Data processing method, device and equipment of high-voltage cable and storage medium

Abstract

The invention discloses a data processing method, device and equipment for a high-voltage cable and a storage medium, and is applied to the technical field of data processing, and the method comprises the steps: obtaining data, and constructing a thermal-electric-mechanical coupling topological graph; obtaining the physical coupling weight of the thermal-electric-mechanical coupling topological graph, and processing the physical coupling weight and the real-time monitoring data of the target high-voltage cable to obtain graph Laplacian deviation data; processing the electrical circulation data to obtain load characteristic data; performing vibration spectrum analysis on the mechanical topological data to obtain mechanical-structural feature data; obtaining a mutual information contribution degree based on a preset operation risk threshold value; and inputting the graph Laplacian deviation data and the mutual information contribution degree into the Bayesian generation type repair model for processing to obtain a data processing result. According to the method provided by the invention, the technical problem that a traditional single-point data processing method neglects physical topology constraints and cannot effectively distinguish real anomalies and noise can be solved.

Description

A data processing method, apparatus, equipment, and storage medium for high-voltage cables. Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a data processing method, apparatus, equipment, and storage medium for high-voltage cables. Background Technology

[0002] In high-voltage power transmission and distribution systems, cross-linked polyethylene insulated cables are key equipment, and their operating status directly affects the safety and stability of the power grid.

[0003] In existing technologies, high-voltage cable data relies on single-point data processing methods, which often confuse physical anomalies with missing data or interference. This masks early risk indicators, such as local overheating or structural deformation, thereby increasing the probability of sudden power grid failures. Consequently, this leads to increased power grid safety hazards, delayed fault warnings, and inaccurate maintenance decisions. Summary of the Invention

[0004] This invention provides a data processing method, apparatus, equipment, and storage medium for high-voltage cables to solve the technical problems of traditional single-point data processing methods that ignore physical topology constraints and cannot effectively distinguish between real anomalies and noise. It achieves high-precision anomaly identification, reliable repair, and feature-preserving compression based on multi-source data fusion and physical mechanism-driven methods, significantly improving the cable status perception capability and the reliability of operation and maintenance decisions.

[0005] To address the aforementioned technical problems, this invention provides a data processing method for high-voltage cables. The method includes: acquiring real-time monitoring data, electrical current data, and mechanical topology data of a target high-voltage cable; constructing a thermo-electric-mechanical coupled topology diagram based on the real-time monitoring data, the electrical current data, and the mechanical topology data; acquiring the physical coupling weights of the thermo-electric-mechanical coupled topology diagram; processing the physical coupling weights and the real-time monitoring data of the target high-voltage cable to obtain graph Laplace deviation data; processing the electrical current data to obtain load characteristic data; performing vibration spectrum analysis on the mechanical topology data to obtain mechanical-structural characteristic data; processing the load characteristic data and the mechanical-structural characteristic data based on a preset operational risk threshold for the target high-voltage cable to obtain the mutual information contribution of several data points in the target high-voltage cable; and inputting the graph Laplace deviation data and the mutual information contribution into a Bayesian generative repair model constructed from the thermo-electric-mechanical coupled topology diagram for processing to obtain the data processing result of the target high-voltage cable.

[0006] As one preferred embodiment, constructing a thermo-electric-mechanical coupled topology map based on the real-time monitoring data, the electrical current data, and the mechanical topology data includes: performing time alignment processing on the real-time monitoring data, the electrical current data, and the mechanical topology data to obtain preprocessed real-time monitoring data, electrical current data, and mechanical topology data; using the monitoring points of the target high-voltage cable as topology node data, performing spatial encoding processing on the topology node data to obtain node coordinate positions; determining topology edge data based on the node coordinate positions and the preprocessed electrical current data and mechanical topology data; and performing edge weight calibration processing on the topology edge data based on the preprocessed real-time monitoring data to obtain the thermo-electric-mechanical coupled topology map.

[0007] As one preferred embodiment, the step of obtaining the physical coupling weights of the thermo-electro-mechanical coupling topology graph and processing the physical coupling weights and the real-time monitoring data of the target high-voltage cable to obtain graph Laplace deviation data includes: processing the thermal resistance information in the obtained real-time monitoring data of the target high-voltage cable to obtain thermal coupling strength; processing the resistance information in the obtained electrical circulating current data of the target high-voltage cable to obtain electrical coupling strength; processing the mechanical topology data in the obtained electrical circulating current data of the target high-voltage cable to obtain mechanical coupling strength; obtaining the physical coupling weights of the thermo-electro-mechanical coupling topology graph based on the thermal coupling strength, the electrical coupling strength, and the mechanical coupling strength; processing the real-time monitoring data of the target high-voltage cable according to the physical coupling weights to obtain neighborhood prediction values; and processing the neighborhood prediction values ​​and the real-time monitoring data to obtain the graph Laplace deviation data.

[0008] As one preferred embodiment, the process of processing the electrical circulating current data to obtain load characteristic data includes: performing symmetrical component decomposition on the electrical circulating current data to obtain positive-sequence component energy data and negative-sequence component energy data; and processing the positive-sequence component energy data and the negative-sequence component energy data to obtain load characteristic data.

[0009] As one preferred embodiment, the step of performing vibration spectrum analysis on the mechanical topology data to obtain mechanical-structural characteristic data includes: acquiring settlement data and vibration data from the mechanical topology data; performing slope analysis on the settlement data to obtain step characteristic data; performing spectral energy analysis on the vibration data to obtain frequency band characteristic data; and obtaining the mechanical-structural characteristic data based on the step characteristic data and the frequency band characteristic data.

[0010] As one preferred embodiment, the step of inputting the mutual information contribution of the graph Laplace bias data and several data points of the target high-voltage cable into a Bayesian generative repair model constructed from the thermo-electro-mechanical coupled topology graph for processing to obtain the data processing result of the target high-voltage cable includes: constructing a Bayesian generative repair model based on the thermo-electro-mechanical coupled topology graph; performing conditional sample extraction processing on the mutual information contribution of the graph Laplace bias data and several data points of the target high-voltage cable to obtain a clean sample dataset; and inputting the clean sample dataset into the Bayesian generative repair model for processing to obtain the data processing result of the target high-voltage cable.

[0011] As one preferred embodiment, after obtaining the data processing results, the data processing method for high-voltage cables further includes: performing quality label generation processing on the data processing results to obtain a quality label matrix; and processing the quality label matrix using a joint cost function minimization technique to obtain data cleaning results.

[0012] The present invention also provides a data processing device for high-voltage cables, comprising: an acquisition module for acquiring real-time monitoring data of the target high-voltage cable, electrical circulating current data of the target sheath grounding circuit, and mechanical topology data of the target cable support; a construction module for constructing a thermo-electric-mechanical coupling topology diagram based on the real-time monitoring data, the electrical circulating current data, and the mechanical topology data; a first processing module for acquiring the physical coupling weights of the thermo-electric-mechanical coupling topology diagram, processing the physical coupling weights and the real-time monitoring data of the target high-voltage cable to obtain graph Laplace deviation data; and a second processing module for processing the electrical... The circulating data is processed to obtain load characteristic data; the mechanical topology data is subjected to vibration spectrum analysis to obtain mechanical-structural characteristic data; the third processing module is used to process the load characteristic data and the mechanical-structural characteristic data based on the preset operating risk threshold of the target high-voltage cable to obtain the mutual information contribution of several data in the target high-voltage cable; the fourth processing module is used to input the graph Laplace deviation data and the mutual information contribution of several data of the target high-voltage cable into the Bayesian generative repair model constructed by the thermo-electric-mechanical coupled topology graph for processing to obtain the data processing result of the target high-voltage cable.

[0013] The present invention also provides a data processing device for high-voltage cables, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor executes the computer program to implement the data processing method for high-voltage cables as described above.

[0014] The present invention further provides a computer-readable storage medium storing a computer program, wherein when the device containing the computer-readable storage medium executes the computer program, it implements the data processing method for high-voltage cables as described above.

[0015] Compared to existing technologies, the beneficial effects of this invention are at least one of the following: This invention acquires real-time monitoring data, electrical current data, and mechanical topology data of a target high-voltage cable; constructs a thermo-electric-mechanical coupled topology diagram based on the real-time monitoring data, the electrical current data, and the mechanical topology data; acquires the physical coupling weights of the thermo-electric-mechanical coupled topology diagram; processes the physical coupling weights and the real-time monitoring data of the target high-voltage cable to obtain graph Laplace deviation data; processes the electrical current data to obtain load characteristic data; performs vibration spectrum analysis on the mechanical topology data to obtain mechanical-structural characteristic data; processes the load characteristic data and the mechanical-structural characteristic data based on a preset operational risk threshold of the target high-voltage cable to obtain the mutual information contribution of several data in the target high-voltage cable; inputs the graph Laplace deviation data and the mutual information contribution into a Bayesian generative repair model constructed from the thermo-electric-mechanical coupled topology diagram for processing to obtain the data processing results of the target high-voltage cable.

[0016] Compared with existing technologies, this invention addresses the shortcomings of traditional single-point data processing, which neglects physical topology constraints and easily confuses real anomalies with noise. By collecting real-time monitoring data, electrical circulation data, and mechanical topology data of high-voltage cables, a thermo-electrical-mechanical coupled topology diagram reflecting the physical laws of cable operation is constructed. The physical coupling weight of this topology diagram is calculated and combined with real-time monitoring data to obtain graph Laplace deviation data. Simultaneously, load characteristic data and mechanical-structural characteristic data are extracted from the electrical circulation data and mechanical topology data, respectively. Then, the mutual information contribution of the two types of characteristic data is quantified based on a preset operational risk threshold. Finally, the graph Laplace deviation data and mutual information contribution are input into a Bayesian generative repair model constructed based on the coupled topology diagram for processing. Relying on the multi-source data fusion and probabilistic reasoning capabilities driven by physical mechanisms, this invention achieves accurate differentiation between real anomalies and noise, effective mining of early risk indicators, reliable repair of missing data, and feature-preserving compression, ultimately significantly improving the state perception capability and reliability of operation and maintenance decisions for high-voltage cables. Attached Figure Description

[0017] Figure 1 is a flowchart illustrating a data processing method for high-voltage cables in one embodiment of the present invention; Figure 2 is a structural schematic diagram illustrating a data processing device for high-voltage cables in one embodiment of the present invention; Figure 3 is a structural schematic diagram illustrating a data processing equipment for high-voltage cables in one embodiment of the present invention; Reference numerals: 11, acquisition module; 12, construction module; 13, first processing module; 14, second processing module; 15, third processing module; 16, fourth processing module; 21, processor; 22, memory. Detailed Implementation

[0018] 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. The purpose of providing these embodiments is to make the disclosure of the present invention more thorough and comprehensive. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0019] In the description of this invention, the terms "first," "second," "third," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined with "first," "second," "third," etc., may explicitly or implicitly include one or more of that feature. In the description of this invention, unless otherwise stated, "a plurality of" means two or more.

[0020] In the description of this invention, it should be noted that, unless otherwise defined, all technical and scientific terms used in this invention have the same meaning as commonly understood by one of ordinary skill in the art. The terminology used in this specification is for the purpose of describing specific embodiments only and is not intended to limit the invention. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0021] Traditional single-point data processing relies on monitoring data from only one dimension, which cannot reflect the full physical picture of cable operation. It is easy to misjudge missing measurements or interference as physical anomalies, or to cover up the real early risks.

[0022] To address this, one embodiment of the present invention provides a data processing method for high-voltage cables. Specifically, please refer to Figure 1, which shows a flowchart of the data processing method for high-voltage cables in one embodiment of the present invention. The method includes: S1: acquiring real-time monitoring data, electrical current data, and mechanical topology data of the target high-voltage cable; S2: constructing a thermo-electrical-mechanical coupled topology diagram based on the real-time monitoring data, the electrical current data, and the mechanical topology data; S3: acquiring the physical coupling weights of the thermo-electrical-mechanical coupled topology diagram, and processing the physical coupling weights and the real-time monitoring data of the target high-voltage cable to obtain the figure. S4: Process the electrical circulating current data to obtain load characteristic data; perform vibration spectrum analysis on the mechanical topology data to obtain mechanical-structural characteristic data; S5: Based on the preset operating risk threshold of the target high-voltage cable, process the load characteristic data and the mechanical-structural characteristic data to obtain the mutual information contribution of several data in the target high-voltage cable; S6: Input the graph Laplace deviation data and the mutual information contribution into the Bayesian generative repair model constructed by the thermo-electric-mechanical coupled topology graph for processing to obtain the data processing result of the target high-voltage cable.

[0023] Real-time data of the target high-voltage cable is acquired, including at least: distributed fiber optic temperature measurement data to characterize the temperature distribution along the cable body; sheath grounding circulation current data to characterize the current characteristics of each phase or loop in the sheath grounding loop; partial discharge monitoring data, including discharge pulse amplitude, discharge pulse occurrence phase, and PRPD (phase distribution) spectrum; settlement and displacement monitoring data to characterize the deformation of cable trenches, manholes, or support structures; and vibration monitoring data to characterize the vibration amplitude and spectral energy distribution of support components or the cable body. These monitoring points cover the condition of both the cable body and the cable support structure.

[0024] Based on the real-time monitoring data, the electrical current data, and the mechanical topology data, a thermo-electric-mechanical coupled topology graph is constructed, including: performing time alignment processing on the real-time monitoring data, the electrical current data, and the mechanical topology data to obtain preprocessed real-time monitoring data, electrical current data, and mechanical topology data; using the monitoring points of the target high-voltage cable as topology node data, performing spatial encoding processing on the topology node data to obtain node coordinate positions; determining topology edge data based on the node coordinate positions and the preprocessed electrical current data and mechanical topology data; and performing edge weight calibration processing on the topology edge data based on the preprocessed real-time monitoring data to obtain the thermo-electric-mechanical coupled topology graph.

[0025] By selecting a unified time base, the data from three different sources are matched and synchronized according to the timestamps to ensure that the real-time monitoring data, electrical circulation data and mechanical topology data under the same time dimension can correspond one-to-one. For data segments with mismatched timestamps, linear interpolation or neighboring value filling can be used to complete them, ensuring the continuity and consistency of the data sequence.

[0026] Discrete outliers that clearly do not conform to physical meaning, including saturation values, range overflows, and single-point jumps to extreme values, are marked but not directly deleted for the time being.

[0027] For signals that are significantly affected by high-frequency noise (such as grounding current and vibration signals), smoothing or envelope extraction can be performed within a shorter time window to suppress isolated electromagnetic interference or mechanical shock noise while maintaining trend characteristics.

[0028] The distribution of all monitoring points of the target high-voltage cable is obtained. Each monitoring point is treated as an independent node in the coupled topology graph. Using spatial coordinate system modeling, each node is assigned a corresponding spatial coordinate value based on the actual installation location of the monitoring point on the cable line.

[0029] For distributed monitoring points, one-dimensional or two-dimensional spatial coding can be performed according to the cable laying direction and length. For centralized monitoring points, coordinate parameters can be determined according to the relative positional relationship to ensure that the node coordinates can accurately reflect the spatial distribution characteristics of the monitoring points.

[0030] Based on the spatial coordinates of the nodes, the physical connection relationship between adjacent nodes is determined. For node pairs with direct spatial association, the electrical coupling strength between nodes is calculated based on the preprocessed electrical circulation data, and the mechanical association degree between nodes is calculated based on the mechanical topology data. The electrical coupling strength and mechanical association degree are used as edge attributes between node pairs. At the same time, the direction and connection type of the edges are marked, and finally, topology edge data containing node connection relationship and edge attribute information is formed.

[0031] Key state parameters, such as temperature, current, and vibration amplitude, are extracted from the preprocessed real-time monitoring data. These parameters are used as calibration criteria to correct and adjust the edge attributes of the constructed topology edge data. For node pairs whose real-time monitoring data shows abnormal states, their edge weights are appropriately increased to highlight the correlation strength. For node pairs whose states are stable, the edge weights are determined according to the benchmark values ​​to ensure that the edge weights can simultaneously reflect the multi-dimensional characteristics of electrical machinery and real-time monitoring.

[0032] After completing the closed-loop construction of the thermo-electric-mechanical coupled topology, the edge weights of the topology can be accurately matched to the actual operating state of the cable through calibration of real-time monitoring data. By integrating the coupling relationship between thermal, electrical and mechanical characteristics, the resulting coupled topology has physical mechanism support and can provide a reliable topology model for subsequent deviation analysis and anomaly identification.

[0033] The physical coupling weights of the thermo-electro-mechanical coupling topology are obtained. The physical coupling weights and real-time monitoring data of the target high-voltage cable are processed to obtain graph Laplace deviation data. This includes: processing the thermal resistance information in the real-time monitoring data of the target high-voltage cable to obtain thermal coupling strength; processing the resistance information in the electrical circulating current data of the target high-voltage cable to obtain electrical coupling strength; processing the mechanical topology data in the electrical circulating current data of the target high-voltage cable to obtain mechanical coupling strength; obtaining the physical coupling weights of the thermo-electro-mechanical coupling topology based on the thermal coupling strength, electrical coupling strength, and mechanical coupling strength; processing the real-time monitoring data of the target high-voltage cable according to the physical coupling weights to obtain neighborhood prediction values; and processing the neighborhood prediction values ​​and the real-time monitoring data to obtain the graph Laplace deviation data.

[0034] Thermal resistance-related parameters of each monitoring point are extracted from real-time monitoring data, including but not limited to temperature monitoring values, heat dissipation environment parameters and material thermal resistance coefficients of various parts of the cable. The heat transfer efficiency and influence between different monitoring points are calculated through the heat conduction equation, and the correlation between adjacent nodes due to heat conduction is quantified to obtain the thermal coupling strength value.

[0035] Specifically, based on information such as cable laying method, relative burial depth, and thermal resistance of surrounding soil or protective conduit, the equivalent thermal distance between node i and node j is calculated. And define thermal coupling strength : in: For nodes With nodes The equivalent thermal distance, which takes into account factors such as the distance between the two nodes along the cable line, the difference in burial depth, whether they are in the same pipe hole or trench section, and the equivalent thermal resistance of the surrounding medium, is used to reflect the ease or difficulty of heat transfer between the two nodes.

[0036] The thermal coupling reference coefficient is set according to engineering conditions such as cable voltage level, conductor cross-section, laying method and typical operating current level. It is used to characterize the theoretical maximum thermal coupling strength between two nodes when the equivalent thermal distance approaches zero.

[0037] The characteristic length is used to characterize the range of thermal diffusion and to describe how quickly the thermal effect decays with the equivalent thermal distance. The larger the value, the wider the range of thermal influence and the higher the thermal correlation between distant nodes.

[0038] The resistance parameters of the cable line are extracted from the electrical circulation data, including conductor resistance, contact resistance and equivalent resistance of the insulation layer. By combining Ohm's law and Kirchhoff's law, the correlation of current transmission and the degree of voltage drop between different monitoring points are calculated, the interaction between nodes caused by electrical characteristics is quantified, and the electrical coupling strength is obtained.

[0039] Specifically, based on the wiring method of the sheath grounding circuit, the cross-interconnection scheme, the positional relationship between the metal sheath and the grounding wire, and information such as soil resistivity, the nodes are calculated. With nodes Equivalent mutual impedance between the circuits And define the electrical coupling strength: in: For nodes With nodes The magnitude of the equivalent mutual impedance between the sheath circuit or related conductors; This is the reference coefficient for electrical coupling; This is an electrical coupling correction factor used to reflect the existence of direct electrical connections or strongly coupled return paths: when node With nodes When directly connected through the same sheath grounding wire or a clearly defined metallic conductor Set to 1; when there is only indirect electromagnetic coupling, shielding coupling, or a long-distance induction channel between the two nodes, Take a reduction factor between 0 and 1.

[0040] Based on the above definition, electrical coupling strength It can comprehensively reflect the combined influence of mutual impedance magnitude and wiring relationship on electrical coupling strength.

[0041] The mechanical structure parameters of the cable are extracted from the mechanical topology data, including the vibration amplitude, deformation degree and stress distribution data of each monitoring point. Combined with the material mechanics model, the correlation between different monitoring points caused by structural deformation or vibration transmission is calculated, the mechanical interaction relationship between nodes is quantified, and the mechanical coupling strength value is obtained.

[0042] Specifically, based on the arrangement, stiffness characteristics, span, connection method, and overall continuity of the manhole or pipe gallery structure of the supporting components, the nodes are calculated. With nodes Equivalent mechanical distance between And define the mechanical coupling edge weights: in: For nodes With nodes The equivalent mechanical distance is determined by combining the relative positions of the two nodes in the support system, whether they are located in the same support, the same support beam or the same well chamber structure, and the stiffness and continuity of the intermediate structure, etc., and is used to reflect the ease with which external loads or structural deformations are transferred from one node to another. The mechanical coupling reference coefficient is set according to the cable laying environment, the stiffness level of the support system, typical external loads, and historical deformation response. It is used to characterize the reference mechanical coupling strength between two nodes when the mechanical distance is small and the structural connection stiffness is large. The characteristic length is used to characterize the range of mechanical action attenuation. It describes how quickly structural deformation or load effects decay with equivalent mechanical distance. The larger the value, the more likely that nodes at greater distances within the same structural system may still have strong related deformation responses. The structural connection stiffness coefficient reflects the stiffness of the nodes. With nodes The dimensionless correction factor for the strength of the connection between structural units.

[0043] Based on the actual physical mechanism of cable thermal-electric-mechanical coupling, corresponding weight coefficients are assigned to thermal coupling strength, electrical coupling strength and mechanical coupling strength. The determination of the weight coefficients needs to take into account the priority of the influence of the three types of coupling characteristics on the cable operating state. Then, the comprehensive coupling strength of each topology edge is calculated by weighted summation. This comprehensive coupling strength is the physical coupling weight of the thermal-electric-mechanical coupling topology.

[0044] Specifically, the weights of the three types of coupling edges mentioned above are normalized globally in the graph, so that... , , All were scaled to The intervals are then weighted and superimposed according to preset weight coefficients to obtain node pairs. Comprehensive edge weight: in: For nodes With nodes The comprehensive coupling edge weights between them are used to uniformly represent the coupling strength of thermal, electrical, and mechanical multiphysics fields in the topology graph; These are the weighting coefficients for thermal coupling, electrical coupling, and mechanical coupling, respectively, satisfying... And optionally satisfy It is used to adjust the contribution ratio of different coupling channels in the comprehensive side weight according to the focus of the monitoring task.

[0045] A node neighborhood association model is constructed based on physical coupling weights. Taking a target node as the core, relevant nodes in its topological neighborhood are selected. Using the real-time monitoring data of the neighborhood nodes and the physical coupling weights, the theoretical value of the monitoring data that the target node should have under normal physical coupling law is calculated by weighted averaging or regression analysis. This theoretical value is the neighborhood prediction value.

[0046] Specifically, for each node i in the topological graph G, at time t, based on the set of nodes adjacent to that node... The effective measurement value, according to the physical coupling weight Calculate the neighborhood prediction value of node i: in, Let be the measured value of node j at time t.

[0047] Next, the difference between the real-time monitoring data of the target node and the predicted value of the neighborhood is calculated. Then, combined with the transformation rules of the graph Laplace matrix, the difference is embedded into the matrix operation of the thermo-electric-mechanical coupled topology graph to obtain a value that reflects the degree to which the node data deviates from the topological coupling law. This value is the graph Laplace deviation data.

[0048] It should be noted that three types of characteristic data are recorded for each node and each time point: deviation data; node status; and sampling continuity index.

[0049] After obtaining the feature data, within a given time window, extract the above three types of features from several spatially adjacent or physically tightly coupled nodes to form samples to be classified.

[0050] Topological constraints are introduced during classification, namely, nodes within the same category must form a connected subgraph on the topological graph G, and the residual distribution of nodes within that category should have a small variance.

[0051] Perform spectral clustering with topological connectivity constraints on the above samples. The goals of clustering are: to maintain spatial coherence of clusters on G and to ensure consistent residual variation within clusters.

[0052] Similarly, for a connected subgraph obtained by clustering, if the nodes corresponding to the subgraph experience significant data interruptions or no effective updates for a long time during the time period, and the residuals suddenly increase or change before and after the interruption, then the time-space segment is marked as a "gap segment".

[0053] For another type of connected subgraph obtained by clustering, if the nodes remain online and the data is continuous during the time period, but the residuals deviate significantly from their neighborhood predictions for a long time, it indicates that the region is an observable physical anomaly, and the time-space segment is marked as a "physical anomaly segment".

[0054] After removing the "missing segments", the remaining data is further subdivided and identified, including at least instantaneous impact anomalies, phase deviation anomalies, and mechanical-structural anomalies.

[0055] The electrical circulating current data is processed to obtain load characteristic data, including: performing symmetrical component decomposition on the electrical circulating current data to obtain positive sequence component energy data and negative sequence component energy data; and processing the positive sequence component energy data and the negative sequence component energy data to obtain load characteristic data.

[0056] Within the sliding time window, the sheath grounding circulating current or equivalent multiphase current is decomposed into symmetrical components, and the positive sequence component energy and negative sequence component energy are calculated.

[0057] Specifically, the symmetrical component method is used to decompose the electrical circulating current data of three-phase high-voltage cables into positive sequence components, negative sequence components, and zero sequence components. The positive sequence component corresponds to the balanced operating state where the three-phase current amplitudes are equal and the phases differ by 120 degrees. The negative sequence component corresponds to the unbalanced operating state where the three-phase current amplitudes are unequal or the phase deviations exceed the standard range. The zero sequence component is related to grounding faults.

[0058] After the decomposition is completed, the energy values ​​of the positive-sequence and negative-sequence components are calculated separately. The calculation method is to integrate the square of the component current within a certain time window to obtain the energy data of the positive-sequence and negative-sequence components.

[0059] Based on the normal operating load standard of high-voltage cables, a benchmark ratio range for positive-sequence energy and negative-sequence energy is set. The ratio of the actual calculated positive-sequence energy data and negative-sequence energy data is calculated to obtain the real-time component energy ratio. At the same time, the absolute value of the difference between the positive-sequence energy and negative-sequence energy is calculated. A multi-dimensional load feature vector is constructed by combining the ratio and the absolute value of the difference. This feature vector is the load feature data.

[0060] Vibration spectrum analysis is performed on the mechanical topology data to obtain mechanical-structural characteristic data, including: acquiring settlement data and vibration data from the mechanical topology data; performing slope analysis on the settlement data to obtain step characteristic data; performing spectral energy analysis on the vibration data to obtain frequency band characteristic data; and obtaining the mechanical-structural characteristic data based on the step characteristic data and the frequency band characteristic data.

[0061] Two core data categories directly related to the mechanical structure of the cable were selected from the mechanical topology dataset: settlement data, corresponding to the monitored settlement values ​​of the ground surface or supports along the cable laying path, including settlement displacement values ​​and settlement change time series at different monitoring points; and vibration data, corresponding to the monitored vibration acceleration or amplitude values ​​of the cable body and its ancillary facilities, including time-domain data of vibration signals in different frequency bands. During the process, invalid and redundant information in the dataset must be removed, and complete timestamps and spatial location markers for both types of data must be retained to ensure a one-to-one correspondence between the data and the cable monitoring points.

[0062] Linear fitting is performed on the time series of settlement data to calculate the slope of settlement displacement over time in different time intervals. A threshold range for slope change is set. When the slope exceeds the threshold for a certain time period, a settlement abrupt change is determined to have occurred in that time period. The duration of the abrupt change at the time point of the abrupt change and the amplitude of the slope change are quantified and integrated to form a parameter set that can characterize the settlement abrupt change. This parameter set is the step characteristic data. For stable settlement data with slopes within the threshold range, it is marked as normal structural state data.

[0063] A feature fusion model is constructed to align the step feature data and frequency band feature data in terms of dimensions. The alignment is based on the timestamp of the data and the spatial location of the monitoring point. The two types of data are then weighted and fused according to their respective impact weights on the mechanical structure risk. The weight allocation is based on historical statistical data of cable structure failures. For example, the impact weight of sudden settlement on structural damage is higher than that of general vibration anomalies. After fusion, a multi-dimensional feature vector containing the features of sudden settlement and vibration frequency band is formed. This feature vector is the mechanical-structural feature data.

[0064] Based on the design standards, operation and maintenance specifications and historical fault data of high-voltage cables, risk threshold ranges corresponding to load characteristics and mechanical structure characteristics are defined. For load characteristic data, a safe range for the positive-sequence and negative-sequence energy ratio and a load imbalance threshold need to be set. For mechanical structure characteristic data, a settlement step slope threshold and a characteristic frequency band vibration energy ratio threshold need to be set. At the same time, the risk levels corresponding to various characteristic indicators exceeding the thresholds are clearly defined, such as mild warning, moderate warning and severe warning.

[0065] For each monitoring feature (including at least the load feature data and the mechanical-structural feature data), calculate its task-related mutual information contribution to the aforementioned risk indicators.

[0066] Specifically, for each acquired monitoring feature, feature extraction and discretization are performed first.

[0067] For distributed fiber optic temperature measurement segments, the maximum value of its spatial temperature gradient is extracted as a random variable feature, and it is divided into discrete intervals. Based on historical operating data or statistical data within a sliding time window, the frequency of occurrence of each feature state is counted, and the marginal probability distribution of the monitoring segment features is constructed.

[0068] The preset operational risk indicators are designated as task variables, and each risk indicator is quantified into a discrete set of states.

[0069] The risk indicators are divided into four states: no risk, low risk, medium risk, and high risk, respectively. The prior probability of each risk state occurring within the current time window is calculated to construct the marginal probability distribution of the risk indicators.

[0070] Next, the co-occurrence frequency of risk indicators in specific states while monitoring segment features take specific values ​​is statistically analyzed, and their joint probability distribution is calculated. This joint probability reflects the pattern of simultaneous occurrence of specific data features and specific risk states. Using Shannon's information theory formula, the basic mutual information between monitoring features and risk indicators is calculated. This value characterizes the information gain in eliminating the uncertainty of risk indicators after observing the monitoring segment data.

[0071] Meanwhile, considering the differences in sensitivity of different monitoring features to different fault types, a task-related weight vector is introduced.

[0072] The weighted mutual information contribution is calculated using the following formula: in, The joint information entropy of all risk indicators is used to normalize the results to an interval. To monitor the basic mutual information between features and risk indicators, This is a task-related weight vector. The calculated mutual information contribution.

[0073] like A value close to 1 indicates that the monitoring segment contains a large amount of valid information about core risk indicators, and is considered key data that should be retained with high priority; if... If the value is below the preset threshold (set to 0.1), it means that even if the data segment is observed, it cannot significantly reduce the uncertainty of the system risk assessment (i.e. the data is not related to the current risk task of concern), and is therefore judged as a candidate for redundant data.

[0074] The graph Laplace bias data and the mutual information contribution are input into a Bayesian generative repair model constructed from the thermo-electro-mechanical coupled topology graph for processing to obtain the data processing result of the target high-voltage cable. This includes: constructing a Bayesian generative repair model based on the thermo-electro-mechanical coupled topology graph; performing conditional sample extraction processing on the mutual information contribution of the graph Laplace bias data and several data points of the target high-voltage cable to obtain a clean sample dataset; and inputting the clean sample dataset into the Bayesian generative repair model for processing to obtain the data processing result of the target high-voltage cable.

[0075] The spatial distribution of nodes, topological edge connections, and physical coupling weights of the thermo-electric-mechanical coupled topology graph are embedded into a Bayesian framework as prior knowledge for the model. Multi-source data features under normal operating conditions of high-voltage cables serve as the baseline distribution for the model, defining the generation and inference layers. The generation layer simulates the normal data distribution of the cable under different operating conditions based on the physical coupling rules of the topology graph. The inference layer establishes a probabilistic mapping relationship between data deviations and anomaly types based on Bayesian probability formulas. Simultaneously, a preset operational risk threshold is transformed into a probabilistic discrimination threshold for the model, setting posterior probability intervals corresponding to different risk levels, thus completing the construction of the Bayesian generative repair model.

[0076] Mutual information contribution is used as the weighting criterion for sample selection. A contribution threshold is set, and feature data with mutual information contribution higher than the threshold and their corresponding graph Laplace deviation data are selected. This type of data represents core samples that are strongly related to cable operation risks.

[0077] For the selected samples, based on the magnitude and direction of the Laplace bias data, we distinguish between bias samples originating from genuine physical anomalies and those originating from noise or missing measurements. We retain genuine anomaly samples and label their anomaly types, while initially filtering out noisy bias samples. Simultaneously, we normalize the sample data to eliminate dimensional differences, ultimately integrating them to form a clean sample dataset.

[0078] The clean sample dataset is input into the constructed Bayesian generative repair model. The model's inference layer calculates the posterior probability of the sample based on the input graph Laplace bias data and prior knowledge, and determines whether the sample data belongs to a real anomaly.

[0079] For samples deemed abnormal, the model's generation layer generates repair data that conforms to the normal operating state of the cable based on the physical coupling rules of the topology graph; for samples deemed normal, the model retains their original features.

[0080] Meanwhile, the model combines the weighted allocation of mutual information contribution to perform feature-preserving compression on the repaired data, removes redundant information and retains core risk features, and finally outputs the target high-voltage cable data processing result that integrates the anomaly identification result, the missing data repair result and the feature compression result.

[0081] Similarly, after obtaining the data processing results, physical topology constraints can also be repaired.

[0082] Select the segment that needs repair, such as "gap segment".

[0083] A Bayesian generative repair model with graph structure priors is established based on the topological graph G.

[0084] The distance definition of the covariance kernel in this model is based on the equivalent thermal resistance along the cable path, the impedance path of the sheath circuit, and the mechanical coupling stiffness of the support, bracket, or well chamber, reflecting the comprehensive degree of thermo-electric-mechanical coupling.

[0085] Using spatially and temporally adjacent valid measurement points that are not marked as anomalies as conditional samples, the posterior distribution of the target segment is calculated. The expected value of the posterior distribution is used as the repair value, and the variance or confidence interval of the posterior distribution is used as the repair uncertainty. The original invalid value of the missing segment or the instantaneous impact-type abnormal segment is replaced with the repair value, while retaining the repair uncertainty information.

[0086] For redundant data, the redundant data is compressed, and the compressed fragments are labeled with the reason for compression, the type of features to be retained, and the type of content to be discarded, so as to facilitate subsequent traceability.

[0087] For different monitoring data, only the minimum necessary features sufficient to support subsequent status assessment and risk judgment are retained. Specifically, for distributed fiber optic temperature measurement data, only representative sampling points that can reconstruct the temperature gradient changes of hot spots are retained; for sheath grounding circulation current data, only characteristic quantities that can characterize the evolution trend of negative sequence energy ratio are retained; for vibration data, only the amplitude envelope of the main abnormal resonance frequency bands are retained; for settlement and displacement data, only key inflection points and long-term trend slopes are retained; for partial discharge data, only typical cluster centers or characteristic templates that can represent the pulse phase distribution mode (PRPD mode) are retained.

[0088] After obtaining the data processing results, the data processing method for high-voltage cables further includes: performing quality label generation processing on the data processing results to obtain a quality label matrix; and processing the quality label matrix using a joint cost function minimization technique to obtain data cleaning results.

[0089] First, define a quality labeling system for high-voltage cable data. The label types should cover the core quality dimensions of the data, including anomaly labels, repair labels, normal labels, and compression labels. Anomaly labels are used to mark real physical anomalies identified by the model. Repair labels are used to mark missing or interfering data that have been reliably repaired by the model. Normal labels are used to mark anomaly-free data that conforms to the thermo-electric-mechanical coupling law. Compression labels are used to mark data that has undergone feature-preserving compression.

[0090] Then, according to the category of each monitoring point data in the data processing results, a corresponding quality label is matched for each data sample. The label information is then arranged in a matrix with the spatial location of the monitoring point as the row dimension and the time series as the column dimension, forming a quality label matrix consistent with the original data dimensions.

[0091] The value of each element in the matrix corresponds to the quality label type of the data sample at the corresponding location. At the same time, a confidence parameter can be embedded in the label. The confidence parameter is determined by the posterior probability output by the Bayesian generative repair model and is used to characterize the reliability of the label judgment result.

[0092] A joint cost function suitable for high-voltage cable data is constructed. The function needs to integrate three core cost terms: the cost of misjudging abnormal data, the cost of correcting data deviation, and the cost of losing useful features. The cost of misjudging abnormal data measures the degree of loss caused by marking normal data as abnormal data or vice versa. The cost of correcting data deviation measures the degree of deviation between the corrected data and the actual operating state of the cable. The cost of losing useful features measures the degree of loss of core risk features during data compression.

[0093] Then, the quality label matrix is ​​input into the joint cost function. With minimizing the cost function value as the optimization objective, the label matrix is ​​iteratively optimized using optimization algorithms such as gradient descent. During the optimization process, the physical constraints of the thermo-electric-mechanical coupling topology graph need to be considered. Labels with low confidence are re-judged, data with excessively high repair bias are repaired a second time, and compressed data with excessive feature loss are decompressed in reverse.

[0094] When the cost function converges during iteration, the optimized quality label matrix is ​​output. Then, the data processing results are filtered and corrected based on the optimized label matrix. Mislabeled abnormal data is removed, valid repaired data is retained, and compressed data with lost features is restored. Finally, the cleaned high-voltage cable data is obtained.

[0095] The label includes at least: segment type; whether it is a candidate for a redundant segment; whether it has been repaired; the confidence interval corresponding to the repair value; and the mutual information contribution.

[0096] Another embodiment of the present invention provides a data processing device for high-voltage cables. Specifically, please refer to Figure 2, which shows a schematic diagram of the structure of the data processing device for high-voltage cables in one embodiment of the present invention. The device includes: an acquisition module 11, used to acquire real-time monitoring data of the target high-voltage cable, electrical circulating current data of the target sheath grounding circuit, and mechanical topology data of the target cable support; a construction module 12, used to construct a thermo-electric-mechanical coupling topology diagram based on the real-time monitoring data, the electrical circulating current data, and the mechanical topology data; and a first processing module 13, used to acquire the physical coupling weights of the thermo-electric-mechanical coupling topology diagram and process the physical coupling weights and the real-time monitoring data of the target high-voltage cable. The system obtains the Graph Laplace deviation data; the second processing module 14 processes the electrical circulating current data to obtain load characteristic data; and performs vibration spectrum analysis on the mechanical topology data to obtain mechanical-structural characteristic data; the third processing module 15 processes the load characteristic data and the mechanical-structural characteristic data based on a preset operating risk threshold of the target high-voltage cable to obtain the mutual information contribution of several data in the target high-voltage cable; and the fourth processing module 16 inputs the Graph Laplace deviation data and the mutual information contribution of several data of the target high-voltage cable into a Bayesian generative repair model constructed from the thermo-electric-mechanical coupled topology graph for processing to obtain the data processing result of the target high-voltage cable.

[0097] Referring to Figure 3, which is a schematic diagram of the structure of the high-voltage cable data processing device provided in the embodiment of the present invention, the high-voltage cable data processing device provided in the embodiment of the present invention includes a processor 21, a memory 22, and a computer program stored in the memory 22 and configured to be executed by the processor 21. When the processor 21 executes the computer program, it implements the steps as described in the above embodiments of the high-voltage cable data processing method, such as steps S1 to S6 in Figure 1; or, when the processor 21 executes the computer program, it implements the functions of each module in the above embodiments of the device, such as the acquisition module 11.

[0098] For example, the computer program can be divided into one or more modules, which are stored in the memory 22 and executed by the processor 21 to complete the present invention. The one or more modules can be a series of computer program instruction segments capable of performing specific functions, which describe the execution process of the computer program in the data processing device for the high-voltage cable. For example, the computer program can be divided into an acquisition module 11, a construction module 12, a first processing module 13, etc., with the specific functions of each module as follows: the acquisition module 11 is used to acquire real-time monitoring data of the target high-voltage cable, electrical circulation current data of the target sheath grounding loop, and mechanical topology data of the target cable support; the construction module 12 is used to construct a thermo-electric-mechanical coupling topology diagram based on the real-time monitoring data, the electrical circulation current data, and the mechanical topology data; the first processing module 13 is used to acquire the physical coupling weights of the thermo-electric-mechanical coupling topology diagram, process the physical coupling weights and the real-time monitoring data of the target high-voltage cable, and obtain graph Laplace deviation ... target high-voltage cable, process the physical coupling weights and the real-time monitoring data of the target high-voltage cable, and obtain graph Laplace deviation data; the first processing module 13 is used to acquire the physical coupling weights of the target high-voltage cable, and execute the physical coupling weights and the physical coupling weights of the target high-voltage cable, and obtain graph Laplace deviation data; the first processing module 12 is used to acquire the physical coupling weights of the target high-voltage cable, and execute the physical coupling weights and physical coupling current data of the target high-voltage cable, and execute the physical coupling weights and physical coupling current data of the target high-voltage cable, and execute the physical coupling weights and physical coupling The second processing module 14 is used to process the electrical circulating current data to obtain load characteristic data; and to perform vibration spectrum analysis on the mechanical topology data to obtain mechanical-structural characteristic data; the third processing module 15 is used to process the load characteristic data and the mechanical-structural characteristic data based on the preset operating risk threshold of the target high-voltage cable to obtain the mutual information contribution of several data in the target high-voltage cable; the fourth processing module 16 is used to input the graph Laplace deviation data and the mutual information contribution of several data of the target high-voltage cable into the Bayesian generative repair model constructed by the thermo-electric-mechanical coupled topology graph for processing to obtain the data processing result of the target high-voltage cable.

[0099] The data processing device for the high-voltage cable may include, but is not limited to, a processor 21 and a memory 22. Those skilled in the art will understand that the schematic diagram is merely an example of a data processing device for the high-voltage cable and does not constitute a limitation on the device. It may include more or fewer components than illustrated, or combine certain components, or use different components. For example, the data processing device for the high-voltage cable may also include input / output devices, network access devices, buses, etc.

[0100] The processor 21 can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor. The processor 21 is the control center of the high-voltage cable's data processing equipment, connecting various parts of the high-voltage cable's data processing equipment via various interfaces and lines.

[0101] The memory 22 can be used to store the computer programs and / or modules. The processor 21 implements various functions of the high-voltage cable data processing device by running or executing the computer programs and / or modules stored in the memory 22 and calling the data stored in the memory 22. The memory 22 may mainly include a program storage area and a data storage area. The program storage area may store the operating system, at least one application program required for a function (such as sound playback function, image playback function, etc.), etc.; the data storage area may store data created according to the use of the mobile phone (such as audio data, phonebook, etc.). In addition, the memory 22 may include high-speed random access memory, and may also include non-volatile memory, such as hard disk, memory, plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, at least one disk storage device, flash memory device, or other volatile solid-state storage device.

[0102] If the module integrated into the data processing equipment of the high-voltage cable 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, all or part of the processes in the methods of the above embodiments of the present invention can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying the computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc.

[0103] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc.

[0104] Accordingly, embodiments of the present invention provide a computer-readable storage medium, the computer-readable storage medium including a stored computer program, wherein, when the computer program is executed, it controls the device where the computer-readable storage medium is located to perform steps in the high-voltage cable data processing method of the above embodiments, such as steps S1 to S6 as shown in FIG1.

[0105] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention. Therefore, the scope of protection of this patent should be determined by the appended claims.

Claims

1. A data processing method for high-voltage cables, characterized in that, include: Acquire real-time monitoring data, electrical current data, and mechanical topology data of the target high-voltage cable; Based on the real-time monitoring data, the electrical circulation data, and the mechanical topology data, a thermo-electric-mechanical coupled topology diagram is constructed. The physical coupling weights of the thermo-electric-mechanical coupling topology are obtained. The physical coupling weights and the real-time monitoring data of the target high-voltage cable are processed to obtain the graph Laplace deviation data. The electrical circulating current data is processed to obtain the load characteristic data. Vibration spectrum analysis is performed on the mechanical topology data to obtain the mechanical-structural characteristic data. Based on the preset operating risk threshold of the target high-voltage cable, the load characteristic data and the mechanical-structural characteristic data are processed to obtain the mutual information contribution of several data in the target high-voltage cable. The graph Laplace bias data and the mutual information contribution are input into the Bayesian generative repair model constructed from the thermo-electro-mechanical coupled topology graph for processing to obtain the data processing results of the target high-voltage cable.

2. The data processing method for high-voltage cables as described in claim 1, characterized in that, The step of constructing a thermo-electric-mechanical coupled topology map based on the real-time monitoring data, the electrical circulation data, and the mechanical topology data includes: performing time alignment processing on the real-time monitoring data, the electrical circulation data, and the mechanical topology data to obtain preprocessed real-time monitoring data, electrical circulation data, and mechanical topology data; using the monitoring points of the target high-voltage cable as topology node data, performing spatial encoding processing on the topology node data to obtain node coordinate positions; determining topology edge data based on the node coordinate positions and the preprocessed electrical circulation data and mechanical topology data; and performing edge weight calibration processing on the topology edge data based on the preprocessed real-time monitoring data to obtain the thermo-electric-mechanical coupled topology map.

3. The data processing method for high-voltage cables as described in claim 1, characterized in that, The process of obtaining the physical coupling weights of the thermo-electro-mechanical coupling topology and processing the physical coupling weights and real-time monitoring data of the target high-voltage cable to obtain graph Laplace deviation data includes: processing the thermal resistance information in the real-time monitoring data of the target high-voltage cable to obtain thermal coupling strength; processing the resistance information in the electrical circulating current data of the target high-voltage cable to obtain electrical coupling strength; processing the mechanical topology data in the electrical circulating current data of the target high-voltage cable to obtain mechanical coupling strength; obtaining the physical coupling weights of the thermo-electro-mechanical coupling topology based on the thermal coupling strength, the electrical coupling strength, and the mechanical coupling strength; processing the real-time monitoring data of the target high-voltage cable according to the physical coupling weights to obtain neighborhood prediction values; and processing the neighborhood prediction values ​​and the real-time monitoring data to obtain graph Laplace deviation data.

4. The data processing method for high-voltage cables as described in claim 1, characterized in that, The process of processing the electrical circulating current data to obtain load characteristic data includes: performing symmetrical component decomposition on the electrical circulating current data to obtain positive-sequence component energy data and negative-sequence component energy data; and processing the positive-sequence component energy data and the negative-sequence component energy data to obtain load characteristic data.

5. The data processing method for high-voltage cables as described in claim 1, characterized in that, The step of performing vibration spectrum analysis on the mechanical topology data to obtain mechanical-structural characteristic data includes: acquiring settlement data and vibration data from the mechanical topology data; performing slope analysis on the settlement data to obtain step characteristic data; performing spectral energy analysis on the vibration data to obtain frequency band characteristic data; and obtaining the mechanical-structural characteristic data based on the step characteristic data and the frequency band characteristic data.

6. The data processing method for high-voltage cables as described in claim 1, characterized in that, The process of inputting the mutual information contribution of the graph Laplace bias data and several data points of the target high-voltage cable into a Bayesian generative repair model constructed from the thermo-electro-mechanical coupled topology graph to obtain the data processing result of the target high-voltage cable includes: constructing a Bayesian generative repair model based on the thermo-electro-mechanical coupled topology graph; performing conditional sample extraction processing on the mutual information contribution of the graph Laplace bias data and several data points of the target high-voltage cable to obtain a clean sample dataset; and inputting the clean sample dataset into the Bayesian generative repair model for processing to obtain the data processing result of the target high-voltage cable.

7. The data processing method for high-voltage cables as described in claim 1, characterized in that, After obtaining the data processing results, the data processing method for high-voltage cables further includes: performing quality label generation processing on the data processing results to obtain a quality label matrix; and processing the quality label matrix using a joint cost function minimization technique to obtain data cleaning results.

8. A data processing device for high-voltage cables, characterized in that, include: The acquisition module is used to acquire real-time monitoring data of the target high-voltage cable, electrical circulating current data of the target sheath grounding circuit, and mechanical topology data of the target cable support; The construction module is used to construct a thermo-electric-mechanical coupling topology diagram based on the real-time monitoring data, the electrical circulation data, and the mechanical topology data; The first processing module is used to obtain the physical coupling weight of the thermo-electro-mechanical coupling topology diagram, process the physical coupling weight and the real-time monitoring data of the target high-voltage cable to obtain graph Laplace deviation data. The second processing module is used to process the electrical circulating current data to obtain load characteristic data; and to perform vibration spectrum analysis on the mechanical topology data to obtain mechanical-structural characteristic data. The third processing module is used to process the load characteristic data and the mechanical-structural characteristic data based on the preset operating risk threshold of the target high-voltage cable, and obtain the mutual information contribution of several data in the target high-voltage cable. The fourth processing module is used to input the mutual information contribution of the graph Laplace deviation data and several data of the target high-voltage cable into the Bayesian generative repair model constructed by the thermo-electro-mechanical coupled topology graph for processing, so as to obtain the data processing result of the target high-voltage cable.

9. A data processing device for high-voltage cables, characterized in that, It includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor, when executing the computer program, implements the data processing method for a high-voltage cable as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, wherein when the device containing the computer-readable storage medium executes the computer program, it implements the data processing method for high-voltage cables as described in any one of claims 1 to 7.

Images