Power grid network topology automatic generation and dynamic checking method supporting full voltage level penetration

Abstract

This invention discloses a method for automatically generating and dynamically verifying power grid topology supporting full voltage level connectivity, belonging to the field of power grid topology technology. The method includes the following steps: generating topologies at each level and merging them into an initial topology supporting full voltage level connectivity; establishing an update triggering mechanism to update the initial topology in real time; constructing a multi-dimensional verification index system and verifying the multi-dimensional verification index system; and constructing a visualization platform. This invention solves the problems of existing technologies being unable to achieve full coverage of the physical connections of the entire power grid and unable to monitor changes in the topology structure during power grid operation and correct anomalies in a timely manner. By generating a topology structure, this invention breaks down data silos and voltage level separation, quickly detects anomalies and triggers updates, reduces maintenance intensity and fault repair time through automated processes, and achieves efficient fault handling based on accurate topology.

Description

Technical Field

[0001] This invention relates to the field of power grid topology technology, specifically to a method for automatic generation and dynamic verification of power grid topology that supports the continuity of all voltage levels. Background Technology

[0002] The power grid topology is the core foundational data for power grid dispatching, fault analysis, and operation and maintenance management. Its accuracy and real-time performance directly affect the safety and reliability of power grid operation. With the continuous expansion of the power grid, voltage levels cover 500kV and above ultra-high voltage, 220kV-110kV high voltage, 35kV-10kV medium voltage, and 0.4kV low voltage, making the power grid structure increasingly complex.

[0003] Chinese patent CN119397065A discloses an automatic topology identification method for complex regional power grid structures. The method involves traversing and searching the regional power grid to identify all circuit breakers and isolating switches, assigning them numbers; finding adjacent circuit breakers and assigning them a value of 1, while assigning non-adjacent circuit breakers a value of 0, forming a vector; constructing a circuit breaker association adjacency matrix for the regional power grid, where each pair of circuit breakers is associated with a value of 1 when the corresponding number in the adjacency vector matrix is ​​assigned a value of 1; selecting several flow vectors, choosing the next target node using a roulette wheel algorithm, calculating path pheromones, and updating the roulette wheel probability of the target node; after a certain number of iterations and updates, when the minimum path length is less than a threshold, it indicates that the path is closed-loop; finally, the automatic topology identification of the regional power grid structure is obtained. This method can automatically identify the topology of complex regional power grids or isolated power grids, providing a more accurate basis for judgment in relay protection and stability control strategies.

[0004] In practical use, the aforementioned patents do not address topology processing across all voltage levels. When faced with complex power grids spanning multiple voltage levels, they cannot achieve full coverage of the physical connections of the entire grid structure. Furthermore, the method only focuses on topology identification and lacks a dynamic verification step, making it impossible to monitor changes in the topology structure during power grid operation in real time and correct anomalies promptly. When the power grid undergoes line modifications, equipment additions or removals, topology identification errors are prone to occur. Therefore, it does not meet current requirements. To address this, we propose a method for automatic generation and dynamic verification of power grid topology that supports cross-voltage levels. Summary of the Invention

[0005] The purpose of this invention is to provide an automatic generation and dynamic verification method for power grid topology that supports the continuity of all voltage levels. By generating a topology covering all aspects of power generation, transmission, transformation, distribution, and consumption, it breaks down data silos and voltage level separation, providing a unified topology view to help grasp the overall state of the power grid. By establishing a multi-dimensional indicator system including connectivity and weather adaptability, combined with real-time data for continuous verification, it quickly detects anomalies and triggers updates, ensuring that the topology is consistent with reality. This provides an accurate basis for fault location, power flow calculation, etc. Weather early warning function can reduce disaster losses. Automated processes reduce maintenance intensity and fault repair time, reducing power outage losses and enterprise costs. Through topology process automation, it can reduce manual intervention and maintenance intensity. Relying on accurate topology, it can achieve efficient fault handling, shorten repair time, reduce power outages and enterprise economic losses, and solve the problems mentioned in the background technology.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a method for automatically generating and dynamically verifying the power grid topology across all voltage levels, comprising;

[0007] A method for automatically generating and dynamically verifying power grid topology that supports continuity across all voltage levels includes the following steps:

[0008] S1: Construct a full-voltage-level power grid data fusion system to acquire multi-source data from the main grid dispatching system, distribution network automation system, low-voltage distribution area management system, equipment ledger system, and Internet of Things monitoring platform. The multi-source data includes basic equipment information, operating status data, electrical measurement data, and geospatial data.

[0009] S2: Automatic generation of topology for all voltage levels based on topology rules, constructing a topology association rule library, using geospatial data as a benchmark, automatically identifying device connection relationships according to the rule library, generating topologies at each level and merging them into an initial topology that connects all voltage levels; S3: Dynamic updating of topology based on real-time data, establishing an update triggering mechanism, monitoring data changes and triggering local adjustments to the initial topology, and updating the initial topology in real time;

[0010] S4: Construct a multi-dimensional verification indicator system, which includes topological connectivity indicators, data consistency indicators, alarm correlation accuracy indicators, and meteorological adaptability indicators. Verify the multi-dimensional verification indicator system, and trigger a topological local update mechanism when topological anomalies occur.

[0011] S5: Build a visualization platform to realize multi-scale topological display and interactive linkage with multi-source data, and automatically correct and update topological deviations.

[0012] Preferably, the data fusion system includes a data access layer, a data processing layer, and a data storage layer:

[0013] The data access layer is used to acquire main grid tripping intelligent alarms, main grid SOE data, distribution transformer power data, transmission video data and typhoon meteorological data, and adopts differentiated access strategies according to the characteristics of different data sources;

[0014] The data processing layer is used to clean, standardize, and perform correlation mapping on the data acquired by the data access layer;

[0015] The data storage layer uses a relational database to store basic equipment ledgers and user information structure data, a time-series database to store real-time operating data of main transformer temperature and line current, and a graph database to store topology data, efficiently representing equipment connection relationships in the form of graph nodes and edges.

[0016] Preferably, the step of automatically identifying device connection relationships based on geospatial data and a rule base, generating topologies at each level, and merging them into an initial topology that connects all voltage levels, specifically includes:

[0017] Pre-define equipment physical connection rules, voltage level matching constraint rules, power supply path compliance rules, and geospatial association rules, and import high-precision geospatial data from the GIS system to complete the precise binding of coordinate information for each power grid device;

[0018] Based on the established topology association rule base, an improved breadth-first search algorithm is used to determine the hierarchical association between power generation equipment, transmission lines and substations, and generate a power generation-transmission hierarchical network topology.

[0019] By using precise voltage level matching logic, the internal connection relationships between the main transformer and busbar, and between the busbar and outgoing switch within the substation are identified, generating a power transmission-substation hierarchical network topology.

[0020] By combining the geographical path data of the distribution network lines, the connection relationship between the outgoing switches and the distribution network branches, and between the branches and the distribution transformers, the substation-distribution hierarchical network topology is generated.

[0021] Based on the location information of low-voltage user access points, the connection relationship between distribution transformers and low-voltage branch switches, and between branch switches and end users is identified, and a distribution-low-voltage user hierarchical network topology is generated.

[0022] The topologies of the above voltage levels are vertically integrated to form a continuous initial network topology covering all voltage levels. Each topology node is bound to a unique identifier of the corresponding device, realizing a precise mapping and association between the topology node and the power grid data resource pool of all voltage levels.

[0023] Preferably, the improved breadth-first search algorithm is implemented using the following steps:

[0024] The set of starting points for topology generation is selected from the core hub nodes in the power grid system, with priority given to hub substations with voltage levels of 500kV and above as the core starting points for topology traversal.

[0025] Using a multi-source data fusion framework for the entire voltage level power grid, the basic parameters and inherent connection relationships of the main transformers, busbars, various switches and instrument transformers inside the starting substation are collected to build an initial topology node database.

[0026] Starting from the initial set, a hierarchical traversal is carried out in descending order of voltage level to ensure the logical coherence and structural integrity of the topology generation process.

[0027] Using preset power grid topology logic verification criteria, the compliance verification of the traversed device connection relationships is carried out, and erroneous connections and redundant connections are screened out and eliminated.

[0028] Verified device nodes and their connections are converted into standardized graph structure data and stored in a dedicated graph database for easy access and topology updates.

[0029] Preferably, the implementation process of S3 specifically includes:

[0030] Establish a multi-scenario linkage topology update triggering mechanism, with trigger types covering equipment status change triggering, electrical measurement data mutation triggering, and manual operation triggering;

[0031] Real-time monitoring of equipment status changes such as new additions, decommissioning, and technical upgrades within the standardized data asset database; and monitoring of abnormal fluctuations in real-time measurement data such as main transformer power, line current, and switch open / close status.

[0032] When the update trigger condition is met, extract all information of the abnormal equipment and its associated equipment, make local precise adjustments to the initial power grid topology based on the topology association rule library, and synchronously update the equipment connection relationship and corresponding power supply path.

[0033] If the update is manually triggered, it supports manually editing topology nodes and adjusting connection relationships through a visual interactive interface, and synchronously updating the data mapping association between the device and the topology nodes.

[0034] Preferably, the verification of the multi-dimensional verification indicator system specifically includes:

[0035] Connectivity verification: The continuity status of the topology path is verified by real-time power flow data of the power grid. The topology nodes are fully traversed to verify the continuous connectivity of each power supply path and identify isolated nodes and redundant connections. If the active power transmission value of a certain topology path is zero and there is no corresponding power outage alarm signal, it is determined to be an abnormal topology connectivity.

[0036] Data consistency verification: Compare real-time data collected from different data sources of the same device to verify the energy conservation relationship of electrical quantities such as power and current within the same power supply path. If the data deviation exceeds the preset threshold and lasts for more than 5 seconds, it is judged as data consistency abnormal.

[0037] Status matching verification: The operating status of associated devices and the topology connection relationship are accurately bound to the power grid alarm data and the topology structure. The location of alarm devices in the topology and the matching degree of associated devices are verified to be consistent with the actual on-site conditions. If they are inconsistent, it is determined to be an abnormal status matching.

[0038] Meteorological adaptability verification: Based on external meteorological monitoring data such as typhoons, ice accumulation, and wildfires, analyze the meteorological impact range of key equipment such as transmission lines and towers within the topology, and verify the rationality and risk resistance adaptability of the topology equipment correlation relationship.

[0039] After detecting topology anomalies through multi-dimensional verification, the topology local update mechanism is immediately activated. This mechanism sequentially performs anomaly location, incremental update, and update verification.

[0040] Preferably, the anomaly localization specifically includes:

[0041] The anomaly detection scope is compressed and the anomaly detection efficiency is improved by adopting a partitioned location algorithm. The power grid is divided into several independent topological partitions according to geographical administrative regions and voltage levels, and each topological partition is equipped with a dedicated independent verification unit.

[0042] After detecting a topology anomaly, the topology partition to which the anomaly belongs is first located. Then, within the partition, the root cause of the anomaly is traced back through the device hierarchy relationship to accurately locate the local topology segment that needs to be updated.

[0043] Preferably, the incremental update specifically includes:

[0044] For abnormal topology segments that have been located, local targeted updates are performed. When a new device node is added, a globally unique device identification code is automatically assigned, full parameter information of the device is collected synchronously, and a standardized topology connection with neighboring devices is established.

[0045] When removing or decommissioning a device node, the node information and all associated topology edge information of that node in the graph database are deleted simultaneously.

[0046] When the device connection relationship changes, update the status attributes of the corresponding topology edge, switch the original connected state to the disconnected state, or add a new topology edge to establish a new compliant connection.

[0047] The entire incremental update process records the update content, update time, and information of the person responsible for the operation in real time, forming a complete update log, which facilitates subsequent topology tracing and auditing.

[0048] Preferably, when a sudden change in electrical quantity is detected in the multi-source data, and it is necessary to accurately distinguish whether the sudden change in electrical quantity is caused by a real physical topology break or by a transient electrical inrush current in the system, the real-time update of the initial topology specifically includes:

[0049] The electrical feature sequence of each topological node in the initial topology in the multi-source data is extracted in a continuous historical time period, and the physical impedance parameters and spatial connection relationship between each topological node in the initial topology are extracted based on geospatial data to construct a topological distance attenuation matrix;

[0050] Based on the topology distance decay matrix, determine the total number of target evaluation nodes within the initial topology and the number of adjacent topology nodes within a preset topology hop range from the target evaluation node, and simultaneously establish a sliding data acquisition time window for the adjacent topology nodes;

[0051] The sliding data acquisition time window is used to obtain the real-time electrical measurement data of the adjacent topology nodes at the current sampling time and the real-time electrical measurement data at the previous sampling time, and simultaneously obtain the number of abnormal alarm nodes in the topology partition to which the target evaluation node belongs and the total number of topology nodes.

[0052] Based on the acquired real-time electrical measurement data, the number of abnormal alarm nodes, and the shortest topological path length within the topological distance attenuation matrix, the spatiotemporal disturbance index of the target evaluation node in the current environment is calculated. This spatiotemporal disturbance index quantifies the probability of a real physical topological break or alteration occurring around the target evaluation node, rather than measurement jumps caused by transient electrical inrush currents. The specific formula for calculating the spatiotemporal disturbance index of the target evaluation node within the initial topology in the current environment is as follows:

[0053]

[0054] In the formula, the Indicates the target evaluation node The spatiotemporal disturbance index; This represents the total number of adjacent topology nodes; Indicates the sequence number of the adjacent topological nodes; Indicates the first The voltage level weighting coefficient of each adjacent topology node; the higher the voltage level, the larger the voltage level weighting coefficient. Indicates the first The adjacent topology nodes at the current sampling time The real-time electrical measurement data; Indicates the first The adjacent topology nodes at the pre-sampling time The real-time electrical measurement data; Indicates the relationship with the first The reference electrical constants for the voltage levels corresponding to the adjacent topology nodes; Represents the natural constant; the This represents the spatial attenuation coefficient, with dimensions of the reciprocal of length; The target evaluation node is extracted from the topological distance decay matrix. With the The shortest topological path length between adjacent topological nodes; Indicates the disturbance balance coefficient; the Indicates the alarm gain coefficient; the Indicates the number of abnormal alarm nodes; This represents the total number of nodes in the topology;

[0055] The basic disturbance values ​​of the target evaluation node during its historical steady-state operation are retrieved and combined with the current grid load level to generate a dynamic comparison threshold. The calculated spatiotemporal disturbance index is compared with the dynamic comparison threshold. If the spatiotemporal disturbance index exceeds the dynamic comparison threshold, the target evaluation node is determined to meet the topology update triggering condition, and the initial topology local adjustment is triggered for the target evaluation node.

[0056] Preferably, when a large-scale meteorological disaster warning is received, and it is necessary to defend against massive concurrent device status change data that could cause graph database computing power overload and topology view jitter and crash, the multi-dimensional verification index system is checked, and a topology local update mechanism is triggered when topology anomalies occur. Specifically, this includes:

[0057] Multi-dimensional meteorological impact parameters, including wind speed, ice thickness, and precipitation, are extracted from the multi-source data in real time, and the multi-dimensional meteorological impact parameters are converted into comprehensive disaster intensity values ​​based on a preset meteorological disaster assessment model.

[0058] Determine whether the comprehensive disaster intensity value exceeds the preset extreme disaster critical threshold. If it exceeds the extreme disaster critical threshold, issue a global write transaction latch instruction to the graph database storing the initial topology, forcibly suspending the global breadth-first search process for the initial topology and the dynamic verification and update process of network connectivity.

[0059] Based on the latitude and longitude range of the multi-dimensional meteorological impact parameters, a disaster spatial envelope is generated in the standard geographic coordinate system of the power grid, and the associated device set within the disaster spatial envelope is mapped to the network space, thereby rigidly isolating the associated device set into a disaster-affected closed topology partition.

[0060] Create an independent high-speed cache sandbox in system memory, and clone the current topology connection snapshot of the disaster-affected closed topology partition into the high-speed cache sandbox;

[0061] Inside the cache sandbox, data interaction with the external main network view is blocked. Based only on the edge remote signaling data of the surviving devices in the disaster-affected closed topology partition, a high-frequency local topology reconstruction exercise is performed on the associated device set, and the set of steady-state edge topology nodes generated during the reconstruction exercise is continuously updated and recorded.

[0062] The multi-dimensional meteorological impact parameters are continuously monitored. When it is confirmed that the comprehensive disaster intensity value has fallen below the extreme disaster critical threshold and after a preset system smoothing period, the global write transaction latch instruction of the graph database is released. The final topological association state of the steady-state edge topological node set is extracted and a connection conflict check is performed with the main network edge node in the graph database at this time.

[0063] If the connection conflict verification passes, the final topology association state is written to the initial topology in an incremental merging manner through an asynchronous message queue to eliminate the invalid system computing power consumption and topology view jitter caused by extreme weather disturbances.

[0064] Compared with the prior art, the beneficial effects of the present invention are:

[0065] 1. This invention constructs a multi-source data fusion system, integrating equipment and operational data from the main grid, distribution network, and low-voltage network. It employs an improved breadth-first search algorithm to generate a full-voltage-level topology covering power generation, transmission, substation, distribution, low voltage, and users, achieving topology processing across all voltage levels. This solves the problem that traditional topologies cannot achieve full coverage of the physical connections of the entire network structure when facing complex power grids across voltage levels. It provides dispatchers with a comprehensive and unified view of the power grid topology, helping them to grasp the overall operating status of the power grid.

[0066] 2. This invention effectively reduces human error and improves topology generation efficiency and accuracy through data standardization processing and connection relationship verification. At the same time, it adopts an improved breadth-first search algorithm to realize local topology updates, avoids full reconstruction, shortens the topology update response time, and can reflect changes in the power grid operating status in a timely manner.

[0067] 3. This invention establishes a multi-dimensional verification index system including connectivity, consistency, matching, and meteorological adaptability, and continuously verifies the topology by combining real-time operating data and equipment status. This enables rapid detection of topology anomalies and triggers updates, solving the problem that traditional topologies cannot monitor changes in the topology structure during power grid operation in real time and correct anomalies in a timely manner. When the power grid undergoes line modifications, equipment additions or removals, topology identification deviations are prone to occur. This invention ensures that the topology data is consistent with the actual operating status of the power grid, providing a real-time and accurate topology foundation for applications such as rapid fault location and power flow calculation.

[0068] 4. This invention supports multi-state display of topology data and business applications through the topology visualization platform it builds, which is in operation, historical and planning states. For example, in fault handling, the fault tracing function based on topology can shorten the fault location time from the original 30 minutes to less than 5 minutes, which significantly improves the efficiency of fault handling. In response to meteorological disasters, the meteorological impact assessment function can provide early warning of the impact of disasters such as typhoons and icing on the power grid, buy time for operation and maintenance and repair, and reduce power grid failure losses.

[0069] 5. This invention reduces manual intervention by automating topology generation and dynamic verification, thereby reducing the workload of maintenance personnel. At the same time, based on accurate topology data, fault handling and scheduling decisions shorten fault repair time, reduce power outage losses, and lower maintenance costs and economic losses for power grid companies. Attached Figure Description

[0070] Figure 1 This is a flowchart of the method for automatic generation and dynamic verification of power grid topology supporting the continuity of all voltage levels according to the present invention.

[0071] Figure 2 The flowchart shows the specific implementation of the improved breadth-first search algorithm of this invention. Detailed Implementation

[0072] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0073] To address the shortcomings of existing technologies, which do not address topology processing across all voltage levels, this approach cannot achieve full coverage of the physical interconnections of the entire grid structure when dealing with complex power grids spanning multiple voltage levels. Furthermore, this method only focuses on topology identification and lacks a dynamic verification step, failing to monitor changes in the grid's topology in real time and promptly correct anomalies. When the grid undergoes line modifications or equipment additions / removals, topology identification errors are prone to occur. Please refer to [the relevant documentation / reference]. Figures 1-2 This embodiment provides the following technical solution:

[0074] A method for automatically generating and dynamically verifying the grid topology across all voltage levels includes the following steps:

[0075] S1: Construct a full-voltage-level power grid data fusion system to acquire multi-source data from the main grid dispatching system, distribution network automation system, low-voltage distribution area management system, equipment ledger system and IoT monitoring platform. The multi-source data includes basic equipment information, operating status data, electrical measurement data and geospatial data.

[0076] The data fusion system includes:

[0077] The data access layer employs a multi-protocol adaptation module, supporting various power industry standard protocols and general protocols such as IEC61850, DL / T645, MQTT, and HTTP. It acquires main grid tripping intelligent alarms, main grid SOE data, distribution transformer power data, transmission video data, and typhoon weather data. Different access strategies are adopted based on the characteristics of different data sources: for main grid switch status and distribution network tripping alarms with high real-time requirements, a real-time streaming transmission method is used, with an access cycle controlled within 1 second; for distribution transformer voltage curves and 7-day weather forecast data, a timed batch transmission method is used, with an access cycle configured from 5 minutes to 1 hour.

[0078] The data processing layer is used to clean, standardize, and map the data acquired by the data access layer. Data cleaning includes using outlier detection algorithms to identify and remove outliers from the data, and using interpolation to fill in missing data to ensure data integrity. Data standardization includes establishing unified data specifications and converting device parameters with different formats in different systems into a unified format. Mapping includes establishing relationships between different data sources based on the unique identifier of the device.

[0079] The data storage layer is used to store basic equipment ledgers and user information structure data using relational databases, store real-time operating data of main transformer temperature and line current using time-series databases, support high-concurrency writing and fast time-series querying, and store topology data using graph databases, which efficiently express equipment connection relationships in the form of graph nodes and edges, facilitating topology traversal and correlation analysis.

[0080] The beneficial effects achieved by the above solution are as follows: The data access layer utilizes a multi-protocol adaptation module to support power grid standard protocols such as IEC61850 and DL / T 645, enabling data access from multiple sources, including the main grid dispatching system and distribution network automation system. This achieves standardized access and unified management of multi-source data. The data processing layer deploys data cleaning and standardized processing nodes to achieve integrated data management. By storing basic ledger data in a relational database, real-time operational data in a time-series database, and topology data in a graph database, efficient data storage and rapid querying are achieved. Simultaneously, data backup and disaster recovery mechanisms ensure data security and reliability.

[0081] S2: Automatic generation of topology for all voltage levels based on topology rules. Construct a topology association rule library. Based on geospatial data, automatically identify device connection relationships according to the rule library, generate topology at each level and merge them into an initial topology that connects all voltage levels.

[0082] Based on geospatial data, the system automatically identifies device connection relationships using a rule base, generates topologies at each level, and merges them into an initial topology that connects all voltage levels. Specifically, this includes:

[0083] Determine the equipment connection rules, voltage level matching rules, power supply path constraint rules, and geospatial association rules. At the same time, import the coordinate information of each device based on the geospatial data in the GIS system.

[0084] Based on the rule base, an improved breadth-first search algorithm is used to identify the connection relationships between power generation, transmission lines and substations, and generate a power generation-transmission topology;

[0085] By matching equipment voltage levels, the connection relationships between the substation main transformer and busbar, and between the busbar and outgoing switch are identified, and a transmission-substation topology is generated.

[0086] Based on the path information of the distribution network lines, the connection relationship between the outgoing line switch and the distribution network branch line, and between the branch line and the distribution transformer is identified, and the substation-distribution topology is generated.

[0087] By using the access point information of low-voltage users, the connection relationship between distribution transformers and low-voltage branch switches, and between branch switches and users is identified, and a distribution-low-voltage user topology is generated.

[0088] The topologies of each level are merged to form an initial topology that connects all voltage levels. Each topology node is associated with a unique identifier of the device, realizing the mapping with the power grid data resource pool of all voltage levels.

[0089] The specific implementation steps of the improved breadth-first search algorithm are as follows:

[0090] Core hub nodes in the power grid are selected as the starting set for topology generation, with priority given to hub substations of 500kV and above.

[0091] The basic information and connection relationships of the main transformers, busbars, switches and instrument transformers in the starting substation are obtained through the data fusion system, and an initial topology node library is established.

[0092] Starting from the starting set, perform a layered traversal according to the voltage level from high to low to ensure the logic and completeness of the topology generation;

[0093] The first layer traverses the 220kV-500kV transmission lines and associated substations, obtains the conductor type, length, impedance parameters of the transmission lines and the equipment information of the associated substations, adds these devices as new topology nodes to the node library, and records the connection relationships between nodes.

[0094] The second layer traverses the 110kV-220kV power distribution lines and substations, including equipment such as switches, busbars, and transformers in the substations, and establishes the topological connections of the power distribution links.

[0095] The third layer traverses low-voltage branch lines, distribution transformers, low-voltage switches and user terminals of 10kV and below, completing the topology coverage of the entire voltage level from high voltage to low voltage.

[0096] The device connection relationships obtained through traversal are verified using preset power grid topology logic rules, and erroneous connection relationships are eliminated.

[0097] The verified device nodes and connection relationships are converted into standardized graph structure data and stored in a graph database. In the graph structure, each device is a node, and the node attributes include device ID, device type, voltage level, rated parameters, and installation location information. The connection relationship between devices is an edge, and the edge attributes include connection type, conductor impedance, topology distance, and operating status information. At the same time, metadata information of the topology structure is generated, including topology generation time, data source, and coverage.

[0098] The beneficial effects achieved by the above scheme are as follows: By establishing an initial topology node library, a foundation is provided for subsequent topology traversal; by performing layered traversal according to voltage levels from high to low, the logic and completeness of topology generation can be ensured; during the traversal process, the connection ports of the equipment, conductor types, and topology distance information are recorded in real time, providing data support for subsequent topology verification and application; by eliminating erroneous connection relationships, the accuracy of the topology structure can be ensured. For example, when verifying the matching of the main transformer and the voltage level, a 110kV main transformer can only be connected to a 110kV bus and cannot be directly connected to a 220kV bus; when verifying the compliance of the connection between the switch and the bus, the disconnecting switch should be directly connected to the bus and cannot be connected across voltage levels; when verifying the matching of the line and the equipment, a 10kV line can only be connected to a 10kV distribution transformer and cannot be connected to a 35kV distribution transformer. For the erroneous connection relationships found during verification, they are marked as abnormal and output as logs, which facilitates the investigation of causes by technical personnel; by generating metadata information of the topology structure, topology management and traceability are facilitated.

[0099] S3: Based on real-time data, dynamic topology updates are implemented, an update triggering mechanism is established, data changes are monitored and local adjustments to the initial topology are triggered, and the initial topology is updated in real time.

[0100] Establish a topology update triggering mechanism, which includes equipment status change triggering, measurement data mutation triggering, and manual operation triggering.

[0101] Real-time monitoring of information on the addition, decommissioning, and modification status changes of equipment in the standardized data asset database, as well as sudden changes in real-time measurement data such as main transformer power, line current, and switch status;

[0102] When the topology update condition is triggered, the associated information of the changed equipment is extracted, and the initial power grid topology is locally adjusted according to the topology association rule library to update the equipment connection relationship and power supply path.

[0103] If triggered by manual operation, it supports manual editing of topology nodes and connection relationships through a visual interface, and synchronously updates the associated data mapping.

[0104] S4: Construct a multi-dimensional verification indicator system, which includes topology connectivity indicators, data consistency indicators, alarm correlation accuracy indicators, and meteorological adaptability indicators. Verify the multi-dimensional verification indicator system, and trigger a topology local update mechanism when topology anomalies occur.

[0105] The multi-dimensional verification indicator system was verified, specifically including:

[0106] The connectivity of the topology path is verified by real-time power flow data. The connectivity of the topology path includes: traversing the topology nodes to verify the connectivity of each power supply path, identifying isolated nodes and redundant connections. When the active power transmission value of a certain topology path is 0 and there is no power outage alarm, it is determined to be an abnormal topology connectivity.

[0107] By comparing real-time data obtained from different data sources by the same device, the conservation relationship of power and current electrical quantities on the same power supply path in the topology is verified. When the data deviation exceeds the threshold and lasts for more than 5 seconds, it is judged as an abnormal data consistency.

[0108] The status of associated devices is linked to the topology connection relationship. Alarm data is associated with the topology to verify the location of alarm devices in the topology and whether the associated devices are consistent with the actual situation. If there is a mismatch, it is determined to be an abnormal status matching.

[0109] By combining meteorological data on typhoons, ice accumulation, and wildfires, the meteorological impact range of transmission lines and tower equipment in the topology is analyzed, and the rationality of equipment association in the topology is verified.

[0110] When multi-dimensional verification indicators detect topological anomalies, a topological local update mechanism is triggered. The local update mechanism includes anomaly location, incremental update, and update verification.

[0111] Anomaly localization specifically includes:

[0112] By using a partitioned location algorithm to narrow down the anomaly range and improve location efficiency, the power grid is divided into multiple topological partitions according to geographical area and voltage level, and each partition is equipped with an independent verification unit.

[0113] When an anomaly is detected, the topology partition where the anomaly is located is determined, and the source of the anomaly is traced through the device association within the partition to locate the topology segment that needs to be updated. For example, if an anomaly in the topology connectivity of 10kV line B is detected, the distribution partition where the line is located is first determined, and then the switches, transformers and other equipment associated with the line are traced to locate the abnormal topology node.

[0114] Incremental updates specifically include:

[0115] The abnormal topology segments are locally updated. For newly added device nodes, a unique device ID is automatically assigned, device parameter information is obtained, and a topology connection relationship with surrounding devices is established.

[0116] For removed device nodes, delete their node information and related topology edge information in the graph database;

[0117] For topological segments with changing connectivity, update the state attributes of the relevant topological edges, changing the connected state to the disconnected state;

[0118] During incremental updates, the updated content, update time, and operator information are recorded in real time to facilitate topology traceability.

[0119] The update verification specifically includes:

[0120] The Newton-Raphson method is used to perform power flow calculations on the updated topology segments to verify the rationality of power transmission in the topology path. If the calculation results conform to the power grid operation law, it indicates that the topology update is correct.

[0121] The device status comparison compares the updated topology status with the real-time device status to ensure that the topology structure is consistent with the actual operating status of the devices.

[0122] If the verification passes, the topology update is completed and synchronized to the topology visualization platform. If the verification fails, the process returns to the anomaly location stage to re-investigate the problem.

[0123] The beneficial effects achieved by the above scheme are as follows: The topology connectivity index mainly verifies the connectivity of topology paths through real-time power flow data. During normal grid operation, each connected topology path should have a certain amount of active power transmission. When the active power transmission value of a certain topology path is consistently 0 and there is no corresponding power outage alarm information, it is determined to be an abnormal topology connectivity, which may indicate problems such as switch malfunction or line disconnection. For example, if the topology path of 110kV line A shows a connected state, but real-time power flow data shows its active power is 0 and there is no power outage alarm for that line, then the topology connectivity of that line is determined to be abnormal. The data consistency index is used to compare parameter information obtained from different data sources by the same device to ensure data accuracy. For example, when comparing the remote signaling status data of the same switch in the dispatch automation system and the distribution network automation system, if the switch status (open / closed) in the two systems is inconsistent and the deviation lasts for more than 5 seconds, it is determined to be a data inconsistency anomaly. Similarly, when comparing the rated capacity parameters of the same main transformer in the equipment ledger system and the measurement system, if the deviation exceeds 5%, it is determined to be a data inconsistency anomaly. The status matching index is used to associate the equipment status with the topology connection relationship to ensure that the topology structure can accurately reflect the actual operating status of the equipment. For example, when the remote signaling status of the distribution network switch is tripped, it is checked whether its associated topology branch is marked as disconnected. If the topology branch still shows as connected, it is determined to be an abnormal status matching. When the main transformer trips due to a fault, it is checked whether its associated bus topology is marked as undervoltage. If it is not marked, it is determined to be abnormal. By constructing a multi-dimensional verification index system, the topology structure can be comprehensively verified from different angles. Through three stages of anomaly location, incremental update and update verification, the topology structure can be updated quickly and accurately. By performing secondary verification on the updated topology segments, the accuracy and reliability of the updated topology structure can be ensured. Local updates are performed on the located abnormal topology segments, rather than reconstructing the entire topology structure, which effectively improves the update efficiency.

[0124] S5: Build a visualization platform to realize multi-scale topological display and interactive linkage with multi-source data, and automatically correct and update topological deviations;

[0125] Based on the grid topology that connects all voltage levels, a multi-dimensional visualization display system is constructed, including geographical layout map display, electrical wiring diagram display, and three-dimensional point cloud grid structure display.

[0126] The topology can be displayed in layers, scaled, rotated, and queried through, and the real-time status of devices, measurement data, alarm information and topology can be displayed in linkage.

[0127] Quickly query device ledgers, real-time data, historical curves and related device information through topology nodes; receive topology editing, relationship adjustment and anomaly feedback commands initiated by users through a visual interface;

[0128] For topology deviations identified by dynamic verification, the system automatically completes topology adjustments based on preset correction rules, including adding new device nodes, deleting retired device nodes, and updating device connection relationships. It also generates topology update logs and pushes them synchronously to the power grid dispatching and operation and maintenance management related systems.

[0129] A 3D point cloud network structure demonstration, specifically including:

[0130] The operational status display is used to show the current topology of the power grid in real time, overlaying real-time data on main transformer temperature, line current and switch status. Different colors are used to indicate the operating status of equipment, and clicking on equipment nodes allows you to view detailed ledgers and real-time curves.

[0131] Historical display is used to store topology data from different historical periods. It supports querying historical topology by time dimension, uses colors to distinguish line equipment put into operation in different years, and provides a comparison function between historical topology and current topology.

[0132] The planning display is used to access the distribution network planning data, overlay planned equipment nodes and lines on the current topology, support the separate display and simulation analysis of the planned topology, and predict the operating effect of the planned topology.

[0133] In summary, this invention constructs a multi-source data fusion system, integrating equipment and operational data from the main grid, distribution network, and low-voltage network. It employs an improved breadth-first search algorithm to generate a full-voltage-level topology covering generation, transmission, substation, distribution, low-voltage, and user systems. This solves the problem of voltage-level fragmentation in traditional topology management, providing dispatchers with a comprehensive and unified view of the power grid topology, facilitating a better understanding of the overall grid operation. Through data standardization and connectivity verification, it effectively reduces human error and improves topology generation efficiency and accuracy. Furthermore, the improved breadth-first search algorithm enables partial topology updates, avoiding full reconstruction and shortening the topology update response time, allowing for timely reflection of changes in grid operation. By establishing a multi-dimensional verification index system including connectivity, consistency, matching, and meteorological adaptability, combined with real-time operational data and equipment status, it continuously verifies the topology, enabling rapid detection of topology anomalies and triggering further updates. The new technology addresses the issue of lagging traditional topology updates, ensuring that topology data remains consistent with the actual operating status of the power grid. This provides a real-time and accurate topology foundation for applications such as rapid fault location and power flow calculation. The built topology visualization platform supports multi-state display (operating, historical, and planned states), achieving deep integration of topology data with business applications. For example, in fault handling, the topology-based fault tracing function can reduce fault location time from 30 minutes to less than 5 minutes, significantly improving fault handling efficiency. In meteorological disaster response, the meteorological impact assessment function can provide early warnings of the impact of typhoons, icing, and other disasters on the power grid, buying time for maintenance and repair, and reducing power grid fault losses. Automated topology generation and dynamic verification reduce manual intervention, lowering the workload of maintenance personnel. Simultaneously, fault handling and scheduling decisions based on accurate topology data shorten fault repair time, reduce power outage losses, and lower maintenance costs and economic losses for power grid companies.

[0134] Furthermore, when a sudden change in electrical quantity is detected in the multi-source data, and it is necessary to accurately distinguish whether the sudden change in electrical quantity is caused by a real physical topology break or by a transient electrical inrush current in the system, the real-time update of the initial topology specifically includes:

[0135] The electrical feature sequence of each topological node in the initial topology in the multi-source data is extracted in a continuous historical time period, and the physical impedance parameters and spatial connection relationship between each topological node in the initial topology are extracted based on geospatial data to construct a topological distance attenuation matrix;

[0136] Based on the topology distance decay matrix, determine the total number of target evaluation nodes within the initial topology and the number of adjacent topology nodes within a preset topology hop range from the target evaluation node, and simultaneously establish a sliding data acquisition time window for the adjacent topology nodes;

[0137] The sliding data acquisition time window is used to obtain the real-time electrical measurement data of the adjacent topology nodes at the current sampling time and the real-time electrical measurement data at the previous sampling time, and simultaneously obtain the number of abnormal alarm nodes in the topology partition to which the target evaluation node belongs and the total number of topology nodes.

[0138] Based on the acquired real-time electrical measurement data, the number of abnormal alarm nodes, and the shortest topological path length within the topological distance attenuation matrix, the spatiotemporal disturbance index of the target evaluation node in the current environment is calculated. This spatiotemporal disturbance index quantifies the probability of a real physical topological break or alteration occurring around the target evaluation node, rather than measurement jumps caused by transient electrical inrush currents. The specific formula for calculating the spatiotemporal disturbance index of the target evaluation node within the initial topology in the current environment is as follows:

[0139]

[0140] In the formula, the Indicates the target evaluation node The spatiotemporal disturbance index; This represents the total number of adjacent topology nodes; Indicates the sequence number of the adjacent topological nodes; Indicates the first The voltage level weighting coefficient of each adjacent topology node; the higher the voltage level, the larger the voltage level weighting coefficient. Indicates the first The adjacent topology nodes at the current sampling time The real-time electrical measurement data; Indicates the first The adjacent topology nodes at the pre-sampling time The real-time electrical measurement data; Indicates the relationship with the first The reference electrical constants for the voltage levels corresponding to the adjacent topology nodes; Represents the natural constant; the This represents the spatial attenuation coefficient, with dimensions of the reciprocal of length; The target evaluation node is extracted from the topological distance decay matrix. With the The shortest topological path length between adjacent topological nodes; Indicates the disturbance balance coefficient; the Indicates the alarm gain coefficient; the Indicates the number of abnormal alarm nodes; This represents the total number of nodes in the topology;

[0141] The basic disturbance values ​​of the target evaluation node during its historical steady-state operation are retrieved and combined with the current grid load level to generate a dynamic comparison threshold. The calculated spatiotemporal disturbance index is compared with the dynamic comparison threshold. If the spatiotemporal disturbance index exceeds the dynamic comparison threshold, the target evaluation node is determined to meet the topology update triggering condition, and the initial topology local adjustment is triggered for the target evaluation node.

[0142] The working principle and beneficial effects of the above technical solution are as follows:

[0143] Regarding the real-time update process of the initial topology, this embodiment extracts the electrical feature sequences of each topological node within the initial topology from the multi-source data over a continuous historical time period. Based on geospatial data, it extracts the physical impedance parameters and spatial connectivity between the topological nodes within the initial topology, constructing a topology distance attenuation matrix. It is understood that in the operation of an AC / DC hybrid power grid, the grid topology is in a dynamic state of change. Besides planned switching operations, physical insulation faults in equipment, instantaneous grounding of overhead lines, and communication errors in remote measurement terminals can all cause topology state jumps at the dispatching system master station. To improve the accuracy of the topology update mechanism and the reliability of data traceability, this embodiment introduces electrical feature sequences into the update mechanism. Specifically, the system uses synchronous phasor measurement devices and distribution network automation feeder terminals deployed on the low-voltage side of busbars and distribution substations at all levels. Using the IEEE 1588 high-precision time synchronization protocol as the reference clock, it continuously collects continuous time-series data such as the effective values ​​of three-phase voltage, effective values ​​of three-phase current, active power, reactive power, and absolute values ​​of voltage phase angles for each topology node at a preset high-frequency sampling frequency (e.g., 50 to 100 frames per second). This data array with absolute timestamps constitutes the electrical characteristic sequence, typically configured as double-precision floating-point numbers and stored in a time-series database. Furthermore, since the power grid is a three-dimensional physical spatial network in its topology, the tightness of electrical coupling between any two electrical nodes depends not only on the number of logical edge hops in the graph theory model but also on the physical laying path in real geographic space and the physical impedance parameters constituted by the conductor material. Therefore, the system's backend service calls spatial vector data from the asset ledger management system and the Geographic Information System (GIS) through interface protocols to extract physical impedance parameters such as conductor cross-section type, laying path length, and resistivity of the material under standard conditions for each physical line. The actual branch AC impedance between any two topological nodes with a direct physical connection. Through the complex field formula The calculation yielded the following result: The branch AC resistance... The calculation formula is:

[0144]

[0145] In this resistance formula, This refers to the AC resistance of the branch, with the unit of measurement being ohms (Ω). ); This refers to the reference unit length resistance obtained from the equipment register, with its dimension being ohms per kilometer (Ω). ); This refers to the length of the physical laying path retrieved from the geographic information system, with the unit of measurement being kilometers (km). ); This represents the temperature conversion factor for the conductor, with its unit of measurement being one-tenth of a degree Celsius. ); This represents the real-time ambient temperature collected by a temperature sensor, with the unit of measurement being degrees Celsius (°C). ); The reference temperature is typically configured as 20°C. The constant. Furthermore, the branch AC power frequency reactance... The calculation formula is defined as follows: In this reactance formula, This refers to the branch AC power frequency reactance, whose dimension is ohm ( ). ); This indicates the power grid frequency, which is 50Hz under the Chinese power grid standard. Pi is a constant. The equivalent geometric mean distance of the three-phase conductors, calculated from the phase-to-phase geometric distance of the transmission line, is expressed in millimeters (mm). ); This represents the equivalent calculated radius of the conductor after considering the skin effect, in millimeters (mm). ); constant product term This is used to convert the inductance dimension, Henry per kilometer, to ohms. The system performs graph mapping and fusion of the above impedance parameters with spatial geographic connectivity to construct the topology distance attenuation matrix in memory, reflecting the damping characteristics of transient electrical disturbance propagation. This matrix is ​​typically stored in Compacted Sparse Row (CSR) format to optimize memory usage. An intermediate parameter is defined within the matrix: the equivalent shortest electrical topology path length. Its calculation logic is as follows: ,in This represents the AC impedance constant of the selected baseline type for the entire network, in ohms (Ω). This conversion eliminates the impedance difference between wires of different diameters when directly comparing their physical lengths.

[0146] In this embodiment, based on the topology distance attenuation matrix, the total number of target evaluation nodes within the initial topology and adjacent topology nodes within a preset topology hop range from the target evaluation node is determined, and a sliding data acquisition time window is established for the adjacent topology nodes. Specifically, when the underlying state estimation and verification module detects that the voltage amplitude or current vector of a node deviates from a preset static dead zone range (such as ±5% of the rated operating value), the node is marked as the target evaluation node. The system then uses the target evaluation node as the core root node, based on the topology distance attenuation matrix in memory, and calls the standard breadth-first search (BFS) algorithm to traverse the directed acyclic graph data structure layer by layer outward. In the implementation of the traversal algorithm, the system initializes a queue to store nodes to be visited and maintains a set of visited nodes and a hash table recording the level. The preset topology hop range is configured to be 3 to 5 hops in this embodiment. The physical basis for setting this numerical range is that, in medium and low voltage distribution network branches, electrical disturbances caused by local topological physical breaks or single-phase metallic ground faults, after crossing 3 to 5 electrical impedance nodes with transformer leakage inductance isolation effect and line distributed capacitance filtering effect, typically attenuate their transient wavefront disturbance amplitude to below the background measurement noise level of the system's measurement devices. By setting this hop count cutoff condition, the system determines the total number of adjacent topological nodes affected by the sudden change, thereby dynamically dividing a local edge computing task cluster. Furthermore, the system instantiates and establishes a sliding data acquisition time window in the form of a ring buffer in the server's memory. The absolute time width of this sliding time window... An adaptive mapping configuration is performed based on the design voltage level of the target evaluation node: for ultra-high voltage backbone network hub nodes of 500kV and above, the absolute time width of the window is... The window is set between 20ms and 100ms (corresponding to 1 to 5 power frequency cycles); for busbar nodes in high-voltage substations ranging from 110kV to 220kV, the absolute time width of the window is... The window is set to be between 200ms and 500ms; for the terminal nodes of distribution network areas at 10kV and below, the absolute time width of the window is... The window size is set to be between 1000ms and 3000ms. The sliding step size of the window is set to an integer multiple of its corresponding sampling period, and the overlap rate of the sliding windows is configured to be 50% to ensure the continuity of time series extraction.

[0147] In this embodiment, the real-time electrical measurement data of the adjacent topology nodes at the current sampling time and the real-time electrical measurement data at the previous sampling time are obtained through the sliding data acquisition time window. Simultaneously, the number of abnormal alarm nodes and the total number of topology nodes within the topology partition to which the target evaluation node belongs are also obtained. Since the sliding data acquisition time window is continuously polled in the scheduling system background based on timer interrupts, the system uses a set window sliding step size as the time difference to extract the data of each adjacent topology node before the fault occurs (i.e., at the previous sampling time). Steady-state measurement baseline data, and data at the current state (i.e., the current sampling time). The system calculates transient change measurement data. This differential measurement data is used to remove low-frequency load slowly changing components and extract high-frequency differential components caused by sudden topology changes. Simultaneously, the system receives, in real-time, the main network tripping intelligent alarm event sequence record message and relay protection action message from the scheduling data network via an enterprise service bus or message queue (such as a Kafka cluster). In this data flow stage, the system calls the hash comparison filter module to deduplicate and merge alarm messages with the same device ID and similar timestamps (time difference less than the preset anti-jitter tolerance), eliminating duplicate alarms caused by communication channel errors or device contact jitter. After processing, within the absolute time boundary of the current sliding data acquisition time window, the system counts the number of abnormal alarm nodes that have triggered hardware protection actions or software alarms within the macro-topology partition to which the current target evaluation node belongs. Simultaneously, the system executes an SQL query to obtain the total number of topology nodes within the specific topology partition from the configuration relational database. These two statistical indicators are passed as macro-constraint condition variables to the subsequent algorithm model.

[0148] In this embodiment, based on the acquired real-time electrical measurement data, the number of abnormal alarm nodes, and the shortest topological path length within the topological distance attenuation matrix, the spatiotemporal disturbance index of the target evaluation node in the current environment is calculated. The spatiotemporal disturbance index is used to quantify the probability of a real physical topological break or alteration occurring around the target evaluation node, rather than measurement jumps caused by transient electrical inrush currents.

[0149] In the above formula for calculating the spatiotemporal perturbation index, the definitions and physical meanings of each parameter are as follows:

[0150] The Indicates the target evaluation node The aforementioned spatiotemporal perturbation index is a double-precision floating-point number, with a theoretical value range of [missing value]. The The total number of adjacent topological nodes determined by the aforementioned breadth-first search constitutes the upper bound of the summation operation; This represents the traversal index value of the adjacent topological node in the summation loop; the Indicates the first The voltage level weighting coefficient of adjacent topology nodes is used to reflect the weight ratio of nodes with different voltage levels in the network structure. The specific parameter mapping table is configured as follows: 500kV and above nodes have a value of 1.0; 220kV nodes have a value of 0.8; 110kV nodes have a value of 0.6; 35kV nodes have a value of 0.4; and 10kV and below nodes have a value of 0.2. Indicates the first The adjacent topology nodes at the current sampling time The extracted real-time electrical measurement data, in this embodiment, is configured as the vector amplitude of the effective value of the three-phase fundamental voltage, with the unit being kilovolts (kV). ); the Indicates the first The adjacent topology nodes at the pre-sampling time The extracted real-time electrical measurement data is in kilovolts (kV). ); the Indicates the relationship with the first Each adjacent topology node corresponds to a reference electrical constant that matches its voltage level. For example, for a 10kV node, the... .pass The operation achieves dimensionless processing of data per-unit; the The base (constant) of the natural logarithm ); the This represents the spatial attenuation coefficient, with dimensions equal to the reciprocal of length. For outdoor overhead power lines, the aforementioned The configuration is set to 0.02; for underground cross-linked polyethylene insulated power cables, the... Configured to 0.05. Contains the aforementioned natural exponential term operator Used to calculate the exponential decay of electromagnetic energy along a distributed parameter path; This represents the target evaluation node extracted from the topological distance decay matrix. With the The shortest topological path length between adjacent topological nodes, in kilometers (km) ); the This represents the disturbance balance coefficient, with a theoretical effective range of values. In this embodiment, the default configuration is 1.2; This represents the alarm gain coefficient, with a theoretical effective value range of [value missing]. In this embodiment, the default configuration is 2.5; The number of abnormal alarm nodes after deduplication and merging is represented by a non-negative integer variable; The total number of nodes in the topology is represented by a positive integer. The macroscopic state penalty term in the latter part of the formula uses the natural logarithm function. The logarithmic function can be constructed in the... When the value is small, it provides an approximately linear gain (based on Taylor series expansion). ,when (time), while in the As the value increases significantly, its numerical growth rate gradually decreases and tends to converge smoothly. This design is intended to prevent the penalty term from diverging during a large number of concurrent alarm messages, thus masking the calculation results of the microscopic electrical characteristics in the first half of the formula.

[0151] In this embodiment, the system retrieves the basic disturbance values ​​of the target evaluation node during its historical steady-state operation and generates a dynamic comparison threshold by combining them with the current grid load level. The calculated spatiotemporal disturbance index is compared with the dynamic comparison threshold. If the spatiotemporal disturbance index exceeds the dynamic comparison threshold, the target evaluation node is determined to meet the topology update triggering condition, and the initial local topology adjustment is triggered for the target evaluation node. Specifically, the system retrieves a fault-free operation dataset of the target evaluation node from the time-series database, showing the same time window over the past 30 days and meteorological parameter (temperature, humidity) deviations within a specified tolerance range. The statistical algorithm module is then called to calculate the historical statistical average of the spatiotemporal disturbance index in this dataset. and standard deviation .calculate And set it as the basic disturbance value. Subsequently, the system obtains the total real-time active load of the entire network at the current time point. and the maximum rated load capacity of the space frame design. Calculate the dimensionless real-time load rate of the cross section. The system further reads the load rate from the previous sampling period. and the load factor of the current sampling period. Obtain the time interval between the two samplings. (Unit of measurement is seconds) Generate the dynamic comparison threshold. The calculation formula is:

[0152]

[0153] In this dynamic threshold formula: the The dimensionless threshold variable is the final output; The aforementioned basic disturbance values; The real-time load factor is dimensionless; The static load redundancy factor is configured with a constant of 0.2; The dynamic load change rate penalty time constant, its unit of measurement is seconds (seconds). Differential term Its dimension is the reciprocal of the second ( ),and The product remains dimensionless after multiplication. The function is limited to adjusting the threshold only during periods of increased load. This applies when the kernel calculates the spatiotemporal perturbation index for a specific node. Greater than the dynamically calculated dynamic comparison threshold When the target evaluation node is selected, the system pushes it into the real-time update operation instruction queue, calls the improved breadth-first search algorithm, extracts the remote signaling connectivity status bits of the associated physical devices within a specified hop range centered on the node, and performs a logical update to disconnect or create a new topology edge association.

[0154] Furthermore, when receiving a large-scale meteorological disaster warning, and needing to defend against massive concurrent device status change data that could cause graph database computing power overload and topology view jitter and crash, the multi-dimensional verification index system is verified, and a topology local update mechanism is triggered when topology anomalies occur. Specifically, this includes:

[0155] Multi-dimensional meteorological impact parameters, including wind speed, ice thickness, and precipitation, are extracted from the multi-source data in real time, and the multi-dimensional meteorological impact parameters are converted into comprehensive disaster intensity values ​​based on a preset meteorological disaster assessment model.

[0156] Determine whether the comprehensive disaster intensity value exceeds the preset extreme disaster critical threshold. If it exceeds the extreme disaster critical threshold, issue a global write transaction latch instruction to the graph database storing the initial topology, forcibly suspending the global breadth-first search process for the initial topology and the dynamic verification and update process of network connectivity.

[0157] Based on the latitude and longitude range of the multi-dimensional meteorological impact parameters, a disaster spatial envelope is generated in the standard geographic coordinate system of the power grid, and the associated device set within the disaster spatial envelope is mapped to the network space, thereby rigidly isolating the associated device set into a disaster-affected closed topology partition.

[0158] Create an independent high-speed cache sandbox in system memory, and clone the current topology connection snapshot of the disaster-affected closed topology partition into the high-speed cache sandbox;

[0159] Inside the cache sandbox, data interaction with the external main network view is blocked. Based only on the edge remote signaling data of the surviving devices in the disaster-affected closed topology partition, a high-frequency local topology reconstruction exercise is performed on the associated device set, and the set of steady-state edge topology nodes generated during the reconstruction exercise is continuously updated and recorded.

[0160] The multi-dimensional meteorological impact parameters are continuously monitored. When it is confirmed that the comprehensive disaster intensity value has fallen below the extreme disaster critical threshold and after a preset system smoothing period, the global write transaction latch instruction of the graph database is released. The final topological association state of the steady-state edge topological node set is extracted and a connection conflict check is performed with the main network edge node in the graph database at this time.

[0161] If the connection conflict verification passes, the final topology association state is written to the initial topology in an incremental merging manner through an asynchronous message queue to eliminate the invalid system computing power consumption and topology view jitter caused by extreme weather disturbances.

[0162] The working principle and beneficial effects of the above technical solution are as follows:

[0163] Furthermore, regarding the graph theory topology update mechanism under extreme weather conditions, this embodiment extracts multi-dimensional meteorological impact parameters, including wind speed, icing thickness, and precipitation, from the multi-source data in real time, and converts these multi-dimensional meteorological impact parameters into a comprehensive disaster intensity value based on a preset meteorological disaster assessment model. Specifically, the system periodically (e.g., every 5 minutes) pulls data from micro-weather station sensors deployed on transmission line towers and meteorological radar echo maps through a data interface. The extracted core meteorological feature values ​​include: the real-time environmental wind speed (variable labeled as...). The dimensional unit is The real-time equivalent ice thickness (the variable is labeled as) The dimensional unit is ) and the hourly cumulative precipitation (variable labeled as The dimensional unit is The preset meteorological disaster assessment model incorporates the limit values ​​of engineering physical parameters from the power grid static asset ledger data. The system extracts the design limit wind speed constant (marked as...) of the transmission towers in the affected area from the relational database. The dimensional unit is The maximum ice-resistant load-bearing thickness limit constant of the conductor design (marked as...) The dimensional unit is ), and the foundation's precipitation limit constant (marked as ), The dimensional unit is Calculate the comprehensive disaster intensity value. The formula is defined as:

[0164]

[0165] In the evaluation calculation formula, the These are dimensionless real-valued variables output by the model. , , These are the wind damage hazard weighting coefficient, the ice damage hazard weighting coefficient, and the water damage hazard weighting coefficient, respectively, and the sum of these three weighting coefficients is set to a constant of 1. The wind speed and icing terms in the formula are constructed using square terms to correspond to the wind pressure load ( (and the growth pattern of ice-covered volume in annular cylinders.)

[0166] In this embodiment, it is determined whether the comprehensive disaster intensity value exceeds a preset extreme disaster critical threshold. If it does, a global write transaction latch instruction is issued to the graph database storing the initial topology, forcibly suspending the global breadth-first search process and the network connectivity dynamic verification and update process for the initial topology. Specifically, the extreme disaster critical threshold set by the system is configured as a real number 1.5. When the background computing engine detects that the comprehensive disaster intensity value is greater than this critical threshold, the system master node sends the global write transaction latch instruction to the underlying graph database cluster through a distributed coordination component (such as ZooKeeper or Redis). This instruction implements an exclusive write lock at the database transaction level, intercepting database structure update requests including INSERT, UPDATE, and DELETE operations. At the same time, the operating system kernel scheduler forcibly suspends the background global breadth-first search traversal task thread and the power flow calculation verification thread by sending an interrupt signal. The system's suspension mechanism is designed to maintain the graph database in a stable snapshot of the last verified state.

[0167] In this embodiment, based on the latitude and longitude range of the multi-dimensional meteorological impact parameters, a disaster spatial envelope is generated in the standard geographic coordinate system of the power grid. The associated device set within the disaster spatial envelope is then mapped to the network space, thereby rigidly isolating the associated device set into a disaster-affected closed topology partition. The specific execution steps are as follows: The system's built-in GIS spatial computing engine obtains the coordinate point set of the meteorological disaster center. On the WGS-84 latitude and longitude coordinate system layer, the dynamic convex hull algorithm in computational geometry is called to generate a polygonal region, and a buffer radius of 5km to 10km is extended outward to generate the disaster spatial envelope. Subsequently, the system uses a spatial intersection query command from a graph database (such as the ST_Intersects function) to retrieve all physical entity nodes whose geographic coordinates are located within the envelope. After obtaining the globally unique device identifier (ID) of the aforementioned physical entities, the system maps them to the graph theory logical network space to generate the associated device set. At the logical level, the system sets the weight of the connected edges crossing the boundary of the envelope to infinity, thereby isolating the associated device set into an independent disaster-affected closed topology partition.

[0168] In this embodiment, an independent high-speed cache sandbox is created in the system memory, and the current topology connection snapshot of the disaster-affected closed topology partition is cloned into the high-speed cache sandbox. Inside the high-speed cache sandbox, data interaction with the external main network view is blocked. Based solely on the edge telemetry data of surviving devices within the disaster-affected closed topology partition, a high-frequency local topology reconstruction exercise is performed on the associated device set, and the set of steady-state edge topology nodes generated during the reconstruction exercise is continuously updated and recorded. In this step, the system allocates a physically contiguous and independent memory block in the dynamic random access memory (DRAM) of the core server as the high-speed cache sandbox. The system writes the topology connection state of the disaster-affected closed topology partition at the moment before isolation as a seed snapshot into the high-speed cache sandbox. The network communication ports inside the sandbox operating environment are configured in one-way inbound mode by the firewall policy module, rejecting outbound database update requests across the sandbox boundary. The graph computing engine within this sandbox environment only receives edge switch telemetry change data and short-circuit over-limit alarm messages uploaded by automated remote terminals with normal communication within the disaster-affected partition. Within the sandbox, the algorithm engine performs local topological connectivity traversal and state deduction at a set high frequency. The engine monitors the degree of each graph node after each iteration and calculates the degree of continuous nodes. The time series variance of the node's degree within each calculation time period. When the variance calculation result is less than the set convergence threshold (approaching zero), the node's state is determined to have converged, it is added to the set of steady-state edge topology nodes, and the record is updated in the sandbox log memory area.

[0169] In this embodiment, the multi-dimensional meteorological impact parameters are continuously monitored. When it is confirmed that the comprehensive disaster intensity value has fallen below the extreme disaster critical threshold and after a preset system smoothing period, the global write transaction latch instruction of the graph database is released. The final topology association state of the steady-state edge topology node set is extracted and a connection conflict check is performed with the main network edge nodes in the graph database at this time. If the connection conflict check passes, the final topology association state is written to the initial topology in an incremental merging manner through an asynchronous message queue to eliminate the invalid system computing power consumption and topology view jitter caused by extreme meteorological disturbances. Specifically, when the comprehensive disaster intensity value calculated by the meteorological model is continuously less than the extreme disaster critical threshold, the system kernel triggers a timer, and the start time is configured to be a system smoothing period of 15 to 30 minutes. After the timer countdown ends, the main control process sends an unlock message to the graph database to release the global write transaction latch instruction. The system extracts the structure data of the steady-state edge topology node set from the memory sandbox and performs a connection conflict matching check with the frozen main network edge nodes in the graph database. The set of logical rules for verification includes: comparing whether the three-phase phase sequences on both sides of the quasi-connecting switch are consistent, verifying whether the rated voltage level fields of adjacent nodes are equal, and calculating whether the theoretical short-circuit capacity estimate of the merged local nodes is less than the rated breaking short-circuit current limit of the circuit breaker. If all Boolean values ​​of the verification results return True, the system calls a distributed asynchronous message queue (such as a RabbitMQ cluster) to compress the final residual steady-state topology node structure difference data recorded in the sandbox into a graph structure incremental patch package. The incremental patch package is written to and updates the initial topology structure relationship table of the backbone network in the form of non-blocking incremental merge transaction instructions according to the timestamp order of the message queue. This implementation process avoids the problems of large data throughput and computing resource consumption caused by meteorological parameter fluctuations, and maintains the stability and data consistency of the main network topology view update.

[0170] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0171] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention.

Claims

1. A method for automatically generating and dynamically verifying the topology of a power grid spanning all voltage levels, characterized in that: Includes the following steps: S1: Construct a full-voltage-level power grid data fusion system to acquire multi-source data from the main grid dispatching system, distribution network automation system, low-voltage distribution area management system, equipment ledger system, and IoT monitoring platform; S2: Automatic generation of topology for all voltage levels based on topology rules. Construct a topology association rule library. Based on geospatial data, automatically identify device connection relationships according to the rule library, generate topology at each level and merge them into an initial topology that connects all voltage levels. S3: Based on real-time data, dynamic topology updates are implemented, an update triggering mechanism is established, data changes are monitored and local adjustments to the initial topology are triggered, and the initial topology is updated in real time. S4: Construct a multi-dimensional verification indicator system, verify the multi-dimensional verification indicator system, and trigger a topology local update mechanism when topology anomalies occur; S5: Build a visualization platform to realize multi-scale topological display and interactive linkage with multi-source data, and automatically correct and update topological deviations.

2. The method for automatic generation and dynamic verification of power grid topology supporting full voltage level continuity as described in claim 1, characterized in that, The data fusion system includes: The data access layer is used to acquire main grid tripping intelligent alarms, main grid SOE data, distribution transformer power data, transmission video data, and typhoon meteorological data, and adopts differentiated access strategies based on the characteristics of different data sources: The data processing layer is used to clean, standardize, and perform correlation mapping on the data obtained from the data access layer. The data storage layer is used to store basic equipment ledgers and user information structure data using relational databases, real-time operating data of main transformer temperature and line current using time-series databases, and topology data using graph databases, so as to efficiently express the equipment connection relationships in the form of graph nodes and edges.

3. The method for automatic generation and dynamic verification of power grid topology supporting full voltage level continuity as described in claim 1, characterized in that, The process of using geospatial data as a benchmark, automatically identifying device connection relationships based on a rule base, generating topologies at each level, and merging them into an initial topology that connects all voltage levels specifically includes: Determine the equipment connection rules, voltage level matching rules, power supply path constraint rules, and geospatial association rules. At the same time, import the coordinate information of each device based on the geospatial data in the GIS system. Based on the rule base, an improved breadth-first search algorithm is used to identify the connection relationships between power generation, transmission lines and substations, and generate a power generation-transmission topology; By matching equipment voltage levels, the connection relationships between the substation main transformer and busbar, and between the busbar and outgoing switch are identified, and a transmission-substation topology is generated. Based on the path information of the distribution network lines, the connection relationship between the outgoing line switch and the distribution network branch line, and between the branch line and the distribution transformer is identified, and the substation-distribution topology is generated. By using the access point information of low-voltage users, the connection relationship between distribution transformers and low-voltage branch switches, and between branch switches and users is identified, and a distribution-low-voltage user topology is generated. By merging the topologies of each level, an initial topology that connects all voltage levels is formed. Each topology node is associated with a unique identifier of the device, realizing the mapping with the power grid data resource pool of all voltage levels.

4. The method for automatic generation and dynamic verification of power grid topology supporting full voltage level continuity according to claim 3, characterized in that, The specific implementation steps of the improved breadth-first search algorithm are as follows: Core hub nodes in the power grid are selected as the starting set for topology generation, with priority given to hub substations of 500kV and above. The basic information and connection relationships of the main transformers, busbars, switches and instrument transformers in the starting substation are obtained through the data fusion system, and an initial topology node library is established. Starting from the starting set, perform a layered traversal according to the voltage level from high to low to ensure the logic and completeness of the topology generation; The device connection relationships obtained through traversal are verified using preset power grid topology logic rules, and erroneous connection relationships are eliminated. The verified device nodes and connection relationships are converted into standardized graph structure data and stored in the graph database.

5. The method for automatic generation and dynamic verification of power grid topology supporting full voltage level continuity according to claim 4, characterized in that, The implementation process of S3 specifically includes: Establish a topology update triggering mechanism, which includes equipment status change triggering, measurement data mutation triggering, and manual operation triggering. Real-time monitoring of information on the addition, decommissioning, and modification status changes of equipment in the standardized data asset database, as well as sudden changes in real-time measurement data such as main transformer power, line current, and switch status; When the topology update condition is triggered, the associated information of the changed equipment is extracted, and the initial power grid topology is locally adjusted according to the topology association rule library to update the equipment connection relationship and power supply path. If triggered by manual operation, it supports manual editing of topology nodes and connection relationships through a visual interface, and synchronously updates the associated data mapping.

6. The method for automatic generation and dynamic verification of power grid topology supporting full voltage level continuity according to claim 1, characterized in that, The verification of the multi-dimensional verification indicator system specifically includes: The connectivity of the topology path is verified by real-time power flow data. The connectivity of the topology path includes: traversing the topology nodes to verify the connectivity of each power supply path, identifying isolated nodes and redundant connections. When the active power transmission value of a certain topology path is 0 and there is no power outage alarm, it is determined to be an abnormal topology connectivity. By comparing real-time data obtained from different data sources by the same device, the conservation relationship of power and current electrical quantities on the same power supply path in the topology is verified. When the data deviation exceeds the threshold and lasts for more than 5 seconds, it is judged as an abnormal data consistency. The status of associated devices is linked to the topology connection relationship. Alarm data is associated with the topology to verify the location of alarm devices in the topology and whether the associated devices are consistent with the actual situation. If there is a mismatch, it is determined to be an abnormal status matching. By combining meteorological data on typhoons, ice accumulation, and wildfires, the meteorological impact range of transmission lines and tower equipment in the topology is analyzed, and the rationality of equipment association in the topology is verified. When multi-dimensional verification indicators detect topological anomalies, a topological local update mechanism is triggered. The local update mechanism includes anomaly location, incremental update, and update verification.

7. The method for automatic generation and dynamic verification of power grid topology supporting full voltage level continuity according to claim 6, characterized in that, The anomaly location specifically includes: By using a partitioned location algorithm to narrow down the anomaly range and improve location efficiency, the power grid is divided into multiple topological partitions according to geographical area and voltage level, and each partition is equipped with an independent verification unit. When an anomaly is detected, the topology partition where the anomaly is located is determined, and the source of the anomaly is traced through device associations within the partition to locate the topology segment that needs to be updated.

8. The method for automatic generation and dynamic verification of power grid topology supporting full voltage level continuity according to claim 6, characterized in that, The incremental update specifically includes: The abnormal topology segments are locally updated. For newly added device nodes, a unique device ID is automatically assigned, device parameter information is obtained, and a topology connection relationship with surrounding devices is established. For removed device nodes, delete their node information and related topology edge information in the graph database; For topological segments with changing connectivity, update the state attributes of the relevant topological edges, changing the connected state to the disconnected state; During incremental updates, the updated content, update time, and operator information are recorded in real time to facilitate topology traceability.

9. The method for automatic generation and dynamic verification of power grid topology supporting full voltage level continuity according to claim 1, characterized in that, When a sudden change in electrical quantity is detected in the multi-source data, and it is necessary to accurately distinguish whether the sudden change in electrical quantity is caused by a break in the actual physical topology or by a transient electrical inrush current in the system, the real-time update of the initial topology specifically includes: The electrical feature sequence of each topological node in the initial topology in the multi-source data is extracted in a continuous historical time period, and the physical impedance parameters and spatial connection relationship between each topological node in the initial topology are extracted based on geospatial data to construct a topological distance attenuation matrix; Based on the topology distance decay matrix, determine the total number of target evaluation nodes within the initial topology and the number of adjacent topology nodes within a preset topology hop range from the target evaluation node, and simultaneously establish a sliding data acquisition time window for the adjacent topology nodes; The sliding data acquisition time window is used to obtain the real-time electrical measurement data of the adjacent topology nodes at the current sampling time and the real-time electrical measurement data at the previous sampling time, and simultaneously obtain the number of abnormal alarm nodes in the topology partition to which the target evaluation node belongs and the total number of topology nodes. Based on the acquired real-time electrical measurement data, the number of abnormal alarm nodes, and the shortest topological path length within the topological distance attenuation matrix, the spatiotemporal disturbance index of the target evaluation node in the current environment is calculated. This spatiotemporal disturbance index quantifies the probability of a real physical topological break or alteration occurring around the target evaluation node, rather than measurement jumps caused by transient electrical inrush currents. The specific formula for calculating the spatiotemporal disturbance index of the target evaluation node within the initial topology in the current environment is as follows: ； In the formula, the Indicates the target evaluation node The spatiotemporal disturbance index; This represents the total number of adjacent topology nodes; Indicates the sequence number of the adjacent topological nodes; Indicates the first The voltage level weighting coefficient of each adjacent topology node; the higher the voltage level, the larger the voltage level weighting coefficient. Indicates the first The adjacent topology nodes at the current sampling time The real-time electrical measurement data; Indicates the first The adjacent topology nodes at the pre-sampling time The real-time electrical measurement data; Indicates the relationship with the first The reference electrical constants for the voltage levels corresponding to the adjacent topology nodes; Represents the natural constant; the This represents the spatial attenuation coefficient, with dimensions of the reciprocal of length; The target evaluation node is extracted from the topological distance decay matrix. With the The shortest topological path length between adjacent topological nodes; Indicates the disturbance balance coefficient; the Indicates the alarm gain coefficient; the Indicates the number of abnormal alarm nodes; This represents the total number of nodes in the topology; The basic disturbance values ​​of the target evaluation node during its historical steady-state operation are retrieved and combined with the current grid load level to generate a dynamic comparison threshold. The calculated spatiotemporal disturbance index is compared with the dynamic comparison threshold. If the spatiotemporal disturbance index exceeds the dynamic comparison threshold, the target evaluation node is determined to meet the topology update triggering condition, and the initial topology local adjustment is triggered for the target evaluation node.

10. The method for automatic generation and dynamic verification of power grid topology supporting full voltage level continuity according to claim 1, characterized in that, When a large-scale meteorological disaster warning is received, and it is necessary to defend against massive concurrent device status change data that could cause graph database computing power overload and topology view jitter and crash, the multi-dimensional verification index system is verified, and a topology local update mechanism is triggered when topology anomalies occur. Specifically, this includes: Multi-dimensional meteorological impact parameters, including wind speed, ice thickness, and precipitation, are extracted from the multi-source data in real time, and the multi-dimensional meteorological impact parameters are converted into comprehensive disaster intensity values ​​based on a preset meteorological disaster assessment model. Determine whether the comprehensive disaster intensity value exceeds the preset extreme disaster critical threshold. If it exceeds the extreme disaster critical threshold, issue a global write transaction latch instruction to the graph database storing the initial topology, forcibly suspending the global breadth-first search process for the initial topology and the dynamic verification and update process of network connectivity. Based on the latitude and longitude range of the multi-dimensional meteorological impact parameters, a disaster spatial envelope is generated in the standard geographic coordinate system of the power grid, and the associated device set within the disaster spatial envelope is mapped to the network space, thereby rigidly isolating the associated device set into a disaster-affected closed topology partition. Create an independent high-speed cache sandbox in system memory, and clone the current topology connection snapshot of the disaster-affected closed topology partition into the high-speed cache sandbox; Inside the cache sandbox, data interaction with the external main network view is blocked. Based only on the edge remote signaling data of the surviving devices in the disaster-affected closed topology partition, a high-frequency local topology reconstruction exercise is performed on the associated device set, and the set of steady-state edge topology nodes generated during the reconstruction exercise is continuously updated and recorded. The multi-dimensional meteorological impact parameters are continuously monitored. When it is confirmed that the comprehensive disaster intensity value has fallen below the extreme disaster critical threshold and after a preset system smoothing period, the global write transaction latch instruction of the graph database is released. The final topological association state of the steady-state edge topological node set is extracted and a connection conflict check is performed with the main network edge node in the graph database at this time. If the connection conflict verification passes, the final topology association state is written to the initial topology in an incremental merging manner through an asynchronous message queue to eliminate the invalid system computing power consumption and topology view jitter caused by extreme weather disturbances.

Images