includes at least one processor 501 and a memory 502. Optionally, the electronic device 50 further includes a communication component 503. The processor 501, memory 502, and communication component 503 are connected via a bus.

[0162] In a specific implementation, at least one processor 501 executes computer execution instructions stored in memory 502, causing at least one processor 501 to perform the above-described method.

[0163] The specific implementation process of processor 501 can be found in the above method embodiments, and its implementation principle and technical effect are similar. It will not be repeated here.

[0164] In the above embodiments, it should be understood that the processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), etc. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the method disclosed in this invention can be directly implemented by a hardware processor, or implemented by a combination of hardware and software modules within the processor.

[0165] The memory may include random access memory (RAM) and may also include non-volatile memory (NVM), such as at least one disk storage device.

[0166] The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of illustration, the buses shown in the accompanying drawings are not limited to a single bus or a single type of bus.

[0167] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the above-described method.

[0168] This application also provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the above-described method.

[0169] The aforementioned readable storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. The readable storage medium can be any available medium accessible to a general-purpose or special-purpose computer.

[0170] An exemplary readable storage medium is coupled to a processor, enabling the processor to read information from and write information to the readable storage medium. Of course, the readable storage medium can also be a component of the processor. The processor and the readable storage medium can reside in an Application Specific Integrated Circuit (ASIC). Alternatively, the processor and the readable storage medium can exist as discrete components in the device.

[0171] The division of units is merely a logical functional division; in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be indirect coupling or communication connection through some interfaces, devices, or units, and may be electrical, mechanical, or other forms.

[0172] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0173] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0174] If a function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0175] Those skilled in the art will understand that all or part of the steps of the above-described method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When executed, the program performs the steps of the above-described method embodiments; and the aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.

[0176] Finally, it should be noted that other embodiments of the invention will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This invention is intended to cover any variations, uses, or adaptations of the invention that follow the general principles of the invention and include common knowledge or customary techniques in the art not disclosed herein, and is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of the invention is limited only by the appended claims.

Claims

1. A method for identifying the location of damage in pipelines of an underwater production system, characterized in that, include: Receive the first reflected ultrasonic guided wave signal from any underwater production system pipeline acquired by the ultrasonic guided wave detection system; A preset noise reduction algorithm is used to perform noise reduction processing on the first reflected ultrasonic guided wave signal to obtain the second reflected ultrasonic guided wave signal; The underwater production system pipeline is divided into multiple pipeline location segments; The second reflected ultrasonic guided wave signal is input into a trained one-dimensional convolutional neural network model for the following processing: damage features related to damage are extracted from the second reflected ultrasonic guided wave signal, wherein the damage features include time-domain damage features and frequency-domain damage features; based on a preset algorithm and weight allocation, the time-domain damage features and frequency-domain damage features are compared and matched with different damage levels and corresponding probability patterns learned during the training of the one-dimensional convolutional neural network model, the damage probability and damage type of each pipeline segment are calculated, and the damage probability of each pipeline segment is output; the damage types include no damage, cracks, corrosion, and fracture. The pipeline segment whose damage probability meets the preset condition is determined as the damage location of the underwater production system pipeline; the preset condition is that the damage probability reaches a preset damage probability threshold; wherein, the preset damage probability threshold is determined based on the fatigue limit of the material, the corrosion rate and the operating environment, and the preset damage probability threshold is different for some pipeline segments. The method further includes: Establish j one-dimensional convolutional neural network models with different structures, where j is an integer greater than 1; A simulation model of the underwater production system pipeline is established based on the preset pipeline parameters and preset damage types. Based on the sinusoidal function modulated by the Hanning window, the preset period and the preset center frequency, the simulated excitation signal is determined. Based on the simulated excitation signal, the simulated ultrasonic guided wave is excited and sent to the underwater production system pipeline simulation model. The underwater production system pipeline simulation model is subjected to extremely refined predefined processing, and the simulated reflected ultrasonic guided wave is collected to obtain the damage dataset. The damage dataset is input into the j one-dimensional convolutional neural network models with different structures. Based on the modified linear unit activation function and the softmax output layer activation function, the j one-dimensional convolutional neural network models with different structures are trained to obtain j trained one-dimensional convolutional neural network models with different structures. The test dataset from the damage dataset is input into j pre-trained one-dimensional convolutional neural network models with different structures. The output results of the one-dimensional convolutional neural network models are compared with the test dataset to obtain the classification accuracy, precision, recall and F1 score of the j pre-trained one-dimensional convolutional neural network models with different structures. The F1 score is the harmonic mean of precision and recall. Based on classification accuracy, precision, recall, and F1 score, a one-dimensional convolutional neural network model that meets preset requirements is selected from j trained one-dimensional convolutional neural network models with different structures, and is determined as a trained one-dimensional convolutional neural network model. The preset requirements are related to the detection requirements of pipeline damage detection in the underwater production system. The ultrasonic guided wave detection system includes an excitation signal source device and a piezoelectric transducer; Accordingly, before receiving the first reflected ultrasonic guided wave signal from any underwater production system pipeline acquired by the ultrasonic guided wave detection system, the following steps are included: Establish a cylindrical coordinate system for the cross-section of the underwater production system pipeline; Based on the cylindrical coordinate system of the cross-section of the underwater production system pipeline, the wave equation and dispersion equation of the underwater production system pipeline are established. The excitation signal is determined based on the wave equation and dispersion equation. An excitation signal is sent to the piezoelectric transducer through an excitation signal source device; In response to the excitation signal, the piezoelectric transducer is triggered to transmit an ultrasonic guided wave signal to the underwater production system pipeline in order to collect the reflected ultrasonic guided wave signal of the underwater production system pipeline.

2. The method according to claim 1, characterized in that, The step of using a preset noise reduction algorithm to denoise the first reflected ultrasonic guided wave signal to obtain the second reflected ultrasonic guided wave signal includes: The reflected ultrasonic guided wave signal is input into a first preset algorithm to decompose the reflected ultrasonic guided wave signal, and the output obtained is n signal components, where n is an integer greater than 1; Calculate the permutation entropy of n signal components; Select signal components whose permutation entropy value is less than a preset threshold to obtain effective signal components; Select signal components whose permutation entropy value is greater than or equal to a preset threshold to obtain noise signal components; The noise signal components are decomposed and reconstructed using wavelet decomposition, and then subjected to wavelet threshold denoising to obtain the denoised signal components. The effective signal component and the denoised signal component are combined to obtain the denoised reflected ultrasonic guided wave signal.

3. The method according to claim 2, characterized in that, Before inputting the reflected ultrasonic guided wave signal into the first preset algorithm, the following steps are included: Based on the second preset algorithm, within a preset numerical range, a first preset number of first parameters and second parameters are generated; The first parameter and the second parameter are combined in pairs, and the envelope entropy value of the combination is calculated. Select combinations whose envelope entropy values ​​are less than a first preset threshold to establish the first preset algorithm.

4. A damage location identification device for pipelines in an underwater production system, characterized in that, include: The receiving module is used to receive the first reflected ultrasonic guided wave signal from any underwater production system pipeline collected by the ultrasonic guided wave detection system; The noise reduction module is used to perform noise reduction processing on the first reflected ultrasonic guided wave signal using a preset noise reduction algorithm to obtain the second reflected ultrasonic guided wave signal. The processing module is used to divide the underwater production system pipeline into multiple pipeline location segments; The processing module is further configured to input the second reflected ultrasonic guided wave signal into a trained one-dimensional convolutional neural network model for the following processing: extracting damage-related features from the second reflected ultrasonic guided wave signal, wherein the damage features include time-domain damage features and frequency-domain damage features; comparing and matching the time-domain damage features and frequency-domain damage features with patterns of different damage degrees and corresponding probabilities learned during the training of the one-dimensional convolutional neural network model based on a preset algorithm and weight allocation, calculating the damage probability and damage type of each pipeline segment, and outputting the damage probability of each pipeline segment; the damage types include no damage, cracks, corrosion, and fracture. The determination module is used to determine the pipeline location segment where the damage probability meets the preset condition as the damage location of the underwater production system pipeline; the preset condition is that the damage probability reaches a preset damage probability threshold; wherein, the preset damage probability threshold is determined based on the fatigue limit of the material, the corrosion rate and the operating environment, and the preset damage probability threshold is different for some pipeline location segments. The processing module is also used to establish j one-dimensional convolutional neural network models with different structures, where j is an integer greater than 1; A simulation model of the underwater production system pipeline is established based on the preset pipeline parameters and preset damage types. Based on the sinusoidal function modulated by the Hanning window, the preset period and the preset center frequency, the simulated excitation signal is determined. Based on the simulated excitation signal, the simulated ultrasonic guided wave is excited and sent to the underwater production system pipeline simulation model. The underwater production system pipeline simulation model is subjected to extremely refined predefined processing, and the simulated reflected ultrasonic guided wave is collected to obtain the damage dataset. The damage dataset is input into the j one-dimensional convolutional neural network models with different structures. Based on the modified linear unit activation function and the softmax output layer activation function, the j one-dimensional convolutional neural network models with different structures are trained to obtain j trained one-dimensional convolutional neural network models with different structures. The test dataset from the damage dataset is input into j pre-trained one-dimensional convolutional neural network models with different structures. The output results of the one-dimensional convolutional neural network models are compared with the test dataset to obtain the classification accuracy, precision, recall and F1 score of the j pre-trained one-dimensional convolutional neural network models with different structures. The F1 score is the harmonic mean of precision and recall. Based on classification accuracy, precision, recall, and F1 score, a one-dimensional convolutional neural network model that meets preset requirements is selected from j trained one-dimensional convolutional neural network models with different structures, and is determined as a trained one-dimensional convolutional neural network model. The preset requirements are related to the detection requirements of pipeline damage detection in the underwater production system. The ultrasonic guided wave detection system includes an excitation signal source device and a piezoelectric transducer; Accordingly, before receiving the first reflected ultrasonic guided wave signal from any underwater production system pipeline acquired by the ultrasonic guided wave detection system, the receiving module is further configured to: Establish a cylindrical coordinate system for the cross-section of the underwater production system pipeline; Based on the cylindrical coordinate system of the cross-section of the underwater production system pipeline, the wave equation and dispersion equation of the underwater production system pipeline are established. The excitation signal is determined based on the wave equation and dispersion equation. An excitation signal is sent to the piezoelectric transducer through an excitation signal source device; In response to the excitation signal, the piezoelectric transducer is triggered to transmit an ultrasonic guided wave signal to the underwater production system pipeline in order to collect the reflected ultrasonic guided wave signal of the underwater production system pipeline.

5. An electronic device, characterized in that, include: Memory, processor; The memory stores computer-executed instructions; The processor executes computer execution instructions stored in the memory, causing the processor to perform the method as described in any one of claims 1-3.

6. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, are used to implement the method as described in any one of claims 1-3.

7. A computer program product comprising a computer program that, when executed by a processor, implements the method of any one of claims 1-3.

Images