Real-time on-line transmission line defect identification system

Abstract

The application discloses a power transmission line defect real-time online identification system and relates to the technical field of power transmission line defect detection. The system comprises a global control unit, which is used for controlling insulators and various monitoring devices in a power transmission line; a defect feature acquisition unit, which is used for acquiring infrared image data of the insulators in the same time period by using a drone to form a test set and judge data features when the insulators produce cracks; a correlation relationship judgment unit, which is used for simulating changes of the insulators under different meteorological conditions according to the data features and judging a correlation relationship between meteorological factors and cracks; a risk probability prediction unit, which is used for constructing a defect risk prediction model by using the correlation relationship and predicting a probability of the insulators producing crack defects; and a real-time online identification unit, which is used for connecting the real-time monitoring device and the risk probability prediction unit and identifying a state of the insulators according to monitoring data. The application helps users to discover and prevent cracks in advance, thereby improving the safety of the power transmission line.

Description

Technical Field

[0001] This invention relates to the field of power transmission line defect detection technology, and in particular to a real-time online identification system for power transmission line defects. Background Technology

[0002] A power transmission line is an electrical line used to transmit electrical energy. It typically consists of conductors, insulators, and poles (or cables). The main function of a power transmission line is to transmit electrical energy generated by power plants to users at different locations. Power transmission lines are classified into high-voltage transmission lines and ultra-high-voltage transmission lines based on their voltage levels. High-voltage transmission lines generally refer to power lines with a voltage of 110 kV or higher, while ultra-high-voltage transmission lines refer to power lines with a voltage of 1000 kV or higher. In general, power transmission lines are an indispensable component of the power system, playing a vital role in transmitting electrical energy generated by power plants to users, providing a reliable power supply for people's production and daily life.

[0003] Insulators are a crucial component of power transmission lines, primarily responsible for supporting conductors and isolating them from the ground to prevent current leakage. The performance and condition of insulators are vital for the stable operation of the power system. Therefore, monitoring and maintaining insulators is an important part of power system operation and maintenance. Insulator defect detection is essential in power system operation and maintenance. Defects or damage to insulators can lead to current leakage and even trigger power system faults such as flashovers and short circuits, seriously threatening the stable operation of the power system. Insulator defect detection is a critical link in power system operation and maintenance. Regular inspections allow for timely detection and repair of problems, preventing faults, ensuring the stable operation of the power system, and improving system safety and reliability.

[0004] With the development of technology, the application of drones in insulator inspection has become an increasingly popular method. Drones are equipped with infrared thermal imaging detectors and other sensors, enabling them to inspect insulators of transmission lines remotely and efficiently. At the same time, using drones for inspection can avoid direct contact between personnel and high-voltage equipment, reducing work risks. The onboard infrared thermal imaging detectors can record detailed change data of the insulators, facilitating subsequent analysis and long-term maintenance records.

[0005] However, most existing technologies can only determine whether a transmission line has a defect when identifying defects, but cannot predict in advance whether a defect will exist. Since transmission lines and their components are exposed to the outdoor environment for a long time, they are extremely susceptible to extreme weather events such as rainstorms and typhoons. Transmission lines are prone to defects, safety hazards, and dangers. If it is not possible to predict in advance whether a transmission line has cracks or defects, when a defect is detected in real time, it may lead to power outages, short circuits, and other faults in the power system, which will seriously affect the stability and safe operation of the power system.

[0006] No effective solutions have yet been proposed to address the problems in the relevant technologies. Summary of the Invention

[0007] To address the shortcomings of the existing technologies, this invention provides a real-time online identification system for transmission line defects. This system can construct a prediction model by analyzing the correlation between infrared image data, vibration data, and meteorological factors, thereby enabling the prediction of the risk of insulator cracks under specific meteorological conditions.

[0008] The specific technical solution is as follows: A real-time online defect identification system for power transmission lines includes: The global control unit is used to control various monitoring devices for the corresponding insulators in each unit and transmission line. The defect feature acquisition unit is used to acquire infrared image data and vibration data of insulators within the same time period to form a test set, and to determine the data features when cracks occur in the insulators based on the test set; The correlation determination unit is used to simulate the changes of insulators under different meteorological conditions based on the data characteristics, and determine the correlation between meteorological factors and cracks. The risk probability prediction unit is used to construct a probability prediction model using the aforementioned correlation to predict the probability of crack defects occurring in the insulator. The real-time online identification unit is used to acquire real-time monitoring data from the monitoring equipment and send it to the risk probability prediction unit, and identify the status of the insulator online based on the feedback prediction results.

[0009] Preferably, the defect feature acquisition unit includes a data acquisition module, a data processing module, and a feature analysis unit; The data acquisition module is used to collect infrared image data of insulators within the same time period using an infrared thermal imager and to acquire vibration data using a vibration sensor to form a test set. The data processing module is used to perform temperature distribution analysis on infrared image data to obtain the temperature gradient change results of the insulator, and to perform frequency domain analysis on vibration data. The feature analysis unit is used to analyze the vibration data and temperature composition crack data characteristics when the insulator develops cracks based on the temperature gradient change results.

[0010] Preferably, the correlation judgment unit includes a model building module, a data simulation module, a change display module, and a correlation judgment module; The model building module is used to solve the general solution of the infrared image data and vibration data based on vibration data and infrared image data as basic variables, and to reconstruct the insulator simulation model based on the general solution. The data simulation module is used to input different meteorological factors as conditional data into the simulation model to output vibration and infrared image data and obtain simulation results; the meteorological factors include temperature, precipitation, snowfall and wind speed; The change display module is used to visualize the simulation results, analyze the changes in meteorological factors, vibration and infrared image data when cracks occur in the insulator, and output the change results. The correlation judgment module is used to analyze the correlation between meteorological factors and the cracks in the insulators based on the change results.

[0011] Preferably, the step of using vibration data and infrared image data as basic variables to solve for a general solution of the infrared image data and vibration data, and reconstructing the insulator simulation model based on the general solution, includes: Infrared image data is used as a reference sequence and vibration data is used as a cooperative sequence to form a basic vector. The correlation coefficient between the reference sequence and the cooperative sequence is calculated using correlation analysis. The correlation coefficients between the reference sequence and the co-occurrence sequence are set as constraints, and the average of these constraints is used to obtain the general solution value of the correlation coefficients. An optimization algorithm is used to perform a local search for the general solution value by relying on the reference sequence and the cooperative sequence. When the general solution value reaches a local optimum, a Gaussian random walk strategy is introduced to select the optimal general solution value. After selecting the optimal general solution, the insulator simulation model is reconstructed using an optimization algorithm and a kernel limit learner.

[0012] Preferably, the step of reconstructing the insulator simulation model using an optimization algorithm and a kernel limit learner after selecting the optimal general solution includes: The optimal general solution is used as the input vector of the kernel limit learner to train the kernel limit learner; The weight decay coefficient and kernel mapping function in the kernel limit learner are optimized using an optimization algorithm, and the optimal weight decay coefficient and kernel mapping function are selected to reconstruct the kernel limit learner. Determine whether the optimized kernel limit learner meets the termination condition. If the termination condition is met, use the optimal general solution as input and combine it with the kernel limit learner to construct an insulator simulation model. If the termination condition is not met, the optimization algorithm is repeatedly used to optimize the kernel limit learner until the termination condition is met.

[0013] Preferably, the step of analyzing the correlation between meteorological factors and insulator crack formation based on the changes includes: The changes in meteorological factors, vibration, and infrared image data when insulator cracks occur are sorted out to distinguish the predicted size, location, and shape of the cracks, and the corresponding meteorological factors are recorded. Topological information on the predicted size, location, and shape of cracks is extracted based on the topological network, and clusters related to cracks are obtained based on meteorological factors. Based on the clustering data, vibration data, and infrared image data, the association set with the strongest correlation to the cluster is obtained, and the association relationship within the association set is mined using the topological hierarchical information mining method.

[0014] Preferably, the risk probability prediction unit includes an optimization and collaboration module, a prediction model construction module, a prediction result analysis module, and a comparison and optimization module; The optimization and collaboration module is used to construct uniformly distributed nodes based on the association relationship and to introduce a collaboration strategy to construct a prediction algorithm. The prediction model construction module is used to optimize the prediction machine according to the prediction algorithm and to complete the construction of the probabilistic prediction model based on the prediction machine. The prediction result analysis module is used to perform visual analysis on the output of the risk probability model and extract the probability of cracks in the insulator. The comparison and optimization module is used to determine the accuracy of the prediction results. If the accuracy is lower than the threshold, the prediction model is optimized; if the accuracy is higher than the threshold, it remains unchanged.

[0015] Preferably, the step of constructing uniformly distributed nodes based on association relationships and introducing a collaborative strategy to construct a prediction algorithm includes: Initialize nodes by using the optimal point set to initialize the association relationship, so that the nodes are evenly distributed in the solution space, and divide the nodes into several child nodes; Set child nodes to move within the solution space, calculate the fitness value of each child node within the solution space, record the fitness value, and select the child node with the largest fitness value as the global optimum; Record the globally optimal child node state and update the child node position, and introduce a node cooperation strategy at its position so that different child nodes can interact with each other; A target function is constructed using the positions of child nodes as vectors, and the target function is used to optimize and update the distribution and cooperative strategies of nodes to obtain a prediction algorithm.

[0016] Preferably, the step of optimizing the predictor based on the prediction algorithm and constructing the probabilistic prediction model based on the predictor includes: The structural parameters of the forecasting machine are optimized using a forecasting algorithm. The optimized forecasting machine is then combined with a probability model to build a probability forecaster. Meteorological factors, vibration data, and infrared image feature variables are added to the probability forecaster. Preset feature variable parameters are input into the probability predictor to predict the probability of insulator crack generation, and a residual sequence is constructed based on the prediction results; The residual sequence is optimized using chaos theory, and the final optimization result is input into the probability predictor to complete the construction of the probability prediction model.

[0017] Preferably, the expression for the probability prediction model is: ; In the formula, t represents the probability that the insulator will develop a crack; Indicates a point in time; Indicates the first Vibration data values ​​at each time point; Indicates the first The temperature points reflected in the infrared images at each time point; Indicates the first Meteorological factor characteristic points at each time point; This indicates the number of factors influencing the formation of cracks in the insulator; Indicates the service life of the insulator; This indicates the loss factor of the insulator.

[0018] The beneficial effects of this invention are as follows: 1. This invention uses historical data to determine the vibration data and infrared image data at the time of insulator crack formation to form feature values. It then analyzes the correlation between these feature values ​​and meteorological factors. Based on this correlation, a probability prediction model is constructed and connected to the monitoring equipment. This allows for real-time online identification of whether insulators have crack defects. This real-time online identification enables early detection of the possibility of insulator cracks. Furthermore, the prediction model based on the correlation between feature values ​​and meteorological factors can predict the risk of insulator cracks under specific meteorological conditions, allowing for proactive measures to improve equipment operational stability. This invention helps users detect and prevent cracks in advance, thereby enhancing the safety of transmission lines.

[0019] 2. This invention uses vibration data and infrared image data as basic variables to solve for the general solution value, which reflects the actual state of the insulator. Based on the general solution, a simulation model is constructed to simulate the behavior of the insulator under different conditions. At the same time, different meteorological factor data are input into the simulation model to predict the vibration and infrared image data of the insulator under different meteorological conditions, and then predict the possibility of cracks in the insulator. By visualizing and analyzing the output results, we can gain a deeper understanding of the changes in various factors when cracks occur, laying the foundation for subsequent online identification.

[0020] 3. This invention constructs nodes with uniform and distributed distribution based on correlation relationships, introduces a collaborative strategy to construct a prediction algorithm to build a risk probability model and complete probability prediction, which can provide probability information about the occurrence of cracks in insulators, enabling real-time online identification of insulator defects on transmission lines, thereby improving the safety of transmission lines. Attached Figure Description

[0021] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. 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.

[0022] Figure 1 This is a functional architecture diagram of a real-time online defect identification system for power transmission lines according to an embodiment of the present invention.

[0023] Figure 2 This is a block diagram illustrating the specific architectural principle of an embodiment of the present invention.

[0024] Figure 3 This is a block diagram of the principle of the real-time online identification unit according to an embodiment of the present invention.

[0025] Figure 4 This is a schematic diagram of the steps of a real-time online defect identification method for transmission lines according to an embodiment of the present invention.

[0026] In the picture: 1. Global Control Unit; 2. Defect Feature Acquisition Unit; 201. Data Acquisition Module; 202. Data Processing Module; 203. Feature Analysis Unit; 3. Correlation Judgment Unit; 301. Model Building Module; 302. Data Simulation Module; 303. Change Display Module; 304. Correlation Judgment Module; 4. Risk Probability Prediction Unit; 401. Optimization and Collaboration Module; 402. Prediction Model Building Module; 403. Prediction Result Analysis Module; 404. Comparison and Optimization Module; 5. Real-time Online Identification Unit. Detailed Implementation

[0027] To further illustrate the various embodiments, the present invention provides accompanying drawings, which are part of the disclosure of the present invention. These drawings are mainly used to illustrate the embodiments and can be used in conjunction with the relevant descriptions in the specification to explain the operating principles of the embodiments. With reference to these drawings, those skilled in the art should be able to understand other possible implementation methods and the advantages of the present invention.

[0028] like Figure 1As shown, the real-time online identification system for transmission line defects according to an embodiment of the present invention includes a global control unit 1, a defect feature acquisition unit 2, a correlation judgment unit 3, a risk probability prediction unit 4, and a real-time online identification unit 5.

[0029] Global control unit 1 is used to control various monitoring devices for the corresponding insulators in each unit and transmission line; The defect feature acquisition unit 2 is used to acquire infrared image data and vibration data of the insulator within the same time period to form a test set, and to determine the data features when the insulator develops cracks based on the test set; The correlation judgment unit 3 is used to simulate the changes of insulators under different meteorological conditions based on the data characteristics and judge the correlation between meteorological factors and cracks. Risk probability prediction unit 4 is used to construct a probability prediction model using the correlation to predict the probability of crack defects in the insulator; The real-time online identification unit 5 is used to acquire real-time monitoring data from the monitoring equipment and send it to the risk probability prediction unit 4, and identify the status of the insulator online based on the feedback prediction results.

[0030] Based on the above scheme, we will elaborate in detail, such as Figure 2 As shown: (1) Global control unit 1, used to control various monitoring devices of each unit and corresponding insulators in the transmission line.

[0031] The global control unit 1 can effectively control insulators and various monitoring devices, promptly detect and address problems, reduce the risk of faults, ensure the safe operation of transmission lines, and obtain equipment operating status information in a timely manner through monitoring and control equipment. It can also predict and address potential problems in advance, avoid sudden equipment failures, and reduce unnecessary maintenance costs and time.

[0032] (2) Defect feature acquisition unit 2 is used to acquire infrared image data and vibration data of insulators within the same time period to form a test set, and to judge the data features when cracks occur in insulators based on the test set.

[0033] In this embodiment, the defect feature acquisition unit 2 includes a data acquisition module 201, a data processing module 202, and a feature analysis unit 203.

[0034] The data acquisition module 201 is used to acquire infrared images of insulators within the same time period using an infrared thermal imager mounted on a drone, and to acquire vibration data using a vibration sensor to form a test set.

[0035] The use of drones equipped with infrared thermal imaging detectors to capture infrared images of insulators is based on the principle of infrared imaging. The infrared thermal imaging detector can capture the infrared radiation emitted by the insulators and convert it into a visual temperature image for subsequent identification and processing.

[0036] It should be noted that using drones equipped with infrared thermal imaging detectors to capture infrared images of insulators enables remote operation, avoiding direct contact with high-voltage equipment and greatly reducing safety risks. At the same time, drones can fly to areas that are difficult for humans to reach and use infrared thermal imaging detectors to capture subtle temperature changes in insulators, providing high-precision data. In addition, drones can transmit images and data in real time, facilitating immediate analysis and handling of potential problems. Therefore, drones equipped with infrared thermal imaging detectors provide an efficient, safe, and accurate new method for the inspection and maintenance of insulators.

[0037] Specifically, the process of using a drone equipped with an infrared thermal imager to collect infrared images of insulators within the same time period, and using vibration sensors to acquire vibration data to form a test set includes the following steps: Ensure the vibration sensor and infrared thermal imager are functioning properly and calibrate the equipment. Install the vibration sensor and infrared thermal imager onto the insulator to ensure that the sensor and the detector can accurately collect vibration and infrared image data of the insulator. To ensure the usability of the drone, vibration sensor, and infrared thermal imager during severe weather, protective measures can be taken to protect the vibration sensor and infrared thermal imager to ensure their normal operation.

[0038] Simultaneously activate the vibration sensor and infrared thermal imager to collect vibration and infrared image data of the insulator. It is important to ensure that the acquisition time of the two devices is synchronized to guarantee the correspondence of the data. Record the data collected by the vibration sensor and infrared thermal imager, including vibration data, infrared image data, and acquisition timestamps, to ensure the accuracy and completeness of the data recording.

[0039] The collected vibration data and infrared image data are preprocessed to match the vibration data and infrared image data, ensuring that the vibration data and infrared image data at the same time point can be correctly matched. The corresponding vibration data and infrared image data are organized into a test set, and the vibration data and corresponding infrared image data at each time point are used as a sample to form the test set.

[0040] The data processing module 202 is used to perform temperature distribution analysis on infrared image data, obtain the temperature gradient change of the insulator, and perform frequency domain analysis on vibration data.

[0041] Specifically, temperature distribution analysis is performed on infrared image data to obtain the temperature gradient changes of the insulator, and frequency domain analysis is performed on vibration data, including: First, temperature distribution analysis of the infrared image data is performed, including: The acquired infrared image data is preprocessed, including adjusting the brightness and contrast of the image. Then, based on the infrared image data acquired by the infrared thermal imager, each pixel in the image is converted into a corresponding temperature value. Based on the converted temperature data, the temperature distribution on the surface of the insulator is analyzed, and a temperature image or thermal map is drawn to ensure that the temperature distribution can be displayed intuitively.

[0042] The gradient change of the insulator surface temperature is obtained by performing gradient calculation on the temperature image, and the temperature gradient is obtained by calculating the temperature difference between adjacent pixels.

[0043] Frequency domain analysis of vibration data includes: The collected vibration data is preprocessed, including noise removal and filtering. By performing time-domain analysis on the preprocessed vibration data, the time-domain characteristics of the vibration signal, including amplitude and frequency, can be obtained. Specifically, the time-domain characteristics of the vibration signal can be displayed by plotting a time-domain waveform.

[0044] Applying Fourier transform or other frequency domain analysis methods to time-domain signals transforms the signals into the frequency domain, yielding a spectrum that displays the energy distribution of the signal at different frequencies. Analyzing the spectrum identifies the main frequency components and frequency bands. Based on the frequency characteristics obtained from the frequency domain analysis, combined with the vibration characteristics and fault mode knowledge of the corresponding insulators, fault diagnosis and identification are performed.

[0045] The feature analysis unit 203 is used to analyze the vibration data and temperature composition crack data features when the insulator develops cracks based on the temperature gradient change results.

[0046] Specifically, based on the analysis of vibration data and temperature composition crack data characteristics during insulator crack formation according to temperature gradient changes, the following are included: The temperature gradient change results and vibration data have the same timestamp or time period to ensure data correspondence. Features are extracted from the temperature gradient change results and vibration data to reflect the data characteristics at the time of crack formation.

[0047] The extracted features are analyzed and correlated to explore the relationship between temperature and vibration data, and to determine the presence of characteristic patterns of cracks. Based on the data analysis results, features with discriminative power are selected.

[0048] The above technical solution can be used to determine the data characteristics when cracks occur in the insulator, laying the foundation for subsequent operations.

[0049] (3) Correlation judgment unit 3 is used to simulate the changes of insulators under different meteorological conditions based on the data characteristics and judge the correlation between meteorological factors and cracks.

[0050] In this embodiment, the association judgment unit 3 includes a model construction module 301, a data simulation module 302, a change display module 303, and an association judgment module 304.

[0051] The model building module 301 is used to solve the general solution of the infrared image data and vibration data based on the vibration data and infrared image data as the basic variables, and to reconstruct the insulator simulation model based on the general solution.

[0052] Specifically, using vibration data and infrared image data as basic variables, the general solution for the infrared image data and vibration data is solved, and the insulator simulation model is reconstructed based on the general solution, including: Infrared image data is used as a reference sequence, and vibration data is used as a cooperative sequence to form a basic vector. The correlation coefficient between the reference sequence and the cooperative sequence is calculated using correlation analysis. The correlation coefficients between the reference sequence and the co-occurrence sequence are set as constraints, and the average of these constraints is used to obtain the general solution value of the correlation coefficients. An optimization algorithm is used to perform a local search for the general solution value by relying on the reference sequence and the cooperative sequence. When the general solution value reaches a local optimum, a Gaussian random walk strategy is introduced to select the optimal general solution value. After selecting the optimal general solution, the insulator simulation model is reconstructed using an optimization algorithm and a kernel limit learner.

[0053] It should be noted that the optimization algorithm can adopt the Harris Eagle algorithm with chaotic elite optimization. It is an optimization algorithm that combines chaotic optimization and Harris Eagle algorithm. Harris Eagle algorithm is an optimization algorithm that simulates the foraging behavior of eagle flocks. It searches for the optimal solution by simulating the exploration and exploitation behaviors of eagle flocks. Chaotic elite optimization is a method of optimization that uses chaotic mapping and elite strategy. By introducing chaotic sequences and retaining elite individuals, it enhances the algorithm's exploration and exploitation capabilities.

[0054] The process of reconstructing the insulator simulation model using an optimization algorithm and a kernel limit learner after selecting the optimal general solution includes: The optimal general solution is used as the input vector of the kernel limit learner to train the kernel limit learner; The weight decay coefficient and kernel mapping function in the kernel limit learner are optimized using an optimization algorithm, and the optimal weight decay coefficient and kernel mapping function are selected to reconstruct the kernel limit learner. Determine whether the optimized kernel limit learner meets the termination condition. If the termination condition is met, use the optimal general solution as input and combine it with the kernel limit learner to construct an insulator simulation model. If the termination condition is not met, the optimization algorithm is repeatedly used to optimize the kernel limit learner until the termination condition is met.

[0055] The data simulation module 302 is used to input different meteorological factors as conditional data into the simulation model and output vibration and infrared image data. Meteorological factors include temperature, precipitation, snowfall, and wind speed.

[0056] Specifically, different temperatures, precipitation, snowfall, and wind speeds are input as conditional data into the simulation model, and the output vibration and infrared image data include: We collected actual observation data on different temperatures, precipitation, snowfall, and wind speeds. We preprocessed the collected data, using temperature, precipitation, snowfall, and wind speed as conditional parameters, and assigned corresponding values ​​to each set of data.

[0057] Inputting the conditional parameters into the insulator simulation model will perform simulation calculations based on the input conditional data, and output the corresponding vibration and infrared image data.

[0058] The change display module 303 is used to visualize the output results and analyze the changes in meteorological factors, vibration and infrared image data when cracks are generated.

[0059] Specifically, visualizing and analyzing the output results, including changes in meteorological factors, vibration, and infrared image data during crack formation, includes: Collect meteorological factor data, vibration data, and infrared image data at the time of crack formation. Time-align the collected data to ensure that the meteorological factors, vibration data, and infrared image data have the same timestamp or time period. Use appropriate visualization tools, such as Python's matplotlib library or Tableau, to plot the changing trends of the meteorological factors, vibration data, and infrared image data.

[0060] By observing and visualizing the patterns and trends among meteorological factors, vibration data, and infrared image data, we can interpret the impact of different meteorological factors on vibration and infrared image data during crack formation based on the data visualization and analysis results.

[0061] The correlation judgment module 304 is used to analyze the correlation between meteorological factors and cracks based on the change results.

[0062] Specifically, the analysis of the correlation between meteorological factors and cracks based on the changes includes: The changes in meteorological factors, vibration, and infrared image data at the time of crack formation are sorted out to distinguish the predicted size, location, and shape of the crack, and the corresponding meteorological factors are recorded. Topological information on the predicted size, location, and shape of cracks is extracted based on the topological network, and clusters related to cracks are obtained based on meteorological factors. Based on the clustering data, vibration data, and infrared image data, the association set with the strongest correlation to the cluster is obtained, and the association relationship within the association set is mined using the topological hierarchical information mining method.

[0063] Using the above technical solution, vibration data and infrared image data are used as basic variables to measure the vibration and infrared image data of insulators under different meteorological conditions, thereby predicting the possibility of cracks in the insulators. By visualizing and analyzing the output results, we can gain a deeper understanding of the changes in various factors when cracks occur, laying the foundation for subsequent online identification.

[0064] (4) Risk probability prediction unit 4, used to construct a probability prediction model using the correlation relationship, and predict the probability of crack defects in the insulator.

[0065] In this embodiment, the risk probability prediction unit 4 includes an optimization and collaboration module 401, a prediction model construction module 402, a prediction result analysis module 403, and a comparison and optimization module 404. The optimization of the collaboration module 401 is used to construct uniformly distributed nodes based on the association relationship and to introduce a collaboration strategy to construct a prediction algorithm.

[0066] Specifically, constructing a uniformly distributed set of nodes based on association relationships and introducing a collaborative strategy to construct a prediction algorithm includes: Initialize nodes by using the optimal point set to initialize the association relationship, so that the nodes are evenly distributed in the solution space, and divide the nodes into several child nodes; Set child nodes to move within the solution space, calculate the fitness value of each child node within the solution space, record the fitness value, and select the child node with the largest fitness value as the global optimum; Record the globally optimal child node state and update the child node position, and introduce a node cooperation strategy at its position so that different child nodes can interact with each other; A target function is constructed using the positions of child nodes as vectors, and the target function is used to optimize and update the distribution and cooperative strategies of nodes to obtain a prediction algorithm.

[0067] The prediction model building module 402 is used to optimize the prediction machine according to the prediction algorithm and to complete the construction of the probabilistic prediction model based on the prediction machine.

[0068] Specifically, optimizing the predictor based on the prediction algorithm and constructing the probabilistic prediction model based on the predictor include: The structural parameters of the forecasting machine are optimized using a forecasting algorithm. The optimized forecasting machine is then combined with a probability model to build a probability forecaster. Meteorological factors, vibration data, and infrared image feature variables are added to the probability forecaster. Preset feature variable parameters are input into the probability predictor to realize the probability prediction of insulator crack generation, and a residual sequence is constructed based on the prediction results; The residual sequence is optimized using chaos theory, and the final optimization result is input into the probability predictor to complete the construction of the probability prediction model.

[0069] The expression for the probabilistic prediction model is as follows: ; In the formula, t represents the probability of the insulator developing cracks, j represents the time point, represents the vibration data value at the j-th time point, represents the temperature point reflected in the infrared image at the j-th time point, represents the meteorological factor characteristic point at the j-th time point, x represents the number of influencing factors that cause cracks in the insulator, represents the service life of the insulator, and w represents the loss factor of the insulator.

[0070] The prediction result analysis module 403 is used to visualize and analyze the output results of the risk probability model and extract the probability of cracks in the insulator.

[0071] Specifically, appropriate visualization tools can be used to visualize the probability of insulators developing cracks, and the relationship between the probability of insulators developing cracks and various factors can be explained based on the probability analysis results.

[0072] The comparison optimization module 404 is used to determine the accuracy of the prediction results. If the accuracy is lower than the threshold, the prediction model is optimized; if the accuracy is higher than the threshold, it remains unchanged.

[0073] By means of the above implementation method, the probability information about the occurrence of cracks in insulators can be provided through the correlation, which enables real-time online identification of insulator defects on transmission lines, thereby improving the safety of transmission lines.

[0074] (5) Real-time online identification unit 5, used to acquire real-time monitoring data from the monitoring equipment and send it to the risk probability prediction unit 4, and based on the feedback prediction results, identify the status of the insulator online, such as Figure 3 As shown.

[0075] In this embodiment, connecting the real-time monitoring device to the risk probability prediction unit and identifying the insulator's state online based on the monitoring data includes: The monitoring data collected by the monitoring equipment is transmitted to the risk probability prediction unit 4 via a network or other adaptation method. The risk probability prediction unit 4 preprocesses the received monitoring data and extracts useful features from the preprocessed data. The monitoring equipment includes vibration sensors and infrared thermal imagers.

[0076] Using a pre-built risk probability prediction model, the extracted features are input into the model for prediction. Based on the output of the prediction model, the state of the insulator is identified. The state of the insulator is judged according to the threshold setting of the prediction result, such as normal, with cracks or defects, abnormal, or large cracks with high risk.

[0077] Based on the identified insulator status, corresponding feedback information or alarms are generated, such as sending alarm information or triggering maintenance plans. Real-time monitoring and updates are performed to continuously monitor the insulator status and update the risk probability prediction model based on new monitoring data.

[0078] Specifically, by connecting the real-time monitoring equipment to the risk probability prediction unit 4 and identifying the status of the insulator online based on the monitoring data, real-time monitoring and prediction of the insulator status can be achieved, abnormal situations can be detected in a timely manner, and corresponding measures can be taken to improve the safety and reliability of the insulator.

[0079] A method for real-time online identification of transmission line defects based on the above embodiments specifically includes: S101: Acquire infrared image data and vibration data of insulators within the same time period to form a test set, and determine the data characteristics when cracks occur in the insulators based on the test set.

[0080] S102: Simulate the changes of insulators under different meteorological conditions based on the data characteristics, and determine the correlation between meteorological factors and cracks.

[0081] S103: Construct a probabilistic prediction model using the aforementioned correlation to predict the probability of crack defects in the insulator.

[0082] S104: Real-time acquisition of monitoring data from monitoring equipment and transmission of probability prediction model to obtain prediction results, thereby enabling online identification of the insulator's status.

[0083] In summary, by utilizing the above-mentioned technical solution of the present invention, the real-time online identification system for transmission line defects provided by the present invention determines the characteristic values ​​formed by vibration data and infrared image data when insulator cracks occur based on historical data, analyzes the correlation between the characteristic values ​​and meteorological factors, constructs a probability prediction model based on the correlation, connects it to the monitoring equipment, and identifies insulators in real time whether crack defects have occurred. This allows for early detection of the possibility of cracks in insulators through real-time online identification. At the same time, the prediction model constructed based on the correlation between characteristic values ​​and meteorological factors can predict the risk of insulator cracks under specific meteorological conditions, thereby enabling early intervention and improving the operational stability of the equipment. This system can help users detect and prevent cracks in advance, thereby improving the safety of transmission lines.

[0084] This invention uses vibration data and infrared image data as basic variables to solve for a general solution that reflects the actual state of the insulator. Based on the general solution, a simulation model is constructed to simulate the behavior of the insulator under different conditions. At the same time, different meteorological factor data are input into the simulation model to predict the vibration and infrared image data of the insulator under different meteorological conditions, thereby predicting the possibility of cracks in the insulator. By visualizing and analyzing the output results, a deeper understanding of the changes in various factors when cracks occur can be achieved, laying the foundation for subsequent online identification.

[0085] This invention constructs a uniformly distributed node based on association relationships, introduces a collaborative strategy to construct a prediction algorithm to build a risk probability model and complete probability prediction, which can provide probability information about the occurrence of cracks in insulators, enabling real-time online identification of insulator defects on transmission lines, thereby improving the safety of transmission lines.

[0086] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A real-time online defect identification system for power transmission lines, characterized in that, include: The global control unit (1) is used to control various monitoring devices for each unit and the corresponding insulators in the transmission line; The defect feature acquisition unit (2) is used to acquire infrared image data and vibration data of the insulator within the same time period to form a test set, and to judge the data features when the insulator develops cracks based on the test set; The correlation judgment unit (3) is used to simulate the changes of insulators under different meteorological conditions based on the data characteristics and judge the correlation between meteorological factors and cracks. Risk probability prediction unit (4) is used to construct a probability prediction model using the correlation to predict the probability of crack defects in the insulator; The real-time online identification unit (5) is used to acquire real-time monitoring data of the monitoring equipment and send it to the risk probability prediction unit (4), and identify the status of the insulator online based on the feedback prediction results; The association judgment unit (3) includes a model building module (301), a data simulation module (302), a change display module (303), and an association judgment module (304). The model building module (301) is used to solve the general solution of the infrared image data and vibration data based on the vibration data and infrared image data as basic variables, and to reconstruct the insulator simulation model based on the general solution. The data simulation module (302) is used to input different meteorological factors as conditional data into the simulation model to output vibration and infrared image data and obtain simulation results; the meteorological factors include temperature, precipitation, snowfall and wind speed; The change display module (303) is used to visualize the simulation results, analyze the changes in meteorological factors, vibration and infrared image data when cracks occur in the insulator, and output the change results. The correlation judgment module (304) is used to analyze the correlation between meteorological factors and cracks in insulators based on the change results; The process involves using vibration data and infrared image data as basic variables to solve for a general solution based on the infrared image data and vibration data, and then reconstructing the insulator simulation model based on this general solution; including: Infrared image data is used as a reference sequence and vibration data is used as a cooperative sequence to form a basic vector. The correlation coefficient between the reference sequence and the cooperative sequence is calculated using correlation analysis. The correlation coefficients between the reference sequence and the cooperating sequence are set as limited values, and the average value of the limited values ​​is used to obtain the general solution value of the correlation coefficient. An optimization algorithm is used to perform a local search for the general solution value by relying on the reference sequence and the cooperative sequence. When the general solution value reaches a local optimum, a Gaussian random walk strategy is introduced to select the optimal general solution value. After selecting the optimal general solution, the insulator simulation model is reconstructed using an optimization algorithm and a kernel limit learner. The expression for the probability prediction model is: ； In the formula, t represents the probability that the insulator will develop a crack; Indicates a point in time; Indicates the first Vibration data values ​​at each time point; Indicates the first The temperature points reflected in the infrared images at each time point; Indicates the first Meteorological factor characteristic points at each time point; This indicates the number of factors influencing the formation of cracks in the insulator; Indicates the service life of the insulator; This indicates the loss factor of the insulator.

2. The real-time online defect identification system for transmission lines according to claim 1, characterized in that, The defect feature acquisition unit (2) includes a data acquisition module (201), a data processing module (202), and a feature analysis unit (203). The data acquisition module (201) is used to acquire infrared image data of insulators within the same time period using an infrared thermal imager and to acquire vibration data using a vibration sensor to form a test set. The data processing module (202) is used to perform temperature distribution analysis on infrared image data to obtain the temperature gradient change results of the insulator, and to perform frequency domain analysis on vibration data. The feature analysis unit (203) is used to analyze the characteristics of the vibration data and temperature composition crack data when the insulator cracks based on the temperature gradient change results.

3. The real-time online defect identification system for transmission lines according to claim 1, characterized in that, The process of reconstructing the insulator simulation model using an optimization algorithm and a kernel limit learner after selecting the optimal general solution includes: The optimal general solution is used as the input vector of the kernel limit learner to train the kernel limit learner; The weight decay coefficient and kernel mapping function in the kernel limit learner are optimized using an optimization algorithm, and the optimal weight decay coefficient and kernel mapping function are selected to reconstruct the kernel limit learner. Determine whether the optimized kernel limit learner meets the termination condition. If it does, use the optimal general solution as input and combine it with the kernel limit learner to construct an insulator simulation model. If the termination condition is not met, the optimization algorithm is repeatedly used to optimize the kernel limit learner until the termination condition is met.

4. The real-time online defect identification system for transmission lines according to claim 1, characterized in that, The method for analyzing the correlation between meteorological factors and insulator crack formation based on the changes includes: The changes in meteorological factors, vibration, and infrared image data when insulator cracks occur are sorted out to distinguish the predicted size, location, and shape of the cracks, and the corresponding meteorological factors are recorded. Topological information on the predicted size, location, and shape of cracks is extracted based on the topological network, and clusters related to cracks are obtained based on meteorological factors. Based on the clustering data, vibration data, and infrared image data, the association set with the strongest correlation to the cluster is obtained, and the association relationship within the association set is mined using the topological hierarchical information mining method.

5. The real-time online defect identification system for transmission lines according to claim 4, characterized in that, The risk probability prediction unit (4) includes an optimization and collaboration module (401), a prediction model construction module (402), a prediction result analysis module (403), and a comparison and optimization module (404). The optimization and collaboration module (401) is used to construct uniformly distributed nodes based on the association relationship and to introduce a collaboration strategy to construct a prediction algorithm. The prediction model construction module (402) is used to optimize the prediction machine according to the prediction algorithm and to complete the construction of the probability prediction model according to the prediction machine; The prediction result analysis module (403) is used to perform visual analysis on the output results of the risk probability model and extract the probability of cracks in the insulator. The comparison optimization module (404) is used to determine the accuracy of the prediction result. If the accuracy is lower than the threshold, the prediction model is optimized; if the accuracy is higher than the threshold, it remains unchanged.

6. The real-time online defect identification system for transmission lines according to claim 5, characterized in that, The step of constructing uniformly distributed nodes based on association relationships and introducing a collaborative strategy to construct a prediction algorithm includes: Initialize nodes by using the optimal point set to initialize the association relationship, so that the nodes are evenly distributed in the solution space, and divide the nodes into several child nodes; Set child nodes to move within the solution space, calculate the fitness value of each child node within the solution space, record the fitness value, and select the child node with the largest fitness value as the global optimum; Record the globally optimal child node state and update the child node position, and introduce a node cooperation strategy at its position so that different child nodes can interact with each other; A target function is constructed using the positions of child nodes as vectors, and the target function is used to optimize and update the distribution and cooperative strategies of nodes to obtain a prediction algorithm.

7. The real-time online defect identification system for transmission lines according to claim 6, characterized in that, The step of optimizing the predictor machine based on the prediction algorithm and constructing the probabilistic prediction model based on the predictor machine includes: The structural parameters of the forecasting machine are optimized using a forecasting algorithm. The optimized forecasting machine is then combined with a probability model to build a probability forecaster. Meteorological factors, vibration data, and infrared image feature variables are added to the probability forecaster. Preset feature variable parameters are input into the probability predictor to predict the probability of insulator crack generation, and a residual sequence is constructed based on the prediction results; The residual sequence is optimized using chaos theory, and the final optimization result is input into the probability predictor to complete the construction of the probability prediction model.

Images