Fault diagnosis method and device, computer equipment, readable storage medium and program product

Abstract

The invention relates to a fault diagnosis method and device, computer equipment, a computer readable storage medium and a computer program product. The method comprises the following steps: obtaining detection data of each structural layer of the cable, wherein the detection data is obtained by detection in the production process of the structural layers; constructing a three-dimensional data model of the cable by fusing the detection data; after the cable is connected to a target power grid, obtaining operation monitoring data of the target power grid; and performing fault diagnosis on the target power grid in combination with the operation monitoring data and the three-dimensional data model to obtain a fault diagnosis result. By adopting the method, the fault diagnosis accuracy can be improved.

Description

Technical Field

[0001] This application relates to the field of power grid operation and maintenance technology, and in particular to a fault diagnosis method, device, computer equipment, computer-readable storage medium, and computer program product. Background Technology

[0002] In recent years, with the rapid increase in electricity load in large cities across the country and the increasing requirements of modern urban construction for land use and urban environment, power transmission cables have been widely used in urban power grids due to their advantages such as small footprint and good concealment.

[0003] With the continuous expansion of the cable network, power grid operation and maintenance face challenges related to the broad scope of supervision and the large volume of data. For example, in Shanghai, the length of 110kV and above transmission cables has exceeded 4,000 km, with an annual growth rate of approximately 12%, while the number of operation and maintenance personnel is only about 200. Traditional manual inspections and periodic testing methods are insufficient to meet actual needs. Therefore, online monitoring devices have been deployed extensively. By deploying various sensors, real-time data collection and preliminary anomaly warnings can be provided for cable operation, improving the timeliness and coverage of operation and maintenance to a certain extent.

[0004] However, current online monitoring technologies still have relatively low accuracy in fault diagnosis. In actual operation, the system frequently experiences false alarms, missed alarms, or situations where the cause of alarms cannot be clearly determined. This makes it difficult for maintenance personnel to accurately locate and analyze the root causes of faults, affecting the timeliness of fault handling and the scientific nature of maintenance decisions. Summary of the Invention

[0005] Therefore, it is necessary to provide a fault diagnosis method, apparatus, computer equipment, computer-readable storage medium, and computer program product that can improve the accuracy of fault diagnosis in response to the above-mentioned technical problems.

[0006] Firstly, this application provides a fault diagnosis method, including:

[0007] The test data of each structural layer of the cable is obtained during the production process of the structural layer; a three-dimensional data model of the cable is constructed by integrating the test data; after the cable is connected to the target power grid, the operation monitoring data of the target power grid is obtained; the fault diagnosis of the target power grid is performed by combining the operation monitoring data and the three-dimensional data model to obtain the fault diagnosis results.

[0008] Secondly, this application also provides a fault diagnosis device, comprising:

[0009] The first acquisition module is used to acquire the test data of each structural layer of the cable. The test data is obtained during the production process of the structural layers. The generation module is used to construct a three-dimensional data model of the cable by integrating the test data. The second acquisition module is used to acquire the operation monitoring data of the target power grid after the cable is connected to the target power grid. The fault diagnosis module is used to perform fault diagnosis on the target power grid by combining the operation monitoring data and the three-dimensional data model to obtain the fault diagnosis results.

[0010] Thirdly, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to perform the following steps:

[0011] The process involves acquiring test data for each structural layer of the cable, obtained during the production process of each layer; constructing a three-dimensional data model of the cable by fusing the test data; acquiring operational monitoring data of the target power grid after the cable is connected to the target power grid; and performing fault diagnosis on the target power grid by combining the operational monitoring data and the three-dimensional data model to obtain the fault diagnosis results.

[0012] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, performs the following steps:

[0013] The process involves acquiring test data for each structural layer of the cable, obtained during the production process of each layer; constructing a three-dimensional data model of the cable by fusing the test data; acquiring operational monitoring data of the target power grid after the cable is connected to the target power grid; and performing fault diagnosis on the target power grid by combining the operational monitoring data and the three-dimensional data model to obtain the fault diagnosis results.

[0014] Fifthly, this application also provides a computer program product, including a computer program that, when executed by a processor, performs the following steps:

[0015] The process involves acquiring test data for each structural layer of the cable, obtained during the production process of each layer; constructing a three-dimensional data model of the cable by fusing the test data; acquiring operational monitoring data of the target power grid after the cable is connected to the target power grid; and performing fault diagnosis on the target power grid by combining the operational monitoring data and the three-dimensional data model to obtain the fault diagnosis results.

[0016] The aforementioned fault diagnosis methods, devices, computer equipment, computer-readable storage media, and computer program products, by constructing a data fusion diagnostic system covering the entire cable lifecycle, can effectively solve the technical problem of low fault diagnosis accuracy. Specifically, even cable products that meet quality standards inevitably have differences in material properties and minor internal defects. These minor defects, along with the cable's inherent characteristics, can potentially interact with other defects or external operating conditions to cause faults. However, traditional online monitoring methods primarily rely on collected operational status data for fault diagnosis, making it difficult to trace whether the anomaly originates from cable inherent characteristics, manufacturing defects, on-site installation problems, or real-time operating conditions. This results in insufficient diagnostic evidence and low fault diagnosis accuracy.

[0017] Based on this, the present invention first acquires the inspection data obtained during the production process of each structural layer of the cable, and then constructs a three-dimensional data model that accurately reflects the characteristics of the cable itself and manufacturing defects by fusing these inspection data. After the cable is connected to the power grid and put into operation, real-time monitoring of the power grid's operating status is achieved by acquiring the operational monitoring data of the target power grid during actual operation. Subsequently, by jointly analyzing and comparing the operational monitoring data with the pre-constructed three-dimensional data model, the correlation between power grid operational anomalies and the characteristics and manufacturing defects of the cable itself is traced. In this way, by fusing data from both the manufacturing and operational stages, supplementary judgment criteria from a manufacturing perspective can be provided for fault diagnosis, thereby enhancing the accuracy and reliability of diagnostic conclusions and contributing to more scientific operation and maintenance decisions. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the description of the embodiments of this application or related technologies will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a flowchart illustrating a fault diagnosis method in one embodiment of this application;

[0020] Figure 2 This is a flowchart illustrating the process of constructing a three-dimensional data model in one embodiment of this application;

[0021] Figure 3 This is a flowchart illustrating the fault diagnosis process in one embodiment of this application;

[0022] Figure 4 This is a structural block diagram of a fault diagnosis device in one embodiment of this application;

[0023] Figure 5 This is an internal structural diagram of a computer device in one embodiment of this application. Detailed Implementation

[0024] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0025] In recent years, with the rapid increase in electricity load in large cities across China and the ever-increasing demands of modern urban construction on land use and urban environment, power transmission cables have been widely used in urban power grids due to their advantages such as small footprint and good concealment. Statistics show that the total length of 6-500kV cable lines in my country has exceeded 8,000 km, of which the operational mileage of 66-500kV high-voltage and ultra-high-voltage cables has surpassed 30,000 km. To support the high-capacity power transmission needs of large cities, large-section, high-current-carrying high-voltage and ultra-high-voltage cables have become key carriers for urban power transmission, and many large cities have planned, constructed, or put into operation 500kV cable networks. Against this backdrop, ensuring the safe and stable operation of power transmission cables throughout their entire life cycle has become a crucial foundation for reliable power supply from the power grid.

[0026] With the continuous expansion of cable infrastructure and the widespread deployment of various monitoring systems, power grid operation and maintenance face challenges related to the broad scope of supervision and the massive volume of data. For example, the length of 110kV and above transmission cables in Shanghai has exceeded 4,000 km, with an annual growth rate of approximately 12%, while the number of maintenance personnel is only about 200. Traditional manual inspections and periodic testing methods are insufficient to meet actual needs. Therefore, online monitoring devices have been deployed extensively, currently numbering in the thousands, with some systems achieving integrated visualization. However, the existing monitoring system still suffers from problems such as the separation of manufacturing and transmission cable operation data, and a lack of in-depth analysis capabilities, hindering the accurate assessment of cable conditions and the realization of predictive maintenance.

[0027] In the cable manufacturing process, traditional quality inspection mainly relies on manual visual inspection or localized 3D imaging and inspection technology. These methods generally suffer from low inspection efficiency, significant influence of subjective judgment, and susceptibility to missed or false detections due to visual fatigue, making it difficult to meet the high standards of quality consistency and stability required in large-scale cable production.

[0028] In actual production, even cables that meet factory standards inevitably have differences in material properties and minor defects in their internal structure. These minor defects, along with the cable's inherent characteristics, can potentially combine with other defects or external operating conditions to cause malfunctions. However, on the one hand, the testing systems for each process in cable production are usually provided by different suppliers, with varying data formats and interfaces, making it difficult to integrate manufacturing data and isolating it from subsequent power grid operation monitoring systems. On the other hand, cables have a multi-layered composite structure with significant differences in materials and processes between layers, making it difficult for existing non-destructive testing methods to perform complete and accurate three-dimensional reconstruction and feature extraction of their internal structure.

[0029] Therefore, current online monitoring methods based on operational data mainly rely on operational status data collected during power grid operation for fault diagnosis. Due to the lack of records and characteristic analysis of the structural status during the cable manufacturing stage, it is difficult to accurately trace the root cause of the abnormality when a fault occurs. It is also difficult to effectively distinguish whether the cause of the fault is due to the inherent characteristics of the cable itself, manufacturing defects, on-site installation process, or real-time operating conditions, which limits the accuracy of the diagnostic conclusions.

[0030] In summary, existing power grid fault diagnosis methods are still limited to isolated analysis of local surface features of cables and data from operational phases, failing to form a data fusion diagnostic system covering the entire lifecycle of cable manufacturing and operation. This hinders the improvement of fault diagnosis accuracy. Therefore, there is an urgent need to construct a data fusion diagnostic system that integrates manufacturing and operational status monitoring across the entire cable lifecycle to improve the reliability of cable condition assessment and fault early warning, thereby supporting the safe and stable operation of the power grid.

[0031] To address the aforementioned problems, this application provides a fault diagnosis method. This fault diagnosis method is applied to a terminal. The terminal can be, but is not limited to, various personal computers, laptops, smartphones, tablets, servers, IoT devices, and portable wearable devices. Portable wearable devices can be smartwatches, smart bracelets, head-mounted devices, etc. Head-mounted devices can be virtual reality (VR) devices, augmented reality (AR) devices, smart glasses, etc.

[0032] like Figure 1 As shown, the fault diagnosis method provided in this application includes:

[0033] Step 102: Obtain the test data of each structural layer of the cable. The test data is obtained by testing during the production process of the structural layers.

[0034] Among them, structural layers can refer to the functional layers that make up the cable body, such as conductor layer, insulation layer, buffer layer, metal sheath, armor layer, outer sheath layer, etc.

[0035] Test data can refer to digital data that reflects the geometric, physical, or electrical characteristics of structural layers, obtained through testing equipment during the manufacturing process of each structural layer. Examples include the geometry of a conductor layer composed of multiple stranded metal wires obtained by a camera, the thickness data of an insulation layer obtained by a laser diameter gauge, the uniformity data of a shielding layer obtained by an X-ray imaging system, and the partial discharge signal data obtained during factory testing by a partial discharge testing system.

[0036] As an example, the detection data is obtained by the detection device set up at the detection station to acquire images of the structural layer placed at the detection station.

[0037] In some feasible embodiments, different testing stations can be set for different structural layers. For example, a conductor stranding testing station can be set for the conductor layer, an insulation layer testing station can be set for the insulation layer, a buffer layer testing station can be set for the buffer layer, a metal sheath testing station can be set for the metal sheath, and an outer sheath testing station can be set for the outer sheath layer.

[0038] In some feasible embodiments, the detection device may include at least one of an image acquisition device, a laser diameter gauge, an X-ray imaging system, a partial discharge testing system, etc.

[0039] In some feasible embodiments, at the conductor stranding inspection station, information such as the geometry, parameters, and quality defects of the conductor layer formed by multiple stranded metal wires can be acquired using industrial cameras and 3D (3D) cameras. Information such as stranding pitch and conductor surface quality can also be recorded, while real-time conductor resistance can be measured using an online inspection device. Some or all of this information can be used as the conductor layer's inspection data.

[0040] In some feasible embodiments, at the insulation layer inspection station, information such as the geometry, parameters, and surface defects of the insulation layer, such as eccentricity and ellipticity, can be collected using industrial cameras and 3D cameras; information such as internal defects of the insulation layer can be collected using an X-ray imaging system; and information such as the aging degree, cross-linking degree, and tensile strength of the insulation layer material can be measured using inspection devices. Some or all of this information can be used as the inspection data of the insulation layer.

[0041] In some feasible embodiments, at the buffer layer inspection station, industrial cameras and 3D cameras can be used to collect three-dimensional structural data (such as tape thickness, width, wrapping pitch, wrapping outer diameter, etc.), geometry, parameters, and surface defects of the buffer layer; the wrapping tightness, uniformity, appearance defects (such as wrapping gaps, wrinkles), and appearance dimensions of the buffer layer can be recorded; and the moisture content and tensile strength of the buffer layer can be measured using inspection devices. Some or all of this information can be used as the inspection data for the buffer layer.

[0042] In some feasible embodiments, at the metal coating inspection station, information such as the geometry, parameters, and surface defects of the metal coating can be collected using industrial cameras and 3D cameras, focusing on the overlap rate, rolled surface dimensions, thickness, and weld defects. Information such as lateral pressure and tensile strength of the metal coating can also be measured using inspection devices. Some or all of this information can be used as inspection data for the metal coating.

[0043] In some feasible embodiments, at the outer sheath layer inspection station, two-dimensional images and three-dimensional geometric data can be acquired using a multi-camera system to detect information such as the thickness, uniformity, appearance defects (e.g., bubbles, wrinkles), and abrasion resistance of the outer sheath layer; the aging degree and tensile strength of the insulation material can also be measured using the inspection device. Some or all of this information can be used as the inspection data for the outer sheath layer.

[0044] In some feasible embodiments, to overcome the limited perspective of single-point detection, the detection station can be designed as a multi-angle platform. This platform is essentially a ring-shaped detection fixture where detection devices can be installed at each vertex. For example, placing image acquisition devices at each corner is equivalent to deploying multiple surrounding detection points, enabling coordinated imaging of the cable structure layer from different directions. In this way, through a spatial deployment strategy, the global nature of the acquired data is ensured in hardware, laying a solid foundation for the subsequent generation of a complete and accurate 3D data model.

[0045] As an example, the inspection station can employ a multi-angle platform with 6-8 corners, with wide-angle cameras positioned at each corner to achieve 360° coverage of each structural layer without blind spots. The technical parameters of the wide-angle cameras may include: adjustable resolution from 0.3MP to 600MP; frame rate of 30-60fps; infrared anti-interference light source; and IP67 protection rating.

[0046] For example, during the extrusion, cross-linking, cooling, and molding processes of each structural layer, such as the conductor layer and insulation layer, after the molding of each structural layer, and during the assembly of the structural layers into a cable, specified parameters of each structural layer are measured using testing devices integrated on the production line (such as image acquisition devices, thickness gauges, eccentricity detectors, diameter measuring instruments, etc.) or staged offline testing equipment (such as high-voltage testing equipment, etc.). The acquired test data is recorded and stored in digital format, forming the original test dataset associated with the cable. These datasets characterize the initial state and potential risk characteristics of each structural layer when the cable is manufactured.

[0047] Step 104: Construct a three-dimensional data model of the cable by integrating the various test data.

[0048] In this context, a 3D data model refers to a digital model constructed in a computer based on inspection data, reflecting the internal and external three-dimensional geometry and key physical property distribution of a cable. This model not only includes shape information but can also associate or embed characteristic parameters reflected in the inspection data. The 3D data model is a precise digital mirror of the cable's manufacturing state. By integrating discrete, layered inspection data into a unified, intuitive, and computable 3D data model, the cable's internal structure, dimensional variations, and potential defect locations can be visualized and quantitatively analyzed in three dimensions. This overcomes the limitations of traditional inspection reports or two-dimensional drawings, providing crucial geometric and attribute benchmarks for spatial positioning and feature association during the operational phase.

[0049] For example, after obtaining the test data, the spatial distribution of each structural layer can be determined according to the cable's structural design. Then, the test data corresponding to each structural layer is transformed into the boundary conditions and attribute parameters of that layer in three-dimensional space, resulting in a three-dimensional data sub-model of each structural layer. Subsequently, the three-dimensional data sub-models of each structural layer are fused according to their spatial relationships, ultimately generating a spatially continuous, attribute-checkable, and feature-traceable three-dimensional cable data model. This realizes the visualization and structured expression of manufacturing test data from one-dimensional sequences or two-dimensional cross-sections to three-dimensional physical space.

[0050] In some feasible embodiments, during the display of the three-dimensional data model, the user can click on any location of the three-dimensional data model (e.g., the 500th meter, the inner surface of the armor layer) to query all relevant original inspection data (e.g., geometric dimensions, defect images, electrical performance, etc.) and production defect analysis conclusions (e.g., "This point has scratches, which are moderate defects") at that location.

[0051] In some feasible embodiments, the three-dimensional data sub-model can be the spatial distribution of detection data in the three-dimensional entity space; the process of fusing the three-dimensional data sub-models of each structural layer may include aligning and associating the detection data in each three-dimensional data sub-model.

[0052] Alignment methods can include at least one of spatial alignment and temporal alignment. Spatial alignment can be further divided into length direction alignment and cross-sectional position alignment. Length direction alignment means that all detection data must be synchronized with the real-time meter mark during the cable production process. The production length information recorded by the encoder is a common time-space axis. For example, if a laser thickness gauge detects an insulation thickness of 9.1 mm at the meter mark "205.3 meters", then this thickness data is precisely anchored at 205.3 meters on the cable length coordinate. Cross-sectional position alignment means that for detection data with the same meter mark position, its radial and circumferential positions are further located using a unified cross-sectional coordinate system. Temporal alignment means that for dynamically acquired detection data, its time series needs to be converted with the production speed, and finally the time information is also mapped onto the length coordinate to eliminate timing deviations.

[0053] The association methods can include at least one of structural topology association and feature-based causal association. Structural topology association can include: establishing an ideal theoretical hierarchical topology based on the cable's structural design, and then filling or binding the aligned detection data to the corresponding layer of this theoretical hierarchical topology. Feature-based causal association can include: identifying and establishing potential causal relationships between detection data of different layers and types at a unified location. For example, if insulation eccentricity (geometric data) is detected at 500 meters, and the background value of the partial discharge signal (electrical data) at that location is high, these two potential risk features can be associated and marked as "a potential risk of electric field distortion due to uneven insulation exists at 500 meters".

[0054] In some feasible implementations, in addition to incorporating detection data, order data and meter information can also be incorporated into the 3D data model.

[0055] Order data can be obtained from the order management module through a communication connection. Order data may include cable specifications (such as section meter marks, cable specifications, cable structure, etc.), production batches, and quality traceability codes.

[0056] Meter label information can be assigned by the meter label management system through communication. The meter label management system is responsible for unique meter label assignment, and the format of meter label information can be "Cable Number - Start Meter Label - End Meter Label". In the 3D data model, the inspection data of each structural layer can be associated with the order data according to the meter label information.

[0057] In cable production, the meter mark is the unique identifier for length. By synchronously collecting cable production length information through an encoder and spatiotemporally correlating it with inspection data, the inspection data of each structural layer of the cable are then fused and superimposed according to a set meter mark interval. This allows for a precise mapping of "defect location → cable length coordinates" for each structural layer of the cable. For example, it can be shown that "a cable has a moderate defect of 'insulation layer thickness insufficient' at 500 meters," facilitating rapid location and traceability in subsequent operation and maintenance.

[0058] In some feasible embodiments, in addition to incorporating inspection data, the three-dimensional data model can also incorporate production defect identification results. It should be clarified that minor defects may exist in the structural layers that are allowed to remain due to compliance with factory standards; the defect characteristics of these minor defects are already contained in the inspection data. Production defect identification results, on the other hand, are obtained through further defect identification based on the inspection data. They are specific production anomalies determined through comprehensive analysis of multiple defect characteristics and material properties, such as loose conductor stranding, uneven insulation thickness, wrinkled buffer layers, scratches on the metal sheath, or damage to the outer sheath. The production defect identification results can also include defect ratings based on power grid industry standards (such as GB / T 12706 and DL / T 401), such as "minor," "moderate," and "severe," thereby providing a grading basis for production quality control.

[0059] In some feasible embodiments, defects in the structural layer can be identified based on detection data using deep learning-based defect detection models, geometric feature matching algorithms, etc., to obtain production defect identification results.

[0060] In some feasible embodiments, during the production process, two-dimensional image data of the structural layer can be acquired from multiple perspectives using a camera. Computer vision algorithms (such as stereo matching, point cloud reconstruction technology, etc.) are then used to map the dispersed two-dimensional image data into three-dimensional space to generate 3D point cloud data of the structural layer. The 3D point cloud data can contain spatial information such as the geometric shape, size parameters, and surface texture of the structural layer. Subsequently, based on the 3D point cloud data, defects in the structural layer are identified using a deep learning-based defect detection model, geometric feature matching algorithm, etc., to obtain the production defect identification results.

[0061] In some feasible implementations, data preprocessing such as denoising and distortion correction can be performed before various types of data are fused.

[0062] In some feasible implementations, before various types of data are fused, the integrity and accuracy of the data can be checked, such as missing images, interrupted sensor data, duplicate meter markings, and incorrect meter markings, to filter invalid data and ensure data quality.

[0063] Step 106: After the cable is connected to the target power grid, acquire the operation monitoring data of the target power grid.

[0064] The target power grid can refer to the specific power network to which the cable is connected, such as a city distribution network, a regional transmission network, or an independent power supply system. This embodiment does not limit this.

[0065] Operational monitoring data refers to various physical quantities that reflect the operating status of a cable after it has been put into operation in the target power grid, continuously collected by online monitoring devices (such as distributed fiber optic temperature measurement systems, partial discharge monitoring systems, and grounding current monitoring devices) installed on the cable body, joints, or auxiliary equipment. Examples include temperature data, partial discharge data, current carrying capacity data, and sheath circulation current data at different locations on the cable.

[0066] For example, after the cable is laid, installed, and energized, various operational monitoring data reflecting the cable's operating status are continuously collected at a set sampling frequency by deploying or utilizing existing online monitoring sensor networks along the cable line. Terminals can receive these operational monitoring data transmitted from the online monitoring sensor network via wired or wireless communication networks. The received operational monitoring data can be time-series data, dynamically reflecting changes in the cable's electrical, thermal, and mechanical states under actual load and environmental conditions.

[0067] In some feasible embodiments, a database can be constructed, specifically including a 3D cable database and a defect database. After the cable is connected to the target power grid, the terminal can communicate with the database and the data interface service of the target power grid respectively, obtain operational monitoring data from the target power grid, and query and read the 3D data model and defect information of the cable from the database.

[0068] The 3D cable database stores 3D data models of cables with multiple layers stacked on top of each other, supporting features such as fast querying and version management. The defect database stores defect information throughout the entire cable lifecycle, supporting defect statistics and trend analysis.

[0069] The target power grid's data interface service acts as a bridge between the terminal and the target power grid's existing system, enabling protocol conversion, data uploading, and command reception. The target power grid's data interface service can initially interface with SCADA (Supervisory Control and Data Acquisition) systems, EMS (Energy Management System), and monitoring equipment (such as inspection robots and online monitoring terminals) to collect basic cable information (such as model, years of operation, and geographical location) and operating parameters (such as load current, ambient temperature, and partial discharge values).

[0070] The terminal can then use this data interface service to push the cable's three-dimensional data model, production defect analysis results, monitoring alarms, etc., to the power grid production management system to support operation and maintenance decisions.

[0071] The data interface service can also receive remote control commands (such as "trigger partial discharge retesting of a certain section of cable" or "adjust defect monitoring frequency") and link the front-end detection equipment to perform operations.

[0072] To enable the deep application of cable 3D data models in power grid operation and maintenance, they can be integrated with the power grid operator's operating system through data interface services. Specific integration methods include: interfacing with existing power grid business systems (such as production management systems, dispatching systems, and testing equipment management systems) to synchronously acquire basic cable data (including model, commissioning time, location, etc.) and upload monitoring and analysis results; at the communication level, supporting compatibility and adaptation between power communication protocols and internal data formats through standard protocol conversion (such as IEC 61850 standard, modem bus protocol, open platform unified communication architecture, IoT protocols, and representational status transfer application programming interfaces, etc.) to ensure interconnectivity between heterogeneous systems; at the data format level, supporting structured data interfaces such as Extensible Markup Language (EXPLAIN) and JavaScript object representation, as well as binary stream transmission; and for transmission security, employing Secure Sockets Layer / Transport Layer security protocols for encrypted transmission to ensure data security. It also supports bidirectional command interaction, can receive remote control commands (such as triggering specific section detection tasks), and can proactively report monitoring anomalies and alarm information to the business system.

[0073] The generated 3D data supports multiple output formats to adapt to different application scenarios, including a general 3D object file format and point cloud data in a polygon file format that can carry color information. In terms of interaction, the system can provide real-time data stream push via the WebSocket protocol, and also provides system function libraries in multiple programming languages ​​such as C, C++, C#, Visual Basic, and Delphi, facilitating deep integration and functional expansion by third parties. Finally, all complete 3D cable model data and fused defect information are stored in a 3D cable professional database. This database provides a complete application programming interface, supporting interactive operations such as calling, querying, rotating, and scaling the 3D model, providing a visualized and analyzable data foundation for intelligent power grid operation and maintenance.

[0074] In some feasible embodiments, the terminal may include a comprehensive intelligent monitoring and visualization system for the overall condition of power transmission cables and a data fusion processing center. The data fusion processing center acquires detection data from each structural layer of the cable and constructs a three-dimensional data model of the cable by fusing the detection data. The comprehensive intelligent monitoring and visualization system for the overall condition of power transmission cables connects and synchronizes data with the data fusion processing center through a data interface service, obtaining the three-dimensional data model of the cable from the data fusion processing center. The comprehensive intelligent monitoring and visualization system for the overall condition of power transmission cables can be presented through a unified monitoring screen interface. This interface integrates a geographic information system map display area, a three-dimensional cable visualization area, a real-time operating data panel, a defect warning and alarm area, a risk analysis and prediction area, and a control operation area, providing users with a panoramic and immersive monitoring experience. The comprehensive intelligent monitoring and visualization system for the overall condition of power transmission cables can also integrate a standardized intelligent monitoring system for power grid cables operated by telecom operators. This allows for centralized monitoring and management of the production capacity, product yield rate, and production equipment status of cable manufacturers, and, based on built-in algorithms, optimizes production scheduling and provides early warnings of equipment failures, thereby improving the collaborative efficiency of the supply chain.

[0075] Step 108: Combine operational monitoring data and a three-dimensional data model to perform fault diagnosis on the target power grid and obtain the fault diagnosis results.

[0076] Among them, the fault diagnosis result can refer to the conclusions drawn from the analysis regarding whether the cable status in the target power grid is abnormal at the current moment or in a preset future time period, the type of abnormality, the severity, possible causes, and the location.

[0077] For example, after acquiring the operational monitoring data, abnormal features can be identified first. When an abnormal feature is identified (e.g., excessive temperature at a certain location, sudden increase in partial discharge signal), the diagnostic process is initiated. First, based on the location information in the operational monitoring data (e.g., the distance to the fiber optic temperature measurement point, the time difference of arrival of the partial discharge signal), the corresponding three-dimensional spatial coordinates are accurately located in the three-dimensional data model. Then, the detection data of the manufacturing stage associated with this coordinate location in the three-dimensional data model is retrieved and analyzed. By performing correlation analysis between the abnormal operational features and the detection data at the corresponding location, and excluding interference factors such as installation and environment, a comprehensive judgment is made as to whether the current anomaly is related to the inherent features or defects of the manufacturing stage. Based on the correlation between the two, the probability of failure, the risk level, and the diagnostic conclusion that the root cause of the failure tends to be manufacturing defects, operational overload, or other external factors are further determined. Finally, a diagnostic report containing location, cause, risk level, etc., is output as the fault diagnosis result.

[0078] In some feasible embodiments, to enhance the system's analytical and predictive capabilities, the intelligent monitoring and visualization system for integrated power transmission cable status can incorporate a machine learning module. This module, combined with operational monitoring data and a 3D data model, can diagnose potential faults in the target power grid within a preset future timeframe, yielding fault diagnosis results. The machine learning module encompasses a defect evolution model and a risk prediction algorithm: the defect evolution model analyzes historical data to learn defect development patterns and reveals the potential correlation between defects during the production phase and cable operational performance; the risk prediction algorithm comprehensively utilizes various machine learning methods, including random forest classifiers for cable risk level determination, support vector machines for operational anomaly detection, and deep learning networks for intelligent identification of image-based defects, thereby achieving an intelligent upgrade from data to decision-making.

[0079] Among the aforementioned fault diagnosis methods, constructing a data fusion diagnostic system covering the entire cable lifecycle can effectively address the technical problem of low fault diagnosis accuracy. Specifically, even cable products that meet quality standards inevitably exhibit differences in material properties and minor internal defects. These subtle defects, along with the cable's inherent characteristics, can potentially interact with other defects or external operating conditions to cause faults. However, traditional online monitoring methods primarily rely on collected operational status data for fault diagnosis, aiming to trace whether the anomaly stems from cable characteristics, manufacturing defects, on-site installation issues, or real-time operating conditions. This results in insufficient diagnostic evidence and low fault diagnosis accuracy.

[0080] Based on this, the present invention first acquires the inspection data obtained during the production process of each structural layer of the cable, and then constructs a three-dimensional data model that accurately reflects the characteristics of the cable itself and manufacturing defects by fusing these inspection data. After the cable is connected to the power grid and put into operation, real-time monitoring of the power grid's operating status is achieved by acquiring the operational monitoring data of the target power grid during actual operation. Subsequently, by jointly analyzing and comparing the operational monitoring data with the pre-constructed three-dimensional data model, the correlation between power grid operational anomalies and the characteristics and manufacturing defects of the cable itself is traced. In this way, by fusing data from both the manufacturing and operational stages, supplementary judgment criteria from a manufacturing perspective can be provided for fault diagnosis, thereby enhancing the accuracy and reliability of diagnostic conclusions and contributing to more scientific operation and maintenance decisions.

[0081] In one exemplary embodiment, combined with Figure 1 ,like Figure 2 As shown, the above-described method constructs a three-dimensional data model of the cable by fusing various detection data, including steps 202 to 204. Wherein:

[0082] Step 202: Perform three-dimensional reconstruction of each structural layer based on the detection data to obtain the three-dimensional data sub-model corresponding to each structural layer.

[0083] It should be noted that in the field of cable monitoring and diagnosis, traditional modeling methods typically focus only on the overall outer diameter or simple geometric parameters of the cable, failing to accurately characterize the manufacturing details and spatial distribution characteristics of each internal structural layer. Even when acquiring detection data for each structural layer, existing technologies struggle to integrate this scattered and heterogeneous data into a complete three-dimensional model that accurately reflects the cable's true internal structure. This disconnect between data and model prevents the precise correlation between operational anomalies and the manufacturing characteristics of specific structural layers during fault diagnosis, reducing the accuracy and reliability of the diagnosis.

[0084] Three-dimensional reconstruction refers to the process of constructing a continuous and complete three-dimensional geometric shape and structure of an object in digital space by using a set of discrete detection data that reflects the surface or internal features of the object through calculation and interpolation. For example, by using densely collected insulation layer thickness data along the length of a cable, the three-dimensional curved surfaces of the inner and outer surfaces of the insulation layer can be reconstructed.

[0085] A 3D data sub-model refers to a digital model generated after independently reconstructing a specific structural layer in a cable in 3D, which contains only the geometric and attribute information of that layer.

[0086] For example, for each structural layer, the acquired detection data corresponding to that structural layer can be input into the corresponding 3D reconstruction algorithm to generate a corresponding 3D data sub-model. Each 3D data sub-model is constructed in a unified spatial coordinate system, accurately representing the geometry, material distribution, and manufacturing characteristics of that structural layer in the cable, providing basic data units for subsequent model fusion.

[0087] Step 204: Construct a three-dimensional data model of the cable by fusing the various three-dimensional data sub-models.

[0088] For example, after determining the 3D data sub-model of each structural layer, a unified global coordinate system is first established. Then, based on the actual positional relationship of each structural layer in the cable, the 3D data sub-models are spatially aligned. Next, the inter-layer interface data is processed to ensure the data continuity and consistency of adjacent structural layers at the boundaries. Finally, through a data integration algorithm, the geometric parameters, material properties, and manufacturing features of each 3D data sub-model are uniformly encoded to generate a complete 3D data model of the cable. This model can accurately reflect the spatial distribution, material properties, and microstructural features and production defects formed during the manufacturing process of each structural layer of the cable from the inside out, providing a precise ontological structural reference for subsequent correlation analysis with operational monitoring data.

[0089] In this embodiment, by using a layered reconstruction and fusion approach, the three-dimensional data model of the cable maintains the manufacturing specificity of each structural layer and fully reflects the structural relationship between layers, significantly improving the accuracy and completeness of the model. This greatly enhances the accuracy and relevance of fault diagnosis, providing reliable technical support for precise operation and maintenance decisions.

[0090] In an exemplary embodiment, three-dimensional reconstruction is performed on each structural layer based on the detection data to obtain a three-dimensional data sub-model corresponding to each structural layer, including:

[0091] By using a multi-view stereo vision algorithm, each detection data is projected onto the target coordinate system to obtain the three-dimensional data sub-model corresponding to each structural layer. The multi-view stereo vision algorithm includes direct linear transformation and inverse projection transformation algorithm based on view parameters.

[0092] It should be noted that in the field of cable 3D modeling, traditional reconstruction methods typically rely on single-view inspection data or simplified geometric assumptions, failing to accurately characterize the complex 3D morphology of each structural layer within the cable. Particularly for features such as stranded conductor structures, the distribution of microscopic defects in the insulation layer, and material inhomogeneities, single-view data struggles to provide complete spatial information. Furthermore, data acquired by different inspection devices often exist in their own independent coordinate systems, lacking an effective multi-view data fusion mechanism. This results in low accuracy and missing details in the reconstructed model, failing to provide a reliable structural reference for accurate fault diagnosis.

[0093] Among them, multi-view stereo vision algorithm refers to the technology of recovering the three-dimensional geometric information of an object's surface by matching corresponding feature points and using the principle of triangulation from two-dimensional images taken from multiple different viewpoints.

[0094] The target coordinate system can be a predefined spatial reference system used to uniformly represent all three-dimensional data.

[0095] Direct linear transformation refers to a linear camera calibration method that solves for the projection matrix by establishing a direct linear relationship between image point coordinates and object space coordinates, without relying on the initial values ​​of the camera's internal parameters. For example, the camera's projection model parameters can be solved using control points with known spatial locations and their corresponding image points in the image.

[0096] The inverse projection transformation algorithm based on viewpoint parameters refers to the process of mapping the coordinates of a two-dimensional image back to three-dimensional space and calculating the spatial light rays that the corresponding object points may be located, given the intrinsic parameters (such as focal length, principal point, etc.) and extrinsic parameters (such as position and attitude, etc.) of the camera.

[0097] For example, firstly, synchronous two-dimensional image data acquired from multiple fixed viewpoints of the structural layer is used as the main detection data input. Then, a multi-view stereo vision algorithm is applied for processing. Specifically, the direct linear transformation method is first used to solve the preliminary projection matrix of each camera based on the known three-dimensional coordinates of the calibrated object or feature point in the target coordinate system and its two-dimensional projection coordinates in each viewpoint image, establishing a direct mapping relationship from the image plane to the target coordinate system. On this basis, a reverse projection transformation algorithm based on viewpoint parameters is further used for refinement and three-dimensional reconstruction. That is, using the calibrated or optimized precise camera viewpoint parameters (intrinsic and extrinsic parameters), the feature points (e.g., surface texture points, edge points, marker points, etc.) identified in each two-dimensional image belonging to the structural layer are inversely projected. By finding the optimal intersection or nearest point among multiple spatial rays projected from different viewpoints corresponding to the same object point, the precise three-dimensional coordinates of the feature point in the target coordinate system are calculated. This process is repeated for a large number of feature points to generate a three-dimensional point cloud on the surface of the structural layer. Then, through surface fitting or mesh generation, a continuous three-dimensional surface model of the structural layer is formed, i.e., a three-dimensional data sub-model.

[0098] In some feasible embodiments, the methods for constructing a three-dimensional data sub-model may include:

[0099] First, the two-dimensional image data acquired from multiple fixed perspectives are preprocessed, including removing noise using methods such as Gaussian filtering, correcting lens distortion based on camera calibration parameters, and highlighting the structural features of the cable through edge detection and texture enhancement, in order to improve the accuracy of subsequent processing.

[0100] The next step is feature processing. The scale-invariant feature transformation algorithm is used to extract stable key feature points from the preprocessed two-dimensional image data. Feature points are matched between two-dimensional image data from different perspectives using a matcher such as fast nearest neighbor search. Algorithms such as random sampling consistency are applied to eliminate mismatched pairs to ensure the reliability of feature correspondence.

[0101] After obtaining reliable feature point matching, the two-dimensional image coordinates of the feature points are converted into three-dimensional spatial coordinates of the target coordinate system through spatial transformation calculations. The core algorithms of this process include direct linear transformation and inverse projection transformation based on viewpoint parameters. Direct linear transformation provides a method for solving projection relationships based on linear equations, which is suitable for the rapid reconstruction of regular surface models; while inverse projection transformation based on viewpoint parameters utilizes precise camera intrinsic and extrinsic parameters to achieve higher-precision three-dimensional point localization through back projection and intersection calculations, which is particularly suitable for the fine reconstruction of cable surfaces and complex internal curved surfaces.

[0102] The generated 3D point cloud data will then enter the point cloud processing stage. This stage involves statistical filtering to remove outlier noise points, using voxel meshes for simplification, and employing Poisson reconstruction techniques to fill in missing parts, thereby optimizing the quality and completeness of the point cloud. Finally, in the 3D model generation stage, based on the optimized point cloud data and feature matching relationships, triangulation and surface reconstruction algorithms are used to construct continuous 3D surface models of each structural layer of the cable. This accurately recreates its geometric morphology and layered spatial structure, forming a 3D data sub-model that can be used for subsequent analysis and diagnosis.

[0103] This multi-view stereo vision processing workflow, through systematic image preprocessing, feature matching, spatial coordinate transformation, point cloud optimization and surface reconstruction, achieves a reliable conversion from multi-source two-dimensional images to high-fidelity three-dimensional data sub-models, providing an accurate hierarchical geometric basis for constructing cable three-dimensional data models.

[0104] In this embodiment, by employing multi-view stereo vision technology that incorporates direct linear transformation and inverse projection transformation algorithms based on viewpoint parameters, it is possible to fully utilize cable structure layer image data acquired from multiple angles to accurately reconstruct three-dimensional spatial point coordinates from two-dimensional image information. This method effectively overcomes the inherent problems of incomplete models, missing occluded areas, and geometric distortion inherent in single-view reconstruction, achieving a more complete and accurate three-dimensional digital reconstruction of the surface morphology of each cable structure layer. The resulting three-dimensional data sub-model has higher geometric fidelity and can more realistically reflect subtle features such as insulation surface roughness, shielding layer coverage integrity, and sheath layer concavity and convexity defects, laying a solid and reliable geometric foundation for subsequent generation of a high-precision overall three-dimensional data model of the cable and achieving accurate three-dimensional quantification and spatial positioning of defects.

[0105] In one exemplary embodiment, a three-dimensional data model of the cable is constructed by fusing the various three-dimensional data sub-models, including:

[0106] The three-dimensional data sub-models are fused using the symbolic distance field technique and smooth Boolean operation to obtain the three-dimensional data model of the cable. The structural layer includes the armor layer. During the fusion of the three-dimensional data sub-models, smooth difference operation is used to simulate the indentation effect of the armor layer on the inner structural layer.

[0107] Among them, the signed distance field technique can refer to the method of implicitly representing three-dimensional shapes using mathematical functions. For any point in space, the function gives the shortest signed distance from the point to the surface of the object. The distance is positive when the point is outside the object, zero when it is on the surface, and negative when it is inside the object.

[0108] Smooth Boolean operations refer to methods for performing union, intersection, or difference operations on two three-dimensional shapes using a signed distance field representation. Unlike traditional Boolean operations that produce sharp edges, it generates smooth, natural transition regions at the boundaries of operations through a smoothing function.

[0109] Smooth difference is a type of smooth Boolean operation used to eliminate one shape from another, producing a smooth transition at the elimination boundary. For example, using smooth difference to eliminate a sphere from a cuboid will create a pit with smooth inner walls within the cuboid, rather than a sharp-edged hole.

[0110] The armor layer can refer to the protective layer in a cable formed by spirally winding metal wires or metal strips, used to enhance the cable's mechanical strength and resistance to external forces, such as steel tape armor or steel wire armor structures.

[0111] The indentation effect refers to the phenomenon that, during cable manufacturing, when the metal armor layer is wound or wrapped, the tension and contact pressure cause the inner layer structure to produce adaptive micro-indentations or deformations, resulting in a tight fit between the inner and outer layers.

[0112] For example, the three-dimensional data sub-models of each structural layer are first converted into symbolic distance field representations. This involves calculating the signed distance value from each sampling point in space to the surface of each layer, forming a continuous distance field function. Then, in a unified global coordinate system, based on the actual structural order of the cable from the inside out, these symbolic distance fields are subjected to ordered smooth Boolean operations to complete the fusion. For the inner layers except the armor layer, a smooth union operation is used to merge them into a single internal structural field. When introducing the armor layer, instead of simply smoothing and unioning the symbolic distance field of the armor layer with the internal structural field, a smooth difference operation is used. The symbolic distance field representing the internal structure is used as the subtraction object, and the symbolic distance field representing the armor layer is used as the subtraction tool. By adjusting the smoothing parameters to control the indentation depth and transition region width, the indentation effect of the armor layer's spiral structure on the inner layer surface is simulated, causing the inner structural layers to form small deformations at the armor gaps that conform to physical reality. Finally, all the calculation results are converted back to an explicit three-dimensional geometric representation, resulting in a cable three-dimensional data model that fuses all layers and reflects the actual interlayer contact state.

[0113] In one exemplary embodiment, a three-dimensional data model of the cable is constructed by fusing the various three-dimensional data sub-models, including:

[0114] First, in the structural modeling stage, based on cable design specifications or historical data, each structural layer, such as conductor layer, insulation layer, and sheath layer, is defined, and its material properties (such as insulation dielectric constant), theoretical thickness, and spatial topological relationship are clarified, and each layer is assigned initial spatial coordinates.

[0115] Secondly, in the layer relationship configuration stage, based on the physical behavior of the cable in actual operation (such as thermal expansion, mechanical stress, etc.), the spatial constraints and dynamic relationships between each layer are set. For example, the allowable gap between the insulation layer and the conductor, the displacement coordination relationship between the sheath layer and the armor layer due to the adhesive force, and the deformation coupling coefficient of different materials under temperature changes are defined.

[0116] Subsequently, geometric overlay calculations are performed. Based on the interlayer relationships defined in the previous steps, coordinate transformation techniques are used to unify the 3D data sub-models of each layer into the same coordinate system: axial alignment is performed with the conductor layer axis as the reference, and point cloud registration or surface fitting ensures seamless connection of the radial boundaries of each layer. During the overlay process, the system automatically performs conflict detection, identifies and corrects geometric inconsistencies such as interlayer overlap or abnormal gaps, and finally completes the geometric fusion of each layer model through Boolean operations to form a complete and consistent 3D geometric skeleton of the cable.

[0117] Finally, defect data integration is performed. The terminal is associated with a database storing various detected defects, and the spatial coordinates, types (e.g., air gaps, cracks, etc.) and severity levels of the defects are accurately mapped to the corresponding structural layers in the aforementioned fusion model, generating a 3D cable data model with visual defect annotations (e.g., highlighting damaged areas).

[0118] The above steps, through systematic structural definition, relational configuration, geometric fusion, and information annotation, integrate the hierarchical three-dimensional data sub-models into a computable and traceable three-dimensional cable data model containing complete structural information and defect distribution status.

[0119] In this embodiment, by employing signed distance field technology combined with smoothed Boolean operations, particularly using smoothed difference operations to simulate the pressing effect of the armor layers, it goes beyond simple geometric assembly. By fusing the 3D data sub-models of each structural layer, high-fidelity mathematical modeling of key physical interactions (interlayer mechanical pressing) in the cable manufacturing process can be achieved. This effectively overcomes the shortcomings of traditional 3D models, which only contain geometric parameters and lack physical realism. The 3D data model generated in this embodiment not only includes the precise geometric features of each structural layer but also realistically reproduces the contact morphology and spatial relationships formed between layers during the manufacturing process. This allows the model to provide geometric and boundary conditions that more closely resemble the actual cable structure during subsequent fault diagnosis, thereby improving the reliability and accuracy of fault diagnosis.

[0120] In some feasible embodiments, a complete 3D reconstruction and fusion algorithm for multi-layer cable structures is proposed. The overall idea is: "From the inside out, layered construction; process-driven, intelligent fusion." The core of the algorithm is based on image preprocessing such as denoising, distortion correction, and feature enhancement. It can also achieve feature matching through SIFT (Scale-Invariant Feature Transform) feature extraction, feature point matching, and mismatch removal. Combined with edge computing processing, direct linear transformation, and cloud point data processing, it accurately reflects the 3D model generation (geometric features) constructed after each layer is stacked. It can also efficiently handle the interaction between layers (such as compression and embedding). Through layered construction and fusion strategies, high-fidelity digital twin modeling of multi-layer cable structures is achieved.

[0121] 3D reconstruction can include fusion algorithms, SDF (sign distance function) calculations, defect ensemble algorithms, and real-time rendering optimization algorithms.

[0122] As an example, the process of constructing a 3D data model of a cable may include:

[0123] Construct the sign distance function SDF1 for conductor layer L1. Construct the sign distance function SDF2 for insulating layer L2, but ignore the indentation effect for now. Construct the sign distance function SDF3 for buffer layer L3. Construct the sign distance function SDF4 for metallic sheath layer L4. Apply the indentation effect to SDF2 and SDF3: use SDF4 to perform a smooth difference operation on SDF2 and SDF3 to obtain the modified SDF2' and SDF3'. Construct the sign distance function SDF5 for outer sheath layer L5, and merge SDF4 with SDF1, SDF2', and SDF3' into the overall SDF of the inner layer. inner SDF5=SDF inner -T5, where T5 is the outer sheath thickness. Construct the symbolic distance function SDF6 for the temperature-sensing fiber L6: Place SDF6 between SDF2 and SDF4, and embed SDF6 within SDF3.

[0124] The conductor layer L1 (SDF1) includes a wire and a conductive strip; the insulation layer L2 (SDF2) includes an inner shielding layer, an insulation layer, and an outer shielding layer; and the outer sheath layer L5 (SDF5) includes an anti-corrosion layer, an insulating outer sheath, and a semi-conductive outer sheath.

[0125] The conductor layer is constructed, including:

[0126] Construction target: typically a circular single-core or multiple stranded split conductors.

[0127] Algorithm construction: For a single-core conductor, a simple parameterized cylinder can be used. The central axis path C(t) (C(t) is the parametric equation of the cable's central axis, t∈[0,1]) and the conductor radius R are used as parameters for scanning or rotation shaping.

[0128] For conductor layers formed by stranding split conductors, the "instantiation along helical path placement" algorithm can be used. First, a 3D model (slender cylinder) of a single guide wire is created as a primitive; then, within the total radius of the conductor, multiple layers and multiple spatial helices are defined to determine the helical path of each guide wire, with its pitch and radius following the stranding process; then, instantiation is performed through an instance transformation matrix, that is, the guide wire primitives are copied and transformed along their respective paths to fill the entire conductor cross-section, forming a tightly stranded structure.

[0129] Using the instantiation-based helical path placement algorithm, the helical path of the guidewire can be represented as:

[0130]

[0131] Among them, P strand (t, Φ) represents the center position of the monofilament in three-dimensional space under parameters t and Φ; t is a scalar parameter that varies along the central curve C(t), which can be simply understood as a proportionality or arc length parameter in the length direction, t∈[0,1]; Φ is the circumferential rotation angle. Pitch is the twist pitch; Φ0 is the initial phase angle; r is the helical radius of the guide wire; N(t) is the normal vector of the Fleury frame; B(t) is the secondary normal vector of the Fleury frame.

[0132] Instance transformation matrix M mathcal It can be represented as:

[0133]

[0134] Where Scale(r) wire The symbol ) represents a scaling transformation, scaling the unit cylinder to the actual size of a single filament. wire It is the radius of the monofilament itself; Translate(P) strand (t,Φ)) represents a translation transformation, moving the scaled cylinder from the world origin (0,0,0) to point P. strand (t,Φ); Rotate(Φ) represents a rotation transformation.

[0135] The insulation layer is constructed, including:

[0136] Construction objective: An extruded layer that is uniformly or non-uniformly wrapped around a conductor.

[0137] Algorithm construction: Ideally, the outer surface of the conductor is used as a reference to perform a radial offset or shell operation, with the offset amount being the nominal thickness of the insulation layer.

[0138] The SDF offset based on the conductor layer can be expressed as:

[0139]

[0140] Among them, T insulation Nominal thickness of the insulation layer; SDF insulation (P) represents the sign distance function of the insulating layer, where P represents any point in space; SDF conductor (P) represents the symbolic distance function of the conductor layer.

[0141] In the presence of eccentricity, an outer cylindrical surface that is not concentric with the conductor can be constructed. A smooth thickness transition can be achieved using convolutional surfaces or variable distance offsets, which can be represented as:

[0142]

[0143] Where e(t) is the eccentric vector function.

[0144] The insulation layer and the conductor layer are "wrapped" together and are usually directly bonded. When fusing, the slight unevenness on the inner surface of the insulation layer that may be caused by conductor stranding must be taken into account.

[0145] The buffer layer is constructed, including:

[0146] Construction target: usually a buffer strip wrapped around the package, with overlapping gaps.

[0147] Algorithm Construction: First, simplify to a thin layer. For the macroscopic model, it can be considered as a very thin cylindrical shell. Then, refine the modeling. To represent the wrapping details, use a "strip primitive instantiation along a spiral path" similar to the armor layer. Then, construct the primitives. The primitive can be a flat cuboid representing a buffer strip. Then, form the path. Calculate the overlap along the corresponding spiral on the outer surface of the insulation layer. Then, instantiate. Arrange the strip primitives closely along the path, ensuring overlap between adjacent strips.

[0148] The buffer layer is very thin and primarily affects the contact boundaries with the inner and outer layers. In the fusion model, it can be treated as a special texture or attribute, marked on the inner surface of the insulating layer or metal sheath.

[0149] The metallic sheath comprises a copper wire / copper strip shielding layer, a steel strip / steel wire armor layer, and a continuous metallic sheath layer. The metallic sheath construction includes:

[0150] Construction of copper wire / copper strip shielding layer.

[0151] Construction Algorithm: Similar to the segmented conductor, it adopts "instantiation along the spiral winding structure path", with the spiral path equation as follows:

[0152]

[0153] in, Pitch is the winding pitch; R inner The radius of the shielding layer is the spiral radius.

[0154] Copper wire can be densely spirally wound on the outer surface of the insulation layer using a thin cylinder as the basic unit.

[0155] Copper strips can be made using flat cuboids as basic units, with a certain overlap rate, and spirally wrapped around them, i.e., the direction of the strip basic units.

[0156] The copper wire / strip shielding layer contains thousands of discrete elements. Direct Boolean operations are not feasible; implicit fusion must be performed using the symbolic distance field (SDF).

[0157] Construction of steel strip / wire armor layer.

[0158] Construction Algorithm: Similar to the shielding layer principle, but on a larger scale. It employs "instantiation placed along a spiral path." The primitives are steel strips (cubic prisms) or steel wires (cylinders). Accurate simulation of gaps and pitch is required.

[0159] The steel strip / wire armor layer is the main cause of the inner layer's "indentation effect." Simply performing a Boolean difference between the inner layer model and the armor model will produce unrealistic sharp edges. Therefore, a smooth Boolean or physically based deformation field is used to simulate the smooth indentation of the inner layer insulation / shielding caused by the armor strip.

[0160] Construction of continuous metal sheaths (aluminum sheath / lead sheath (argon arc welding, aluminum extrusion, lead extrusion)).

[0161] Construction Goal: A continuous, sealed metal tube. Construction Algorithm: Basic Shape: A parametric cylinder, formed by extrusion or rotation. The basic cylindrical SDF can be represented as:

[0162]

[0163] Among them, SDF sheath (P) represents the signed distance function of the metal sheath layer; P projected Let R be the projection of point P onto the central axis; sheath The radius of the metal sheath.

[0164] Modeling of aluminum sleeve (argon arc welding) weld seam:

[0165]

[0166] Among them, SDF weld (P) represents the sign distance function of the aluminum bushing weld; SDF sheath (P) represents the signed distance function of the metal sheath layer; d axial σ is the axial distance from point P to the weld line; A is the control of the bulge amplitude; σ is the standard deviation of the Gaussian function, which controls the width of the weld influence. The larger σ is, the wider and smoother the axial range of the weld influence; the smaller σ is, the narrower and sharper the weld.

[0167] For longitudinal welding (such as argon arc welding), an axial, slightly raised stripe or a stripe with different material properties needs to be created on the surface of the cylinder to represent the weld.

[0168] For extrusion (aluminum extrusion, lead extrusion), it can be regarded as an ideal seamless cylinder.

[0169] The metal sheath is a continuous layer, simply enclosing the inner layer. The fusion process is relatively simple, mainly involving checking for interference with the inner layer (e.g., whether wrinkles in the buffer layer cause gaps).

[0170] Construction of the outer sheath layer.

[0171] Construction goal: A continuous plastic layer that encases all internal structures, including the uneven armor layer.

[0172] Construction algorithm: Directly offset by a thickness outside the metal sheath to generate a smooth cylinder.

[0173] Regarding the fusion algorithm, global precise Boolean operations are impractical for such a heterogeneous multi-layered structure. Therefore, this embodiment adopts a hybrid strategy.

[0174] (1) Implicit modeling and symbolic distance field.

[0175] Idea: Use the mathematical function SDF(P) to represent space, where the value is the shortest distance from point P to the surface of the object.

[0176] Application: Basic Boolean operations.

[0177] Union .

[0178] Here, A and B represent two independent symbolic distance field functions participating in the union operation. This is used to merge all copper wires and all steel strips.

[0179] Smooth union. A smoothing function is introduced based on the min function to avoid sharp seams and is used to handle effects such as the embedding of armor layers into insulation layers. It can be expressed as:

[0180]

[0181] Where k is a coefficient used to control the smoothness.

[0182] difference set

[0183] For materials such as copper wire and steel strip, grid Boolean operations are not performed in the CPU (Central Processing Unit). The CPU only stores primitives and path parameters.

[0184] (2) Modeling of armor layer indentation effect.

[0185] Smooth difference set simulation push can be expressed as:

[0186]

[0187] Among them, SDF indentation (P) represents the sign distance function of the indentation; SDF smooth-substract (P) represents the smoothed difference set; SDF armor (P) represents the symbolic distance function of the armor layer.

[0188] The smooth difference set can be defined as:

[0189]

[0190] (3) Formula for constructing the outer sheath layer.

[0191] Based on the convolutional surface of all internal layers, this is the preferred and most efficient method.

[0192] First, all internal layers (especially the armor layer) are converted into a unified SDF, i.e. .

[0193] The surface of the outer sheath is made of a new definition:

[0194]

[0195] in, This refers to the thickness of the outer sheath layer.

[0196] A high-fidelity model is then constructed, which can at least reflect the texture of the armor layer.

[0197] From a distance: It can be represented by a smooth cylinder with a normal map.

[0198] Mid-range: Enables instantiation to render geometric details.

[0199] Close-up / Cross-section: The full-detail SDF fusion model of the area can be loaded for high-fidelity display and interaction.

[0200] Then, data mapping and attribute transfer are performed to generate unified two-dimensional texture mapping coordinates or three-dimensional parametric coordinates for the final fused model. The physical properties (such as material and thickness), process parameters (such as pitch and overlap ratio), and detected defects (such as bubble and burr location) of each layer are used as attribute textures or vertex attributes and mapped onto the fused three-dimensional model.

[0201] Through this "divide and conquer-integrate" strategy, we can automatically construct a full-layer digital twin of the cable, starting from a simple central axis and a series of process parameters. This twin can macroscopically display the overall structure, microscopically examine local details, and carry rich physical and testing information.

[0202] This operation is equivalent to "growing" a uniformly thick "shell" on the outside of all the internal structures. It can perfectly replicate the concave and convex shape of the internal armor, while the outer surface is smooth.

[0203] In real-time rendering, to maximize performance, high-precision normal maps can be used on smooth cylinders to simulate armor textures. This results in a precise, uniformly thick outer shell. It can be represented as:

[0204]

[0205] in, This represents the shortest distance from point P to the inner surface of the outer sheath.

[0206] Then operate through the casing: It can generate an outer sheath for packages with complex internal structures.

[0207] The multi-layer superposition and fusion adopts SDF operation, especially by simulating the indentation effect of the armor layer on the inner structure through smooth difference set, to achieve high-fidelity fusion of the multi-layer cable structure.

[0208] In some exemplary embodiments, surface defects include bulges and depressions. Gaussian bulge defects. It can be represented as:

[0209]

[0210] Among them, SDF layer (P) represents the signed distance function of the structural layer containing the defect; P0 is the coordinate of the center point of the defect. Numericalized depression defect. It can be represented as:

[0211]

[0212] Where H(·) is the Heaviside step function.

[0213] In one exemplary embodiment, such as Figure 3 As shown, fault diagnosis of the target power grid is performed by combining operation monitoring data and a three-dimensional data model to obtain fault diagnosis results, including steps 302 to 308. Wherein:

[0214] Step 302: If an operational anomaly is detected in the target meter section of the cable based on the operation monitoring data, locate the target data set corresponding to the target meter section from the three-dimensional data model.

[0215] It should be noted that in the field of cable fault diagnosis, traditional monitoring systems can only make simple judgments based on operating parameter thresholds, and cannot correlate operational anomalies with the structural characteristics and potential defects in the cable manufacturing process. When cables exhibit operational anomalies such as abnormal temperature or increased partial discharge, maintenance personnel find it difficult to determine whether the anomaly is caused by manufacturing defects in the cable itself or by external environmental factors or changes in operating conditions. This disconnect between operational data and manufacturing data leads to low accuracy in fault diagnosis, high false alarm and false negative rates, and an inability to provide a scientific basis for precise maintenance decisions.

[0216] In this context, the target meter segment can refer to a continuous length range of cable where a specific operational anomaly has occurred, as determined by the meter-marking function in the operational monitoring data. For example, if a distributed fiber optic temperature measurement system detects an abnormal temperature rise in the section from meter mark 495 to 505 meters, this 10-meter range is considered a target meter segment.

[0217] An abnormal operation can refer to a characteristic or event that deviates from the normal operating state of the cable, as identified by an online monitoring system. Examples include a partial discharge monitoring system detecting a continuous exceedance of pulse amplitude, a distributed temperature sensing system detecting a sudden change in temperature gradient at a certain point, and a grounding current monitoring system detecting an abnormal increase in circulating current.

[0218] The target data set can refer to a subset of data extracted from the three-dimensional data model of the cable that is spatially completely corresponding to the target meter segment and contains the geometric features, material properties, and labeled production defect information of all structural layers of the cable segment.

[0219] For example, after obtaining the operation monitoring data, abnormal features can be identified first. When abnormal operation features are identified (such as excessive temperature at a certain location or a sudden increase in partial discharge signal), the diagnostic process is initiated.

[0220] First, the target meter segment containing the abnormal operation characteristics is located. Based on the start and end meter values ​​of this target meter segment, a query is performed in the spatial index of the cable's 3D data model. Utilizing the meter-spatial coordinate mapping relationship built into the 3D data model, the 3D geometric portion and all its associated attribute data that perfectly correspond to the target meter segment are precisely extracted from the 3D data model, forming a target data set. This set serves as the data foundation for subsequent analysis.

[0221] Step 304: Identify the target production defects corresponding to the target meter segment based on the target data set.

[0222] Among them, target production defects can refer to defects that exist in the cable manufacturing stage and are recorded in the three-dimensional model, as confirmed by model query and identification functions in the target data set. For example, target production defects such as "insulation thickness too thin", "shield surface burrs", and "armor layer gap too large" identified within the corresponding meter section.

[0223] For example, the obtained target dataset is parsed and features are extracted. By querying the production defect information integrated into the 3D data model, all target production defects existing within the target meter segment are automatically identified and listed.

[0224] Step 306: Perform a correlation analysis on the target operational anomaly and the target production defect to obtain the correlation degree between the target operational anomaly and the target production defect.

[0225] Among them, correlation degree can refer to a numerical or grade index calculated through correlation analysis, which is used to quantify the degree of correlation between target operational anomalies and target production defects.

[0226] For example, the characteristic parameters of the target operational anomaly (e.g., anomaly type, location, intensity, time pattern, etc.) are matched and analyzed with the characteristic parameters of the target production defect (e.g., defect type, precise location within the meter range, geometric dimensions, severity level, etc.). A quantitative index characterizing the strength of the correlation between the two is calculated using a preset analysis algorithm (e.g., spatial distance weighted correlation, feature similarity calculation, physical influence propagation model simulation, etc.).

[0227] Step 308: If the correlation degree is higher than the preset correlation degree threshold, generate the fault diagnosis result of the target power grid by combining the target operation anomaly and the target production defect.

[0228] The preset correlation threshold refers to a pre-defined critical value used to determine whether the correlation is sufficient to support diagnostic decisions. When the calculated correlation is higher than the preset correlation threshold, a strong correlation is determined between the operational anomaly and the corresponding production defect, thus incorporating them as mutually corroborating causal elements into the final diagnostic result. Conversely, when the calculated correlation is lower than the preset correlation threshold, the operational anomaly is considered to lack sufficient correlation with the currently identified production defect. In this case, the system can initiate supplementary diagnostic paths: first, based on the current operational anomaly characteristics, combined with historical operational data, typical fault modes, and other background information, a preliminary fault judgment is made; second, other potential fault causes (such as installation defects, external damage, abnormal operating environment, etc.) are investigated one by one until the exact fault cause with a correlation higher than the threshold is found, thereby completing the diagnosis.

[0229] For example, the correlation is compared with a preset correlation threshold. If the correlation is higher than the threshold, it is determined that the target's operational anomaly is strongly correlated with the identified target production defect, and the diagnostic conclusion will integrate the information from both. The generated fault diagnosis results may include: confirming the fault or abnormal state, indicating its spatial location (target meter segment), analyzing its possible root causes (related target production defect types), assessing the risk level, and providing preliminary handling suggestions (such as key monitoring, planned maintenance, or emergency handling). If the correlation is lower than the threshold, the conclusion may point to other causes (such as external damage, overload, installation problems, etc.), which may indicate the need for further investigation.

[0230] In this embodiment, by establishing a quantitative correlation mechanism between operational anomalies and production defects, the cause of operational anomalies can be scientifically determined, thereby effectively improving the accuracy and reliability of fault diagnosis. This enables maintenance personnel to make precise interventions for potential hazards caused by actual production defects, avoiding unnecessary power outages for maintenance, while reducing the risk of overlooking potential high-risk faults, and effectively improving the safety and economy of power grid operation.

[0231] The fault diagnosis method provided in this embodiment has the following significant advantages compared to the prior art:

[0232] 1. Full Lifecycle Quality Management: Achieve full lifecycle data management for cables, from raw materials to operation and maintenance, and establish a complete quality traceability system. Through a unified data model, link every defect in the manufacturing process with the operational status, providing data support for quality improvement.

[0233] 2. Precise Defect Location and Analysis: Through multi-layer data fusion technology, the system can accurately display the defect status of each layer at any location on the cable, supporting 3D visualization query and analysis. Maintenance personnel can intuitively view the three-dimensional distribution of defects through the system, providing accurate data for fault analysis.

[0234] 3. Improved power grid operation reliability: Real-time comparison and analysis of manufacturing data and operational data are achieved. When anomalies occur during operation, relevant defects in the manufacturing process can be quickly located, greatly shortening the fault diagnosis time and improving power supply reliability.

[0235] 4. Predictive Maintenance Capability: Based on machine learning algorithms, a defect evolution model is established to predict the remaining lifespan and failure risk of cables, enabling a shift from reactive maintenance to predictive maintenance and significantly reducing operation and maintenance costs. Practical applications show that this can reduce cable quality loss by more than 30%, reduce operation and maintenance costs by more than 40%, and improve power supply reliability by more than 15%.

[0236] 5. Data-driven decision support: Provides raw data support for cable operation in the power grid, and discovers the root cause of cable operation problems through real-time power grid monitoring data and raw data analysis, thereby promoting the continuous improvement of power grid operation and maintenance technology.

[0237] 6. Standardization and Scalability: The system adopts standardized interfaces and protocols, facilitating integration with equipment and systems from different power grid operators. Modular design supports functional expansion, adapting to future technological advancements.

[0238] It should be understood that although the steps in the flowcharts of the above embodiments are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the above embodiments may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0239] Based on the same inventive concept, this application also provides a fault diagnosis device for implementing the fault diagnosis method described above. The solution provided by this device is similar to the solution described in the above method; therefore, the specific limitations in one or more fault diagnosis device embodiments provided below can be found in the limitations of the fault diagnosis method described above, and will not be repeated here.

[0240] In one exemplary embodiment, such as Figure 4 As shown, a fault diagnosis device is provided, including: a first acquisition module 402, a generation module 404, a second acquisition module 406, and a fault diagnosis module 408, wherein:

[0241] The first acquisition module 402 is used to acquire the detection data of each structural layer of the cable. The detection data is obtained during the production process of the structural layer. The generation module 404 is used to construct a three-dimensional data model of the cable by fusing the detection data. The second acquisition module 406 is used to acquire the operation monitoring data of the target power grid after the cable is connected to the target power grid. The fault diagnosis module 408 is used to perform fault diagnosis on the target power grid by combining the operation monitoring data and the three-dimensional data model to obtain the fault diagnosis result.

[0242] In an exemplary embodiment, the generation module 404 is further configured to: perform three-dimensional reconstruction of each structural layer based on each detection data to obtain a three-dimensional data sub-model corresponding to each structural layer; and construct a three-dimensional data model of the cable by fusing the three-dimensional data sub-models.

[0243] In an exemplary embodiment, the generation module 404 is further configured to: project each detection data onto the target coordinate system using a multi-view stereo vision algorithm to obtain a three-dimensional data sub-model corresponding to each structural layer, wherein the multi-view stereo vision algorithm includes a direct linear transformation and an inverse projection transformation algorithm based on viewpoint parameters.

[0244] In an exemplary embodiment, the generation module 404 is further configured to: fuse each three-dimensional data sub-model according to the symbolic distance field technique and smooth Boolean operation to obtain a three-dimensional data model of the cable, wherein the structural layer includes an armor layer, and during the fusion of each three-dimensional data sub-model, a smooth difference operation is used to simulate the indentation effect of the armor layer on the inner structural layer.

[0245] In an exemplary embodiment, the fault diagnosis module 408 is further configured to: locate the target data set corresponding to the target meter segment from the three-dimensional data model when a target operational anomaly is detected in the target meter segment of the cable based on the operation monitoring data; identify the target production defect corresponding to the target meter segment based on the target data set; perform correlation analysis on the target operational anomaly and the target production defect to obtain the correlation degree between the target operational anomaly and the target production defect; and generate the fault diagnosis result of the target power grid by combining the target operational anomaly and the target production defect when the correlation degree is higher than the preset correlation degree threshold.

[0246] In an exemplary embodiment, the detection data includes visual detection data; the visual detection data is obtained by image acquisition device set on the detection station to acquire images of the structural layer corresponding to the detection station; the detection station is a multi-angle platform, and the image acquisition device is set at the corner of the multi-angle platform; the detection station includes conductor stranding detection station, insulation layer detection station, buffer layer detection station, metal sheath detection station and outer sheath layer detection station.

[0247] Each module in the aforementioned fault diagnosis device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in the processor of a computer device in hardware form or independent of it, or stored in the memory of the computer device in software form, so that the processor can call and execute the corresponding operations of each module.

[0248] In one exemplary embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as follows: Figure 5 As shown, the computer device includes a processor, memory, input / output interfaces, a communication interface, a display unit, and an input device. The processor, memory, and input / output interfaces are connected via a system bus, and the communication interface, display unit, and input device are also connected to the system bus via the input / output interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The input / output interfaces are used for exchanging information between the processor and external devices. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, mobile cellular networks, Near Field Communication (NFC), or other technologies. When the computer program is executed by the processor, it implements a fault diagnosis method. The display unit is used to form a visually visible image and can be a display screen, a projection device, or a virtual reality imaging device. The display screen can be an LCD screen or an e-ink screen. The input device of the computer device can be a touch layer covering the display screen, or buttons, trackballs, or touchpads set on the casing of the computer device, or external keyboards, touchpads, or mice, etc.

[0249] Those skilled in the art will understand that Figure 5 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0250] In one embodiment, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above method embodiments.

[0251] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon that, when executed by a processor, implements the steps in the above method embodiments.

[0252] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above method embodiments.

[0253] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0254] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, database, or other media used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, artificial intelligence (AI) processors, etc., and are not limited to these.

[0255] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0256] The above embodiments are merely illustrative of several implementation methods of this application, and their descriptions are relatively specific and detailed. However, they should not be construed as limiting the scope of this application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A fault diagnosis method, characterized in that, The method includes: The test data of each structural layer of the cable are obtained, wherein the test data is obtained by testing during the production process of the structural layer; A three-dimensional data model of the cable is constructed by fusing the various test data. After the cable is connected to the target power grid, the operation monitoring data of the target power grid is acquired; By combining the operational monitoring data and the three-dimensional data model, fault diagnosis is performed on the target power grid to obtain fault diagnosis results.

2. The method according to claim 1, characterized in that, The step of constructing a three-dimensional data model of the cable by fusing the various detection data includes: Based on the detection data, three-dimensional reconstruction is performed on each of the structural layers to obtain the three-dimensional data sub-models corresponding to each of the structural layers. A three-dimensional data model of the cable is constructed by fusing the various three-dimensional data sub-models.

3. The method according to claim 2, characterized in that, The step of performing three-dimensional reconstruction of each structural layer based on the detection data to obtain a three-dimensional data sub-model corresponding to each structural layer includes: The detection data are projected onto the target coordinate system by a multi-view stereo vision algorithm to obtain the three-dimensional data sub-model corresponding to each structural layer. The multi-view stereo vision algorithm includes direct linear transformation and inverse projection transformation algorithm based on view parameters.

4. The method according to claim 2, characterized in that, The step of constructing a three-dimensional data model of the cable by fusing the various three-dimensional data sub-models includes: The three-dimensional data sub-models are fused using the symbolic distance field technique and smooth Boolean operation to obtain the three-dimensional data model of the cable. The structural layer includes an armor layer. During the fusion of the three-dimensional data sub-models, smooth difference operation is used to simulate the indentation effect of the armor layer on the inner structural layer.

5. The method according to any one of claims 1 to 4, characterized in that, The method of combining the operational monitoring data and the three-dimensional data model to perform fault diagnosis on the target power grid, and obtaining fault diagnosis results, includes: If an operational anomaly is detected in the target meter section of the cable based on the operational monitoring data, the target data set corresponding to the target meter section is located from the three-dimensional data model; Identify the target production defects corresponding to the target meter segment based on the target data set; A correlation analysis is performed on the target operational anomaly and the target production defect to obtain the correlation degree between the target operational anomaly and the target production defect; If the correlation degree is higher than a preset correlation degree threshold, the fault diagnosis result of the target power grid is generated by combining the target operational anomaly and the target production defect.

6. The method according to any one of claims 1 to 4, characterized in that, The detection data includes visual detection data; the visual detection data is obtained by acquiring images of the structural layer corresponding to the detection station through an image acquisition device set on the detection station; the detection station is a polygonal platform, and the image acquisition device is set at the corner of the polygonal platform; the detection station includes a conductor stranding detection station, an insulation layer detection station, a buffer layer detection station, a metal sheath detection station, and an outer sheath layer detection station.

7. A fault diagnosis device, characterized in that, The device includes: The first acquisition module is used to acquire the test data of each structural layer of the cable. The test data is obtained by testing during the production process of the structural layer. A generation module is used to construct a three-dimensional data model of the cable by fusing the various detection data; The second acquisition module is used to acquire the operation monitoring data of the target power grid after the cable is connected to the target power grid; The fault diagnosis module is used to perform fault diagnosis on the target power grid by combining the operation monitoring data and the three-dimensional data model, and obtain the fault diagnosis results.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

Images