A smart grid data communication optimization method based on edge computing

Abstract

The present invention provides a smart grid data communication optimization method based on edge computing, which relates to the field of smart grid technology, including: obtaining the inspection topology network of the target power grid; performing multi-angle communication analysis to obtain K sets of transmission equipment communication feature tuples and L inspection site communication feature tuples; performing edge jump identification to obtain M edge topology sub-networks; performing sub-network service resource balancing configuration to obtain M edge computing servers; using inspection drones and inspection robots to perform inspections to obtain L site inspection video data; distributing to M edge computing servers for data preprocessing, uploading to a cloud computing center for inspection data analysis, and obtaining inspection analysis results. The present invention solves the technical problem that the prior art power grid monitoring usually relies on centralized data processing, resulting in data transmission delays and excessive network load, which in turn leads to low data processing efficiency.

Description

Technical Field

[0001] The present invention relates to the field of smart grid technology, and in particular to a smart grid data communication optimization method based on edge computing. Background Art

[0002] In traditional smart grids, data typically needs to be transmitted from edge devices to a central cloud computing center for centralized processing. This approach results in high transmission latency, especially in scenarios with high real-time monitoring and response requirements. For example, status monitoring and fault diagnosis of transmission lines require fast feedback, and the high latency in traditional methods may affect the security and stability of the power grid. This centralized processing mode often leads to excessive use of network bandwidth and increased network load. In smart grids, especially when large-scale data collection and processing are required, the demand for data transmission increases dramatically, which leads to reduced transmission efficiency. In addition, bandwidth limitations affect data transmission speed and reliability, which in turn affects the real-time monitoring and decision-making capabilities of the power grid. Summary of the Invention

[0003] This application provides a smart grid data communication optimization method based on edge computing, aiming to solve the technical problem that power grid monitoring in the existing technology usually relies on centralized data processing, resulting in data transmission delays and excessive network load, which in turn leads to low data processing efficiency.

[0004] The present application discloses a smart grid data communication optimization method based on edge computing, the method comprising: obtaining a patrol topology network of a target power grid, wherein the patrol topology network comprises K transmission lines, L patrol sites and a cloud computing center, and the K transmission lines comprise K transmission equipment sets and K transmission paths; performing multi-angle communication analysis on the K transmission equipment sets and the L patrol sites to obtain K transmission equipment communication feature tuple sets and L patrol site communication feature tuples; performing edge jump identification on the patrol topology network based on the K transmission equipment communication feature tuple sets and the L patrol site communication feature tuples to obtain M edge topology subnetworks, wherein the M edge topology subnetworks comprise M edge communication coefficients; performing edge jump identification on the patrol topology network based on the K transmission equipment communication feature tuple sets and the L patrol site communication feature tuples to obtain M edge topology subnetworks, wherein the M edge topology subnetworks comprise M edge communication coefficients; performing edge jump identification on the patrol topology network based on the K transmission equipment communication feature tuple sets and the L patrol site communication feature tuples M edge communication coefficients are used to balance the sub-network service resources of the M edge topology sub-networks to obtain M configured edge computing servers; K inspection drones are used to inspect the K transmission lines according to the K transmission paths respectively to obtain K line inspection video data sets, and L inspection robots are used to inspect the L inspection sites according to the preset station inspection routes to obtain L site inspection video data; based on the M edge topology sub-networks, the K line inspection video data sets and the L site inspection video data are distributed to M edge computing servers for data preprocessing, and the preprocessed data are uploaded to the cloud computing center for inspection data analysis according to the preset communication protocol to obtain the inspection analysis results of the target power grid.

[0005] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0006] By obtaining the inspection topology network of the target power grid, including the layout of transmission lines, inspection sites and cloud computing centers, a comprehensive understanding of the power grid structure is provided, which lays the foundation for subsequent communication analysis and edge computing optimization, and ensures the accuracy of data collection and processing; multi-angle communication analysis of transmission equipment and inspection sites is carried out to obtain a set of communication feature tuples. This multi-dimensional data analysis can identify the communication characteristics of equipment and sites, establish comprehensive communication characteristics, and help optimize data flow and network load; edge jump identification of the inspection topology network based on the communication feature tuple set can divide the power grid into multiple edge topology sub-networks and calculate the edge communication coefficient. This method can effectively identify key edge nodes and communication paths in the network, optimize the path selection of data transmission, and improve data processing efficiency and transmission stability; sub-network service resources are balanced based on the edge communication coefficient, and the resource allocation of edge computing servers is optimized, which ensures the balanced use of computing resources and improves The server's processing capacity and response speed are improved; inspection drones and inspection robots are used to inspect transmission lines and inspection sites to obtain inspection video data. This method can provide real-time equipment status and safety monitoring, and improve the automation level and accuracy of inspections. Drones cover a wide area of ​​the transmission lines, while robots conduct detailed inspections inside the sites, ensuring the comprehensiveness and meticulousness of the data; data preprocessing is performed through edge computing servers, and the processed data is uploaded to the cloud computing center, which optimizes the data processing flow. Edge computing reduces data transmission delays and completes preliminary processing locally, thereby reducing the pressure on the cloud computing center and improving the response speed and data processing efficiency of the overall system; the cloud computing center analyzes the uploaded preprocessed data to generate inspection analysis results for the target power grid. This centralized analysis can integrate data from different sources, provide a comprehensive assessment of the power grid's operating status and potential problems, and improve the scientific nature and effectiveness of power grid management.

[0007] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] Figure 1 A flow chart of a smart grid data communication optimization method based on edge computing is provided for an embodiment of the present application;

[0009] Figure 2 An embodiment of the present application provides a flow chart of constructing a set of K transmission equipment communication feature tuples and L inspection site communication feature tuples in a smart grid data communication optimization method based on edge computing. DETAILED DESCRIPTION

[0010] The embodiments of the present application provide a smart grid data communication optimization method based on edge computing, which solves the technical problem that power grid monitoring in the prior art generally relies on centralized data processing, resulting in data transmission delays and excessive network load, which in turn leads to low data processing efficiency.

[0011] After introducing the basic principles of the present application, various non-limiting implementation methods of the present application will be specifically introduced in conjunction with the drawings in the specification.

[0012] like Figure 1 As shown, an embodiment of the present application provides a smart grid data communication optimization method based on edge computing, the method comprising:

[0013] An inspection topology network of a target power grid is obtained, wherein the inspection topology network includes K transmission lines, L inspection sites, and a cloud computing center, and the K transmission lines include K transmission equipment sets and K transmission paths.

[0014] First, clarify the physical scope and equipment distribution of the target power grid, including all transmission lines, transmission equipment, and inspection stations. The target power grid can be a regional power grid or a sub-part of a large power system.

[0015] Determine the target power grid's K transmission lines, L inspection sites, and cloud computing center, where K and L are positive integers. The K transmission lines are all the transmission lines in the target power grid that need to be inspected. These lines include not only the actual physical lines but also all related transmission equipment, such as transformers, transmission towers, switchgear, etc.; the L inspection sites are sites in the power grid that require regular or irregular inspections. These sites usually include key nodes such as substations, distribution stations, and control centers; the cloud computing center is connected to the power grid and is responsible for processing and storing data obtained from power grid inspections and providing computing resources for data analysis.

[0016] For the K transmission lines, determine all the transmission equipment contained in each transmission line, including cables, transformers, switchgear, etc. These devices play a key role in the operation of the power grid. K transmission lines include K sets of transmission equipment. At the same time, determine the specific path of each transmission line, including information such as the physical layout of the line, path length, and the terrain environment it passes through. The detailed information of the path is very important for subsequent inspection planning. K transmission lines include K transmission paths.

[0017] The information of K transmission lines, L inspection stations and cloud computing centers is integrated into a complete inspection topology network. This topology network can display all key components of the power grid and their interrelationships, which can be used to plan and execute power grid inspection tasks.

[0018] Multi-angle communication analysis is performed on the K power transmission equipment sets and the L inspection sites to obtain K power transmission equipment communication feature tuple sets and L inspection site communication feature tuples.

[0019] For a set of K transmission devices, their surrounding environmental characteristics, location coordinates, and channel power losses are analyzed separately to obtain K sets of transmission device environmental characteristics, K sets of transmission device location coordinates, and K sets of transmission device channel power losses. These are then integrated to obtain a set of K transmission device communication feature tuples. Each tuple contains the unchangeable three-dimensional communication characteristics of each transmission device, which cannot be changed in practical applications. For each inspection station, the surrounding environmental characteristics, location coordinates, and channel power losses are similarly analyzed to obtain L inspection station communication feature tuples.

[0020] The inspection topology network is subjected to edge jump identification based on the K power transmission equipment communication feature tuple sets and the L inspection site communication feature tuples to obtain M edge topology subnetworks, wherein the M edge topology subnetworks include M edge communication coefficients.

[0021] The obtained set of K transmission equipment communication feature tuples and L inspection station communication feature tuples serve as the basic data for analyzing the inspection topology network. In graph theory, jump identification involves detecting key turning points or edge locations between nodes in a network. Edge jump identification is used here to locate points where communication characteristics in the inspection topology network change significantly. These changes mark the division of different regions or subnetworks. Based on the results of edge jump identification, the inspection topology network is divided into M edge topology subnetworks, where M is a positive integer, and each subnetwork represents an independent part of the inspection network. The M edge topology subnetworks include M edge communication coefficients, which represent the communication efficiency, stability, and other relevant communication performance indicators within the subnetwork and between the subnetwork and external networks.

[0022] Based on the M edge communication coefficients, sub-network service resources of the M edge topology sub-networks are balancedly configured to obtain M configured edge computing servers.

[0023] Using the calculated M edge communication coefficients, evaluate the computing resource requirements of each edge topology subnetwork, including computing power, storage space, network bandwidth, response time, etc. Determine the balanced configuration goals, such as ensuring that each subnetwork can meet its service requirements while avoiding over-allocation or under-allocation of resources.

[0024] Based on the balanced configuration goal and the edge communication coefficient, computing resources are rationally allocated to evenly distribute the load across each subnetwork. For example, subnetworks with higher service demands are prioritized for more computing resources to ensure communication quality and service performance. Based on the edge communication coefficient and demand assessment results, resources are allocated to the M edge topology subnetworks. Each subnetwork receives resources including computing power, storage, and network bandwidth. Based on the resource allocation results, a corresponding edge computing server is deployed for each edge topology subnetwork, resulting in a total of M configured edge computing servers. These edge computing servers are responsible for handling computing tasks and data processing within the subnetwork, providing low-latency, high-efficiency computing services.

[0025] By balancing the service resources of the sub-networks, the computing efficiency and stability of the inspection topology network can be effectively improved, ensuring that the edge computing server can operate stably under high load conditions, and providing reliable computing support for power grid inspection and maintenance.

[0026] K inspection drones are used to inspect the K transmission lines along the K transmission paths, respectively, to obtain K line inspection video data sets, and L inspection robots are used to inspect the L inspection sites along the preset station inspection routes, to obtain L site inspection video data.

[0027] Each transmission line is assigned a patrol drone. This drone is an aerial device used for transmission line inspections. Equipped with a high-definition camera, it automatically flies along a pre-set route to capture video data on the line's operation and safety. The drone automatically patrols the transmission path, using its onboard high-definition camera to record the operation and safety of the transmission line and its equipment in real time. This detailed video data is then transmitted to the ground control center in real time, resulting in K sets of line inspection video data.

[0028] Each inspection station is equipped with a patrol robot. This ground-based mobile device is used for internal inspections within the station. Equipped with a high-definition camera, it automatically patrols within a pre-set route, collecting video data on the operating status and safety of equipment within the station. Based on the structural layout of each inspection station, a pre-set inspection route is established, covering all key equipment and areas within the station to ensure comprehensiveness and accuracy. The patrol robot follows the pre-set route and conducts automatic inspections within the inspection station. Its camera captures the operating status and safety of equipment within the station in real time, collects video data from the inspection, and uploads it to the control center in real time, generating video data for L inspections of each station.

[0029] Based on the M edge topology sub-networks, the K line inspection video data sets and the L site inspection video data are distributed to M edge computing servers for data preprocessing, and the preprocessed data are uploaded to the cloud computing center for inspection data analysis according to the preset communication protocol to obtain the inspection analysis results of the target power grid.

[0030] Based on M edge topology sub-networks, K line inspection video data sets and L site inspection video data are distributed to M edge computing servers. Each edge computing server corresponds to an edge topology sub-network and is responsible for processing the allocated video data.

[0031] After receiving the distributed data, the edge computing server uses its computing resources to pre-process the data, including video compression, denoising, image enhancement, data format conversion, and preliminary fault detection. Among them, preliminary fault detection can use preset detection algorithms to perform preliminary analysis of inspection video data to identify possible equipment failures, line anomalies or safety hazards. Preprocessing aims to reduce the amount of data transmission, improve data quality, and preliminarily identify possible safety hazards or failures.

[0032] The preprocessed data is uploaded to the cloud computing center according to the preset communication protocol. After receiving the preprocessed data from M edge computing servers, the cloud computing center uses advanced analysis algorithms and big data processing tools to conduct in-depth analysis of all inspection video data. The analysis process includes a comprehensive inspection of the equipment operating status, identification of safety hazards, trend prediction, etc., and generates inspection analysis results for the target power grid. The analysis results include information such as equipment health status, risk assessment, and fault prediction to support power grid operation and maintenance decisions.

[0033] Furthermore, if Figure 2 As shown, a multi-angle communication analysis is performed on the K power transmission equipment sets and the L inspection sites to obtain K power transmission equipment communication feature tuple sets and L inspection site communication feature tuples, including:

[0034] A local scene sensor is used to identify the surrounding environmental features of the K power transmission equipment sets and the L inspection sites to obtain K power transmission equipment environmental feature sets and L inspection site features; in combination with the inspection topology network, a K power transmission equipment position coordinate set, L inspection site position coordinates, and cloud computing center position coordinates of the K power transmission equipment sets, L inspection site position coordinates, and cloud computing center position coordinates are obtained; based on a channel power loss function, a channel power loss analysis is performed on the K power transmission equipment position coordinate set, L inspection site position coordinates, and cloud computing center position coordinates to obtain K power transmission equipment channel power loss amounts and L site channel power loss amounts; based on the K power transmission equipment environmental feature sets, L inspection site features, K power transmission equipment position coordinate sets, L inspection site position coordinates, K power transmission equipment channel power loss amounts, and L site channel power loss amounts, the K power transmission equipment communication feature tuple sets and the L inspection site communication feature tuples are constructed.

[0035] Local scene sensors are devices used to perceive environmental information within a specific area, such as high-definition cameras, lidars, and temperature and humidity sensors. Local scene sensors are installed at key locations in K power transmission equipment sets and L inspection stations to enable them to cover the environment around the equipment.

[0036] Local scene sensors collect environmental data around power transmission equipment and inspection stations. This data includes visual images, temperature distribution, humidity levels, and obstacle locations. Using computer vision and sensor data processing algorithms, they extract key environmental features from this data. For example, image processing algorithms identify physical features such as trees, buildings, and terrain changes; temperature sensors record temperature changes near the equipment; and lidar creates a three-dimensional model of the environment around the equipment. This environmental feature extraction generates a set of K power transmission equipment environmental features and L inspection station features.

[0037] Using GPS or BeiDou navigation systems, obtain the precise geographic coordinates of each power transmission device and inspection site. For the cloud computing center, also obtain its geographic coordinates. Combine the geographic coordinate data for K power transmission devices into a single set, representing their specific locations within the inspection topology network. This yields a set of K power transmission device location coordinates. Similarly, combine the geographic coordinate data for L inspection sites into a single set, yielding L inspection site location coordinates. Similarly, obtain the location coordinates of the cloud computing center. These location coordinates are integrated with the inspection topology network to determine the relative positions and distances of each device and site within the network, laying the foundation for subsequent inspection tasks and data processing.

[0038] A channel power loss function is set, and the channel power loss is calculated for the K transmission equipment location coordinate sets and the cloud computing center location coordinates using the channel power loss function to obtain the channel power loss of the K transmission equipment; the channel power loss is calculated for the L inspection site location coordinates and the cloud computing center location coordinates using the channel power loss function to obtain the channel power loss of the L sites.

[0039] The K sets of transmission equipment environmental features, the K sets of transmission equipment location coordinates, and the K transmission equipment channel power loss values ​​are integrated to obtain the K sets of transmission equipment communication feature tuples. The L inspection station features, the L inspection station location coordinates, and the L station channel power loss values ​​are integrated to obtain the L inspection station communication feature tuples. The communication feature tuple contains the three-dimensional communication features of a transmission equipment or inspection station that cannot be changed, including environmental features, location coordinates, and channel power loss values, and is used to determine the communication characteristics of the transmission equipment or inspection station.

[0040] Furthermore, based on the channel power loss function, a channel power loss analysis is performed on the K power transmission device location coordinate sets, the L inspection site location coordinates, and the cloud computing center location coordinates to obtain the K power transmission device channel power loss amounts and the L site channel power loss amounts, including:

[0041] Construct the channel power loss function, wherein the channel power loss function is:

[0042] ;

[0043] in, is the channel power loss, The height from the horizontal plane of the signal transmission device of the inspection drone or inspection robot for inspecting power transmission equipment or inspection stations, The location coordinates of the power transmission equipment or the inspection station. is the location coordinate of the cloud computing center, , is the reference distance When the transmission power is 1w, the receiving power is Transmission loss of the signal transmission device of the inspection drone or inspection robot for inspecting power transmission equipment or inspection stations, The receiving loss of the cloud computing center, represents the signal transmission wavelength; the K power transmission equipment location coordinate sets and the cloud computing center location coordinates are respectively input into the channel power loss function to obtain the channel power loss amounts of the K power transmission equipment; the L inspection site location coordinates and the cloud computing center location coordinates are respectively input into the channel power loss function to obtain the channel power loss amounts of the L sites.

[0044] Specifically, the channel power loss function is:

[0045] ;

[0046] The channel power loss is a measure of the power loss of the signal from the source to the receiver, that is, the power lost by the signal during transmission from the inspection equipment or site to the cloud computing center. The greater the power loss, the worse the signal quality. This formula calculates the power loss of the signal during transmission. The numerator is As the benchmark value of signal loss, it represents the signal attenuation at the reference distance; the denominator combines the height factor and the actual transmission distance to adjust this benchmark value, thereby calculating the actual channel power loss, thereby determining the signal propagation performance in different environments and conditions, and providing a basis for optimizing communication networks and adjusting equipment configurations.

[0047] The K transmission equipment location coordinate sets and the cloud computing center location coordinates are respectively input into the channel power loss function, and after calculation, the channel power loss of the K transmission equipment is obtained; the L inspection site location coordinates and the cloud computing center location coordinates are respectively input into the channel power loss function, and after calculation, the channel power loss of the L sites is obtained.

[0048] Furthermore, a local scene sensor is used to identify the surrounding environmental features of the K power transmission equipment sets and the L inspection sites, thereby obtaining K power transmission equipment environmental feature sets and L inspection site features, including:

[0049] Utilize video surveillance equipment to collect monitoring videos of the K sets of power transmission equipment and the L inspection sites, and obtain K sets of power transmission equipment monitoring screens and L inspection site monitoring screens; call the interest recognition unit in the local scene sensor to extract interested targets from the K sets of power transmission equipment monitoring screens and the L inspection site monitoring screens, and obtain K sets of power transmission equipment monitoring targets and L inspection site monitoring targets; use the K sets of power transmission equipment monitoring targets and the L inspection site monitoring targets as dynamic semantic recognition objects, call the semantic recognition unit in the local scene sensor to perform environmental dynamic feature recognition, and obtain the K sets of power transmission equipment environmental features and the L inspection site features.

[0050] Based on the geographical environment, equipment layout, and monitoring requirements of the transmission equipment and inspection sites, appropriate video surveillance equipment, such as high-definition cameras, is deployed to ensure coverage of key areas of the transmission equipment and inspection sites. Once activated, the video surveillance equipment continuously monitors K sets of transmission equipment and L inspection sites, generating K sets of transmission equipment monitoring images and L inspection site monitoring images.

[0051] The interest recognition unit (IRI) in the local scene sensor is primarily used to automatically analyze monitoring footage and identify dynamic disturbances around the equipment, such as whether people are approaching the equipment or whether there are any nearby hazards. The IRI analyzes the monitoring footage using image processing algorithms, deep learning models, and other technologies to accurately extract objects of interest, resulting in a set of K monitoring targets for power transmission equipment and L monitoring targets for inspection stations.

[0052] The K sets of transmission equipment monitoring targets and the L inspection station monitoring targets are used as dynamic semantic recognition targets, which is the basis for identifying dynamic environmental features. The semantic recognition unit of the local scene sensor is called upon to extract dynamic environmental features from the monitoring video. It uses deep learning or other advanced image processing techniques to identify changes in the environment and identify dynamic features, such as object movement, human proximity, and changes around equipment.

[0053] Dynamic feature recognition is performed on the monitoring target. For example, it detects whether there is anyone approaching the power transmission equipment, whether there are any dangerous sources (such as flames or smoke) around the equipment, etc. The recognition results are processed into environmental feature data, including the environmental features of each power transmission device and the environmental features of each inspection site, such as whether there is dynamic interference. These features provide dynamic information about the environment around the equipment and the site. Each power transmission device corresponds to a power transmission device environmental feature.

[0054] Furthermore, edge jump identification is performed on the inspection topology network based on the K power transmission equipment communication feature tuple sets and the L inspection site communication feature tuples to obtain M edge topology sub-networks, including:

[0055] N communication feature tuples are randomly selected from the K power transmission equipment communication feature tuple sets and the L inspection site communication feature tuples for edge jump starting point authentication. When the authentication is passed, N edge jump starting points are obtained, wherein N is an integer greater than or equal to 3; edge jump identification is performed in the inspection topology network based on the N edge jump starting points to obtain N first edge topology sub-networks, wherein the N first edge topology sub-networks include N first edge communication coefficients; N updated edge jump starting points are obtained again to perform edge jump identification on the inspection topology network, and after multiple updates and identifications, until the K power transmission equipment communication feature tuple sets and the L inspection site communication feature tuples are all divided into edge topology sub-networks, N updated edge topology sub-network sets are obtained; the M edge topology sub-networks and the M edge communication coefficients are obtained according to the N first edge topology sub-networks and the N updated edge topology sub-network sets.

[0056] From the K sets of communication feature tuples of transmission equipment and the L sets of communication feature tuples of inspection stations, N communication feature tuples are randomly extracted, where N is an integer greater than or equal to 3. The extracted N communication feature tuples are verified to be valid edge jump starting points. The edge jump starting point refers to the initial node selected in the edge computing network. These nodes serve as the starting point for data transmission and processing in the network. The verification criteria can be based on factors such as the stability of the communication characteristics, the acceptable range of the channel power loss, and the safety of the environmental characteristics. Only feature tuples that meet these criteria can pass the certification. After the certification is passed, N edge jump starting points are obtained.

[0057] Edge hop identification involves determining the optimal path and structure for edge computing in a patrol topology network based on N selected edge hop starting points by analyzing the communication characteristics between each starting point and other nodes. Specifically, the N edge hop starting points are used to analyze the communication paths with other nodes in the patrol topology network, focusing on factors such as channel power loss, device location, and environmental characteristics in the communication feature tuples to identify stable and efficient communication paths. Based on the results of the path analysis, the patrol topology network is divided into several subnetworks. Each subnetwork is centered around an edge hop starting point, forming N independent first edge topology subnetworks. Each first edge topology subnetwork is centered around the edge hop starting point and includes other nodes directly or indirectly connected to the core node. These subnetworks will handle data processing tasks in the local area. The communication paths within each subnetwork are analyzed, and the communication efficiency of each path is calculated to obtain the subnetwork's communication coefficient, which is used to quantify the communication performance of the first edge topology subnetwork.

[0058] According to the previous identification results, a new jump starting point is selected based on the communication feature tuple of the uncovered area. The edge jump identification of the inspection topology network is performed again using the updated jump starting point. This process is iterated continuously until all the communication feature tuples of the entire inspection topology network are divided into edge topology subnetworks, and finally a set of N updated edge topology subnetworks is formed. Each subnetwork in the set covers a specific communication feature tuple.

[0059] N first edge topology subnetworks and N updated edge topology subnetworks are integrated to eventually generate M edge topology subnetworks. Each edge topology subnetwork is independent and has efficient communication capabilities, which can maximize the effectiveness in inspection tasks. Each edge topology subnetwork corresponds to an edge communication coefficient, which represents key parameters such as communication quality, channel stability, etc. between devices or sites within the network.

[0060] Further, including:

[0061] According to the coordinate positions of the N edge jump starting points, neighborhood identification is performed in the inspection topology network according to a preset identification bandwidth to obtain N edge jump starting point neighborhoods, wherein each edge jump starting point neighborhood includes multiple transmission equipment communication feature tuples and multiple inspection site communication feature tuples; the N edge jump starting points and the corresponding N edge jump starting point neighborhoods are similarly identified using a cosine similarity calculation formula, and the N edge jump starting point neighborhoods are cleaned according to the identification results to obtain N cleaned edge jump starting point neighborhoods; the N first edge topology sub-networks are constructed according to the N cleaned edge jump starting point neighborhoods, wherein the N first edge topology sub-networks include N first transmission equipment communication feature tuple sets and N inspection site communication feature tuple sets; communication coefficients are centrally analyzed based on the N first transmission equipment communication feature tuple sets and the N inspection site communication feature tuple sets to obtain N first edge communication coefficients.

[0062] The preset identification bandwidth refers to the range of communication feature tuples covered from an edge jump starting point in the inspection topology network, which is similar to the radius in a multidimensional space. The size of this bandwidth determines the number of transmission equipment and inspection sites that can be included in the same neighborhood. The preset identification bandwidth is determined based on the scale of the inspection topology network, the equipment distribution density, and the similarity of communication characteristics. A larger bandwidth covers a wider area but may introduce more heterogeneous communication feature tuples; a smaller bandwidth focuses on more compact equipment and sites with similar communication characteristics.

[0063] Using the coordinate positions of the N edge jump starting points obtained previously, the specific position of each jump starting point in the inspection topology network is clarified. Starting from each edge jump starting point, the neighborhood range is identified with the preset identification bandwidth as the radius. Within this range, multiple transmission equipment communication feature tuples and inspection site communication feature tuples close to the starting point position are included. Based on the identification results, N edge jump starting point neighborhoods are formed.

[0064] For each edge jump starting point, the cosine similarity calculation formula is used to calculate the cosine similarity between it and all the communication feature tuples in its neighborhood to obtain a set of similarity values. The closer the similarity value is to 1, the higher the similarity of the two vectors and the closer the communication characteristics. According to the actual application requirements, a similarity threshold is set, and the communication feature tuples with similarity values ​​lower than the threshold are marked as low-similarity tuples, while those above the threshold are marked as high-similarity tuples. Cleaning is performed based on the marking results. Specifically, in each neighborhood, the communication feature tuples marked as low-similarity tuples are removed. These tuples may introduce communication interference or instability. According to the cleaning results, N cleaned edge jump starting point neighborhoods are obtained. The cleaned edge jump starting point neighborhood retains the communication feature tuples that are most similar to the edge jump starting point, and the communication between these tuples is more stable.

[0065] Based on N cleaning edge jump starting point neighborhoods, communication links are constructed between nodes. Each link corresponds to the communication relationship between the transmission equipment communication feature tuple and the inspection site communication feature tuple. According to the determined nodes and communication links, N first edge topology sub-networks are constructed. The topological structure of each sub-network reflects the communication path and association relationship between the nodes in the edge jump starting point neighborhood. In the constructed first edge topology sub-network, corresponding communication feature sets are generated, including N first transmission equipment communication feature tuple sets and N inspection site communication feature tuple sets. These sets contain valid communication feature information within the sub-network and support subsequent data transmission and processing.

[0066] Based on N sets of first transmission equipment communication feature tuples and N sets of inspection station communication feature tuples, feature vectors are constructed. The extracted feature vectors are used for centralized analysis. The communication coefficient of each edge topology subnetwork is calculated according to the analysis results to obtain N first edge communication coefficients. The communication coefficient is a key indicator reflecting factors such as communication efficiency, channel loss, and interference degree between nodes in the subnetwork, and is used to evaluate and optimize network performance.

[0067] Furthermore, based on the N first power transmission equipment communication feature tuple sets and the N inspection site communication feature tuple sets, a centralized analysis of communication coefficients is performed to obtain N first edge communication coefficients, including:

[0068] Perform weighted calculations on the communication characteristics of the N first power transmission equipment communication feature tuple sets and the N inspection site communication feature tuple sets according to preset weights to obtain N first power transmission equipment communication coefficient sets and N first inspection site communication coefficient sets; randomly extract N centralized analysis starting points from the N first power transmission equipment communication coefficient sets, iterate the N first power transmission equipment communication coefficient sets according to a preset iteration step size, and obtain N iterative first power transmission equipment communication coefficients; determine whether the concentration difference between the N iterative first power transmission equipment communication coefficients and the N centralized analysis starting points meets the preset set concentration difference. Moderate difference: If not, the N iterative first transmission device communication coefficients are updated to the N iterative centralized analysis starting points, and the iteration is continued according to the preset iteration step size; if so, the iteration is stopped, and the N iterative first transmission device communication coefficients are used as the N first target transmission device communication coefficients; a communication coefficient centralized analysis is performed on the N first inspection site communication coefficient set to obtain N first target inspection site communication coefficients; and a weighted calculation is performed on the N first target transmission device communication coefficients and the N first target inspection site communication coefficients according to preset weights to obtain the N first edge communication coefficients.

[0069] The weights of each feature in each feature tuple are set based on the requirements of the actual application scenario and the importance of the communication features. These weights can be determined through historical data analysis or experimental results. For example, in power transmission equipment communication, the channel power loss is more important than the device location coordinates, so its weight is higher, and the sum of the feature weights is 1.

[0070] According to the preset weights, each feature in each communication feature tuple is multiplied by the corresponding weight value, and then the weighted feature values ​​are added together to obtain the communication coefficient of a single feature tuple. The above weighted calculation is performed on all tuples in the N first power transmission device communication feature tuple sets and the N inspection site communication feature tuple sets to obtain N first power transmission device communication coefficient sets and N first inspection site communication coefficient sets, respectively.

[0071] Based on the calculated set of N first transmission device communication coefficients, N starting points are randomly selected as the initial positions for the centralized analysis. A preset iteration step size is set based on actual application requirements and the changing patterns of device communication characteristics. The iteration step size refers to the magnitude of the communication coefficient adjustment in each iteration. A larger step size results in faster iteration speed but may lack accuracy; a smaller step size results in higher iteration accuracy but may require more iterations.

[0072] Using N randomly selected centralized analysis starting points as initial points, in the first round of iteration, these starting points are adjusted according to a preset iteration step size, and new communication coefficients are calculated. The coefficients in the set of N first transmission device communication coefficients are adjusted over multiple iterations. In each iteration, the updated communication coefficients are used as input for the next iteration, and these coefficients are gradually optimized until the preset convergence condition is met or the maximum number of iterations is reached, thereby obtaining the final N iterative first transmission device communication coefficients.

[0073] Calculate the concentration difference between the N iterative first transmission device communication coefficients and the N centralized analysis starting points. The concentration difference is the difference between the communication coefficient updated each time during the iteration and the initial centralized analysis starting point, which can be obtained by calculating the difference between each iterative communication coefficient and its corresponding starting point coefficient.

[0074] A preset concentration difference is set to determine whether the iteration has converged. This value can be set based on the application scenario and algorithm characteristics. The calculated concentration difference is compared with the preset concentration difference. If the calculated difference is greater than the preset concentration difference, it indicates that the current iteration result has not converged. In this case, the communication coefficient of the first transmission device of N iterations is updated to the starting point of the N iterations of concentration analysis, and the next round of iterations continues according to the preset iteration step size.

[0075] If the difference is less than or equal to the preset concentration difference, the iterative process has converged and the iteration can be stopped. At this time, the current N iterative first transmission device communication coefficients are used as the final N first target transmission device communication coefficients as the basis for subsequent analysis and decision-making.

[0076] The same analysis method as that for power transmission equipment is used to analyze and obtain the communication coefficients of the N first target inspection sites. For the sake of brevity, this will not be repeated here.

[0077] According to actual application requirements and network optimization strategies, preset weights are set. The weights can reflect the degree of influence of different communication coefficients on edge communication performance. According to the preset weights, the communication coefficients of the N first target transmission devices and the communication coefficients of the N first target inspection sites are weightedly calculated to obtain the final N first edge communication coefficients.

[0078] Further, including:

[0079] The operation status of the M edge computing servers is monitored to obtain M operation coefficients; when the M operation coefficients are greater than or equal to the preset operation coefficients, an early warning instruction is obtained and the early warning instruction is sent to the cloud computing center for alarm.

[0080] The operating coefficient is a quantitative value of the operating status and performance indicators of each edge computing server. Specifically, through hardware monitoring tools and sensors, indicators such as CPU utilization, memory usage, and disk read and write speed are collected. The collected monitoring data is standardized to ensure the comparability of various indicators. The data of each indicator is weighted and summed according to the weights of different indicators to obtain M operating coefficients. The weights can be adjusted according to system requirements and performance importance.

[0081] Define the maximum allowable performance index value of the edge computing server under normal operating conditions, set a preset operating coefficient, compare the operating coefficient of each edge computing server with the preset operating coefficient, and determine whether it exceeds the preset threshold. When the operating coefficient is greater than or equal to the preset operating coefficient, trigger an early warning, generate corresponding early warning instructions, and send the early warning instructions to the cloud computing center for alarm to ensure stable and efficient operation of the system.

[0082] In summary, the embodiment of the present application provides a smart grid data communication optimization method based on edge computing, which has the following technical effects:

[0083] By obtaining the inspection topology network of the target power grid, including the layout of transmission lines, inspection sites and cloud computing centers, a comprehensive understanding of the power grid structure is provided, which lays the foundation for subsequent communication analysis and edge computing optimization, and ensures the accuracy of data collection and processing; multi-angle communication analysis of transmission equipment and inspection sites is carried out to obtain a set of communication feature tuples. This multi-dimensional data analysis can identify the communication characteristics of equipment and sites, establish comprehensive communication characteristics, and help optimize data flow and network load; edge jump identification of the inspection topology network based on the communication feature tuple set can divide the power grid into multiple edge topology sub-networks and calculate the edge communication coefficient. This method can effectively identify key edge nodes and communication paths in the network, optimize the path selection of data transmission, and improve data processing efficiency and transmission stability; sub-network service resources are balanced based on the edge communication coefficient, and the resource allocation of edge computing servers is optimized, which ensures the balanced use of computing resources and improves The server's processing capacity and response speed are improved; inspection drones and inspection robots are used to inspect transmission lines and inspection sites to obtain inspection video data. This method can provide real-time equipment status and safety monitoring, and improve the automation level and accuracy of inspections. Drones cover a wide area of ​​the transmission lines, while robots conduct detailed inspections inside the sites, ensuring the comprehensiveness and meticulousness of the data; data preprocessing is performed through edge computing servers, and the processed data is uploaded to the cloud computing center, which optimizes the data processing flow. Edge computing reduces data transmission delays and completes preliminary processing locally, thereby reducing the pressure on the cloud computing center and improving the response speed and data processing efficiency of the overall system; the cloud computing center analyzes the uploaded preprocessed data to generate inspection analysis results for the target power grid. This centralized analysis can integrate data from different sources, provide a comprehensive assessment of the power grid's operating status and potential problems, and improve the scientific nature and effectiveness of power grid management.

[0084] The above description of the disclosed embodiments is intended to enable one skilled in the art to implement or use the present application. Various modifications to these embodiments will be readily apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application is not limited to the embodiments shown herein, but is intended to conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A smart grid data communication optimization method based on edge computing, characterized in that: The method comprises: Obtaining an inspection topology network of a target power grid, wherein the inspection topology network includes K transmission lines, L inspection sites, and a cloud computing center, and the K transmission lines include K transmission equipment sets and K transmission paths; Perform multi-angle communication analysis on the K power transmission equipment sets and the L inspection sites to obtain K power transmission equipment communication feature tuple sets and L inspection site communication feature tuples; Based on the K transmission equipment communication feature tuple sets and the L inspection site communication feature tuples, edge jump identification is performed on the inspection topology network to obtain M edge topology sub-networks, wherein the M edge topology sub-networks include M edge communication coefficients; Based on the M edge communication coefficients, the M edge topology sub-networks are configured with sub-network service resources in a balanced manner to obtain M configured edge computing servers; Using K inspection drones to inspect the K transmission lines according to the K transmission paths, respectively, to obtain K line inspection video data sets, and using L inspection robots to inspect the L inspection sites according to preset in-station inspection routes, to obtain L site inspection video data; Based on the M edge topology subnetworks, the K line inspection video data sets and the L site inspection video data are distributed to M edge computing servers for data preprocessing, and the preprocessed data are uploaded to the cloud computing center for inspection data analysis according to a preset communication protocol to obtain the inspection analysis results of the target power grid; Based on the K power transmission equipment communication feature tuple sets and the L inspection site communication feature tuples, edge jump identification is performed on the inspection topology network to obtain M edge topology sub-networks, including: Randomly extract N communication feature tuples from the K power transmission equipment communication feature tuple sets and the L inspection site communication feature tuples to perform edge jump starting point authentication, and when the authentication passes, obtain N edge jump starting points, where N is an integer greater than or equal to 3; Based on the N edge jump starting points, edge jump identification is performed in the inspection topology network to obtain N first edge topology sub-networks, wherein the N first edge topology sub-networks include N first edge communication coefficients; Obtaining N updated edge jump starting points again to perform edge jump identification on the inspection topology network, and performing multiple update identifications until the K transmission equipment communication feature tuple sets and the L inspection site communication feature tuple sets are all divided into edge topology sub-networks, thereby obtaining N updated edge topology sub-network sets; Obtaining the M edge topology subnetworks and the M edge communication coefficients according to the N first edge topology subnetworks and the N updated edge topology subnetwork sets; According to the coordinate positions of the N edge jump starting points, neighborhood identification is performed in the inspection topology network according to a preset identification bandwidth to obtain N edge jump starting point neighborhoods, wherein each edge jump starting point neighborhood includes a plurality of transmission equipment communication feature tuples and a plurality of inspection site communication feature tuples; Using a cosine similarity calculation formula, similarity recognition is performed on N edge step starting points and corresponding N edge step starting point neighborhoods, and the N edge step starting point neighborhoods are cleaned according to the recognition result to obtain N cleaned edge step starting point neighborhoods; Constructing the N first edge topology sub-networks according to the N cleaning edge jump starting point neighborhoods, wherein the N first edge topology sub-networks include N first power transmission equipment communication feature tuple sets and N inspection site communication feature tuple sets; Based on the N first power transmission equipment communication feature tuple sets and the N inspection site communication feature tuple sets, a centralized analysis of communication coefficients is performed to obtain N first edge communication coefficients.

2. A smart grid data communication optimization method based on edge computing as claimed in claim 1, characterized in that: Performing multi-angle communication analysis on the K power transmission equipment sets and the L inspection sites to obtain K power transmission equipment communication feature tuple sets and L inspection site communication feature tuples, including: Using a local scene sensor to identify the surrounding environmental features of the K power transmission equipment sets and the L inspection sites, to obtain K power transmission equipment environmental feature sets and L inspection site features; In combination with the inspection topology network, obtain the K power transmission equipment sets, the L inspection sites, and the K power transmission equipment location coordinate sets, the L inspection site location coordinates, and the cloud computing center location coordinates of the cloud computing center; Based on the channel power loss function, a channel power loss analysis is performed on the K power transmission equipment location coordinate sets, the L inspection site location coordinates and the cloud computing center location coordinates to obtain the K power transmission equipment channel power loss amounts and the L site channel power loss amounts; The K power transmission equipment communication feature tuple sets and the L inspection site communication feature tuples are constructed based on the K power transmission equipment environmental feature sets, the L inspection site features, the K power transmission equipment location coordinate sets, the L inspection site location coordinates, the K power transmission equipment channel power loss amounts and the L site channel power loss amounts.

3. A smart grid data communication optimization method based on edge computing as claimed in claim 2, characterized in that: Based on the channel power loss function, a channel power loss analysis is performed on the K power transmission equipment location coordinate sets, the L inspection site location coordinates and the cloud computing center location coordinates to obtain the K power transmission equipment channel power loss amounts and the L site channel power loss amounts, including: Construct the channel power loss function, wherein the channel power loss function is: ; in, is the channel power loss, The height from the horizontal plane of the signal transmission device of the inspection drone or inspection robot for inspecting the power transmission equipment or inspection station. is the location coordinates of the power transmission equipment or the inspection station. is the location coordinate of the cloud computing center, , The reference distance When the transmission power is 1w, the receiving power is The transmission loss of the signal transmission device of the inspection drone or inspection robot for inspecting power transmission equipment or inspection stations, The receiving loss of the cloud computing center, Indicates the signal transmission wavelength; Inputting the K power transmission equipment location coordinate sets and the cloud computing center location coordinates into the channel power loss function respectively to obtain the channel power loss amounts of the K power transmission equipment; The position coordinates of the L inspection sites and the position coordinates of the cloud computing center are respectively input into the channel power loss function to obtain the channel power loss amounts of the L sites.

4. A smart grid data communication optimization method based on edge computing as claimed in claim 2, characterized in that: The local scene sensor is used to identify the nearby environmental features of the K power transmission equipment sets and the L inspection sites, and obtain the K power transmission equipment environmental feature sets and the L inspection site features, including: Using video monitoring equipment to collect monitoring videos of the K power transmission equipment sets and the L inspection sites, to obtain K power transmission equipment monitoring screen sets and L inspection site monitoring screens; Calling the interest identification unit in the local scene sensor to extract the interested targets from the K power transmission equipment monitoring screen sets and the L inspection site monitoring screens, and obtaining the K power transmission equipment monitoring target sets and the L inspection site monitoring targets; The K power transmission equipment monitoring target sets and the L inspection site monitoring targets are taken as dynamic semantic recognition objects, and the semantic recognition unit in the local scene sensor is called to perform environmental dynamic feature recognition to obtain the K power transmission equipment environmental feature sets and the L inspection site features.

5. The method for optimizing smart grid data communication based on edge computing according to claim 1, characterized in that: Based on the N first power transmission equipment communication feature tuple sets and the N inspection site communication feature tuple sets, a centralized analysis of communication coefficients is performed to obtain N first edge communication coefficients, including: Performing weighted calculation of communication characteristics on the N first power transmission device communication feature tuple sets and the N inspection site communication feature tuple sets according to preset weights, to obtain N first power transmission device communication coefficient sets and N first inspection site communication coefficient sets; Randomly extract N centralized analysis starting points from the N first power transmission device communication coefficient sets, iterate the N first power transmission device communication coefficient sets according to a preset iteration step length, and obtain N iterated first power transmission device communication coefficients; Determine whether the concentration difference between the N iterative first power transmission device communication coefficients and the N concentrated analysis starting points meets a preset concentration difference, and if not, update the N iterative first power transmission device communication coefficients to the N iterative concentrated analysis starting points, and continue to iterate according to a preset iteration step size; If yes, stop the iteration and use the N iterative first power transmission device communication coefficients as N first target power transmission device communication coefficients; Performing centralized communication coefficient analysis on the communication coefficient sets of the N first inspection sites to obtain communication coefficients of the N first target inspection sites; The N first target power transmission equipment communication coefficients and the N first target inspection site communication coefficients are weightedly calculated according to preset weights to obtain the N first edge communication coefficients.

6. A smart grid data communication optimization method based on edge computing as claimed in claim 1, characterized in that: include: Performing operation status monitoring on the M edge computing servers to obtain M operation coefficients; When the M operating coefficients are greater than or equal to the preset operating coefficients, an early warning instruction is obtained and the early warning instruction is sent to the cloud computing center for alarm.

Images