Power transmission line monitoring method based on edge calculation

Abstract

The invention provides a power transmission line monitoring method based on edge calculation, and the method comprises the steps: deploying an intelligent edge calculation unit at a key node of a power transmission line, and collecting the icing thickness, environment temperature and humidity, cable tension and other multi-source data in real time; processing the data by using an edge calculation fusion algorithm, generating a real-time early warning signal and an abnormal data set, and calculating a risk assessment index by using an adaptive neural network optimization algorithm in combination with a historical environment; when the risk assessment index reaches an early warning threshold value, data are uploaded to a remote monitoring center through a communication module, a fault point is quickly positioned, a fault diagnosis report is generated, and real-time monitoring of the state of the power transmission line and intelligent management of safe operation are achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of power transmission lines, and in particular to a power transmission line monitoring method based on edge computing. Background Art

[0002] Transmission lines are an important part of power intelligent edge computing units, responsible for transmitting electricity from power plants to user terminals. Their safe and stable operation is crucial to ensuring power supply. In traditional transmission line monitoring intelligent edge computing units, periodic manual inspections or simple sensor monitoring are usually used, which have obvious limitations. Manual inspections are costly and inefficient, and cannot achieve continuous real-time monitoring of line status. Simple sensor monitoring can often only provide limited data and lacks the ability to comprehensively analyze complex environmental factors, making it difficult to accurately predict and assess the safety risks of transmission lines. When a transmission line fails, the fault location technology is not accurate enough, resulting in time-consuming and inefficient maintenance work.

[0003] Existing technologies also show slow response speed and low processing efficiency when dealing with sudden environmental events such as ice cover, lightning strikes, etc. on power transmission lines. Therefore, there is an urgent need for a monitoring intelligent edge computing unit that can monitor the status of power transmission lines in real time, respond quickly, and accurately handle various complex situations, so as to improve the reliability of power transmission lines and the overall security of the power grid. Summary of the invention

[0004] 1. Technical issues to be solved

[0005] In view of the shortcomings of the prior art, the present invention provides a transmission line monitoring method based on edge computing, which solves the problem of how to achieve real-time monitoring of transmission lines, accurately assess risk status and quickly respond to abnormal events.

[0006] (II) Technical solution

[0007] To achieve the above objectives, the present invention is implemented through the following technical solutions: A transmission line monitoring method based on edge computing, comprising:

[0008] S1. Deploy intelligent edge computing units at key nodes of the transmission line, wherein the intelligent edge computing units include a multimodal sensing module, a data processing module, and a low-power communication module; collect multi-source data through the multimodal sensing module, including ice thickness, ambient temperature and humidity, cable tension, vibration frequency, and lightning current, and add timestamps to the collected data for time synchronization and dynamic analysis;

[0009] S2. The intelligent edge computing unit processes the collected multi-source data in real time and uses edge computing fusion algorithms to generate real-time warning signals and abnormal data sets;

[0010] S3. In the edge computing unit, based on the generated warning signals, abnormal data sets and multimodal data streams, and combined with historical environmental data, a risk assessment index reflecting the overall operating status of the transmission line is calculated through an adaptive neural network optimization algorithm;

[0011] S4. When the risk assessment index reaches the warning threshold, the risk assessment index, warning signal and abnormal data set are uploaded to the remote monitoring center through the low-power communication module, and the emergency signal is transmitted to the surrounding edge computing units to achieve distributed collaborative response;

[0012] S5. When an abnormal line interruption is detected, the fault point is quickly located by combining the abnormal data set and risk assessment index, and a fault type diagnosis report is generated and uploaded to the remote monitoring center;

[0013] S6. In the remote monitoring center, risk indexes, warning signals and abnormal data sets are combined to build a panoramic picture of the safety of transmission line operation, and the power supply path is dynamically adjusted according to the panoramic picture of safety.

[0014] Preferably, in the intelligent edge computing unit for monitoring transmission lines, key nodes are located at branch points of the line, sections susceptible to environmental influences, and areas with frequent historical failures; for example, at branch points where transmission lines are connected to substations or distribution networks, these locations are prone to failures due to complex structures and stress concentration; mountainous areas, coastal areas or forest areas susceptible to environmental influences will also become key nodes for monitoring due to harsh natural conditions such as strong winds, lightning strikes, and salt spray erosion; in addition, through historical data analysis, areas where failures frequently occur are identified, and transmission lines in these areas require more detailed monitoring; at these key nodes, the intelligent edge computing unit is deployed on the tower or pole of the transmission line, installed in a position that is easy to maintain and can monitor the status of the line; a multimodal sensing module is integrated inside the unit, including but not limited to temperature and humidity sensors, vibration sensors, tension sensors, and current sensors, to collect various status information of the line in real time; the data collected by each sensor will be given a timestamp to record the exact time when the data was collected, ensuring the time consistency of the data and the accuracy of subsequent analysis.

[0015] Preferably, the processing process of the intelligent edge computing unit includes: dynamically synchronizing multi-source data with different timestamps through an adaptive time weighting factor to generate a multimodal data stream with a unified time series for feature extraction and modeling; based on the multimodal data stream after dynamic synchronization, using Gaussian variational distribution modeling technology to convert the nonlinear relationship of multi-source data into a joint probability distribution model, and output a dynamic probability matrix to represent the real-time risk status of the transmission line; using the risk status data in the probability matrix, combined with sliding window analysis to extract local fluctuation characteristics, and dynamically identify sudden lightning strikes or abnormal ice cover changes; when the local fluctuation exceeds the preset threshold, a real-time warning signal is generated, and an abnormal data set related to the abnormality is output at the same time.

[0016] Preferably, in the transmission line monitoring intelligent edge computing unit, the edge computing unit is responsible for analyzing the timestamps of each data source; the intelligent edge computing unit collects data from different sensors, such as temperature, humidity and cable tension, each of which carries a timestamp indicating the time when the data was collected, and the intelligent edge computing unit identifies the relative position of each data point in the time dimension by calculating the difference between these timestamps, i.e., the deviation; for example, if the data timestamp of a temperature sensor is 0.5 seconds earlier than the data timestamp of a humidity sensor, the intelligent edge computing unit will record this deviation; the intelligent edge computing unit calculates an adaptive time weighting factor based on the degree of these deviations and the characteristics of each data source; the calculation of the weighting factor takes into account the time correlation and acquisition frequency of the data, and its core lies in dynamically adjusting the weights of different data sources to alleviate the time differences caused by acquisition speed or Alignment problems caused by different accuracies; for example, if the tension sensor has a low data collection frequency but high real-time requirements, the intelligent edge computing unit will give it a larger weighting factor to ensure that it has a higher influence in the data synchronization process; after completing the weighted calculation, the intelligent edge computing unit maps these multi-source data to a unified time series through interpolation and time window adjustment methods to form a multimodal data stream; such as linear interpolation or polynomial interpolation, and dynamic adjustment of the time window to ensure the continuity and consistency of the data on the time axis; for example, if a sensor is missing data within a time window, the intelligent edge computing unit will estimate the data value at the missing time through interpolation based on the previous and subsequent timestamps and data values ​​to ensure the integrity of the data stream; convert heterogeneous data from different sensors into a synchronized, unified format data stream.

[0017] Preferably, the Gaussian variational distribution modeling technique makes a Gaussian distribution assumption for each data source to obtain a probability density function of each data source. Where i is the number of data sources, x is the set of multi-source data, μ i is the mean of the ith data source, is the variance; through the Gaussian process regression method, the nonlinear relationship of x is modeled as a joint probability distribution Where θ is the model parameter, including the mean and variance of all data sources; the kernel function k(x,x') is used to represent the similarity between data points; the covariance matrix is ​​calculated by the kernel function in is the noise level, I is the identity matrix; the joint probability distribution model is combined with the covariance matrix to output the dynamic probability matrix where |C| is the determinant of the covariance matrix, C -1 It is the inverse of the covariance matrix; it converts the nonlinear relationship of multi-source data into a joint probability distribution model and outputs a dynamic probability matrix.

[0018] By monitoring the data in the probability matrix in real time, local fluctuation characteristics are extracted through sliding window analysis; the data change trend of the transmission line within a certain period of time is captured to identify abnormal behavior; when the monitored local fluctuation characteristics exceed the preset safety threshold, the intelligent edge computing unit will immediately generate a real-time warning signal and send the signal to the remote monitoring center and surrounding intelligent edge computing units through the low-power communication module. The intelligent edge computing unit extracts data related to the anomaly to form an abnormal data set.

[0019] In the intelligent edge computing unit for monitoring transmission lines, the intelligent edge computing unit sets a time window, which slides in the data stream to capture and analyze local data fluctuations within the window; for example, the intelligent edge computing unit sets a 30-second sliding window, which slides forward every 5 seconds to continuously monitor slight changes in the line status; by calculating the mean and variance statistical characteristics of the data in the window, the intelligent edge computing unit identifies the trend of data changes; if the variance of cable tension suddenly increases within a certain period of time, it indicates that there may be line abnormalities caused by icing or strong winds; the intelligent edge computing unit presets a series of safety thresholds based on historical data and expert experience. For temperature data, it is assumed that the normal operating temperature range of the transmission line is -10 degrees Celsius to 40 degrees Celsius; by analyzing historical data, the intelligent edge computing unit sets the threshold are the upper and lower boundaries of this range; if the temperature of a certain section of the line suddenly drops to -15 degrees Celsius or rises to 45 degrees Celsius, indicating extreme weather impact or equipment failure, the intelligent edge computing unit will trigger a warning signal; when local fluctuation characteristics, such as the temperature mutation rate, exceed these thresholds, the intelligent edge computing unit will consider that there is abnormal behavior in the transmission line and immediately trigger a real-time warning signal; for example, if the tension variance of three consecutive sliding windows exceeds the preset threshold, the intelligent edge computing unit will generate a warning signal; this warning signal and data related to the abnormality, such as tension value, timestamp and location information, are quickly sent to the remote monitoring center and surrounding intelligent edge computing units through the low-power communication module, and timely response measures are taken; real-time monitoring of the status of the transmission line, rapid response to potential safety hazards, and ensuring the stable operation of the power grid.

[0020] The generated warning signals, abnormal feature data sets and multimodal data streams are input into the adaptive neural network optimization algorithm. The input layer of the neural network receives the trigger frequency of the warning signal, the risk state parameters in the probability matrix and the collection of historical multi-source data. These data are input into the neural network and converted through the activation function to output a comprehensive risk assessment index R = ω(f + ∑P ij +H), where f is the triggering frequency of the warning signal, H is the collection of historical multi-source data, and R reflects the real-time risk status of the transmission line.

[0021] When an abnormal interruption occurs in a transmission line, the remote monitoring center mobilizes various intelligent edge computing units; assuming that after a strong storm, an abnormality is detected at a key node of the transmission line, the intelligent edge computing unit immediately begins to analyze the multi-source data when the abnormality occurs, including ice thickness, cable tension, vibration frequency, etc. By analyzing these data and comparing them with the risk status in the dynamic probability matrix, it is found that the readings of ice thickness and cable tension rise sharply when the abnormality occurs, and the vibration frequency also shows an abnormal peak; these data and results will be comprehensively analyzed. If it is found that the increase in ice thickness and the abnormal change in cable tension are significantly different from the normal state, this indicates that the fault is likely to be a line overload caused by ice; through the abnormal data collected by the specific intelligent edge computing unit, the data related to the fault is quickly locked, and the fault point is determined to be located near a branch point of the transmission line.

[0022] The intelligent edge computing unit generates a detailed fault type diagnosis report, which includes the nature, location, possible causes and recommended repair measures of the fault; this report is uploaded to the remote monitoring center through the low-power communication module; at the monitoring center, the operator combines the risk index, early warning signals and abnormal data sets to build a panoramic view of the transmission line operation safety; through this panoramic view, the operator can intuitively see the operating status of the entire transmission line and dynamically adjust the power supply path according to the safety panoramic view, such as switching to the backup line, to ensure the continuity and safety of power supply.

[0023] (III) Beneficial effects

[0024] The present invention provides a transmission line monitoring method based on edge computing, which has the following beneficial effects:

[0025] 1. The present invention realizes real-time monitoring of line status by deploying intelligent edge computing units at key nodes of transmission lines, and adopts Gaussian variational distribution modeling technology to construct a joint probability distribution model, output a dynamic probability matrix, and accurately evaluate the real-time risk status of transmission lines; it provides more comprehensive and accurate risk assessment results, and provides reliable guarantee for the safe operation of power intelligent edge computing units.

[0026] 2. The present invention identifies sudden lightning strikes or abnormal ice cover changes and generates real-time warning signals; calculates a risk assessment index that reflects the overall operating status of the transmission line; when the risk assessment index reaches the warning threshold, the intelligent edge computing unit will immediately upload the risk index, warning signal and related abnormal data sets to the remote monitoring center, and transmit emergency signals to the surrounding edge computing units to achieve distributed collaborative response, thereby quickly responding to abnormal events and avoiding accidents.

[0027] 3. The present invention quickly locates the fault point based on the abnormal data set and risk assessment index. In the remote monitoring center, it combines the risk index, early warning signal and abnormal data set to build a safe panoramic view of the transmission line operation, and dynamically adjusts the power supply path according to the safe panoramic view, thereby improving the intelligence level of the intelligent edge computing unit and providing more accurate decision-making support for the safe operation of the power intelligent edge computing unit. DETAILED DESCRIPTION

[0028] The technical solutions in the embodiments of the present invention will be described clearly and completely below. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0029] The embodiment of the present invention provides a transmission line monitoring method based on edge computing to achieve comprehensive and accurate monitoring and effective management of the transmission line status.

[0030] Deploying intelligent edge computing units at key nodes of transmission lines is the foundation of the entire monitoring system. The selection of these key nodes is targeted. For example, due to changes in current distribution and line direction, branch points in the line have complex stress distribution and are prone to problems such as loose connections or discharges. Areas that are susceptible to environmental influences, such as some lines crossing mountainous areas, often face severe weather such as strong winds, heavy rains, low temperatures and freezing, as well as areas with frequent lightning activity, and are more likely to be struck by lightning. There are also areas with frequent historical failures, which are derived from in-depth analysis of historical data. These places have some potential risk factors that have not yet been fully resolved. In a section of transmission line in a certain mountainous area, there is a branch point located at the wind outlet of the valley, and there are many trees around it, with frequent lightning activity. This place is identified as one of the key nodes. The intelligent edge computing unit is installed on the transmission line tower at the branch point, and the installation position is about 15 meters above the ground, which is convenient for maintenance personnel to carry out daily inspections and equipment maintenance, and can ensure that the sensor accurately collects line status information.

[0031] The multimodal sensor module integrated inside the intelligent edge computing unit starts working and collects multi-source data; the temperature and humidity sensor collects ambient temperature and humidity data every 10 minutes. For example, during the high temperature period in summer, the ambient temperature is recorded to be 40 degrees Celsius and the humidity is 60%. These data can reflect the heat and humidity conditions of the line environment, which is of great significance for judging whether there is a risk of overheating of the line and the risk of insulation moisture; the vibration sensor monitors the vibration frequency of the line in real time. When strong winds pass by, the vibration frequency is detected to increase significantly from the normal 5Hz to 15Hz, which indicates that the transmission line is subject to greater external The tension sensor continuously measures the cable tension, which is maintained at around 5000N during normal operation. If the tension changes suddenly due to icing or other factors, such as if the tension rises to 6000N after a slight icing, this needs to be paid attention to. The current sensor closely monitors the current during lightning weather. In a lightning strike, a peak current of 100kA was detected. This data helps to assess the impact of lightning on the line. Each item of data collected is timestamped to ensure the time synchronization of the data in subsequent processing and the accuracy of dynamic analysis.

[0032] The intelligent edge computing unit processes the collected multi-source data in real time. First, the data with different timestamps are dynamically synchronized through an adaptive time weighting factor. Assuming that the timestamp of the temperature sensor data lags behind the timestamp of the tension sensor data by 2 seconds, the weighting factor of the temperature sensor data is calculated to be 0.3 and the weighting factor of the tension sensor data is 0.7 according to the data source characteristics and acquisition frequency of the two. After weighted calculation and interpolation processing, they are mapped to a unified time series to form a multimodal data stream. Then, based on the multimodal data stream after dynamic synchronization, the Gaussian variational distribution modeling technology is adopted. A Gaussian distribution assumption is made for each data source to calculate the probability density function. For example, for the cable tension data, it is assumed that its mean is 5000N and the variance is 500. According to The probability density function is calculated by the formula; then the nonlinear relationship of multi-source data is modeled as a joint probability distribution through the Gaussian process regression method, and the covariance matrix is ​​calculated using the kernel function, and finally the dynamic probability matrix is ​​output; if an element in the dynamic probability matrix shows that the risk status value suddenly increases and exceeds the normal range, it indicates that the line may have an abnormal situation; at the same time, the local fluctuation characteristics are extracted by combining the sliding window analysis, and the sliding window is set to 20 seconds, sliding every 4 seconds, and calculating the mean and variance statistical characteristics of the data in the window; when the variance of the cable tension in a certain sliding window increases from the normal 100 to 500, exceeding the preset threshold, the intelligent edge computing unit immediately generates a real-time warning signal, and extracts the tension data, timestamp and location information related to the anomaly to form an abnormal data set.

[0033] In the edge computing unit, based on the generated warning signals, abnormal data sets and multimodal data streams, combined with historical environmental data, an adaptive neural network optimization algorithm is used to calculate the risk assessment index; the input layer of the neural network receives the triggering frequency of the warning signal, such as the warning signal was triggered 3 times in the past hour; the risk status parameters in the probability matrix, such as the risk status value of a key node is 0.8; and a collection of historical multi-source data, such as the line operation data of the area in a similar environment in the past week; after these data are input into the neural network, they are converted by the activation function and a comprehensive risk assessment index is output. Assuming that the calculated risk assessment index is 0.75, it indicates that the transmission line is at a high risk and timely measures need to be taken.

[0034] When the risk assessment index reaches the warning threshold, assuming it is set to 0.7, the intelligent edge computing unit uploads the risk index, warning signal and related abnormal data set to the remote monitoring center through the low-power communication module, and transmits emergency signals to the surrounding intelligent edge computing units at the same time; for example, in a rainstorm, the risk assessment index reaches 0.8, and the intelligent edge computing unit quickly uploads the data. After receiving the emergency signal, the surrounding intelligent edge computing units strengthen the monitoring of their responsible areas and make preparations for coordinated response; after receiving the data, the remote monitoring center has a more comprehensive understanding of the operating status of the entire transmission line.

[0035] When an abnormal line interruption is detected, the fault point is quickly located by combining the abnormal data set and the risk assessment index. For example, the intelligent edge computing unit at a key node detects an abnormal interruption. By analyzing the abnormal data set, it is found that the ice thickness data increases rapidly before the interruption and the cable tension rises sharply. Combined with the risk assessment index, it shows that the area is in a high-risk state, and it is judged that the fault point may be near the node. Further analysis shows that the ice thickness in the area reaches 10mm, which exceeds the normal range, causing the cable tension to exceed the limit and eventually causing the line to be interrupted. A detailed fault type diagnosis report is generated, including the nature of the fault being ice overload causing line interruption, the location being near a specific branch point, the possible cause being bad weather causing excessive ice growth, and the recommended maintenance measures being to clear the ice first and then check the line damage, etc., and uploaded to the remote monitoring center.

[0036] In the remote monitoring center, a panoramic view of the transmission line operation safety is constructed by combining risk indexes, warning signals and abnormal data sets. Through the panoramic view, operators can intuitively see the operating status of each key node of the entire transmission line, including which areas are at high risk and where warnings have appeared. The power supply path is dynamically adjusted according to the safety panoramic view. For example, when a high-risk warning appears on a major transmission line, part of the load is promptly switched to the backup line to ensure the continuity and safety of the power supply, realize the intelligent management of the transmission line, and effectively improve the reliability and stability of the transmission line operation.

[0037] 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 monitoring method based on edge computing, characterized in that; The following steps are involved: S1. Deploy intelligent edge computing units at key nodes of the transmission line, wherein the intelligent edge computing units include a multimodal sensing module, a data processing module, and a low-power communication module; collect multi-source data through the multimodal sensing module, including ice thickness, ambient temperature and humidity, cable tension, vibration frequency, and lightning current, and add timestamps to the collected data for time synchronization and dynamic analysis; S2. The intelligent edge computing unit processes the collected multi-source data in real time and uses edge computing fusion algorithms to generate real-time warning signals and abnormal data sets; S3. In the edge computing unit, based on the generated warning signals, abnormal data sets and multimodal data streams, and combined with historical environmental data, a risk assessment index reflecting the overall operating status of the transmission line is calculated through an adaptive neural network optimization algorithm; S4. When the risk assessment index reaches the warning threshold, the risk assessment index, warning signal and abnormal data set are uploaded to the remote monitoring center through the low-power communication module, and the emergency signal is transmitted to the surrounding edge computing units to achieve distributed collaborative response; S5. When an abnormal line interruption is detected, the fault point is quickly located by combining the abnormal data set and risk assessment index, and a fault type diagnosis report is generated and uploaded to the remote monitoring center; S6. In the remote monitoring center, risk indexes, warning signals and abnormal data sets are combined to build a panoramic picture of the safety of transmission line operation, and the power supply path is dynamically adjusted according to the panoramic picture of safety.

2. The power transmission line monitoring method based on edge computing according to claim 1, characterized in that: The key nodes of the transmission line are branch points of the line, sections susceptible to environmental impacts, and areas with frequent historical failures; intelligent edge computing units are deployed at the selected key nodes; the intelligent edge computing units internally integrate a multi-modal sensing module, a data processing module, and a low-power communication module; The multimodal sensing module collects multi-source data and adds a timestamp to each collected data.

3. The power transmission line monitoring method based on edge computing according to claim 1, characterized in that: The processing process of the intelligent edge computing unit includes: S2.1: Dynamically synchronize multi-source data with different timestamps through adaptive time weighting factors to generate a multimodal data stream with a unified time series for feature extraction and modeling; S2.2: Based on the multimodal data stream after dynamic synchronization, the Gaussian variational distribution modeling technology is used to convert the nonlinear relationship of multi-source data into a joint probability distribution model, and the dynamic probability matrix is ​​output to represent the real-time risk status of the transmission line; S2.3: Through the risk status in the probability matrix, combined with the sliding window analysis, the local fluctuation characteristics are extracted to dynamically identify sudden lightning events or abnormal ice cover changes; when the local fluctuation exceeds the preset threshold, a real-time warning signal is generated and the abnormal data set is output at the same time.

4. The power transmission line monitoring method based on edge computing according to claim 3 is characterized in that: The edge computing unit performs a preliminary analysis on the timestamps of each data source, identifies the relative positions of different data in the time dimension by calculating the deviations between the timestamps, and calculates an adaptive time weighting factor based on the degree of deviation and the characteristics of the data source; The weighting factor dynamically adjusts the weights of different data sources according to the time of the data; after completing the weighted calculation, the multi-source data is mapped to a multimodal data stream with a unified time series through interpolation and time window adjustment.

5. The power transmission line monitoring method based on edge computing according to claim 3 is characterized in that: The Gaussian variational distribution modeling technique makes a Gaussian distribution assumption on each data source to obtain the probability density function of each data source. Where i is the number of data sources, x is the set of multi-source data, μ i is the mean of the ith data source, is the variance; through the Gaussian process regression method, the nonlinear relationship of x is modeled as a joint probability distribution Where θ is the model parameter, including the mean and variance of all data sources; the kernel function k(x,x') is used to represent the similarity between data points; the covariance matrix is ​​calculated by the kernel function in is the noise level, I is the identity matrix; the joint probability distribution model is combined with the covariance matrix to output the dynamic probability matrix (x-μ)), where |C| is the determinant of the covariance matrix, C -1 It is the inverse of the covariance matrix; it converts the nonlinear relationship of multi-source data into a joint probability distribution model and outputs a dynamic probability matrix.

6. The power transmission line monitoring method based on edge computing according to claim 5, characterized in that: The data in the probability matrix is ​​monitored in real time, and local fluctuation characteristics are extracted through sliding window analysis; the data change trend of the transmission line within a certain period of time is captured to identify abnormal behavior; when the monitored local fluctuation characteristics exceed the preset safety threshold, the intelligent edge computing unit immediately generates a real-time warning signal, and sends the signal to the remote monitoring center and the surrounding edge computing units through the low-power communication module. The intelligent edge computing unit extracts data related to the anomaly to form an abnormal data set.

7. The power transmission line monitoring method based on edge computing according to claim 1, characterized in that: The generated warning signals, abnormal data sets and multimodal data streams are input into the adaptive neural network optimization algorithm. The input layer of the neural network receives the trigger frequency of the warning signal, the risk state parameters in the probability matrix and the collection of historical multi-source data. The data is input into the neural network and converted through the activation function to output a comprehensive risk assessment index R = ω(f + ∑P ij +H), where f is the triggering frequency of the warning signal, H is the collection of historical multi-source data, and R reflects the real-time risk status of the transmission line.

8. The power transmission line monitoring method based on edge computing according to claim 1, characterized in that: When an abnormal line interruption is detected, the abnormal data set and risk assessment index are generated by the intelligent edge computing unit deployed at the key nodes of the transmission line; the time synchronization analysis is performed on the timestamps attached to the multi-source data collected by the multimodal sensing module, combined with the real-time risk status reflected by the dynamic probability matrix; through a comprehensive analysis of the data and results, according to the characteristics of different data when the fault occurs and its difference from the normal state, the data related to the fault is quickly locked, so as to locate the fault point, and then generate a fault type diagnosis report and upload it to the remote monitoring center.

Images