A Method and System for Constructing a Dataset of Acoustic Imaging of Cable Joints Ingress into Water

Abstract

This invention discloses a method and system for constructing a acoustic fingerprint dataset for cable joints exposed to water ingress based on acoustic simulation, belonging to the field of cable accessory condition monitoring technology. The method includes: constructing a three-dimensional model of the cable joint and defining medium properties; configuring an acoustic-structural coupling boundary at the solid-fluid interface; setting a PML perfectly matched layer outside the computational domain and setting fixed constraints at both ends of the cable to simulate a real open environment and impact detection state; subsequently setting the water volume variable, deploying excitation sources and monitoring points; performing multi-field transient solutions on the pressure acoustic equation and structural dynamics equation to obtain the sound pressure time-domain signal; finally, extracting waveform features and combining them with operating condition labels to generate initial samples, and aggregating them to construct a training dataset. This invention solves the problems of difficulty in obtaining actual water ingress fault samples and high measurement costs. Through refined boundary settings, it ensures the high fidelity of the simulated data, providing high-quality data support for the acoustic fingerprint recognition model and enhancing the detection accuracy and generalization ability.

Description

Technical Field

[0001] This invention belongs to the field of cable accessory condition monitoring technology, specifically relating to a method and system for constructing a cable joint water ingress acoustic pattern dataset based on acoustic simulation. Background Technology

[0002] During cable laying, due to limitations in channel routing, spatial layout, and equipment interface locations, elbow-type plug-in cable terminals (hereinafter referred to as "elbow terminals") are widely used for corner connections or equipment access. This type of connector is widely used in cable systems with voltage levels from 10kV to 35kV due to its compact structure, convenient installation, and adaptability to multi-directional connections.

[0003] In actual operating environments, the reliability of elbow joints is affected by various factors, among which water accumulation is a prominent issue. Cable trenches and cable wells are generally characterized by high humidity and poor ventilation. Factors such as groundwater infiltration, rainwater backflow, and condensation can easily lead to water accumulation within these trenches. After installation, the elbow joint forms a multi-layered interface structure with the cable body and equipment bushings. If it remains in a humid environment for a long time, moisture may seep in along the interface or form a water film on the joint surface. Because the relative permittivity of water is much higher than that of air and insulating materials, the presence of moisture alters the electric field distribution inside the joint, leading to localized electric field distortion. Over time, this can cause surface discharge, electrical treeing degradation, and even insulation breakdown.

[0004] If such faults occur in cable trenches in core areas such as substations and power plants, they may not only cause damage to individual equipment, but also trigger grounding faults or phase-to-phase short circuits, affecting the coordination of system relay protection and expanding the scope of the accident. At the same time, underground cable trenches are narrow and have a complex environment, making repair work difficult and time-consuming once a fault occurs, often resulting in prolonged power outages.

[0005] Currently, monitoring methods for the operational status of cable elbow joints remain relatively limited. Power maintenance units mainly rely on regular manual inspections, visually inspecting the joint's appearance or conducting spot checks using handheld partial discharge meters. This method has the following shortcomings: First, the inspection cycle is long, making it difficult to detect intermittent or slowly developing water accumulation defects; second, the inspection coverage is limited, and many elbow joints located deep in cable trenches cannot be effectively inspected; third, there is a lack of quantitative analysis methods for the electric field distribution under water accumulation conditions, making it difficult to assess the specific impact of water accumulation on insulation performance and failing to provide data support for maintenance decisions. Summary of the Invention

[0006] To address the shortcomings of existing technologies, the present invention aims to provide a method and system for constructing a water inlet acoustic texture dataset for cable joints based on acoustic simulation, thereby solving the problems in existing technologies.

[0007] The objective of this invention can be achieved through the following technical solutions: A method for constructing a water ingress acoustic signature dataset for cable joints based on acoustic simulation includes the following steps: Obtain the three-dimensional geometric model of the cable joint, and configure the acoustic physical parameters of the solid medium entity and the fluid medium entity in the three-dimensional geometric model respectively; An acoustic-structural coupling boundary is configured at the interface between the solid medium entity and the fluid medium entity; acoustic simulation boundary conditions are set, including: configuring a perfectly matched layer at the outermost computational domain boundary of the three-dimensional geometric model, and configuring fixed constraints at both ends of the cable entity of the three-dimensional geometric model; The water volume variable is set in the preset water inlet area inside the three-dimensional geometric model, and excitation sound sources and virtual monitoring points located in the air medium entity are set up at the specified spatial coordinate positions. Based on the acoustic simulation boundary conditions, the set water volume variable, and the acoustic-structure coupling boundary, a multiphysics transient solution is performed. For fluid media, a transient pressure acoustic equation containing a damping attenuation term is solved, and for solid media, a structural transient dynamic equation containing stress tensor divergence is solved to obtain the sound pressure time-domain signal at the virtual monitoring point. The sound pressure time-domain signal is truncated and its features are extracted. Initial acoustic samples with water inflow condition labels are generated by combining them with the corresponding water volume variables. Multiple sets of initial acoustic samples are aggregated to generate an acoustic print dataset for training the water inflow detection model.

[0008] Further, configuring the acoustic physical parameters includes: Define the corresponding solid acoustic physical parameters for the EPDM insulating rubber, stainless steel, copper, and epoxy resin components in the three-dimensional geometric model respectively; The corresponding fluid acoustic physical parameters are defined for the air domain and the water domain in the three-dimensional geometric model, respectively.

[0009] Furthermore, the process of configuring a perfectly matched layer at the outermost computational domain boundary of the three-dimensional geometric model includes: introducing a computational region with absorption characteristics near the boundary of the computational domain, the absorption characteristics of which are related to the propagation direction and frequency of the sound wave, so that the wave propagating into the computational region is attenuated.

[0010] Furthermore, the process of configuring fixed constraints at both ends of the cable entity in the three-dimensional geometric model includes: applying full displacement constraints to the geometric entities at both ends of the cable entity, so that the translational displacement in all directions on the selected geometric entity is zero, and if there is rotational degree of freedom, the rotational degree of freedom in all directions is also zero.

[0011] Furthermore, the water volume variable includes a gradient sequence of multiple preset inlet volume nodes set according to volume; The methods for setting the water volume variable include: representing different water inflow volumes by changing the size or material properties of the medium in the preset water inflow area in the three-dimensional geometric model; Furthermore, the expression for the transient pressure acoustic equation is as follows: In the formula, c is the density of the medium; c is the speed of sound; Total transient pressure; For the sound source term, it represents the intensity of the sound source or excitation; For damping; It describes the rate of change of sound pressure over time; This is a term describing the distribution of sound pressure in space.

[0012] Furthermore, the expression for the transient dynamic equation of the structure is as follows: in, The density of the medium; Let be the displacement vector, representing the position change of a particle in the material at time t. Let be the second time derivative of the displacement, and let represent the acceleration of the particle. Let be the divergence of the stress tensor. Body force density vector.

[0013] Furthermore, the process of truncating and extracting features from the sound pressure time-domain signal includes: The amplitude of the truncated sound pressure time-domain signal is normalized. Based on the normalized signal, peak amplitude features, root mean square (RMS) value features, decay rate features, and rise time features are extracted respectively, and the extracted features are combined into an acoustic feature vector.

[0014] The system for constructing a water ingress acoustic signature dataset for cable joints based on acoustic simulation executes the above-mentioned methods, including: The physics configuration module is used to acquire the three-dimensional geometric model of the cable joint and configure the acoustic physical parameters of the solid medium entity and the fluid medium entity in the three-dimensional geometric model respectively. The coupling and constraint configuration module is used to configure acoustic-structural coupling boundaries at the interface between solid and fluid media entities; and to set acoustic simulation boundary conditions, including configuring a perfectly matched layer at the outermost computational domain boundary of the three-dimensional geometric model and configuring fixed constraints at both ends of the cable entity in the three-dimensional geometric model. The variable and source deployment module is used to set the water volume variable in the preset water inlet area inside the three-dimensional geometric model, and to deploy the excitation sound source and the virtual monitoring point located in the air medium entity at the specified spatial coordinate position. The multiphysics solution module is used to perform multiphysics transient solutions based on the acoustic simulation boundary conditions, the set water volume variables, and the acoustic-structure coupling boundary. Specifically, for fluid media, it solves the transient pressure acoustic equation containing damping attenuation terms, and for solid media, it solves the structural transient dynamic equation containing stress tensor divergence, thereby obtaining the sound pressure time-domain signal at the virtual monitoring point. The dataset generation module is used to extract features from the sound pressure time-domain signal and generate initial acoustic samples with water inflow condition labels by combining them with the corresponding water volume variables. Multiple sets of the initial acoustic samples are aggregated to generate an acoustic print dataset for training the water inflow detection model.

[0015] An electronic device includes: a processor, a memory, a communication interface, and a communication bus, wherein the processor, the memory, and the communication interface communicate with each other through the communication bus; The memory is used to store at least one executable instruction, which causes the processor to perform the operation corresponding to the above-described method for constructing a cable joint water ingress acoustic fingerprint dataset based on acoustic simulation.

[0016] The beneficial effects of this invention are: 1. This invention sets different water inflow rates from 0ml to 50ml in a three-dimensional geometric model and solves the sound pressure time-domain signal under different conditions using multiphysics transient equations; it effectively overcomes the difficulty in collecting samples of water immersion faults in actual cable trenches; and can generate acoustic response data covering various water immersion levels in a low-cost, batch manner under precise control of variables, providing effective data support for the formulation of operation and maintenance strategies and the quantitative assessment of fault levels.

[0017] 2. In acoustic simulation modeling, this invention introduces a PML (Perfect Matching Layer) at the outermost boundary of the computational domain, utilizing its absorption characteristics related to the direction and frequency of sound wave propagation to handle outwardly radiated sound waves. This effectively attenuates sound waves propagating to the simulation boundary, effectively avoiding sound wave reflection caused by the truncation of the computational domain boundary. This eliminates the interference of reflected waves on the internal sound field distribution, making the simulation results closer to the real open or semi-open acoustic environment of cable trenches.

[0018] 3. In the automatic impact detection simulation process, the present invention sets fixed constraint boundary conditions at both ends of the cable geometry, which restricts the translational displacement and rotational degrees of freedom of the entity in all directions, effectively preventing the overall rigid body displacement of the cable terminal head during the simulated impact collision; it also eliminates the interference of displacement noise on the local sound pressure signal, and provides double protection that the simulation model closely resembles the constrained state of the actual installation environment at the mechanical level, thereby improving the objectivity of the acquired signal.

[0019] 4. This invention performs time-domain truncation and amplitude normalization on the acquired sound pressure time-domain signal, further extracting features such as peak amplitude, RMS value, and attenuation rate, and attaching corresponding water inflow condition labels, dividing it into training and testing sets; thereby changing the outdated method of relying on manual handheld instruments for sampling in the prior art; through a standardized data processing flow, the original waveform is transformed into a structured feature vector that can be directly absorbed by the machine learning model, effectively improving the training accuracy and generalization ability of the subsequent water inflow detection soundprint recognition model.

[0020] 5. All physical parameters (such as air and water density and sound velocity) and three-dimensional geometric model dimensions of this invention are defined parametrically. When faced with products of different voltage levels or different manufacturers, technicians only need to adjust the corresponding geometric dimensions and material parameter matrices to transfer the simulation dataset construction method to different models and specifications of cable elbow plugs at low cost, which has a good engineering applicability. Attached Figure Description

[0021] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0022] Figure 1 This invention is a three-dimensional model of a cable elbow-type plug-in end; Figure 2 This is a schematic diagram of water accumulation at an elbow-type adapter in finite element analysis software. Figure 3 This is a distribution diagram of reflected acoustic patterns corresponding to different water accumulation conditions according to the present invention; Figure 4 This is a schematic diagram of the fixed constraint setting of the elbow-type plug-in head in the finite element analysis software of this invention; Figure 5 This is a flowchart of the method for constructing a water ingress acoustic pattern dataset for cable joints based on acoustic simulation, as described in this invention. Figure 6 This is a schematic diagram of the striking position of the elbow-type cable plug-in end of the present invention. Detailed Implementation

[0023] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0024] Example 1 In this embodiment, a cable elbow plug is used as the cable connector. The method for constructing a cable connector water inlet acoustic signature dataset based on acoustic simulation is specifically described. Figure 5 As shown, it includes the following steps: S1. Obtain the three-dimensional geometric model of the cable joint, and configure the acoustic physical parameters of the solid medium entity and the fluid medium entity in the three-dimensional geometric model respectively; configure the acoustic-structure coupling boundary at the interface between the solid medium entity and the fluid medium entity; set the acoustic simulation boundary conditions, including: configuring a perfect matching layer at the outermost computational domain boundary of the three-dimensional geometric model, and configuring fixed constraints at both ends of the cable entity in the three-dimensional geometric model. In this embodiment, the three-dimensional geometric model of the cable connector (cable elbow plug) is as follows: Figure 1 As shown, the configured acoustic physical parameters specifically include: defining corresponding solid acoustic physical parameters for EPDM insulating rubber, stainless steel, copper, and epoxy resin components in the three-dimensional geometric model; and defining corresponding fluid acoustic physical parameters for the air domain and water domain in the three-dimensional geometric model.

[0025] The acoustic physical parameters include: density, elastic modulus, and sound velocity of the solid portion, as well as density and sound velocity of the fluid medium; In this embodiment, an acoustic-structure coupling boundary is set at the interface between the solid domain and the fluid domain, and an acoustic-structure coupling boundary is configured at the interface between the solid medium entity and the fluid medium entity. This boundary satisfies the kinematic continuity condition and the dynamic equilibrium condition, and the specific constraint equations are as follows: in, The unit normal vector representing the fluid-structure interaction interface. For the displacement vector of the solid structure, The density of the fluid (acoustic medium) is indicated. It is the sound pressure (pressure field) in the fluid. For the Cauchy stress tensor of a solid, This represents the gradient of the fluid pressure field. It is the acceleration vector of the solid structure, that is, the second partial derivative of displacement with respect to time.

[0026] To prevent boundary wave reflections from affecting the calculation results of the internal region, a Perfectly Matched Layer (PML) is added to the finite element analysis simulation software as the outermost layer of the model; the specific material properties are shown in Table 1. In this model, the thickness of the PML layer is designed according to the wavelength of the sound wave corresponding to the simulation center frequency, and is set to be no less than 1 / 4 of that wavelength; in the finite element analysis software, the PML domain adopts a structured swept mesh, with 5 to 8 layers of elements uniformly set along the thickness direction of the PML, in conjunction with the built-in polynomial coordinate stretching and scaling function, to ensure the non-reflection absorption effect of incident sound waves in the target frequency band; the specific material properties are shown in Table 1. Table 1 Acoustic and mechanical parameters of model materials In acoustic simulation, Propagation Layer (PML) is used to simulate the propagation of sound waves in the air, preventing sound waves from reflecting at the boundaries of the computational domain and affecting the sound field distribution in the inner region. The basic principle of PML is to introduce a special absorption region near the boundary of the simulation area. The absorption characteristics of this region are related to the direction and frequency of wave propagation, allowing waves to be rapidly attenuated after entering the PML layer, thus preventing reflected waves from returning to the computational domain. The PML layer can effectively absorb waves propagating to the boundary of the simulation domain, preventing boundary reflections from affecting the calculation results of the inner region and making the simulation closer to a real open environment. On the other hand, in automatic impact detection, to prevent cable terminations from shifting during impact and affecting the detection results, the simulation sets fixed boundary conditions at both ends of the cable. For example... Figure 4 As shown. Fixed constraints completely constrain the geometric entity, meaning that displacements in all directions on the selected geometric entity are zero, and rotational degrees of freedom, if any, will also be zero. This also makes the simulation closer to a real open environment.

[0027] S2, set the water volume variable in the preset water inlet area inside the three-dimensional geometric model, and set up the excitation sound source and the virtual monitoring point located in the air medium entity at the specified spatial coordinate position. In the geometric model, different inflow rates (water volume variables) are set, and different inflow rates are simulated by changing the size or material properties of the preset inflow area. Figure 2 As shown; excitation sound sources are set up at specified locations on the model, and virtual monitoring points (virtual sensors) are set up at predetermined locations. The water inflow rate mentioned in S2 is set by volume, including 0ml, 10ml, 20ml, 30ml, 40ml, and 50ml; the excitation sound source is selected as a Gaussian pulse, and the geometric position of the excitation sound source is as follows. Figure 6 The monitoring location is 10cm away from the impact point within the air domain of the model, and is on the same horizontal plane as the impact point.

[0028] The mathematical expression for the Gaussian pulse is: In the formula, A is the peak pulse amplitude, which is 1×10⁻⁶. 4 Pa; t0 is the pulse center time, set to 0.5 ms; σ is the pulse width parameter, with a value of 0.1 ms, corresponding to a center frequency of approximately 1.6 kHz, to ensure that the sound wave has suitable penetration depth and interface resolution in the EPDM insulating rubber; the excitation sound source is applied to the geometric center position of the outer surface of the stainless steel flange of the cable joint, such as... Figure 1 As shown; the monitoring position is located in the air domain inside the model. The virtual monitoring point is set at a spatial coordinate with a radial distance of 10 cm from the center of the excitation sound source and an angle of 45° with the axis of the connector, so as to simulate the typical sound pickup position of the sensor close to the outer shell of the connector in actual impact testing.

[0029] S3. Based on the acoustic simulation boundary conditions, the set water volume variable and the acoustic-structure coupling boundary, perform multi-physics transient solution. For fluid media, solve the transient pressure acoustic equation containing damping attenuation terms. For solid media, solve the structural transient dynamic equation containing stress tensor divergence to obtain the sound pressure time domain signal at the virtual monitoring point. In finite element analysis software, the propagation of sound waves in fluid media is mainly analyzed through the transient pressure acoustic equation, the expressions of which are shown in (1) and (2).

[0030] In the formula, c is the density of the medium; c is the speed of sound; This is the total transient pressure, usually the sum of sound pressure levels; This is the dynamic pressure component, representing the sound pressure that changes over time; This is the background pressure or static pressure, which may be the reference pressure for a certain initial state; For the sound source term, it represents the intensity of the sound source or excitation; This is a damping term, which is usually related to the absorption or dissipation effects in the medium. This describes the rate of change of sound pressure over time. This term indicates that the change in sound pressure over time affects the propagation of sound waves; The term describes the sound pressure distribution in space; it includes the sound pressure gradient and a damping term, indicating that the sound wave attenuates due to damping during propagation.

[0031] In finite element analysis software, the propagation of sound waves in solid media is mainly analyzed through the structural transient dynamic equation, the expression of which is shown in (3).

[0032] In the formula The density of the medium; Let be the displacement vector, representing the position change of a particle in the material at time t. Let be the second time derivative of the displacement, and let represent the acceleration of the particle. The divergence of the stress tensor describes the distribution of internal stress in a material; The force density vector represents the external force acting on the material.

[0033] Solve the above equations to obtain the time-domain sound pressure signals under different water inflow rates.

[0034] S4, the sound pressure time-domain signal is truncated and its features are extracted, and initial acoustic samples with water inflow condition labels are generated by combining the corresponding water volume variables. Multiple sets of initial acoustic samples are aggregated to generate an acoustic print dataset for training the water inflow detection model.

[0035] The sound pressure time-domain signals acquired by S3 under different inflow rates were truncated in the time domain and normalized in amplitude to form the following: Figure 3 The initial acoustic samples are shown below; label information reflecting the influent flow conditions is added to each initial acoustic sample, as shown in Table 2 below: Table 2. Statistical analysis of water accumulation characteristics over time. The labeled acoustic samples are then divided into training, validation, and test sets according to a preset ratio and stored in a computer-readable medium. The training set is used to train the water inflow detection neural network model, the validation set is used to adjust the model hyperparameters, and the test set is used to evaluate the model accuracy; a table of temporal characteristic changes for different water inflow volumes is then generated.

[0036] The water ingress detection neural network model adopts a one-dimensional convolutional neural network (1D-CNN) architecture, specifically comprising: an input layer that receives a 4-dimensional acoustic feature vector consisting of peak amplitude, RMS value, decay rate, and rise time; three alternating stacked one-dimensional convolutional layers and max-pooling layers for extracting local temporal features and higher-order abstract features of the acoustic signal; two fully connected layers, with the first fully connected layer containing 128 neurons and the second fully connected layer containing 64 neurons, both using the ReLU activation function; and an output layer using a Softmax classifier to output the probability distribution of six water ingress levels (0ml, 10ml, 20ml, 30ml, 40ml, 50ml). During model training, a cross-entropy loss function and an Adam optimizer are used, with a learning rate set to 0.001, a batch size of 32, and 100 training iterations.

[0037] To verify the performance improvement effect of the simulation dataset generated by this method on the model, the following comparative experiment was conducted: The same 1D-CNN model was trained using only a small number of measured acoustic samples (20 sets for each type of water accumulation condition, totaling 120 sets), and the recognition accuracy of the model on the test set was 65.3%. After training the high-fidelity simulation dataset generated by this method (1000 sets for each type of water accumulation condition, totaling 6000 sets) with the above measured samples, the recognition accuracy of the same model on the same test set increased to 92.8%, verifying that the acoustic dataset generated by this invention can significantly enhance the accuracy and generalization ability of the water ingress detection model.

[0038] Example 2 In this embodiment, a system for constructing a water ingress acoustic signature dataset for cable joints based on acoustic simulation is proposed, including: The physics configuration module is used to acquire the three-dimensional geometric model of the cable joint and configure the acoustic physical parameters of the solid medium entity and the fluid medium entity in the three-dimensional geometric model respectively. The coupling and constraint configuration module is used to configure acoustic-structural coupling boundaries at the interface between solid and fluid media entities; and to set acoustic simulation boundary conditions, including configuring a perfectly matched layer at the outermost computational domain boundary of the three-dimensional geometric model and configuring fixed constraints at both ends of the cable entity in the three-dimensional geometric model. The variable and source deployment module is used to set the water volume variable in the preset water inlet area inside the three-dimensional geometric model, and to deploy the excitation sound source and the virtual monitoring point located in the air medium entity at the specified spatial coordinate position. The multiphysics solution module is used to perform multiphysics transient solutions based on the acoustic simulation boundary conditions, the set water volume variables, and the acoustic-structure coupling boundary. Specifically, for fluid media, it solves the transient pressure acoustic equation containing damping attenuation terms, and for solid media, it solves the structural transient dynamic equation containing stress tensor divergence, thereby obtaining the sound pressure time-domain signal at the virtual monitoring point. The dataset generation module is used to extract features from the sound pressure time-domain signal and generate initial acoustic samples with water inflow condition labels by combining them with the corresponding water volume variables. Multiple sets of the initial acoustic samples are aggregated to generate an acoustic print dataset for training the water inflow detection model.

[0039] Based on a similar inventive concept, this embodiment of the invention also provides a computer storage medium storing a readable program that, when run by a processor, can execute the above-described method for constructing a cable joint water ingress acoustic fingerprint dataset based on acoustic simulation.

[0040] Based on a similar inventive concept, this invention provides an electronic device, including: a processor, a memory, a communication interface, and a communication bus, wherein the processor, the memory, and the communication interface communicate with each other through the communication bus; The memory is used to store at least one executable instruction, which causes the processor to perform the operation corresponding to the above-described method for constructing a cable joint water ingress acoustic texture dataset based on acoustic simulation.

[0041] Based on a similar inventive concept, this invention also provides a computer program product, including computer instructions, which instruct a computing device to perform the operations corresponding to the above-described method for constructing a cable joint water inlet acoustic signature dataset based on acoustic simulation.

[0042] The methods of the present invention can be implemented in hardware, firmware, or as software or computer code that can be stored in a recording medium (such as a CD-ROM, RAM, floppy disk, hard disk, or magneto-optical disk), or as computer code originally stored on a remote recording medium or a non-transitory machine-readable medium and subsequently stored on a local recording medium, downloaded via a network. Thus, the methods described herein can be processed by software stored on a recording medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware (such as an ASIC or FPGA). It is understood that the computer, processor, microprocessor controller, or programmable hardware includes storage components (e.g., RAM, ROM, flash memory, etc.) capable of storing or receiving software or computer code that, when accessed and executed by the computer, processor, or hardware, implements the methods described herein. Furthermore, when a general-purpose computer accesses the code used to implement the methods shown herein, the execution of the code transforms the general-purpose computer into a dedicated computer for performing the methods shown herein.

[0043] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the claimed invention.

Claims

1. A method for constructing a water inlet acoustic signature dataset for cable joints based on acoustic simulation, characterized in that, Includes the following steps: Obtain the three-dimensional geometric model of the cable joint, and configure the acoustic physical parameters of the solid medium entity and the fluid medium entity in the three-dimensional geometric model respectively; An acoustic-structural coupling boundary is configured at the interface between the solid medium and the fluid medium. Setting acoustic simulation boundary conditions includes: configuring a perfect matching layer at the outermost computational domain boundary of the three-dimensional geometric model, and configuring fixed constraints at both ends of the cable entity of the three-dimensional geometric model; The water volume variable is set in the preset water inlet area inside the three-dimensional geometric model, and excitation sound sources and virtual monitoring points located in the air medium entity are set at the specified spatial coordinate positions. Based on the acoustic simulation boundary conditions, the set water volume variable, and the acoustic-structure coupling boundary, a multiphysics transient solution is performed. For fluid media, a transient pressure acoustic equation containing a damping attenuation term is solved, and for solid media, a structural transient dynamic equation containing stress tensor divergence is solved to obtain the sound pressure time-domain signal at the virtual monitoring point. The sound pressure time-domain signal is truncated and its features are extracted. Initial acoustic samples with water inflow condition labels are generated by combining them with the corresponding water volume variables. Multiple sets of initial acoustic samples are aggregated to generate an acoustic print dataset for training the water inflow detection model.

2. The method for constructing a cable joint water inlet acoustic signature dataset based on acoustic simulation according to claim 1, characterized in that, Configuring the acoustic physical parameters includes: Define the corresponding solid acoustic physical parameters for the EPDM insulating rubber, stainless steel, copper, and epoxy resin components in the three-dimensional geometric model; The corresponding fluid acoustic physical parameters are defined for the air domain and the water domain in the three-dimensional geometric model, respectively.

3. The method for constructing a cable joint water inlet acoustic signature dataset based on acoustic simulation according to claim 1, characterized in that, The process of configuring a perfectly matched layer at the outermost computational domain boundary of a three-dimensional geometric model includes: introducing a computational region with absorption characteristics near the boundary of the computational domain. The absorption characteristics of the computational region are related to the propagation direction and frequency of the sound wave, so that the wave propagating into the computational region is attenuated.

4. The method for constructing a cable joint water inlet acoustic signature dataset based on acoustic simulation according to claim 1, characterized in that, The process of configuring fixed constraints at both ends of the cable entity in the three-dimensional geometric model includes: applying full displacement constraints to the geometric entities at both ends of the cable entity, so that the translational displacement in all directions on the selected geometric entities is zero, and if there is rotational degree of freedom, the rotational degree of freedom in all directions is also zero.

5. The method for constructing a cable joint water inlet acoustic signature dataset based on acoustic simulation according to claim 1, characterized in that, The water volume variable includes a gradient sequence of multiple preset inflow volume nodes set according to volume; The methods for setting the water volume variable include: representing different water inflow volumes by changing the size or material properties of the medium in the preset water inflow area in the three-dimensional geometric model.

6. The method for constructing a cable joint water inlet acoustic signature dataset based on acoustic simulation according to claim 1, characterized in that, The expression for the transient pressure acoustic equation is: In the formula, c is the density of the medium; c is the speed of sound; Total transient pressure; For the sound source term, it represents the intensity of the sound source or excitation; For damping; It describes the rate of change of sound pressure over time; This is a term describing the distribution of sound pressure in space.

7. The method for constructing a cable joint water inlet acoustic signature dataset based on acoustic simulation according to claim 1, characterized in that, The expression for the transient dynamic equation of the structure is: in, The density of the medium; Let be the displacement vector, representing the position change of a particle in the material at time t. Let be the second time derivative of the displacement, and let represent the acceleration of the particle. Let be the divergence of the stress tensor. Body force density vector.

8. The method for constructing a cable joint water inlet acoustic signature dataset based on acoustic simulation according to claim 1, characterized in that, The process of truncating and extracting features from the sound pressure time-domain signal includes: The amplitude of the truncated sound pressure time-domain signal is normalized. Based on the normalized signal, peak amplitude features, root mean square (RMS) value features, decay rate features, and rise time features are extracted respectively, and the extracted features are combined into an acoustic feature vector.

9. A system for constructing a water inlet acoustic signature dataset for cable joints based on acoustic simulation, comprising the method described in any one of claims 1-8, characterized in that, include: The physics configuration module is used to acquire the three-dimensional geometric model of the cable joint and configure the acoustic physical parameters of the solid medium entity and the fluid medium entity in the three-dimensional geometric model respectively. The coupling and constraint configuration module is used to configure acoustic-structural coupling boundaries at the interface between solid and fluid media entities. Setting acoustic simulation boundary conditions includes: configuring a perfect matching layer at the outermost computational domain boundary of the three-dimensional geometric model, and configuring fixed constraints at both ends of the cable entity of the three-dimensional geometric model; The variable and source deployment module is used to set the water volume variable in the preset water inlet area inside the three-dimensional geometric model, and to deploy the excitation sound source and the virtual monitoring point located in the air medium entity at the specified spatial coordinate position. The multiphysics solution module is used to perform multiphysics transient solutions based on the acoustic simulation boundary conditions, the set water volume variables, and the acoustic-structure coupling boundary. Specifically, for fluid media, it solves the transient pressure acoustic equation containing damping attenuation terms, and for solid media, it solves the structural transient dynamic equation containing stress tensor divergence, thereby obtaining the sound pressure time-domain signal at the virtual monitoring point. The dataset generation module is used to extract features from the sound pressure time-domain signal and generate initial acoustic samples with water inflow condition labels by combining them with the corresponding water volume variables. Multiple sets of the initial acoustic samples are aggregated to generate an acoustic print dataset for training the water inflow detection model.

10. An electronic device, characterized in that, include: The processor, memory, communication interface, and communication bus are provided, wherein the processor, memory, and communication interface communicate with each other via the communication bus. The memory is used to store at least one executable instruction, which causes the processor to perform the operation corresponding to the method for constructing a cable joint water ingress acoustic texture dataset based on acoustic simulation as described in any one of claims 1-8.

Images