A cable elbow terminal head water entry acoustic print detection method and system based on digital-analog cooperative driving

Abstract

The application discloses a cable elbow terminal head water inlet soundprint detection method and system based on digital-analog cooperative driving, relates to the technical field of cable state monitoring, and comprises the following steps: constructing an acoustic propagation numerical model of an elbow terminal head, and obtaining mode conversion parameters; based on standard sound excitation, the mode conversion parameters are inversely calibrated to obtain a propagation path consistent with a real structure; different water inlet conditions are simulated on the propagation path to generate a water inlet offset characteristic vector; the running state soundprint of the cable elbow terminal is obtained, and structured matching is performed based on the water inlet offset characteristic vector to obtain a water inlet coupling degree, and whether water inlet occurs is determined. The application reconstructs a real propagation path through a numerical model, simulates different water inlet conditions, and combines structured matching of a running soundprint to generate a water inlet coupling degree, so that the problem that a soundprint cannot be accurately identified in the prior art due to interface reflection and cavity mode interference is solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of cable condition monitoring, and more particularly, to a cable elbow terminal head water ingress acoustic signature detection method and system based on digital-analog collaborative driving. BACKGROUND

[0002] The power cable elbow terminal head is prone to internal water seepage due to aging of the seal, condensation or operating stress during long-term operation. Once partial water ingress occurs, it will change the local acoustic and electrical characteristics of the main insulation medium, eventually inducing partial discharge accumulation and surface flashover accidents. Existing online detection relies mainly on electrical parameter monitoring or simple acoustic signature comparison. However, the elbow terminal head has a "three-state coupled structure" composed of a semi-conductive shielding layer, an interface transition layer and a structural cavity. Acoustic waves in this structure will undergo multiple reflections, scattering and mode conversion. The water ingress acoustic signature features are structurally mixed with interface reflection signals and cavity inherent modes during transmission, making it difficult for traditional acoustic signature detection to separate the water ingress-related shift response from the mixed signals. Therefore, there is an urgent need for a digital-analog collaborative detection method that can demix the transmission path, explicitly depict the mode conversion behavior, and establish a correspondence between the numerical model and the actual acoustic signature, to realize the observability and online identification of the elbow terminal head water ingress state.

[0003] In the above disclosed technical solution, there are at least the following technical problems: the elbow terminal head has a three-state coupled structure of shielding-interface-cavity, which causes structural mixing of the acoustic signature transmission path after water ingress, making the real water ingress acoustic signature features obscured by interface reflection and cavity mode conversion, and traditional acoustic signature detection cannot directly observe the water ingress features. SUMMARY

[0004] To overcome the above-mentioned defects of the prior art, embodiments of the present application provide a cable elbow terminal head water ingress acoustic signature detection method and system based on digital-analog collaborative driving, which reconstructs the real transmission path through a numerical model and simulates different water ingress scenarios, and generates water ingress coupling degree by combining structured matching of the running acoustic signature, to solve the problem of inaccurate water ingress identification due to acoustic signature interference by interface reflection and cavity mode in the prior art.

[0005] To achieve the above-mentioned purpose, the present application provides the following technical solutions:

[0006] On the one hand, a cable elbow terminal head water ingress acoustic signature detection method based on digital-analog collaborative driving includes the following steps: constructing an acoustic propagation numerical model of the elbow terminal head and obtaining mode conversion parameters;

[0007] The mode conversion parameters are inversely calibrated based on a standard acoustic excitation signal to obtain a propagation path consistent with the real structure; different water inflow conditions are simulated on the propagation path to generate a water inflow offset characteristic vector; an operating state acoustic fingerprint of the cable elbow terminal is obtained, and a water inflow coupling degree is obtained based on the water inflow offset characteristic vector to determine whether water inflows.

[0008] In a preferred embodiment, the acoustic propagation numerical model of the elbow terminal head is constructed, including: obtaining a three-dimensional geometric model and a material list of the elbow terminal head, performing local encrypted grid division at geometric mutations and interface regions and assigning medium parameters to construct a spatial medium parameter field; setting boundary conditions including an adjustable adhesion degree boundary between the shielding-insulating layers; solving acoustic control equations based on the medium parameter field and the boundary conditions to obtain propagation modes; extracting mode conversion indicators at geometric mutations and interfaces to form an initial mode conversion parameter set; and packaging the medium parameter field, the propagation modes and the parameter set to construct the numerical model.

[0009] In a preferred embodiment, the local encrypted grid division at the geometric mutation and interface regions includes: constructing a geometric gradient field based on the three-dimensional geometric model to mark the joint turning and gap cavity regions; setting a regionalized shrinkage coefficient according to the gradient field to mark the areas to be refined; generating local refined grids in the areas to be refined and generating an extended gradual transition unit; and splicing the refined grids and the transition unit into a complete spatial discrete unit grid.

[0010] In a preferred embodiment, the mode conversion parameter includes: obtaining a basic propagation mode set based on the acoustic propagation numerical model simulation; performing local analysis on the basic propagation mode at the geometric mutation and material interface to extract energy distribution and phase change characteristics to form a local mode response set; mapping the local mode response set and the interface attribute to generate a preliminary parameter subset; performing gradient perturbation simulation on the preliminary parameter subset to generate a parameter sensitivity matrix; screening significant parameter items based on the sensitivity matrix and performing quantitative fitting to form the mode conversion parameter set.

[0011] In a preferred embodiment, the mode conversion parameters are inversely calibrated based on a standard acoustic excitation signal to obtain a propagation path consistent with the real structure, including: constructing an observability matrix and generating a reversible partial derivative operator based on the difference between the model predicted and measured propagation response vectors; updating the mode conversion parameters based on the local error weighting, and recalculating the propagation response in the updated model;

[0012] The global consistency reconstruction is performed based on the residual error of the propagation responses before and after the update to obtain a propagation path consistent with the real structure.

[0013] In a preferred embodiment, the reversible partial derivative operator is used to establish a quantitative mapping relationship between the propagation response difference and the local parameter perturbation.

[0014] In a preferred embodiment, the generating reversible partial derivative operator further comprises: obtaining equivalent sensitivity coefficients of each grid cell by passing the model prediction and the measured propagation response difference through the reversible partial derivative operator; generating a perturbed propagation response vector by inversely perturbing the mode conversion parameters of the local grid cell according to the equivalent sensitivity coefficients; obtaining the propagation difference contribution value of each grid cell by differentiating the perturbed propagation response vector and the reference propagation response vector; and establishing an error backtracking mapping based on the contribution value and the equivalent sensitivity coefficients, and obtaining the local error source distribution by weight normalization.

[0015] In a preferred embodiment, the simulating different water inflow conditions on the propagation path to generate a water inflow offset feature vector comprises: based on the calibrated propagation path, setting equivalent water-containing areas of different thicknesses and positions at the interface layer, the insulating layer and the geometric mutation of the elbow terminal head to construct multiple water inflow working condition models; applying a standard acoustic excitation to the water inflow working condition models and solving to obtain corresponding water inflow propagation response vectors; performing difference operation on the water inflow propagation response vectors and the water-free reference propagation response vector to extract low-frequency damping change components caused by the interface water film and high-frequency mode migration components caused by liquid redistribution; and performing normalized feature coding on the low-frequency damping change components and the high-frequency mode migration components to form the water inflow offset feature vector.

[0016] In a preferred embodiment, the obtaining an operating state acoustic fingerprint of the cable elbow terminal and performing structured matching based on the water inflow offset feature vector to obtain a water inflow coupling degree comprises: performing time-frequency rearrangement processing based on the propagation path on the operating state acoustic fingerprint to obtain an acoustic fingerprint representation vector after path calibration; constructing a sparse constraint dictionary based on the propagation path and projecting the acoustic fingerprint representation vector into the dictionary space; constructing a sparse target function containing the acoustic fingerprint representation vector and the water inflow offset feature vector and solving to obtain an optimal sparse excitation vector; and calculating the water inflow coupling degree based on the correlation between the optimal sparse excitation vector and the water inflow offset feature vector.

[0017] In another aspect, a cable elbow terminal head water entry acoustic fingerprint detection system based on digital-analog collaborative driving comprises the following modules: an acoustic propagation modeling module for constructing an acoustic propagation numerical model of the elbow terminal head and obtaining mode conversion parameters; a propagation path reconstruction module for reverse calibration of the mode conversion parameters based on a standard acoustic excitation signal to obtain a propagation path consistent with the real structure; an offset feature generation module for simulating different water entry situations on the propagation path to generate a water entry offset feature vector; and a structured matching module for obtaining an operating state acoustic fingerprint of the cable elbow terminal head and performing structured matching based on the water entry offset feature vector to obtain a water entry coupling degree to determine whether water entry has occurred.

[0018] The technical effects and advantages of the cable elbow terminal head water entry acoustic fingerprint detection system based on digital-analog collaborative driving of the present application are as follows:

[0019] The present application establishes a one-to-one correspondence between the numerical acoustic model and the actual operating acoustic fingerprint through the propagation path, so that the acoustic fingerprint mode aliasing caused by the shield-interface-cavity three-state structure after water entry is explicitly decoupled; by calibrating the mode conversion parameters and simulating the offset features under different water entry amounts, the extraction and matching of the weak water entry response in the real acoustic fingerprint are realized, so that the early water entry state can be reliably determined. Compared with the traditional method which relies on empirical features, the present application can maintain detection stability in the elbow terminal head with complex structure and significant signal aliasing, and improve the accuracy and applicability of water entry identification. BRIEF DESCRIPTION OF DRAWINGS

[0020] Figure 1 FIG. 1 is a flowchart of the cable elbow terminal head water entry acoustic fingerprint detection method based on digital-analog collaborative driving of the present application;

[0021] Figure 2 FIG. 2 is a structural diagram of the cable elbow terminal head water entry acoustic fingerprint detection system based on digital-analog collaborative driving of the present application. DETAILED DESCRIPTION

[0022] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative labor fall within the scope of protection of the present application.

[0023] Embodiment 1, Figure 1 The cable elbow terminal head water entry acoustic fingerprint detection method based on digital-analog collaborative driving of the present application is given, comprising the following steps:

[0024] S1, constructing an acoustic propagation numerical model of the elbow terminal head and obtaining mode conversion parameters;

[0025] In this embodiment, the acoustic propagation numerical model is used to describe the propagation law of sound waves in the cable elbow terminal head, which is a multi-medium, non-uniform, geometric mutation structure. By partitioning modeling the semi-conductive shielding layer, insulating layer, metal contact, potential cavity and possible thin water film in the terminal head, and assigning material parameters such as sound speed, density, impedance to each region, the model can calculate the propagation mode between different media, interface reflection and refraction, energy distribution, and mode conversion and phase disturbance caused by structural mutation. The final output of the numerical model is the propagation path, mode conversion parameter and acoustic fingerprint response that can represent the real internal structure characteristics of the elbow terminal head, providing a physical basis for subsequent reverse calibration, water entry simulation and acoustic fingerprint comparison.

[0026] The acoustic propagation numerical model of the elbow terminal head is constructed, specifically:

[0027] Obtain the three-dimensional geometric model and material list of the elbow terminal head, including the semi-conductive shielding layer, insulating layer, metal contact, possible cavity and joint position, and mark the potential water entry sensitive area;

[0028] Perform local encryption finite element meshing on the three-dimensional geometry, and preferentially refine at the geometric mutation (joint turning, gap, cavity boundary) to form a spatial discrete unit grid that meets the acoustic analysis accuracy requirements;

[0029] Assign medium parameters such as sound speed, density, acoustic impedance and internal friction factor on each grid element according to the material list to construct a spatially distributed medium parameter field; the parameter field is input as a coefficient for finite element solving, and is used in subsequent steps to describe the influence of different regions on wave speed and damping;

[0030] Apply continuity boundary conditions at the material interfaces of the meshed model, including acoustic pressure continuity, normal velocity continuity, and adjustable adhesion degree variation boundary conditions at the interface between the semi-conductive shielding layer and the insulating layer to simulate the interface adsorption / desorption effect; the adhesion degree parameter is a variable sub-item of the mode conversion parameter set, which will be used as the adjustment amount for reverse calibration in the subsequent step; the adhesion degree variation is realized by adjusting the equivalent acoustic impedance or boundary damping coefficient at the interface, which is used to represent the influence of the water content state of the interface on the reflection and attenuation characteristics of acoustic energy;

[0031] Based on the discretized medium parameter field, interface boundary conditions and external absorption boundary, establish the finite element acoustic control equation, including the wave equation form in time domain or frequency domain, and set appropriate absorption or radiation boundary conditions at the model boundary to avoid artificial reflection;

[0032] Solving the control equation under several representative frequencies or frequency bands, extracting the intrinsic mode of the model, the propagation velocity field and the mode energy distribution, obtaining the basic propagation mode set, including longitudinal wave, transverse wave, surface wave and local mode formed in the cavity, as the initial prediction set for subsequent calibration with standard acoustic excitation;

[0033] Quantifying the calculation mode conversion indicators at the identified geometric discontinuities and interfaces, including the energy distribution ratio, phase shift and transverse-longitudinal wave coupling amplitude at the discontinuities, and summarizing these indicators into an initial mode conversion parameter set, each sub-item will be used as an adjustable variable in the inverse calibration;

[0034] Parameterizing and packaging the medium parameter field, the basic propagation mode set and the mode conversion parameter set to form an acoustic propagation numerical model that can be input by external excitation.

[0035] The local encryption of the three-dimensional geometry is specifically:

[0036] Based on the obtained three-dimensional geometric model, a geometric gradient field is constructed to represent the joint turning, gap and cavity boundary change, and a regionalized shrinkage coefficient of grid size is set according to the gradient field, so that the geometric discontinuity region is marked as a region to be refined;

[0037] Based on the shrinkage coefficient, a local refined grid is generated in the region to be refined, and a gradual transition unit is generated in the extension range of the refined region, so that the local refinement can maintain the continuity and numerical stability of the grid in the overall structure;

[0038] The refined grid and the transition grid are spliced together to form a complete spatial discrete unit grid.

[0039] The mode conversion parameter is obtained, specifically:

[0040] Based on the constructed acoustic propagation numerical model, a standard acoustic excitation signal is input, and the sound pressure field, phase field and energy distribution of each key region in the model are simulated and solved to obtain the basic propagation mode set as the reference baseline for subsequent mode analysis;

[0041] Based on the basic propagation mode set, the sound field distribution at the geometric discontinuity region and the material interface is locally amplified and analyzed, the energy distribution ratio, phase shift and transverse-longitudinal wave coupling amplitude on each propagation path are extracted, and a local mode response set is formed to describe the propagation abnormal characteristics of the model in the structure sensitive region;

[0042] The local mode response set and the corresponding interface attribute (including adhesion degree change boundary, medium impedance difference and local cavity size) are one-to-one mapped, and a mode conversion quantitative relationship is constructed according to the propagation energy attenuation trend and the phase shift direction to obtain a preliminary parameter subset;

[0043] Based on the preset medium parameter field and interface boundary condition in the model, gradient perturbation simulation is performed on key parameters in the preliminary parameter subset which are sensitive to sound pressure phase, and a parameter sensitivity matrix S is generated by comparing the response differences before and after the perturbation, so as to identify the core parameters of the dominant mode conversion;

[0044] According to the sensitivity matrix S, the parameter items that have the greatest influence on energy trapping, interface reflection and mode coupling are screened out, and quantitative fitting is performed thereon to form a converged mode conversion parameter set which can completely represent the modulation law of the elbow terminal head structure on the soundprint propagation path.

[0045] The energy distribution ratio acquisition step: in the constructed acoustic propagation model, a uniform amplitude excitation signal is applied to each pre-defined propagation path, and the sound pressure or velocity response of the sound wave at the key section of the path is recorded; based on the energy response values obtained on different propagation branches, the relative energy proportion carried by each branch is calculated to reflect the energy distribution between different propagation paths, and the corresponding energy distribution ratio parameter is obtained.

[0046] The phase offset acquisition step: under the same excitation condition, the phase information of the sound wave at the starting point and the ending point of the path is extracted respectively, and the cumulative phase offset of the sound wave on the propagation path is obtained by comparing the phase difference between the two; the phase offset is used to represent the influence of structural interface and medium change on the time delay and phase distortion of the sound wave propagation.

[0047] The transverse and longitudinal wave coupling amplitude acquisition step: separate the longitudinal and transverse vibration components in the acoustic simulation results, and count the amplitude variation relationship of the two types of components at the same spatial position and in the same frequency band; according to the energy proportion or amplitude variation trend of the transverse wave component relative to the longitudinal wave component, the coupling strength between the transverse and longitudinal waves is determined as the coupling amplitude parameter describing the structural discontinuity or interface effect.

[0048] S2, reverse calibration of the mode conversion parameters based on the standard sound excitation signal to obtain a propagation path consistent with the real structure;

[0049] In this embodiment, the mode conversion parameters are calibrated based on the standard sound excitation signal to obtain a propagation path consistent with the real structure, specifically:

[0050] A standard sound excitation with controllable amplitude and frequency is applied to the three-dimensional finite element model to obtain the reference sound field distribution predicted by the model, and the reference sound field distribution is quantized as a first propagation response vector;

[0051] Based on the difference between the first propagation response vector and the second propagation response vector measured by the field sensor, an acoustic field observability matrix is ​​constructed in the local encrypted region, and an invertible partial derivative operator is generated to describe the acoustic energy scattering sensitivity at the geometric change point.

[0052] Based on the reversible partial derivative operator, the mode conversion parameters of each local grid cell are back-perturbed and traced back cell by cell to obtain the distribution of local error sources associated with the response differences;

[0053] Based on the local error source distribution, the mode conversion parameters are updated with weights, and the reference propagation response vector is recalculated in the updated three-dimensional finite element model.

[0054] Based on the residual between the reference propagation response vector before and after the update, the residual is projected into the inverse space of the invertible partial derivative operator to achieve global consistency reconstruction of the mode conversion path, so as to obtain the final propagation path consistent with the real structure.

[0055] The observability matrix is ​​specifically as follows:

[0056]

[0057] Differences between response vectors:

[0058] The invertible partial derivative operator is specifically:

[0059] in, To propagate the difference vector, This is the propagation response vector measured by the on-site sensor. The reference propagation response vector predicted by the model. The observability matrix, This is the mode conversion parameter vector for each current grid cell. for The regularized pseudoinverse, i.e., the invertible partial derivative operator. for transpose, This is a regularization factor (used for pseudo-inverse and stabilization, set based on historical experience). It is an identity matrix.

[0060] The method, based on the reversible partial derivative operator, performs a cell-by-cell reverse perturbation backtracking on the mode conversion parameters of each local grid cell to obtain the distribution of local error sources associated with the response differences, specifically as follows:

[0061] The difference between the first propagation response vector and the second propagation response vector is input into the invertible partial derivative operator to obtain the equivalent sensitivity coefficient of each local grid cell to the propagation difference;

[0062] based on the equivalent sensitivity coefficient, applying an amplitude-limited reverse perturbation to the mode conversion parameter of each local grid element, and generating a perturbation propagation response vector from the perturbed three-dimensional finite element model;

[0063] differencing the perturbation propagation response vector and the reference propagation response vector to obtain a contribution value of the grid element perturbation to the global propagation difference;

[0064] based on the contribution value and the equivalent sensitivity coefficient, establishing a per-element error backtracking mapping relationship , and performing weighted normalization processing on the local grid element to form a local error source distribution consistent with the propagation response difference space.

[0065] The equivalent sensitivity coefficient is specifically:

[0066]

[0067] The contribution value is specifically:

[0068]

[0069]

[0070] wherein, is the equivalent sensitivity coefficient, is the contribution value, is the difference of the response vector, is the perturbation propagation response vector.

[0071] The global consistency reconstruction of the mode conversion path is implemented to obtain a final propagation path consistent with the real structure, specifically:

[0072] differencing the propagation response vector of the updated model and the aforementioned reference propagation response vector to form a residual vector for representing the overall propagation deviation;

[0073] inputting the residual vector into the reversible partial derivative operator to obtain an equivalent parameter offset corresponding to the residual;

[0074] based on the equivalent parameter offset, performing global consistency balance adjustment on the local error source distribution according to the per-element error backtracking mapping relationship to generate a mode conversion parameter correction amount for global update;

[0075] superimposing the correction amount to the existing mode conversion parameter, and re-solving the acoustic propagation process in the corrected three-dimensional finite element model until the propagation response difference of adjacent two iterations is lower than a preset threshold, thereby obtaining a final propagation path consistent with the real structure.

[0076] S3, simulate different water ingress situations on the propagation path to generate a water ingress offset feature vector;

[0077] In this embodiment, the water ingress offset feature vector is used to describe the systematic offset rules caused by the water film, water droplets or local moisture in the elbow terminal head on the acoustic propagation path, phase change and energy distribution. The essence is to quantify the influence of different water ingress positions, different water contents and different water film thicknesses on the propagation path into a comparable multi-dimensional structured feature, which is used to reflect the changes of propagation delay, mode conversion ratio and local energy attenuation caused by water ingress.

[0078] The water ingress offset feature vector includes a low-frequency damping change vector caused by interface water content and a high-frequency mode migration vector caused by cavity liquid redistribution.

[0079] The water ingress offset feature vector is generated by simulating different water ingress situations on the propagation path, specifically:

[0080] Based on the final propagation path, equivalent water content areas with different thicknesses and different distribution positions are set at the interface layer, insulating layer and geometric mutation of the elbow terminal head to construct multiple water ingress working condition models;

[0081] The standard acoustic excitation is kept unchanged on each water ingress working condition model, and the acoustic propagation process is solved again to obtain the corresponding water ingress propagation response vector;

[0082] Difference operation is performed on each water ingress propagation response vector and the updated reference propagation response vector to extract low-frequency damping change components caused by interface water film and high-frequency mode migration components caused by liquid redistribution;

[0083] Based on the low-frequency damping change components and the high-frequency mode migration components, normalized feature coding is performed to form a water ingress offset feature vector for representing the differences between different water ingress situations.

[0084] S4, obtain the running state acoustic fingerprint of the cable elbow terminal, and perform structured matching based on the water ingress offset feature vector to obtain the water ingress coupling degree to determine whether there is water ingress.

[0085] In this embodiment, the running state acoustic fingerprint of the cable elbow terminal is obtained, and structured matching is performed based on the water ingress offset feature vector to obtain the water ingress coupling degree, specifically:

[0086] Perform time-frequency rearrangement processing on the running state acoustic fingerprint to obtain an acoustic fingerprint representation vector consistent with the real propagation path;

[0087] A sparse constraint dictionary is constructed based on the aforementioned propagation path, only retaining the reachable propagation mode, and the voiceprint representation vector is projected into the feasible space of the sparse constraint dictionary to suppress the non-physical components caused by the interface reflection and cavity mode;

[0088] The projected voiceprint vector is coupled with the water inlet offset feature vector according to different mode channels, and weight adaptive fusion is performed based on the fitting residual of each mode channel to obtain the final water inlet coupling degree representing the inlet level.

[0089] Based on the propagation path response matrix of the running state voiceprint and the water inlet offset feature vector, a sparse projection target function is constructed to obtain an optimal sparse excitation vector;

[0090] The water inlet coupling degree is calculated based on the correlation between the optimal sparse excitation vector and the water inlet offset feature vector.

[0091]

[0092]

[0093] wherein, is the propagation path response matrix of the running state voiceprint, is the optimal sparse excitation vector, is the actually collected running state voiceprint, is a preset sparse regularization coefficient, is the water inlet coupling degree, is the water inlet offset feature vector.

[0094] The determination of whether water inlet is specific to:

[0095] Based on the water inlet coupling degree, an amplitude threshold criterion is established, and the water inlet coupling degree is compared with a background coupling threshold value obtained by statistical analysis of normal running condition voiceprint samples;

[0096] When the water inlet coupling degree is greater than the background coupling threshold value and meets the confidence level, it is determined that there is water inlet.

[0097] The confidence level is specific to:

[0098]

[0099] wherein, is the confidence level, is the historical maximum health-related coupling degree, is a confidence threshold value (set according to historical data).

[0100] Embodiment 2, Figure 2 A cable elbow terminal head water inlet voiceprint detection system based on digital-analog cooperative driving is given, which includes the following modules:

[0101] an acoustic propagation modeling module configured to build an acoustic propagation numerical model of the elbow terminal head and obtain mode conversion parameters;

[0102] a propagation path reconstruction module configured to calibrate the mode conversion parameters based on a standard acoustic excitation signal to obtain a propagation path consistent with the real structure;

[0103] an offset feature generation module configured to simulate different water ingress scenarios on the propagation path to generate a water ingress offset feature vector;

[0104] a structured matching module configured to obtain an operating state acoustic fingerprint of the cable elbow terminal and perform structured matching based on the water ingress offset feature vector to obtain a water ingress coupling degree to determine whether water ingress occurs.

[0105] The above formulas are all dimensionless numerical calculations, and the formulas are obtained by software simulation of a large amount of data to obtain a formula closest to the real situation. The preset parameters in the formula are set by a person skilled in the art according to the actual situation.

[0106] The above embodiments can be realized wholly or partially by software, hardware, firmware or any combination thereof. When realized by software, the above embodiments can be realized wholly or partially in the form of a computer program product.

[0107] Those skilled in the art can appreciate that the modules and algorithm steps of the examples described in conjunction with the embodiments disclosed herein can be realized by electronic hardware or a combination of computer software and electronic hardware. Whether the functions are realized in hardware or software depends on the specific application and design constraints of the technical solutions. A person skilled in the art can use different methods to realize the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0108] In addition, the functional modules in each embodiment of the present application can be integrated in one processing module, or each module can exist physically alone, or two or more modules can be integrated in one module.

[0109] The above is merely specific embodiments of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art can easily think of changes or replacements within the technical scope disclosed in the present application, which should be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

[0110] Finally: the above only for the preferred embodiments of the present application, and not for limiting the present application, any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the present application, should be included in the scope of protection of the present application.

Claims

1. A cable elbow terminal head water entry acoustic print detection method based on digital-analog cooperative driving, characterized in that, The method comprises the following steps: An acoustic propagation numerical model of the elbow terminal head is constructed, and mode conversion parameters are obtained; The mode conversion parameters are inversely calibrated based on a standard acoustic excitation signal to obtain a propagation path consistent with the real structure; Different water inflow conditions are simulated on the propagation path to generate water inflow offset characteristic vectors; The running state acoustic fingerprint of the cable elbow terminal is obtained, and the water inflow coupling degree is obtained based on the water inflow offset characteristic vectors to determine whether water inflow occurs.

2. The method of claim 1, wherein the method is a digital-analog co-driving based cable elbow terminal head water entry acoustic print detection method. The method of constructing the acoustic propagation numerical model of the elbow terminal head comprises: A three-dimensional geometric model and a material list of the elbow terminal head are obtained, local encrypted grid division is performed at geometric mutations and interface regions, and medium parameters are assigned to construct a spatial medium parameter field; Boundary conditions including an adjustable adhesion degree boundary between the shielding and insulation layers are set; Based on the medium parameter field and the boundary conditions, the acoustic control equation is solved to obtain propagation modes; Mode conversion indicators are extracted at the geometric mutations and interfaces to form an initial mode conversion parameter set; The medium parameter field, the propagation modes and the parameter set are packaged to construct the numerical model.

3. The method of claim 2, wherein the method is based on a digital-analog co-driving cable elbow termination head water ingress acoustic fingerprint detection. The method of performing local encrypted grid division at the geometric mutations and interface regions comprises: A geometric gradient field is constructed based on the three-dimensional geometric model to mark the joint turning and gap cavity regions; A regionalized shrinkage coefficient is set according to the gradient field to mark the regions to be refined; Local refined grids are generated in the regions to be refined, and an extended gradual transition unit is generated; The refined grids and the transition unit are spliced into a complete spatial discrete unit grid.

4. The method of claim 3, wherein the method is based on a digital-analog co-driving cable elbow terminal head water entry acoustic signature detection. The method of obtaining the mode conversion parameters comprises: A basic propagation mode set is obtained by simulation based on the acoustic propagation numerical model; Local mode response sets are formed by extracting energy distribution and phase change characteristics by locally analyzing the basic propagation modes at the geometric mutation regions and material interfaces; Preliminary parameter subsets are generated by mapping the local mode response sets and the interface properties; Parameter sensitivity matrices are generated by gradient perturbation simulation of the preliminary parameter subsets; Significant parameter items are screened and quantitatively fitted based on the sensitivity matrices to form the mode conversion parameter set.

5. The method of claim 4, wherein the method is based on a digital-analog co-driving cable elbow termination head water ingress acoustic fingerprint detection. The method of inversely calibrating the mode conversion parameters based on the standard acoustic excitation signal to obtain a propagation path consistent with the real structure comprises: An observability matrix is constructed and a reversible partial derivative operator is generated based on the difference between the model predicted and measured propagation response vectors; The mode conversion parameters are updated by weighting based on local errors, and the propagation response is recalculated in the updated model; The propagation path consistent with the real structure is obtained by global consistency reconstruction based on the residual error of the propagation responses before and after the update.

6. The method of claim 5, wherein the method is a digital-analog co-driving based cable elbow terminal head water entry acoustic print detection method. The reversible partial derivative operator is used to establish a quantitative mapping relationship between the propagation response difference and the local parameter perturbation.

7. The method of claim 5, wherein the method is based on a digital-analog co-driving cable elbow termination head water ingress acoustic fingerprint detection. The method of generating the reversible partial derivative operator further comprises: Equivalent sensitivity coefficients of each grid unit are obtained by the reversible partial derivative operator based on the difference between the model predicted and measured propagation responses; The mode conversion parameters of the local grid units are inversely perturbed according to the equivalent sensitivity coefficients to generate perturbed propagation response vectors; Propagation difference contribution values of each grid unit are obtained by differentiating the perturbed propagation response vectors and the reference propagation response vectors. An error backtracking mapping is established based on the contribution value and the equivalent sensitivity coefficient, and a local error source distribution is obtained by weight normalization.

8. The method of claim 7, wherein the method is a digital-analog co-driving based cable elbow terminal head water entry acoustic signature detection method. The different water inflow situations are simulated on the propagation path to generate the water inflow offset feature vector, including: Based on the calibrated propagation path, equivalent water-containing areas with different thicknesses and positions are set at the interface layer, insulating layer and geometric mutation of the elbow terminal head to construct multiple water inflow working condition models; A standard acoustic excitation is applied to the water inflow working condition model and solved to obtain the corresponding water inflow propagation response vector; The water inflow propagation response vector and the water-free reference propagation response vector are subjected to difference operation to extract the low-frequency damping change component caused by the interface water film and the high-frequency mode migration component caused by liquid redistribution; The low-frequency damping change component and the high-frequency mode migration component are normalized and feature coded to form the water inflow offset feature vector.

9. The method of claim 8, wherein the method is based on a digital-analog co-driving cable elbow termination head water ingress acoustic fingerprint detection. The operating state acoustic print of the cable elbow terminal is obtained, and the water inflow coupling degree is obtained by structured matching based on the water inflow offset feature vector, including: Time-frequency rearrangement processing based on the propagation path is performed on the operating state acoustic print to obtain the path-calibrated acoustic print representation vector; A sparse constraint dictionary is constructed based on the propagation path, and the acoustic print representation vector is projected into the dictionary space; A sparse objective function containing the acoustic print representation vector and the water inflow offset feature vector is constructed and solved to obtain the optimal sparse excitation vector; The water inflow coupling degree is calculated based on the correlation between the optimal sparse excitation vector and the water inflow offset feature vector.

10. A system for detecting the water entry acoustic signature of a cable elbow termination using a digital-analog co-driven based cable elbow termination water entry acoustic signature detection method according to any one of claims 1-9, wherein, The following modules are included: Acoustic propagation modeling module: used to construct an acoustic propagation numerical model of the elbow terminal head and obtain mode conversion parameters; Propagation path reconstruction module: used to inversely calibrate the mode conversion parameters based on a standard acoustic excitation signal to obtain a propagation path consistent with the real structure; Offset feature generation module: used to simulate different water inflow situations on the propagation path to generate a water inflow offset feature vector; Structured matching module: used to obtain the operating state acoustic print of the cable elbow terminal, and perform structured matching based on the water inflow offset feature vector to obtain the water inflow coupling degree to determine whether there is water inflow.

Images