Power transmission line vibration monitoring and early warning method for dealing with extreme weather

Abstract

The invention relates to the field of power system safety monitoring, and discloses a power transmission line vibration monitoring and early warning method for dealing with extreme weather, which comprises the following steps: arranging a multi-level sensor network to monitor vibration and environmental conditions of a power transmission line; monitoring data are processed and transmitted in real time through edge calculation and a 5G network; performing trend prediction on the vibration data by using a multi-modal deep learning model; an early warning threshold value is adjusted through a distributed self-adaptive dynamic threshold value control method; and when abnormal vibration or a high-risk trend is detected, generating and issuing early warning information. Through a multi-level sensor network, an intelligent early warning algorithm and dynamic threshold control, efficient monitoring of power transmission line vibration and accurate early warning in extreme weather are realized, the real-time performance of data transmission and processing is effectively improved, the accuracy of vibration trend prediction is enhanced, the false alarm rate and the missing report rate are reduced, and the method is suitable for popularization and application. Therefore, the method has higher adaptability and stability in a complex environment.

Description

Technical Field

[0001] The present invention relates to the technical field of power system safety monitoring, and in particular to a transmission line vibration monitoring and early warning method for coping with extreme weather. Background Art

[0002] Power transmission lines are affected by environmental factors during operation, especially under extreme weather conditions such as strong winds, heavy snow, and freezing. The vibration of the lines will increase significantly, which may cause damage or even breakage of the lines, seriously affecting the stability and safety of the power system. Therefore, vibration monitoring and early warning methods for transmission lines have become an important technical means to ensure the safe operation of transmission lines. By monitoring the vibration of the lines in real time, abnormalities can be discovered in time and early warnings can be issued in advance, thereby effectively avoiding disasters and ensuring the safety of transmission lines.

[0003] At present, the commonly used transmission line vibration monitoring and early warning methods include monitoring systems based on single sensors and early warning systems with fixed thresholds. These systems usually use single-type sensors such as accelerometers and displacement sensors to monitor the vibration amplitude, acceleration and other data of the line, and make judgments and early warnings through preset fixed thresholds. When the monitoring data exceeds the threshold, the system will automatically trigger an alarm to remind the operation and maintenance personnel to check the line. This type of method can achieve good monitoring results under normal weather conditions and provide certain guarantees for the daily maintenance of the power grid.

[0004] However, existing technologies have shortcomings when facing extreme weather conditions. The limitations of single sensor monitoring make it difficult for the system to obtain accurate data in complex environments, and the fixed threshold method is difficult to dynamically adjust under extreme weather changes, which is prone to false alarms or missed alarms. In addition, the lack of comprehensive analysis capabilities for environmental data leads to insufficient accuracy and real-time nature of early warnings, making it difficult to predict the impact of extreme weather on the line in advance. Summary of the invention

[0005] In view of the shortcomings of the prior art, the present invention provides a transmission line vibration monitoring and early warning method for extreme weather conditions, which solves the problems in the prior art of single sensor monitoring limitations and the difficulty of dynamic adjustment of fixed threshold methods under extreme weather conditions.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: A transmission line vibration monitoring and early warning method for extreme weather, comprising the following steps:

[0007] Deploy a multi-level sensor network to monitor the vibration and environmental conditions of the transmission line;

[0008] Process and transmit monitoring data in real time through edge computing and 5G networks;

[0009] Use multimodal deep learning models to predict trends in vibration data;

[0010] Adjust the warning threshold through a distributed adaptive dynamic threshold control method;

[0011] Generate and issue warning messages when abnormal vibrations or high-risk trends are detected.

[0012] Preferably, the multi-level sensor network deployment includes the following steps:

[0013] Deploy fiber Bragg grating sensors at high-risk intersections of transmission lines to monitor vibration and stress changes;

[0014] Deploy acceleration sensors and meteorological sensors in areas that are most affected by weather to obtain vibration and meteorological data;

[0015] Low-power vibration sensors are deployed in ordinary line sections to achieve wide-area coverage of line vibration.

[0016] Preferably, the real-time processing and transmission of monitoring data comprises the following steps:

[0017] Deploy edge computing nodes near sensor nodes at each level to process sensor data locally;

[0018] Data cleaning, anomaly detection and feature extraction are performed through edge computing nodes;

[0019] Use the 5G network to transmit processed key data to the cloud central server in real time.

[0020] Preferably, real-time processing and transmission of monitoring data through edge computing and 5G network includes the following steps:

[0021] Clean and denoise the collected sensor data to remove outliers;

[0022] Use Kalman filtering method to smooth the data;

[0023] Extract key features from vibration data including amplitude, jerk, and frequency.

[0024] Preferably, the multimodal deep learning model performs trend prediction on vibration data including the following steps:

[0025] Use LSTM model to process time series data to predict vibration trends;

[0026] Use CNN models to process extreme weather image data to identify weather change characteristics;

[0027] The output results of LSTM and CNN are combined in the fusion layer to generate the final vibration trend prediction results.

[0028] Preferably, the training of the multimodal deep learning model comprises the following steps:

[0029] The model was trained using historical vibration and meteorological data;

[0030] Gradient descent and Adam optimization algorithms are used to optimize model parameters;

[0031] The best model parameters were selected through cross-validation to improve prediction accuracy.

[0032] Preferably, the distributed adaptive dynamic threshold control method comprises the following steps:

[0033] Simulate different extreme weather scenarios based on historical data and set initial vibration thresholds for various weather conditions;

[0034] Use deep Q learning algorithm to optimize threshold and adjust threshold strategy through feedback;

[0035] In actual monitoring, the warning threshold is adaptively adjusted in real time according to weather changes.

[0036] Preferably, the threshold optimized deep Q learning algorithm training comprises the following steps:

[0037] In the simulation environment, the warning threshold is continuously optimized by adjusting the threshold and receiving system feedback;

[0038] Update the reward value in the Q table based on the system's performance in terms of false positive rate and false negative rate;

[0039] Through iterative training, the optimal threshold control strategy is converged.

[0040] Preferably, generating and issuing warning information includes the following steps:

[0041] The system analyzes the comprehensive data received from the central server to determine vibration trends and potential risks;

[0042] Automatically generate warning signals when the vibration trend exceeds the dynamic threshold or the future vibration trend reaches the preset risk value;

[0043] The warning information will be released to relevant operation and maintenance personnel so that preventive measures can be taken in advance.

[0044] Preferably, the method further comprises:

[0045] Receive weather forecast data in real time and feed it into a multimodal deep learning model;

[0046] Adaptively update the model using the latest monitoring data to improve vibration prediction accuracy under extreme weather conditions;

[0047] Regularly use the latest monitoring data as new training samples to re-optimize the deep learning model and dynamic threshold control strategy.

[0048] The present invention provides a transmission line vibration monitoring and early warning method for extreme weather. It has the following beneficial effects:

[0049] 1. The present invention realizes real-time monitoring and data processing of transmission line vibration by utilizing a multi-level sensor network combined with edge computing and 5G transmission, greatly improving the response speed and monitoring accuracy of the system, and ensuring efficient transmission and rapid processing of data even under complex extreme weather conditions.

[0050] 2. The present invention adopts a trend prediction method that integrates vibration data and weather data using a multimodal deep learning model, so that the system has the ability to predict the impact of extreme weather on line vibration, enhances the accuracy and environmental adaptability of the early warning, and provides accurate and early warning support for operation and maintenance personnel.

[0051] 3. The present invention uses a dynamic threshold control strategy based on deep Q learning to adaptively adjust the warning threshold, which can effectively prevent false alarms and missed alarms in different weather scenarios. Combined with the regular adaptive update of the model, the system has long-term stability and flexibility, and realizes continuous and efficient protection of lines in complex environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 The figure is a flow chart of the method of the present invention. DETAILED DESCRIPTION

[0053] The following will be combined with the drawings in the specification of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0054] Please see attached Figure 1 The embodiment of the present invention provides a transmission line vibration monitoring and early warning method for extreme weather, comprising the following steps:

[0055] A multi-level sensor network is deployed to monitor the vibration and environmental conditions of the transmission lines. The deployment strategy of the multi-level sensor network enables the system to implement targeted monitoring layout according to the risk characteristics of different areas. High-precision sensors are used for key monitoring in high-risk areas to ensure the accuracy and real-time nature of the data; low-power sensors are used in general areas to effectively reduce costs and maintenance requirements. In this way, the method can comprehensively and flexibly monitor the vibration status and environmental conditions of the transmission lines under extreme weather conditions, providing reliable data support for subsequent predictions and early warnings;

[0056] The monitoring data is processed and transmitted in real time through edge computing and 5G networks. By deploying edge computing nodes along the transmission lines, data can be processed locally in real time, reducing transmission delays and the burden on central servers. The application of 5G networks further improves the real-time and stability of data transmission, ensuring that the system can quickly respond to sudden weather conditions. The combination of edge computing and 5G networks effectively enhances the system's real-time processing capabilities, ensuring that monitoring data can be obtained in a timely manner in extreme weather conditions, and providing fast and accurate data support for early warnings;

[0057] Multimodal deep learning models are used to predict the trend of vibration data. By fusing time series data and meteorological image data, multimodal deep learning models can capture the impact of extreme weather on line vibration and form more accurate vibration predictions. LSTM can effectively predict the time series trend of vibration, while CNN can identify and extract extreme weather features in weather images. Multimodal fusion improves the accuracy and adaptability of predictions, enabling the system to make effective trend predictions in extreme weather conditions, providing a basis for subsequent dynamic threshold adjustments;

[0058] The warning threshold is adjusted through the distributed adaptive dynamic threshold control method. Through the distributed adaptive dynamic threshold control method, the system can adjust the warning threshold in real time under different extreme weather conditions. The deep Q learning algorithm continuously optimizes and adjusts the threshold control strategy through reinforcement learning, making the system adaptive and intelligent, so as to effectively avoid false alarms and missed alarms in storms, blizzards and other weather conditions, improve the accuracy and adaptability of the warning, ensure that the warning signal can be triggered in a timely and accurate manner in extreme weather, and enhance the disaster prevention capabilities of the power system;

[0059] When abnormal vibration or high-risk trends are detected, early warning information is generated and issued. When the system detects abnormal vibration or high-risk trends, timely generation and issuance of early warning information can effectively prevent major accidents caused by untimely handling of transmission lines in extreme weather. Through comprehensive data analysis and real-time judgment, the system can provide accurate early warnings and provide clear guidance information for operation and maintenance personnel to ensure that effective preventive measures are taken in extreme weather, thereby reducing the risk of line failures and improving the overall system safety and reliability.

[0060] Please refer to the attached Figure 1 In a preferred embodiment of the present invention, the multi-level sensor network deployment includes the following steps:

[0061] Fiber Bragg grating sensors are deployed at high-risk intersections of transmission lines to monitor vibration and stress changes. Fiber Bragg grating sensors detect strain changes in the optical fiber and convert them into vibration and stress measurements. Compared with traditional electrical sensors, fiber Bragg grating sensors have strong anti-interference and high stability, and can maintain high-precision data collection in extreme weather such as strong winds or heavy rains. The deployment of this sensor enhances the system's monitoring capabilities at high-risk intersections, allowing the system to accurately grasp stress changes at key nodes of the line in extreme weather, reduce the risk of line breakage or damage, and ensure the safety and stability of power transmission;

[0062] Accelerometers and meteorological sensors are deployed in areas that are most affected by weather to obtain vibration data and meteorological data. Accelerometers can capture the instantaneous vibration characteristics of the line, including changes in vibration frequency and acceleration, and reflect the stress conditions of the line in real time; meteorological sensors provide meteorological information such as ambient temperature, humidity, and wind speed in the area where the line is located, providing data support for analyzing the relationship between climate factors and line vibration. The combination of meteorological data and vibration data enables the system to more accurately analyze the actual impact of extreme weather on the line, thereby improving the accuracy of early warnings, providing a reliable data basis for subsequent predictions, and effectively reducing the risk of line failures under severe weather conditions;

[0063] Low-power vibration sensors are deployed in ordinary line sections to achieve wide-area coverage of line vibration. The deployment of low-power vibration sensors enables the monitoring system to achieve wide-area coverage without increasing excessive maintenance costs, thereby improving the monitoring capabilities of ordinary line sections. These sensors can provide basic information on line vibration in a long-term and stable manner, and are particularly suitable for vibration monitoring in medium and low-risk areas. By reducing power consumption and intelligently controlling the sampling frequency, they can effectively supplement monitoring data during extreme weather periods, improve the overall monitoring coverage and data integrity, and ensure the breadth and continuity of the monitoring network.

[0064] Please refer to the attached Figure 1In a preferred embodiment of the present invention, the real-time processing and transmission of monitoring data includes the following steps:

[0065] Edge computing nodes are deployed near sensor nodes at all levels to process sensor data locally. The deployment of edge computing nodes allows data to be processed quickly locally, reducing the time delay for uploading data to the cloud. By performing preliminary data screening and processing locally, edge computing nodes can effectively reduce network bandwidth usage and the computing load of central servers.

[0066] Data cleaning, anomaly detection and feature extraction are performed by edge computing nodes. The cleaning and anomaly detection of data by edge computing nodes can significantly improve the accuracy of monitoring data and ensure that the system is not disturbed by external noise. The feature extraction process converts complex raw data into feature data that is easy to analyze, reduces the amount of data and improves data transmission efficiency. At the same time, the anomaly detection function enables the system to quickly identify vibration changes before extreme weather arrives, providing the system with more accurate vibration monitoring information, ensuring the system's accurate early warning capabilities in the face of emergencies, and enhancing the overall monitoring and response capabilities.

[0067] The processed key data is transmitted to the cloud central server in real time using the 5G network. The efficient and low-latency data transmission capability of the 5G network is used to achieve real-time data transmission between the edge computing node and the cloud central server, enabling the system to quickly obtain on-site monitoring data. With the stability guarantee of the 5G network, the system can still maintain smooth data transmission in severe weather conditions, enhancing the system's anti-interference ability. At the same time, data redundancy and caching mechanisms can ensure data integrity when the network is unstable and avoid the loss of key monitoring data.

[0068] Please refer to the attached Figure 1 In a preferred embodiment of the present invention, real-time processing and transmission of monitoring data through edge computing and 5G network includes the following steps:

[0069] Clean and de-noise the collected sensor data to remove outliers. Data cleaning and de-noising can effectively remove invalid information in sensor data and improve data quality and accuracy. In a complex power transmission line monitoring environment, sensor data is susceptible to external interference. Cleaning and removing outliers and high-frequency noise can improve data purity and ensure the accuracy of subsequent data analysis.

[0070] The Kalman filter method is used to smooth the data. The Kalman filter method continuously updates the state estimation value, making the data smoothing more real-time and dynamically adaptable. It can effectively suppress sudden fluctuations in a short period of time, thereby improving the accuracy and reliability of the monitoring data and facilitating the subsequent vibration feature extraction and trend analysis.

[0071] Key features are extracted from vibration data, including vibration amplitude, acceleration rate of change, and frequency. The key feature extraction process converts vibration data into characteristic values ​​that are easy to analyze, providing effective input information for subsequent vibration trend prediction. Features such as vibration amplitude, acceleration rate of change, and frequency can comprehensively characterize the different dimensions of line vibration, enabling the system to keenly identify abnormal vibration patterns. Through feature extraction, edge nodes not only reduce the amount of data, but also improve the efficiency of data analysis, which helps the system to quickly analyze the line status when extreme weather occurs, enhances the responsiveness and prediction accuracy of the monitoring system, and provides effective technical guarantees for the safe and stable operation of the line.

[0072] Please see attached Figure 1 In a preferred embodiment of the present invention, the multimodal deep learning model performs trend prediction on vibration data, including the following steps:

[0073] The LSTM model is used to process time series data to predict vibration trends. The LSTM model captures the time dependency in vibration signals through a long-term memory mechanism and can effectively predict future vibration trends. By using the LSTM model for trend prediction, the system can identify possible line vibration anomalies in advance, providing a scientific basis for early warning and maintenance in extreme weather conditions, and improving the accuracy of prediction and the safety of the system.

[0074] The CNN model is used to process extreme weather image data to identify weather change characteristics. The CNN model can efficiently extract key features from image data through convolution and pooling operations. Extreme weather is usually accompanied by obvious meteorological image features, such as strong convective clouds and frontal structures. CNN can identify these features in images and provide the system with intuitive weather background information. By using CNN to process weather image data, the system can effectively capture weather change trends, improve the contextual understanding ability of vibration monitoring, provide data support for accurately predicting the vibration trend of the line affected by weather, and effectively improve the system's early warning capabilities;

[0075] The output results of LSTM and CNN are combined in the fusion layer to generate the final vibration trend prediction results. LSTM provides the long-term memory characteristics of the vibration time series, and CNN provides the ability to recognize extreme weather characteristics. The combination of the two enables the system to more comprehensively predict the vibration trend of transmission lines under complex weather conditions. Multimodal fusion improves the prediction accuracy and system adaptability, provides a reliable basis for early warning in extreme weather conditions, and significantly improves the overall accuracy and intelligence level of the monitoring system.

[0076] Please see attached Figure 1 In a preferred embodiment of the present invention, the training of the multimodal deep learning model includes the following steps:

[0077] The model is trained using historical vibration data and meteorological data. Through training with a large amount of historical data, the model can identify the typical characteristics of transmission line vibration under different weather conditions, and thus has the ability to predict. This process enhances the model's adaptability in extreme weather conditions, allowing the system to more accurately predict future vibration trends;

[0078] Gradient descent and Adam optimization algorithms are used to optimize model parameters. Through the combination of gradient descent and Adam optimization algorithms, the model can converge to the optimal solution more quickly and effectively during the training process, reducing training time while avoiding falling into local optimality;

[0079] The best model parameters are selected through cross-validation to improve the prediction accuracy. The model can maintain a high generalization ability on different data sets and avoid overfitting or underfitting. Cross-validation ensures that the model not only performs well on the training data, but also has a high prediction accuracy on new data, thereby improving the adaptability and robustness of the system under different extreme weather conditions, making the final multimodal deep learning model more reliable and stable, thereby ensuring the accuracy of vibration prediction and warning.

[0080] Please refer to the attached Figure 1 In a preferred embodiment of the present invention, the distributed adaptive dynamic threshold control method comprises the following steps:

[0081] Based on historical data, different extreme weather scenarios are simulated, and initial vibration thresholds are set for various weather conditions. By analyzing historical data and simulating different extreme weather scenarios, the system can accurately understand the typical impact characteristics of extreme weather on line vibration, thereby setting reasonable initial thresholds for different weather conditions. Setting the initial threshold enables the system to have basic early warning capabilities under different weather conditions, which helps to improve the accuracy of early warnings and the system's rapid response capabilities;

[0082] The deep Q learning algorithm is used to optimize the threshold and adjust the threshold strategy through feedback. The deep Q learning algorithm learns and updates the feedback of different threshold strategies and gradually adjusts the optimal threshold setting scheme. Compared with the fixed threshold setting method, deep Q learning can adapt to weather changes in real time and continuously optimize the threshold strategy, effectively improving the sensitivity and adaptability of the system;

[0083] In actual monitoring, the warning threshold is adaptively adjusted in real time according to weather changes. By adaptively adjusting the warning threshold in real time, the system can respond to changes in different weather conditions more agilely. At the same time, the application of feedback optimization mechanism makes the threshold adaptive adjustment more accurate, improves the reliability and timeliness of the warning, and effectively enhances the system's protection capabilities and safety in extreme weather conditions.

[0084] Please see attached Figure 1 In a preferred embodiment of the present invention, the threshold optimized deep Q learning algorithm training includes the following steps:

[0085] In the simulation environment, the warning threshold is continuously optimized by adjusting the threshold and receiving system feedback. By adjusting the threshold and collecting feedback information in the simulation environment, the system can obtain a large amount of data on the effect of threshold adjustment without actually putting it into operation. This process significantly speeds up the threshold optimization speed, reduces the risk of actual trial and error, and provides stable data support for subsequent Q learning model training, making the system more adaptable and accurate in actual operation.

[0086] According to the performance of the system in terms of false alarm rate and missed alarm rate, the reward value in the Q table is updated. By updating the reward value in the Q table based on the performance of the false alarm rate and missed alarm rate, the system can accurately record the effects of different threshold settings and form a set of highly adaptable threshold optimization strategies. The reward mechanism of the Q learning algorithm enables the model to be gradually optimized in repeated training, and automatically selects the appropriate threshold setting when facing different weather conditions, thereby ensuring the sensitivity of the warning while reducing the probability of false alarms and missed alarms, and improving the reliability and stability of the warning system;

[0087] Through iterative training, the system converges to the optimal threshold control strategy. Through multiple iterative training, the system's deep Q learning model can gradually converge to the optimal threshold control strategy in different weather scenarios, ensuring that the system can automatically select the appropriate warning threshold setting under various extreme weather conditions, thereby effectively reducing the risk of false alarms and missed alarms, and improving the adaptability and reliability of the system in the actual operating environment, making the vibration monitoring and early warning system more stable and efficient in extreme weather conditions.

[0088] Please see attached Figure 1 In a preferred embodiment of the present invention, generating and issuing warning information includes the following steps:

[0089] The system analyzes the comprehensive data received by the central server to determine vibration trends and potential risks. Multimodal data integration and trend prediction can effectively identify abnormal development trends of vibration under extreme weather conditions, providing a scientific basis for early warning, thereby improving the system's early warning effect and the safety of transmission lines;

[0090] When the vibration trend exceeds the dynamic threshold or the future vibration trend reaches the preset risk value, an early warning signal is automatically generated. The preset risk value prediction provides an advance warning, so that the system can issue a warning before extreme weather affects the line, which helps relevant personnel take preventive measures in advance, thereby ensuring the safe operation of the line in extreme weather;

[0091] The early warning information is released to relevant operation and maintenance personnel so that preventive measures can be taken in advance. The multi-channel information push method enhances the reliability of information transmission. Even under poor communication conditions, operation and maintenance personnel can obtain early warning information in a timely manner. The real-time update and feedback mechanism further optimizes the emergency response process of the early warning system, allowing the monitoring center to track the progress of early warning processing in real time, ensuring that operation and maintenance personnel can accurately judge and take necessary protective measures in extreme weather, thereby effectively reducing the risk of line failures and ensuring the safe and stable operation of the power system.

[0092] Please see attached Figure 1 In a preferred embodiment of the present invention, the method further comprises:

[0093] Receive weather forecast data in real time and input it into the multimodal deep learning model. By receiving weather forecast data in real time and inputting it into the multimodal deep learning model, the system can have a more comprehensive understanding of the potential impact of environmental factors on vibration. With the help of weather forecast data, the system has the ability to predict future climate change and can detect potential risks earlier before extreme weather arrives, thereby enhancing the system's early warning efficiency and risk resistance;

[0094] The model is adaptively updated using the latest monitoring data to improve the vibration prediction accuracy under extreme weather conditions. The introduction of adaptive updates enables the system to maintain good prediction performance in complex extreme weather environments and avoids the decline in prediction accuracy due to changes in data distribution. Model adaptive adjustment further enhances the system's adaptability, helps the system to operate stably in extreme weather, and provides continuous protection for the safety of power transmission;

[0095] The latest monitoring data is regularly used as new training samples to re-optimize the deep learning model and dynamic threshold control strategy. Through periodic updates, the system can cope with the impact of long-term environmental changes and line aging, and avoid the model's prediction accuracy being reduced due to data aging. The update of the dynamic threshold strategy makes the system more flexible and robust in different extreme weather scenarios, thereby improving the safety and reliability of the system.

[0096] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A transmission line vibration monitoring and early warning method for extreme weather, characterized in that: The following steps are involved: Deploy a multi-level sensor network to monitor the vibration and environmental conditions of the transmission line; Process and transmit monitoring data in real time through edge computing and 5G networks; Use multimodal deep learning models to predict trends in vibration data; Adjust the warning threshold through a distributed adaptive dynamic threshold control method; Generate and issue warning messages when abnormal vibrations or high-risk trends are detected.

2. The transmission line vibration monitoring and early warning method for extreme weather according to claim 1 is characterized in that: The deployment of a multi-level sensor network includes the following steps: Deploy fiber Bragg grating sensors at high-risk intersections of transmission lines to monitor vibration and stress changes; Deploy acceleration sensors and meteorological sensors in areas that are most affected by weather to obtain vibration and meteorological data; Low-power vibration sensors are deployed in ordinary line sections to achieve wide-area coverage of line vibration.

3. The transmission line vibration monitoring and early warning method for extreme weather according to claim 1 is characterized in that: Real-time processing and transmission of monitoring data includes the following steps: Deploy edge computing nodes near sensor nodes at each level to process sensor data locally; Data cleaning, anomaly detection and feature extraction are performed through edge computing nodes; Use the 5G network to transmit processed key data to the cloud central server in real time.

4. The transmission line vibration monitoring and early warning method for extreme weather according to claim 1 is characterized in that: Real-time processing and transmission of monitoring data through edge computing and 5G networks includes the following steps: Clean and denoise the collected sensor data to remove outliers; Use Kalman filtering method to smooth the data; Extract key features from vibration data including amplitude, jerk, and frequency.

5. The transmission line vibration monitoring and early warning method for extreme weather according to claim 1 is characterized in that: The multimodal deep learning model performs trend prediction on vibration data including the following steps: Use LSTM model to process time series data to predict vibration trends; Use CNN models to process extreme weather image data to identify weather change characteristics; The output results of LSTM and CNN are combined in the fusion layer to generate the final vibration trend prediction results.

6. The transmission line vibration monitoring and early warning method for extreme weather according to claim 1 is characterized in that: The training of a multimodal deep learning model consists of the following steps: The model was trained using historical vibration and meteorological data; Gradient descent and Adam optimization algorithms are used to optimize model parameters; The best model parameters were selected through cross-validation to improve prediction accuracy.

7. The transmission line vibration monitoring and early warning method for extreme weather according to claim 1 is characterized in that: The distributed adaptive dynamic threshold control method comprises the following steps: Simulate different extreme weather scenarios based on historical data and set initial vibration thresholds for various weather conditions; Use deep Q learning algorithm to optimize threshold and adjust threshold strategy through feedback; In actual monitoring, the warning threshold is adaptively adjusted in real time according to weather changes.

8. The transmission line vibration monitoring and early warning method for extreme weather according to claim 1 is characterized in that: The training of the threshold-optimized deep Q-learning algorithm consists of the following steps: In the simulation environment, the warning threshold is continuously optimized by adjusting the threshold and receiving system feedback; Update the reward value in the Q table based on the system's performance in terms of false positive rate and false negative rate; Through iterative training, the optimal threshold control strategy is converged.

9. The transmission line vibration monitoring and early warning method for extreme weather according to claim 1, characterized in that: Generating and publishing warning information includes the following steps: The system analyzes the comprehensive data received from the central server to determine vibration trends and potential risks; Automatically generate warning signals when the vibration trend exceeds the dynamic threshold or the future vibration trend reaches the preset risk value; The warning information will be released to relevant operation and maintenance personnel so that preventive measures can be taken in advance.

10. The transmission line vibration monitoring and early warning method for extreme weather according to claim 1, characterized in that: The method also includes: Receive weather forecast data in real time and feed it into a multimodal deep learning model; Adaptively update the model using the latest monitoring data to improve vibration prediction accuracy under extreme weather conditions; Regularly use the latest monitoring data as new training samples to re-optimize the deep learning model and dynamic threshold control strategy.

Images