A non-destructive testing method for strain clamp of power transmission line based on digital radiography technology

Abstract

The application discloses a kind of based on digital ray technology's transmission line strain clamp nondestructive testing method, it is related to transmission line strain clamp detection field, method includes: collection strain clamp radiographic image and geometric dimension, stress and so on multidimensional parameter, data fusion is carried out by hypergraph structure, multidimensional heterogeneous strain clamp analysis model is calculated and analyzed in edge side using optimized hypergraph neural network and edge computing gateway, to visualize presentation detection result in the form of double screen linkage, while realizing intelligent diagnosis and defect early warning;The application provides an efficient, accurate and reliable solution for the intelligent operation and maintenance of transmission line strain clamp, effectively improves the safe operation level and operation and maintenance efficiency of power grid.

Description

A Non-destructive Testing Method for Tension Clamps of Transmission Lines Based on Digital X-ray Technology Technical Field

[0001] This invention relates to the field of transmission line tension clamp testing, and more particularly to a non-destructive testing method for transmission line tension clamps based on digital X-ray technology. Background Technology

[0002] With the continuous expansion of the power grid and the constant upgrading of transmission line voltage levels, the safe and stable operation of transmission lines has become crucial to ensuring a reliable power supply. Tension clamps, as core components in transmission lines that fix conductors and bear tension, directly affect the safety of the transmission lines. Traditional inspection methods rely on manual inspections, which are not only inefficient in complex geographical environments and high-altitude work scenarios, but also highly susceptible to subjective factors, making it difficult to meet the needs of intelligent power grid operation and maintenance. Although digital technology is gradually being applied to the inspection field, many shortcomings still exist.

[0003] The primary problem with existing tension clamp inspection technologies lies in insufficient data processing capabilities. Most systems can only process X-ray images or a limited number of parameters, failing to effectively integrate multi-source heterogeneous data such as tension clamp geometry, material properties, stress-strain, and temperature. The lack of in-depth analysis of the relationships between data points results in an inability to fully reflect the true condition of the tension clamps, significantly reducing the accuracy of the inspection results. Furthermore, existing systems suffer from weak intelligent analysis and early warning capabilities. Traditional algorithms struggle to accurately identify potential defect features from massive amounts of data, and are unable to make timely and accurate diagnoses when faced with complex defect types and subtle performance changes. Simultaneously, early warning mechanisms are mostly based on fixed thresholds, unable to dynamically adjust according to the operating conditions of the tension clamps, leading to false alarms or missed alarms. This hinders early warning and proactive prevention of defects, posing a threat to the safe operation of transmission lines. Summary of the Invention

[0004] In view of the above-mentioned problems, the present invention is proposed.

[0005] Therefore, the technical problem solved by this invention is: how to overcome the limitations of existing tension clamp detection technology, such as insufficient fusion of multi-source heterogeneous data, low accuracy of defect identification, static and single early warning mechanism, and low detection efficiency, and to achieve integrated non-destructive testing of tension clamp status, deep data fusion, intelligent and accurate diagnosis, and dynamic early warning.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0007] In a first aspect, the present invention provides a non-destructive testing method for transmission line tension clamps based on digital X-ray technology, comprising:

[0008] Collect X-ray images of tension clamps in transmission lines and their multidimensional parameters to construct a standard data package;

[0009] Based on standard data packages, a hypergraph structure is constructed to associate ray image features with multidimensional parameter features, and feature decomposition and fusion are performed based on the hypergraph Laplacian matrix to output deeply fused structured data.

[0010] Based on the deeply fused structured data, the system performs real-time computation and analysis at the edge using a hypergraph neural network model and an edge computing gateway analysis model, and outputs the state analysis results of the tension clamp.

[0011] Based on the state analysis results, the two screens are driven to render separately, generating a dual-screen interactive graphical user interface;

[0012] Based on the condition analysis results, fault identification is performed using a defect classification formula. Then, a dynamic threshold model based on historical data is used for analysis. Based on the analysis results of the dynamic threshold model, defect diagnosis conclusions and early warning information for tension clamps are generated.

[0013] As a preferred method for non-destructive testing of transmission line tension clamps based on digital X-ray technology, wherein:

[0014] The method, based on standard data packets, constructs a hypergraph structure to associate ray image features with multidimensional parameter features, and performs feature decomposition and fusion based on the hypergraph Laplacian matrix, outputting deeply fused structured data including:

[0015] When constructing the hypergraph structure, the ray image features and multi-dimensional parameter features of the tension clamp are used as nodes, and the relationships between different features are connected in the form of hyperedges. At the same time, based on the actual characteristics, design standards, operating conditions and historical data of the tension clamp, each feature node is assigned a corresponding weight.

[0016] As a preferred method for non-destructive testing of transmission line tension clamps based on digital X-ray technology, wherein:

[0017] The process of constructing a hypergraph structure based on standard data packets to associate ray image features with multidimensional parameter features, and performing feature decomposition and fusion based on the hypergraph Laplacian matrix to output deeply fused structured data also includes:

[0018] A fusion algorithm based on the Laplacian matrix of a hypergraph is used for data fusion. The node features in the hypergraph are weighted and aggregated, while the correlation between the features represented by the hyperedges is also considered.

[0019] As a preferred method for non-destructive testing of transmission line tension clamps based on digital X-ray technology, wherein:

[0020] The structured data based on deep fusion is processed and analyzed in real time at the edge using a hypergraph neural network model and an edge computing gateway analysis model. The output of the tension clamp status analysis results includes:

[0021] During runtime, the HyperGraph Neural Network extracts and abstracts node features from the fused data layer by layer, extracting initial image features and parameter features.

[0022] As a preferred method for non-destructive testing of transmission line tension clamps based on digital X-ray technology, wherein:

[0023] The structured data based on deep fusion, through a hypergraph neural network model and an edge computing gateway analysis model, performs real-time computation and analysis at the edge, and outputs the state analysis results of the tension clamp, which also includes:

[0024] The edge computing gateway multidimensional heterogeneous tension clamp analysis model combines the specific parameters of the tension clamp to comprehensively analyze and judge the features extracted by the hypergraph neural network, and establishes corresponding analysis rules and decision logic based on the material properties and design specifications of the tension clamp.

[0025] As a preferred method for non-destructive testing of transmission line tension clamps based on digital X-ray technology, wherein:

[0026] The process of driving the rendering of both screens based on the state analysis results to generate a dual-screen interactive graphical user interface includes:

[0027] The first screen uses computer graphics technology to model the surface texture and shape structure of the tension clamp, and at the same time marks the parameter information of key parts on the 3D model.

[0028] As a preferred method for non-destructive testing of transmission line tension clamps based on digital X-ray technology, wherein:

[0029] The process of driving the rendering of the two screens based on the state analysis results to generate a dual-screen interactive graphical user interface also includes:

[0030] The second screen uses a parameter-related heat map visualization method to generate a parameter-related heat map based on the stress, temperature and material properties of the tension clamp. Different colors in the heat map represent different parameter value ranges.

[0031] As a preferred method for non-destructive testing of transmission line tension clamps based on digital X-ray technology, wherein:

[0032] The process of driving the rendering of the two screens based on the state analysis results to generate a dual-screen interactive graphical user interface also includes:

[0033] The two screens are linked by a unique identifier for the tension clamp. When a specific part of the tension clamp is selected on one screen, the other screen will automatically position itself accordingly.

[0034] The beneficial effects of this preferred technical solution are as follows: The first screen uses computer graphics technology to model and label key parameter information, allowing operators to observe the appearance and structural details of the tension clamp from multiple angles and quickly locate anomalies; the second screen uses a parameter correlation heat map visualization method, which can clearly present the correlation and distribution between various parameters; the two screens are linked by a unique identifier, providing operators with a comprehensive and in-depth data analysis perspective, facilitating quick and accurate decision-making and improving detection and maintenance efficiency.

[0035] As a preferred method for non-destructive testing of transmission line tension clamps based on digital X-ray technology, wherein:

[0036] The process of identifying faults based on the state analysis results using a defect classification formula, and then combining this with a dynamic threshold model based on historical data for analysis, generates defect diagnosis conclusions and early warning information for tension clamps based on the analysis results of the dynamic threshold model. This includes:

[0037] By using a defect classification formula, the learned data feature patterns under normal and defective states are compared and classified with the current detection data to identify the defect situation. At the same time, the size and location of the defect and its impact on the overall performance of the tension clamp are analyzed to assess the severity of the defect.

[0038] As a preferred method for non-destructive testing of transmission line tension clamps based on digital X-ray technology, wherein:

[0039] The process of identifying faults based on condition analysis results using a defect classification formula, and then combining this with a dynamic threshold model based on historical data for analysis, to generate defect diagnosis conclusions and early warning information for tension clamps based on the analysis results of the dynamic threshold model, also includes:

[0040] A dynamic threshold model based on historical parameter data of tension clamps is adopted. Based on the historical parameter data of tension clamps under normal operating conditions, and combined with the dynamic adjustment coefficient, the defect warning threshold is determined. When the detected parameter value exceeds the warning threshold, defect warning information containing basic information of tension clamps, detection time, defect type, and severity is automatically generated.

[0041] The beneficial effects of this preferred technical solution are as follows: by comparing the defect classification formula and the characteristic patterns of the classification data, the defect situation can be accurately identified and the severity can be assessed; the early warning threshold is determined based on the dynamic threshold model of historical parameter data, and early warning information containing detailed information can be generated in a timely manner when the tension clamp parameters are abnormal, so as to buy time for operation and maintenance personnel to deal with the problem, avoid power accidents, and ensure the safe and stable operation of transmission lines.

[0042] The beneficial effects of the present invention: The non-destructive testing method for transmission line tension clamps based on digital X-ray technology has significant practical effects. By collecting X-ray images of tension clamps and constructing standard data packages using multi-dimensional parameters, and utilizing a hypergraph structure to correlate and deeply fuse image and parameter features, an accurate and comprehensive data foundation is provided for subsequent analysis. This solves the problems of single data and insufficient correlation analysis in traditional detection. At the edge, real-time analysis using a hypergraph neural network model and edge computing gateway analysis model enables rapid output of tension clamp status, reducing data transmission latency and bandwidth consumption. This allows for timely detection of potential problems, preventing fault escalation and ensuring the real-time stable operation of transmission lines. The generated dual-screen interactive graphical user interface allows operators to intuitively view the appearance, structure, and parameter correlation of tension clamps from different dimensions, facilitating quick and accurate decision-making, improving the efficiency of tension clamp inspection and maintenance, and reducing labor costs and inspection difficulty. In terms of fault identification and early warning, combining defect classification formulas and dynamic threshold models, defect conditions can be accurately identified, severity assessed, and timely warnings with detailed information generated. This allows maintenance personnel to address tension clamp problems early, effectively preventing power accidents caused by tension clamp failures, ensuring the safe and stable operation of transmission lines, and demonstrating significant practical application value and economic benefits. Attached Figure Description

[0043] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0044] Figure 1 is an overall flowchart of the non-destructive testing method for transmission line tension clamps based on digital X-ray technology provided by the present invention. Detailed Implementation

[0045] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.

[0046] Example 1, referring to Figure 1, is the first embodiment of the present invention. This embodiment provides a non-destructive testing method for transmission line tension clamps based on digital X-ray technology, including:

[0047] S1: Collect X-ray images of tension clamps on transmission lines and multi-dimensional parameters of the tension clamps to construct a standard data package;

[0048] S2: Based on standard data packages, a hypergraph structure is constructed to associate ray image features with multidimensional parameter features, and feature decomposition and fusion are performed based on the hypergraph Laplacian matrix to output deeply fused structured data;

[0049] S3: Based on deeply fused structured data, real-time computation and analysis are performed at the edge through a hypergraph neural network model and an edge computing gateway analysis model, outputting the state analysis results of the tension clamp;

[0050] S4: Based on the state analysis results, drive the two screens to render separately, and generate a graphical user interface for dual-screen interaction;

[0051] S5: Based on the condition analysis results, fault identification is performed using a defect classification formula. Then, a dynamic threshold model based on historical data is used for analysis. Based on the analysis results of the dynamic threshold model, defect diagnosis conclusions and early warning information for tension clamps are generated.

[0052] It should be noted that, through steps S1-S5, a non-destructive testing method for tension clamps of transmission lines based on the deep integration of digital ray technology, hypergraph theory, edge computing, and intelligent early warning was constructed and implemented. This method not only achieves a closed-loop processing chain from synchronous acquisition of multi-dimensional data, fusion of multi-source heterogeneous data driven by the hypergraph structure, and edge-side intelligent model analysis, to dual-screen interactive visualization and dynamic threshold defect diagnosis and early warning, but also significantly improves the accuracy and efficiency of identifying internal defects (such as poor crimping, cracks, corrosion, etc.) of tension clamps through the eigenvalue decomposition of the hypergraph Laplacian matrix and the node update mechanism of the hypergraph neural network. At the same time, the edge computing deployment mode ensures real-time requirements, the dual-screen interactive design enhances the intuitiveness and operability of operation and maintenance decisions, and the dynamic threshold model overcomes the limitations of false alarms and missed alarms in traditional fixed threshold early warning.

[0053] Example 2, referring to Figure 1, is an embodiment of the present invention. Based on the previous embodiment, it provides a non-destructive testing method for transmission line tension clamps based on digital X-ray technology, including:

[0054] In this embodiment, the acquisition of X-ray images of the tension clamps of the transmission line and the multi-dimensional parameters of the tension clamps in step S1 above, and the construction of a standard data package, includes:

[0055] The multidimensional parameter sensing integration module collects X-ray images of tension clamps in transmission lines and their multidimensional parameters to construct a standard data package. The multidimensional parameter sensing integration module integrates a digital X-ray generator and multiple types of sensors to collect X-ray images of tension clamps in transmission lines and simultaneously sense the geometric dimensions, material properties, temperature, and stress multidimensional parameters of the tension clamps.

[0056] Specifically, the digital X-ray generator emits rays of specific intensity and wavelength that penetrate tension clamps, utilizing the differences in ray absorption by different materials to form clear X-ray images. In practical implementation, the emission parameters of the X-ray generator can be precisely adjusted according to the material and thickness of the tension clamp; for example, the ray intensity is adjustable from 50-200kV to ensure high-quality image data acquisition. Multiple types of sensors are responsible for sensing multi-dimensional parameters of the tension clamp, including its geometric dimensions, material properties, temperature, and stress. Geometric dimension sensors can accurately measure the length, width, and thickness of the tension clamp with an accuracy of ±0.1mm; temperature sensors use high-precision thermocouples with a measurement range of -40℃ to 200℃ and an accuracy of ±0.5℃; stress sensors, based on the principle of resistance strain gauges, monitor the stress on the tension clamp in real time with a measurement accuracy of ±1MPa. The synchronous acquisition of these data provides a rich and accurate data foundation for subsequent comprehensive analysis of the tension clamp's condition.

[0057] It should be noted that the multi-dimensional parameter sensing integrated module breaks through the limitations of traditional single detection methods. It no longer relies solely on X-ray images to determine the condition of tension clamps, but instead constructs a complete cognitive system for tension clamps through the sensing of multi-dimensional parameters. In terms of implementation, this module can be flexibly deployed at the transmission line tension clamp inspection site via fixed brackets or drones. During installation, it is essential to ensure the accurate relative position of the digital X-ray generator and the tension clamp, and that multiple types of sensors are closely fitted to key parts of the tension clamp to guarantee the validity and reliability of the collected data. The collected data is encoded and encapsulated according to a specific data protocol and transmitted in real-time to the subsequent processing module via high-speed data transmission lines, laying the foundation for the efficient operation of the entire system.

[0058] In this embodiment, in step S2 above, based on standard data packets, a hypergraph structure is constructed to associate ray image features with multidimensional parameter features, and feature decomposition and fusion are performed based on the hypergraph Laplacian matrix. The output deeply fused structured data includes:

[0059] The data fusion preprocessing module receives standard data packets transmitted by the multidimensional parameter perception integration module. By constructing a hypergraph structure, it correlates the X-ray image features and multidimensional parameter features of the tension clamp, and performs data fusion and preliminary processing.

[0060] The hypergraph uses the ray image features and multi-dimensional parametric features of tension clamps as nodes, and the relationships between different features are connected in the form of hyperedges. During the construction process, each feature node is assigned a corresponding weight according to the actual characteristics of the tension clamp. For example, stress parameter nodes in key stress-bearing parts are given higher weights, while some secondary auxiliary parameter nodes have relatively lower weights. The determination of weights is based on a comprehensive consideration of factors such as the design standards of the tension clamp, operating conditions, and historical data. Through the construction of this hypergraph structure, originally scattered and independent multi-source data can be organically integrated, revealing the potential relationships and inherent patterns between the data.

[0061] Specifically, the formula for constructing a hypergraph structure is: ;

[0062] in, This represents the set of feature vectors from the tension cable clamp X-ray image. With multidimensional parameter feature vector set The set of nodes that make up the structure ;

[0063] This represents the set of hyperedges connecting different nodes. Hyperedges are used to represent the relationships between different features. This represents the weight matrix of the hyperedge, and the elements of the weight matrix of the hyperedge are... According to the material coefficient of tension clamp Size ratio factor and feature similarity The calculation yields the following result, which is expressed as:

[0064] ;

[0065] in, This is a custom weight calculation function.

[0066] The data fusion process employs a hypergraph Laplacian matrix-based approach. The fusion algorithm is expressed as:

[0067] ;

[0068] in, Let be the degree matrix of the hypergraph, and let be the diagonal elements of the degree matrix of the hypergraph. For nodes The sum of the weights of the connected hyperedges, Given the adjacency matrix of the hypergraph, we perform eigenvalue decomposition on the Laplacian matrix of the hypergraph to fuse multi-source heterogeneous data of tension clamps.

[0069] It should be noted that this fusion algorithm performs weighted aggregation of node features in the hypergraph, while also considering the correlations between features represented by hyperedges, achieving deep fusion of different types of data. During processing, the data undergoes normalization and denoising to eliminate noise interference and dimensional differences, making the data more suitable for subsequent analysis. After processing by the data fusion preprocessing module, the originally chaotic multi-source data is transformed into structured, highly correlated fused data, providing high-quality input data for the multi-dimensional heterogeneous analysis model deployment module, effectively improving the accuracy and comprehensiveness of the system's analysis of tension clamp status.

[0070] In this embodiment, in step S3 above, based on the deeply fused structured data, real-time computation and analysis are performed at the edge using a hypergraph neural network model and an edge computing gateway analysis model, outputting the state analysis results of the tension clamp, including:

[0071] The multidimensional heterogeneous analysis model deployment module receives deeply fused structured data processed by the data fusion preprocessing module and performs model calculations and analysis at the edge. The multidimensional heterogeneous analysis model deployment module has a built-in optimized hypergraph neural network and edge computing gateway multidimensional heterogeneous tension clamp analysis model.

[0072] Edge computing gateways provide localized data processing and analysis capabilities, enabling rapid computation of data transmitted from the data fusion preprocessing module at locations close to the data acquisition end. The optimized hypergraph neural network is specifically designed for the characteristics of tension clamp data, automatically learning hidden features and patterns within the data. During actual operation, the optimized hypergraph neural network extracts and abstracts node features layer by layer from the fused data, gradually refining more representative and discriminative high-level features from initial image and parameter features.

[0073] Specifically, the optimized hypergraph neural network includes the node update formula:

[0074] ;

[0075] in, Represents a node In the The feature vector of the layer, Represents nodes In the super-edge Connected nodes In the The feature vector of the layer, For super-edge In the The weight matrix of the layer, For bias vectors, For activation function, Represents nodes The set of associated hyperedges, Indicates the superedge The number of nodes included;

[0076] The parameter optimization formula for the multidimensional heterogeneous tension clamp analysis model of the edge computing gateway is as follows:

[0077] ;

[0078] in, For the optimized model parameters, The set of parameters to be optimized For the sample size, This is the actual parameter matrix of the tension clamp. The parameter matrix predicted by the model. and To balance the weighting coefficients of the two errors, To represent the square of the Euclidean norm, This represents the gradient operator.

[0079] It should be noted that the edge computing gateway's multidimensional heterogeneous tension clamp analysis model further integrates the specific parameters of the tension clamp to comprehensively analyze and judge the features extracted by the hypergraph neural network. Based on information such as the material properties and design specifications of the tension clamp, the edge computing gateway's multidimensional heterogeneous tension clamp analysis model establishes corresponding analysis rules and decision logic. This method of analysis and processing at the edge significantly reduces data transmission latency and bandwidth consumption, improves system response speed and real-time performance, and can promptly detect potential problems with the tension clamp, providing strong protection for the safe operation of the power system.

[0080] In this embodiment, step S4 above, which drives the two screens to render based on the state analysis results and generates a dual-screen interactive graphical user interface, includes:

[0081] The dual-screen interactive decision visualization module presents the analysis results in a dual-screen linkage format, visualizing the detection status and parameter information of the tension clamp from different dimensions; the dual-screen interactive decision visualization module is connected to the multi-dimensional heterogeneous analysis model deployment module.

[0082] Specifically, based on the state analysis results, rendering is driven for both screens respectively;

[0083] During the construction of the first screen, computer graphics technology is used to create a detailed model of the surface texture and shape structure of the tension clamp, allowing operators to observe the appearance and structural details of the tension clamp from multiple angles. Simultaneously, parameter information for key components, such as stress concentration areas and temperature anomalies, can be annotated on the 3D model, facilitating operators to quickly locate and understand any abnormalities in the tension clamp.

[0084] Specifically, the first screen uses a 3D visualization method based on the spatial coordinate mapping of tension clamps to display the geometric dimensional parameters of the tension clamps. By combining the features of X-ray images, a three-dimensional visualization model is constructed, represented as follows:

[0085] ;

[0086] in, Represents a 3D visualization model. For 3D modeling functions, This is X-ray image data.

[0087] The second screen uses a parameter-correlation heatmap visualization method, based on the stress parameters of the tension clamp. Temperature parameters and material property parameters Generate a parameter-correlation heatmap, represented as:

[0088] ;

[0089] in, This represents a parameter-related heatmap. The heatmap generation function establishes a linkage between the two screens through the unique identifier ID of the tension clamp.

[0090] It should be noted that in the heat map generated on the second screen, different colors represent different parameter value ranges. Through the distribution and changes of colors, operators can clearly see the correlation between various parameters and the distribution of parameters on the tension clamp.

[0091] It should also be noted that the two screens are linked through the unique identifier of the tension clamp. When a specific part or parameter of the tension clamp is selected on one screen, the other screen will automatically locate the corresponding position or display relevant parameter information. This dual-screen interactive visualization method provides operators with a comprehensive and in-depth data analysis perspective, facilitating quick and accurate decision-making and improving the efficiency of tension clamp inspection and maintenance. In implementation, this is achieved through visualization software installed on the operating terminal. The software interacts with the analysis model module, updating the displayed content in real time.

[0092] In this embodiment, step S5 above involves fault identification based on the state analysis results using a defect classification formula, followed by analysis using a dynamic threshold model based on historical data. The generation of defect diagnosis conclusions and early warning information for the tension clamp based on the analysis results of the dynamic threshold model includes:

[0093] The intelligent diagnosis and defect early warning generation module generates defect diagnosis conclusions and early warning information for tension clamps based on the status analysis results; the intelligent diagnosis and defect early warning generation module is connected to the multidimensional heterogeneous analysis model deployment module.

[0094] Furthermore, the intelligent diagnosis and defect early warning generation module, based on the analysis results of the multi-dimensional heterogeneous analysis model deployment module, intelligently diagnoses the condition of the tension clamp and generates corresponding defect early warning information. In terms of intelligent diagnosis, it utilizes features extracted by an optimized hypergraph neural network and an established analysis model to determine whether the tension clamp has defects and the type of defects. Based on the learned data feature patterns under normal and defective states, it compares and classifies the current detection data. For example, when the detected data features match a preset crack defect feature pattern to a certain threshold, it can be determined that the tension clamp has a crack defect. Simultaneously, it assesses the severity of the defect by analyzing factors such as the size and location of the defect and its impact on the overall performance of the tension clamp, providing corresponding evaluation results.

[0095] Specifically, a certain threshold can be determined through historical data statistical analysis, collecting historical inspection data of tension clamps, calculating the mean and standard deviation of each defect characteristic value, and setting the threshold accordingly. ( The mean, Standard deviation, (Usually 2 or 3), and use some data not included in the statistics for verification and adjustment; machine learning models can also be used for training, using supervised learning algorithms to train historical data, determining the threshold based on the confidence level of the model output, and ensuring rationality through cross-validation; during evaluation, the size of defects can be determined and graded through direct measurement or image recognition; the location of defects can be determined by establishing a three-dimensional coordinate system, and the impact can be evaluated by combining finite element analysis; the impact on overall performance can be assessed by monitoring key performance indicators, comparing changes between normal and defective states, and establishing a comprehensive evaluation model; the corresponding evaluation results can classify the defect status into mild, moderate, and severe levels, with corresponding treatment recommendations; generate a detailed report containing tension clamp information, detection data, defect status, and treatment recommendations; use visualization technology to mark the location and size of defects on the three-dimensional model, using different colors to represent severity, and display the simulation results of performance impact.

[0096] In terms of defect early warning generation, a dynamic threshold model based on historical parameter data of tension clamps is adopted. Based on historical parameter data of tension clamps under normal operating conditions, the mean and standard deviation of the parameters are calculated, and a defect early warning threshold is determined by combining this with a dynamic adjustment coefficient. When the detected parameter value exceeds the early warning threshold, corresponding defect early warning information is automatically generated. The defect early warning information includes basic information about the tension clamp, detection time, defect type, severity, etc., and is promptly conveyed to relevant maintenance personnel through various means such as audible and visual alarms and SMS notifications. This intelligent diagnosis and defect early warning mechanism can detect and issue alarms at an early stage of tension clamp problems, giving maintenance personnel more time to handle the situation, effectively avoiding power accidents caused by tension clamp failures, and ensuring the safe and stable operation of transmission lines. Its implementation involves writing corresponding diagnostic and early warning algorithm programs, running them in real time on the server side, and monitoring and processing the analysis results.

[0097] Specifically, based on the state analysis results, fault identification is performed using a defect classification formula. The defect classification formula based on the optimized hypergraph neural network is as follows:

[0098] ;

[0099] in, This is the vector representing the defect classification results of tension clamps. This is the classification weight matrix. This refers to the feature vectors of the nodes in the last layer of a hypergraph neural network. This is the classification bias vector.

[0100] Analysis based on a dynamic threshold model using historical data: The defect warning threshold calculation adopts a dynamic threshold model based on historical parameter data of tension clamps, expressed as:

[0101] ;

[0102] in, The defect warning threshold is mean. The historical average parameters of the tension clamp are... The standard deviation of historical parameters for tension clamps, This is the dynamic adjustment coefficient.

[0103] When the detection result exceeds the warning threshold, a corresponding defect warning message is generated.

[0104] In another possible implementation, the system control and collaborative scheduling module uses a priority-based task scheduling algorithm to perform collaborative control and scheduling of each module. The system control and collaborative scheduling module is connected to the multi-dimensional parameter perception integration module, the data fusion preprocessing module, the multi-dimensional heterogeneous analysis model deployment module, the dual-screen interactive decision visualization presentation module, and the intelligent diagnosis and defect early warning module, respectively.

[0105] Furthermore, the system control and collaborative scheduling module is responsible for the collaborative control and scheduling of five other modules, including the multi-dimensional parameter sensing integration module and the data fusion preprocessing module. In terms of collaborative control, this module sends corresponding control commands to each module based on the system's operational flow and task requirements. For example, before starting detection, it sends a start command to the multi-dimensional parameter sensing integration module, controlling it to collect data according to predetermined parameters and procedures; after data collection is complete, it sends a data processing command to the data fusion preprocessing module to ensure timely and accurate data processing. Simultaneously, it monitors the operational status of each module in real time. When a fault or abnormality is detected in a module, it takes timely measures to address it, such as restarting the module or switching to backup equipment, to ensure the normal operation of the system.

[0106] For collaborative scheduling, a priority-based task scheduling algorithm is adopted. This algorithm assigns a corresponding priority to each task based on its urgency and resource requirements. Tasks with high urgency and wide impact, such as re-inspecting tension clamps suspected of serious defects, are given higher priority and resources are allocated more quickly. For routine inspection tasks, execution time is allocated reasonably based on system resource usage. During data transmission, an adaptive compression algorithm based on tension clamp parameter characteristics is used to improve data transmission efficiency and reduce bandwidth consumption. The compression ratio is automatically adjusted according to the importance of the tension clamp parameter characteristics and data transmission bandwidth consumption, ensuring that important data is transmitted quickly and accurately. This collaborative control and scheduling mechanism enables close cooperation and efficient operation of all system modules, fully leveraging the overall system performance to achieve rapid and accurate inspection of transmission line tension clamps.

[0107] Specifically, task priority The calculation formula is:

[0108] ;

[0109] in, As an indicator of the urgency of the task, This refers to the resource requirements for the task. and These are the weighting coefficients.

[0110] During data transmission between modules, an adaptive compression algorithm based on the parameter characteristics of tension clamps is adopted, achieving a compression ratio of [missing information]. The calculation formula is:

[0111] ;

[0112] in, This is the proportionality coefficient. As an indicator of the importance of the parameter characteristics of tension clamps, This is a data transmission bandwidth usage indicator. For the number of parameter features, This represents the number of dimensions related to bandwidth usage.

[0113] It should be noted that this invention, through targeted optimization in multiple aspects, ensures the real-time computational needs of the model under the condition of limited edge computing power. First, a significant amount of workload reduction has been completed in the data preprocessing stage. By associating ray image features with multi-dimensional parameter features through a hypergraph structure, and then performing feature decomposition and fusion through the hypergraph Laplacian matrix, not only is the structured integration of multi-source heterogeneous data achieved, but redundant information is also eliminated through normalization, denoising, and other operations. At the same time, weights are assigned to feature nodes, allowing the model to focus only on key features, significantly reducing the amount of data and computational complexity that needs to be processed on the edge side. Second, the hypergraph neural network is specifically optimized for the tension clamp detection scenario and is not a general-purpose deep neural network. It does not require complex training on the edge side, but only performs inference analysis on the fused structured data based on the pre-trained model. By extracting and abstracting node features layer by layer, it quickly extracts high-level features with discriminative power, avoiding redundant network operations. Furthermore, the multidimensional heterogeneous analysis model of the edge computing gateway, combined with specific parameters such as the material properties and design specifications of the tension clamp, establishes specialized analysis rules and decision logic. This enables targeted comprehensive judgment of the features extracted by the hypergraph neural network, avoiding meaningless generalization calculations and further reducing computing power consumption. Finally, the system rationally allocates computing resources to edge devices through a task scheduling algorithm, prioritizing resources for critical detection tasks. Simultaneously, an adaptive compression algorithm is used during data transmission to reduce bandwidth usage, ensuring that computing resources are concentrated on the core state analysis process. This achieves real-time model computation and analysis with limited computing power.

[0114] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A non-destructive testing method for transmission line tension clamps based on digital X-ray technology, characterized in that, include: Collect X-ray images of tension clamps in transmission lines and their multidimensional parameters to construct a standard data package; Based on standard data packages, a hypergraph structure is constructed to associate ray image features with multidimensional parameter features. Feature decomposition and fusion are then performed based on the hypergraph Laplacian matrix to output deeply fused structured data. When constructing the hypergraph structure, the ray image features and multidimensional parameter features of the tension clamp are used as nodes, and the relationships between different features are connected in the form of hyperedges. At the same time, according to the actual characteristics, design standards, operating conditions and historical data of the tension clamp, corresponding weights are assigned to each feature node. A fusion algorithm based on the hypergraph Laplacian matrix is ​​used for data fusion, which performs weighted aggregation of node features in the hypergraph and considers the correlation between features represented by hyperedges. Based on the deeply fused structured data, real-time computation and analysis are performed at the edge through a hypergraph neural network model and an edge computing gateway analysis model, outputting the state analysis results of the tension clamp. During runtime, the hypergraph neural network extracts and abstracts node features layer by layer from the fused data, extracting the initial image features and parameter features. The edge computing gateway multidimensional heterogeneous tension clamp analysis model combines the specific parameters of the tension clamp to comprehensively analyze and judge the features extracted by the hypergraph neural network, and establishes corresponding analysis rules and decision logic based on the material properties and design specifications of the tension clamp. Based on the state analysis results, the two screens are driven to render separately, generating a dual-screen interactive graphical user interface; Based on the condition analysis results, fault identification is performed using a defect classification formula. Then, a dynamic threshold model based on historical data is used for analysis. Based on the analysis results of the dynamic threshold model, defect diagnosis conclusions and early warning information for tension clamps are generated.

2. The non-destructive testing method for transmission line tension clamps based on digital X-ray technology as described in claim 1, characterized in that, The process of driving two screens to render based on the state analysis results and generating a dual-screen interactive graphical user interface includes: the first screen uses computer graphics technology to model the surface texture and shape structure of the tension clamp, and at the same time annotates the parameter information of key parts on the three-dimensional model.

3. The non-destructive testing method for transmission line tension clamps based on digital X-ray technology as described in claim 2, characterized in that, The process of driving two screens to render based on the state analysis results and generating a dual-screen interactive graphical user interface also includes: the second screen uses a parameter-related heat map visualization method to generate a parameter-related heat map based on the stress, temperature and material properties of the tension clamp, with different colors in the heat map representing different parameter value ranges.

4. The non-destructive testing method for transmission line tension clamps based on digital X-ray technology as described in claim 3, characterized in that, The process of driving two screens to render based on the state analysis results and generating a dual-screen interactive graphical user interface also includes: establishing a linkage relationship between the two screens through the unique identifier of the tension clamp.

5. The non-destructive testing method for transmission line tension clamps based on digital X-ray technology as described in claim 4, characterized in that, The process of identifying faults based on state analysis results using a defect classification formula, combined with analysis using a dynamic threshold model based on historical data, and generating defect diagnosis conclusions and early warning information for tension clamps based on the analysis results of the dynamic threshold model includes: comparing and classifying the learned data feature patterns under normal and defective states with the current detection data using a defect classification formula to identify the defect situation; simultaneously, analyzing the size and location of the defect and its impact on the overall performance of the tension clamp to assess the severity of the defect.

6. The non-destructive testing method for transmission line tension clamps based on digital X-ray technology as described in claim 5, characterized in that, The process of identifying faults based on state analysis results using a defect classification formula, and then combining this with a dynamic threshold model based on historical data for analysis, and generating defect diagnosis conclusions and early warning information for tension clamps based on the analysis results of the dynamic threshold model, further includes: using a dynamic threshold model based on historical parameter data of tension clamps, determining the defect early warning threshold based on historical parameter data under normal operating conditions of the tension clamps, and combining it with a dynamic adjustment coefficient; when the detected parameter value exceeds the early warning threshold, automatically generating defect early warning information containing basic information of the tension clamps, detection time, defect type, and severity.

Images