A power distribution network online security analysis, early warning and auxiliary decision integrated system

Abstract

A power distribution network online security analysis, early warning and auxiliary decision method and system, comprising: collecting multi-source graph model data and real-time operation data of main network, distribution network and microgrid, identifying the topological connection relationship of power grid equipment based on graph theory algorithm, and dynamically maintaining through incremental updating algorithm to construct full-voltage-level power grid topology full graph; performing base state power flow calculation and static N-1 safety analysis on the topology full graph, identifying the risk of section overrun and equipment overload, and generating an enhanced topology full graph with risk level, type and trend by using a load prediction model; based on the enhanced topology full graph, the dynamic adaptive fault credibility evaluation model is used to determine the fault point by fusing the topological influence factor, the operation risk trend factor and the equipment health degree attenuation factor. The application realizes the global dynamic presentation of the power grid structure, and improves the fault positioning accuracy and disposal efficiency.

Description

Technical Field

[0001] This invention belongs to the field of power distribution network safety technology, and in particular relates to a method and system for online safety analysis, early warning and auxiliary decision-making in power distribution networks. Background Technology

[0002] Provincial and local power distribution networks are a key component of the power system, undertaking the core tasks of power transmission and terminal power supply. Their safe and stable operation is directly related to the electricity consumption of social production and people's livelihood.

[0003] In recent years, with the large-scale integration of distributed energy resources and the increasing complexity of power grid topologies, distribution network security analysis and early warning technologies have also made some progress. For example, patent document CN120494489A discloses a distribution network security analysis method, which achieves a certain degree of online monitoring and risk assessment by modeling and analyzing the distribution network topology; patent document CN117610934A discloses a power system auxiliary decision-making method, which uses power flow calculation and early warning models to predict equipment overload risks.

[0004] However, the aforementioned existing technologies still have the following shortcomings: First, in terms of fault location, existing methods mostly use statically weighted comprehensive fault indicators to evaluate candidate fault paths. They only consider the voltage and current deviations and path impedance of the nodes themselves, without considering the differences in the importance of the nodes in the overall network topology. This results in faults at nodes located on critical tie lines or hub substations receiving similar evaluation scores as faults at the end nodes, thus reducing the accuracy of fault location.

[0005] Second, existing fault location methods are not linked with system operation risk warning information, and cannot use identified equipment overload, cross-section exceeding limit and other warning states as prior information for fault judgment, which makes fault credibility assessment lack trend perception ability.

[0006] Third, existing methods lack a mechanism for sensing the real-time health status of equipment, and do not incorporate online monitoring data such as transformer oil temperature, partial discharge, and cable joint temperature into the fault judgment model, thus failing to improve the confidence of fault alarms in advance when the equipment is in a sub-healthy state.

[0007] Fourth, existing fault indicators mostly use simple linear weighted sum models, making it difficult to effectively capture the nonlinear coupling relationships between various factors, which limits the generalization ability and accuracy of fault location models.

[0008] Therefore, there is an urgent need for an online safety analysis, early warning, and auxiliary decision-making method and system for distribution networks that can integrate information on topological location importance, operational risk trends, and equipment health status, and perform dynamic adaptive fault reliability assessment through a nonlinear fusion function. Summary of the Invention

[0009] The purpose of this invention is to provide a method and system for online safety analysis, early warning and auxiliary decision-making in power distribution networks, in order to solve the problems in the existing power distribution network fault location methods mentioned in the background art, such as insufficient perception of the importance of topological location by static weighted indicators, lack of linkage with operational risk early warning information, failure to incorporate equipment health status information and limited generalization ability of linear models.

[0010] This invention utilizes graph theory algorithms and incremental update algorithms to construct and dynamically maintain a complete topology map across all voltage levels. Compared to traditional analysis models that only target a single voltage level or isolated power grids, this invention solves the problems of fragmented data and lagging topology updates in traditional systems. Furthermore, this invention upgrades the comprehensive fault index from a static weighted formula to a dynamic adaptive fault reliability assessment model. It introduces a topology influence factor based on node betweenness centrality and electrical coupling, an operational risk trend factor based on a risk list, and an equipment health decay factor based on online equipment monitoring data. A nonlinear fusion function is used for comprehensive evaluation, significantly improving the accuracy and reliability of fault location.

[0011] The specific steps of this invention are as follows: Step 1: Collect multi-source graph model data and real-time operation data of the main grid, distribution network and microgrid, identify and dynamically update the topological connection relationship of grid equipment, and construct a full topological graph of the grid at all voltage levels; Step 2: Perform ground-state power flow calculation and static N-1 security analysis on the full voltage level power grid topology map to identify cross-section overload and equipment overload risks, and use the load forecasting model to generate an enhanced topology map with risk level, type and trend. Step 3: Based on the enhanced topology full graph, the fault point is determined by the dynamic adaptive fault credibility assessment model, and a fault handling plan is generated based on the knowledge graph reasoning method. The dynamic adaptive fault credibility assessment model integrates topology influence factors, operational risk trend factors and equipment health decay factors, and calculates the dynamic comprehensive fault index of the fault candidate path through a nonlinear fusion function, and determines the fault point based on the dynamic comprehensive fault index.

[0012] More preferably, step 1 involves collecting multi-source graph model data and real-time operational data from the main grid, distribution network, and microgrid; identifying and dynamically updating the topological connection relationships of grid equipment; and constructing a full-scale grid topology map across all voltage levels. Specific implementation methods include: Step 1.1: Based on multi-source graph model data and real-time operation data, traverse the power grid equipment nodes and connection edges, identify and establish the electrical connection relationship of the entire network, and form the initial topology; Step 1.2: Based on the initial topology, and combined with the switch change signals extracted from real-time operating data, the parts of the topology that have changed are locally reconstructed and updated to form a complete topology map of the power grid at all voltage levels.

[0013] More preferably, in step 2, ground-state power flow calculation and static N-1 security analysis are performed on the full-voltage-level power grid topology map to identify cross-section overload and equipment overload risks. An enhanced topology map with risk levels, types, and trends is generated using a load forecasting model. Specific implementation methods include: Step 2.1: Perform ground-state power flow calculation based on the full voltage level grid topology map, and use the static N-1 interruption analysis method to simulate the expected fault set, outputting steady-state and fault-state analysis results including bus voltage and branch power flow; Step 2.2: Based on the steady-state and fault-state analysis results, combined with the equipment's rated parameters and cross-sectional power transmission limits, the risk levels of equipment overload and cross-sectional limit exceedance are obtained, and a structured risk list is generated. Step 2.3: Use the load forecasting model to obtain forecast information, associate the risk list with the forecast information with the corresponding grid equipment in the full voltage level grid topology map, and generate an enhanced topology map with risk level, type and trend.

[0014] More preferably, in step 2.2, the risk levels of equipment overload and cross-sectional exceedance are obtained, and a structured risk list is generated. Specific steps include: For the equipment, the active power and reactive power of each branch are obtained based on the steady-state and fault-state analysis results. The rated capacity of the equipment is introduced, and the equipment load rate is calculated by ratio. For the cross section, the active power of the cross section is obtained based on the steady-state and fault-state analysis results. A cross section transmission power limit is introduced, and the cross section load rate is obtained by calculating the ratio. Based on equipment load rate and cross-section load rate, a preset risk level threshold is introduced to classify the risk status of equipment and cross-section into multiple levels. Iterate through all equipment and cross-sections, determine the risk level, and record any non-normal status as a risk event, generating a structured risk list. The risk list includes the risk equipment number, risk type, risk level, and occurrence time.

[0015] More preferably, in step 3, the implementation method of the dynamic adaptive fault reliability assessment model includes: Based on the connection relationships and path impedances of power grid equipment in the enhanced topology graph, a fault information matrix is ​​constructed; and a breadth-first search algorithm is used to select candidate fault paths starting from the fault reporting origin node. Calculate the basic electrical quantity deviations, including voltage deviation per unit value, for each node on the candidate fault path. Per-unit value of current deviation and path impedance per unit value ; based on , , Calculate the topology impact factor, operational risk trend factor, and equipment health degradation factor for each node: A nonlinear fusion function is used to fuse topology influence factors, operational risk trend factors, equipment health degradation factors, and basic electrical quantity deviations to obtain a dynamic comprehensive fault index.

[0016] More preferably, the topological influence factor is calculated for any node as follows: The topology influence factor is obtained by weighted summing of the normalized values ​​of the betweenness centrality and electrical coupling of the nodes. Betweenness centrality is the frequency at which the node is the intermediate node of the shortest path; electrical coupling is calculated based on the equivalent electrical impedance between the node and other nodes.

[0017] More preferably, the calculation method for the operating risk trend factor is as follows: The operational risk trend factor is obtained based on the risk level mapping.

[0018] More preferably, the method for calculating the equipment health degradation factor is as follows: The normalized result of the difference between the actual measured value and the normal operating reference value of each equipment indicator is calculated, and the weighted normalized result is subtracted to obtain the difference result. Multiply the interpolation results corresponding to all devices to obtain the product result, and subtract the product result to obtain the device health decay factor.

[0019] More preferably, a nonlinear fusion function is used to fuse the topology influence factor, operational risk trend factor, equipment health degradation factor, and basic electrical quantity deviation to obtain a dynamic comprehensive fault index. Specific steps include: The feature vector is composed of voltage deviation per unit value, current deviation per unit value, path impedance per unit value, topology influence factor, operational risk trend factor, and equipment health degradation factor. The dynamic comprehensive fault index is calculated by inputting the feature vector into the nonlinear fusion function obtained by training based on historical fault data.

[0020] More preferably, the fault point is determined based on dynamic comprehensive fault indicators, and a fault handling plan is generated based on knowledge graph reasoning methods. Specific methods include: The candidate fault path with the largest dynamic comprehensive fault index is selected as the most likely fault path, and the fault location is determined by calculating the voltage deviation gradient between adjacent nodes on the most likely fault path. Based on the location of the fault point, the outage area and key load points affected by the fault are identified in the enhanced topology map. The upstream power supply point and downstream load node of the fault point are marked, the location of the disconnecting switch to be operated is identified to form electrical isolation, feasible power transfer paths that meet the load rate threshold are screened, and a fault handling plan is generated after verification.

[0021] This invention also proposes an online security analysis, early warning, and auxiliary decision-making system for power distribution networks, comprising a topology full-map construction unit, a security analysis and early warning unit, and a collaborative scheduling unit: The topology map construction unit is used to collect multi-source map model data and real-time operation data of the main grid, distribution network and microgrid, identify and update the topology connection relationship of grid equipment, and construct a full topology map of the grid at all voltage levels; The safety analysis and early warning unit is used to perform basic state power flow calculation and static N-1 safety analysis on the full voltage level power grid topology map, identify the risk of cross-section overload and equipment overload, and generate an enhanced full topology map with risk level, type and trend using the load forecasting model; The collaborative scheduling unit is used to determine fault points based on the enhanced topology full graph, through a dynamic adaptive fault credibility assessment model, and to generate fault handling plans based on knowledge graph reasoning methods.

[0022] More preferably, the topology map construction unit includes a topology relationship identification module and a dynamic maintenance module; the security analysis and early warning unit includes a security analysis module, an operational risk assessment module, and a risk map generation module.

[0023] More preferably, the dynamic adaptive fault reliability assessment model in the collaborative scheduling unit integrates the topology influence factor, operational risk trend factor and equipment health decay factor of each node, and comprehensively evaluates the candidate fault path through a nonlinear fusion function trained based on historical fault data to determine the fault location.

[0024] More preferably, the topology influence factor quantifies the positional importance of a node in the overall network topology based on node betweenness centrality and electrical coupling; the operational risk trend factor is determined based on the risk level mapping of the risk list in the enhanced topology map; and the device health decay factor is determined based on online device monitoring data.

[0025] More preferably, the collaborative scheduling unit identifies the power outage area based on the location of the fault point, and generates a fault handling plan through electrical isolation and load transfer path screening and verification.

[0026] More preferably, the safety analysis module uses the Newton-Raphson method to calculate the ground-state power flow and uses the static N-1 interruption analysis method to simulate the expected fault set.

[0027] More preferably, the risk map generation module uses a long short-term memory network as a load forecasting model to predict future load changes and associates the risk list with the forecast information to generate an enhanced topology map.

[0028] The present invention also proposes a terminal, including a processor and a storage medium: The storage medium is used to store instructions; The processor is used to perform the steps of the above method according to the instructions.

[0029] The present invention also proposes a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the above-described method.

[0030] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. This invention integrates multi-source data from the main grid, distribution grid, and microgrid through a topology full-map construction unit. It uses graph theory algorithms and incremental update algorithms to construct and dynamically maintain a full-topology map of all voltage levels. Compared with the traditional analysis mode that only targets a single voltage level or isolated grid, it realizes a global and dynamic presentation of the grid structure and solves the problems of data fragmentation, lagging topology updates, and inability to adapt to the needs of multi-grid collaborative management in traditional systems.

[0031] 2. This invention generates an enhanced topology map with risk levels and trends by combining ground-state power flow calculation using the Newton-Raphson method, static N-1 security analysis, and load prediction using long short-term memory networks. It also achieves accurate fault location and scientific handling through a dynamic adaptive fault credibility assessment model and knowledge graph reasoning. Compared with traditional risk judgment and fault handling methods that rely on human experience, this invention avoids the subjectivity and lag of human decision-making and solves the problems of weak risk prediction capability, low fault handling efficiency, and easy expansion of the impact of power outages in traditional systems.

[0032] 3. This invention upgrades the traditional static weighted comprehensive fault index to a dynamic adaptive fault credibility assessment model. It integrates topology influence factors, operational risk trend factors, and equipment health decay factors. Through a nonlinear fusion function trained based on historical fault data, it performs a comprehensive evaluation. It can dynamically perceive the positional importance of nodes in the entire network topology, the operational risk trend of equipment, and real-time health status, which significantly improves the accuracy and robustness of fault location and overcomes the shortcomings of traditional static indicators in adapting to complex power grid fault scenarios. Attached Figure Description

[0033] Figure 1 This is a flowchart of a method for online safety analysis, early warning, and auxiliary decision-making in power distribution networks according to the present invention; Figure 2 This is a system framework diagram of an online safety analysis, early warning and auxiliary decision-making system for power distribution networks according to the present invention. Detailed Implementation

[0034] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of this invention. The embodiments described in this application are merely some embodiments of this invention, and not all embodiments. Based on the spirit of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the protection scope of this invention.

[0035] like Figure 1 As shown, this invention proposes a method for online safety analysis, early warning, and auxiliary decision-making in power distribution networks, comprising the following steps: Step 1: Collect multi-source graph model data and real-time operation data of the main grid, distribution network and microgrid, identify and dynamically update the topological connection relationship of grid equipment, and construct a full topological graph of the grid at all voltage levels; Step 1 involves collecting multi-source graph model data and real-time operational data from the main grid, distribution grid, and microgrids; identifying and dynamically updating the topological connections of grid equipment; and constructing a full-scale grid topology map across all voltage levels. Specific implementation methods include: Step 1.1: Based on multi-source graph model data and real-time operation data, traverse the power grid equipment nodes and connection edges, identify and establish the electrical connection relationship of the entire network, and form the initial topology; Step 1.2: Based on the initial topology, and combined with the switch change signals extracted from real-time operating data, the parts of the topology that have changed are locally reconstructed and updated to form a complete topology map of the power grid at all voltage levels.

[0036] Step 2: Perform ground-state power flow calculation and static N-1 security analysis on the full voltage level power grid topology map to identify cross-section overload and equipment overload risks, and use the load forecasting model to generate an enhanced topology map with risk level, type and trend. Step 2 involves performing ground-state power flow calculations and static N-1 security analysis on the full voltage level power grid topology map to identify cross-section overload and equipment overload risks. An enhanced topology map with risk levels, types, and trends is then generated using a load forecasting model. Specific implementation methods include: Step 2.1: Perform ground-state power flow calculation based on the full voltage level grid topology map, and use the static N-1 interruption analysis method to simulate the expected fault set, outputting steady-state and fault-state analysis results including bus voltage and branch power flow; Step 2.2: Based on the steady-state and fault-state analysis results, combined with the equipment's rated parameters and cross-sectional power transmission limits, the risk levels of equipment overload and cross-sectional limit exceedance are obtained, and a structured risk list is generated. In step 2.2, the risk levels of equipment overload and cross-sectional exceedance are obtained, and a structured risk list is generated. Specific steps include: For the equipment, the active power and reactive power of each branch are obtained based on the steady-state and fault-state analysis results. The rated capacity of the equipment is introduced, and the equipment load rate is calculated by ratio. For the cross section, the active power of the cross section is obtained based on the steady-state and fault-state analysis results. A cross section transmission power limit is introduced, and the cross section load rate is obtained by calculating the ratio. Based on equipment load rate and cross-section load rate, a preset risk level threshold is introduced to classify the risk status of equipment and cross-section into multiple levels. Iterate through all equipment and cross-sections, determine the risk level, and record any non-normal status as a risk event, generating a structured risk list. The risk list includes the risk equipment number, risk type, risk level, and occurrence time.

[0037] Step 2.3: Use the load forecasting model to obtain forecast information, associate the risk list with the forecast information with the corresponding grid equipment in the full voltage level grid topology map, and generate an enhanced topology map with risk level, type and trend.

[0038] Step 3: Based on the enhanced topology full graph, the fault point is determined by the dynamic adaptive fault credibility assessment model, and a fault handling plan is generated based on the knowledge graph reasoning method. The dynamic adaptive fault credibility assessment model integrates topology influence factors, operational risk trend factors and equipment health decay factors, and calculates the dynamic comprehensive fault index of the fault candidate path through a nonlinear fusion function, and determines the fault point based on the dynamic comprehensive fault index.

[0039] In step 3, the implementation method of the dynamic adaptive fault reliability assessment model includes: Based on the connection relationships and path impedances of power grid equipment in the enhanced topology graph, a fault information matrix is ​​constructed; and a breadth-first search algorithm is used to select candidate fault paths starting from the fault reporting origin node. Calculate the basic electrical quantity deviations, including voltage deviation per unit value, for each node on the candidate fault path. Per-unit value of current deviation and path impedance per unit value ; based on , , Calculate the topology impact factor, operational risk trend factor, and equipment health degradation factor for each node: A nonlinear fusion function is used to fuse topology influence factors, operational risk trend factors, equipment health degradation factors, and basic electrical quantity deviations to obtain a dynamic comprehensive fault index.

[0040] For any node, the topological influence factor is calculated as follows: The topology influence factor is obtained by weighted summing of the normalized values ​​of the betweenness centrality and electrical coupling of the nodes. Betweenness centrality is the frequency at which the node is the intermediate node of the shortest path; electrical coupling is calculated based on the equivalent electrical impedance between the node and other nodes.

[0041] The calculation method for the operating risk trend factor is as follows: The operational risk trend factor is obtained based on the risk level mapping.

[0042] The calculation method for the equipment health degradation factor is as follows: The normalized result of the difference between the actual measured value and the normal operating reference value of each equipment indicator is calculated, and the weighted normalized result is subtracted to obtain the difference result. Multiply the interpolation results corresponding to all devices to obtain the product result, and subtract the product result to obtain the device health decay factor.

[0043] A nonlinear fusion function is used to fuse topology influence factors, operational risk trend factors, equipment health degradation factors, and basic electrical quantity deviations to obtain a dynamic comprehensive fault index. The specific steps include: The feature vector is composed of voltage deviation per unit value, current deviation per unit value, path impedance per unit value, topology influence factor, operational risk trend factor, and equipment health degradation factor. The dynamic comprehensive fault index is calculated by inputting the feature vector into the nonlinear fusion function obtained by training based on historical fault data.

[0044] In step 3, the fault point is determined based on the dynamic comprehensive fault index, and a fault handling plan is generated based on the knowledge graph reasoning method. The specific methods include: The candidate fault path with the largest dynamic comprehensive fault index is selected as the most likely fault path, and the fault location is determined by calculating the voltage deviation gradient between adjacent nodes on the most likely fault path. Based on the location of the fault point, the outage area and key load points affected by the fault are identified in the enhanced topology map. The upstream power supply point and downstream load node of the fault point are marked, the location of the disconnecting switch to be operated is identified to form electrical isolation, feasible power transfer paths that meet the load rate threshold are screened, and a fault handling plan is generated after verification.

[0045] This invention also proposes an online security analysis, early warning, and auxiliary decision-making system for power distribution networks, comprising a topology full-map construction unit, a security analysis and early warning unit, and a collaborative scheduling unit: The topology map construction unit is used to collect multi-source map model data and real-time operation data of the main grid, distribution network and microgrid, identify and update the topology connection relationship of grid equipment, and construct a full topology map of the grid at all voltage levels; The topology map construction unit includes a topology relationship identification module and a dynamic maintenance module; the security analysis and early warning unit includes a security analysis module, an operational risk assessment module, and a risk map generation module.

[0046] The safety analysis module uses the Newton-Raphson method to calculate the ground-state power flow and the static N-1 interruption analysis method to simulate the expected fault set.

[0047] The risk map generation module uses a long short-term memory network as a load forecasting model to predict future load changes and associates the risk list with the forecast information to generate an enhanced topology map.

[0048] The safety analysis and early warning unit is used to perform basic state power flow calculation and static N-1 safety analysis on the full voltage level power grid topology map, identify the risk of cross-section overload and equipment overload, and generate an enhanced full topology map with risk level, type and trend using the load forecasting model; The collaborative scheduling unit is used to determine fault points based on the enhanced topology full graph, through a dynamic adaptive fault credibility assessment model, and to generate fault handling plans based on knowledge graph reasoning methods.

[0049] The dynamic adaptive fault reliability assessment model in the collaborative scheduling unit integrates the topology influence factor, operational risk trend factor and equipment health decay factor of each node, and comprehensively evaluates the candidate fault path through a nonlinear fusion function trained based on historical fault data to determine the fault location.

[0050] The topology influence factor quantifies the importance of a node's position in the overall network topology based on node betweenness centrality and electrical coupling; the operational risk trend factor is determined based on the risk level mapping of the risk list in the enhanced topology graph; and the equipment health decay factor is determined based on online equipment monitoring data.

[0051] The collaborative scheduling unit identifies the power outage area based on the location of the fault point, and generates a fault handling plan through electrical isolation and load transfer path screening and verification.

[0052] The present invention also proposes a terminal, including a processor and a storage medium: The storage medium is used to store instructions; The processor is used to perform the steps of the above method according to the instructions.

[0053] The present invention also proposes a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the above-described method.

[0054] Example 1 This invention proposes a method for online security analysis, early warning, and auxiliary decision-making in power distribution networks, comprising: Step 1: Collect multi-source graph model data and real-time operation data of the main grid, distribution network and microgrid; identify the topological connection relationship of grid equipment based on graph theory algorithm; and dynamically maintain the topological relationship through incremental update algorithm to construct a full topological graph of the grid at all voltage levels. Among them, power grid equipment includes all primary physical devices in the power generation, transmission, transformation, distribution and consumption links that constitute the main power grid, distribution network and microgrid, specifically covering generators, transformers, transmission lines, circuit breakers, busbars, capacitors, reactors and various distributed power sources; Step 1.1: Based on multi-source graph model data and real-time operation data, the electrical connection relationships of the entire network are identified and established by traversing the power grid equipment nodes and connection edges through a depth-first search algorithm, thus forming the initial topology.

[0055] Specifically, multi-source graphical model data and real-time operational data of the main grid, distribution grid and microgrid are acquired, and the data is cleaned and verified to remove abnormal records; among them, the real-time operational data includes the current value, voltage value and switch status signal of each monitoring point; Among them, multi-source graphic model data includes graphic and model data from different sources in the main power grid, distribution network and microgrid, such as power grid structure model in CIM / E format, real-time switch status, equipment parameters and electrical connection relationships.

[0056] Configure power grid equipment as a set of nodes. The electrical connections between power grid devices are defined as edge sets. Constructing a graph structure Among them, power grid equipment includes at least generators, transformers, lines, and circuit breakers; The graph structure is traversed using a depth-first search algorithm. Identify all connected subgraphs, forming several connected electrical regions, and record all connection nodes and path impedances for each node. ;in, For node indexing.

[0057] Specifically, depth-first search is a graph traversal algorithm whose strategy is to visit adjacent nodes as deeply as possible along a path until it can no longer proceed and then backtracks, in order to systematically explore all connected parts of the graph; a connected electrical region refers to the area formed by a group of devices that are electrically directly connected by closed switches or lines in a power grid, in which current can flow. Represents a node With nodes The path impedance between nodes, i.e., the current flowing through the nodes in the power grid. With nodes The total resistance encountered during the path is the sum of the impedances of each line, transformer, and other equipment along the path, and is measured in Ω.

[0058] Combining graph structure The connected electrical regions together form the initial topology; Step 1.2: Based on the initial topology, and combined with the switch position change signals extracted from real-time operation data, an incremental update algorithm is used to locally reconstruct and update the parts of the topology that have changed, forming a full-scale power grid topology map of all voltage levels. Specifically, switch position change signals are obtained from real-time operating data, and when a switch state change (such as opening or closing) is detected, the position change device number is extracted. Based on the initial topology, a breadth-first search algorithm is used to traverse the graph structure, centering on the node corresponding to the displacement device number. Only for distance displacement devices (generally =2~3) Reconstruct the nodes and edges within the jump to determine the local subgraph that needs to be updated. ; In local subgraphs Within the scope, the electrical connection relationship between nodes is recalculated according to the switch status. If the switch is open, its corresponding connection edge is removed; if the switch is closed, the corresponding connection edge is added, thereby generating an updated local subgraph. The updated local subgraphs are seamlessly integrated with the unaffected initial topology to form a complete grid topology map across all voltage levels.

[0059] Step 2: Perform ground-state power flow calculation and static N-1 security analysis on the full voltage level power grid topology map. By identifying the risks of cross-section overload and equipment overload, an enhanced topology map with risk level, type and trend is generated using the load forecasting model.

[0060] Step 2.1: Based on the full voltage level grid topology map, the Newton-Raphson method is used to perform ground state power flow calculation, and the static N-1 interruption analysis method is used to simulate the expected fault set, outputting steady state and fault state analysis results including bus voltage and branch power flow.

[0061] Specifically, based on the full-voltage-level power grid topology map, the electrical connections and equipment parameters of the entire network are obtained to form a node admittance matrix. ; Wherein, the node admittance matrix elements in Represents a node With nodes Mutual admittance between them Represents a node The self-admittance is calculated from the impedance parameters of equipment such as lines and transformers in the power grid, specifically as follows: ,in For nodes With nodes Path impedance between; For nodes The sum of the admittances of all connected branches.

[0062] Using real-time operating data as the initial operating state, a node power deviation convergence threshold is set. (e.g., 0.001 per unit value) and node voltage deviation convergence threshold (e.g., 0.001 per unit value); Iterative calculations using the Newton-Raphson method are performed, solving for power imbalance and correcting node voltages in each iteration until the power and voltage deviations of all nodes are less than a set threshold. The node voltages and branch power flows at this point are recorded in the ground state, forming a steady-state analysis result including the bus voltage and branch power flows in the ground state. Node voltages include their amplitude and phase angle; branch power flows include the active and reactive power of each branch; bus voltage refers to the voltage amplitude and phase angle at the bus node in the power grid, which determines the operating state of the equipment connected to that node and the power distribution of the system.

[0063] Specifically, based on the current node voltage amplitude and phase angle Calculate the injection power imbalance at each node. and the imbalance of injected reactive power ; Construct the corrected equation: ,in It is the Jacobian matrix, which is composed of the partial derivatives of nodal power with respect to voltage magnitude and phase angle; Solving the corrected equation yields the node voltage phase angle correction. and node voltage amplitude correction ; Update node voltage: , ;in, This represents the old value of the node voltage phase angle during the iteration process. This is the updated value of the voltage phase angle. This represents the old value of the node voltage amplitude during the iteration process. This is the updated value of the voltage amplitude; Repeat the above process until all nodes are... and Record the voltage amplitude, phase angle, and branch power flow of each node at this time; among them, branch power flow refers to the active and reactive power flowing through the connecting elements (i.e., branches) of the power grid.

[0064] The static N-1 interruption analysis method is used to sequentially simulate the interruption of each critical device (such as a line or transformer) in the anticipated fault set. The nodal admittance matrix is ​​then used. The fault is simulated by changing the admittance value of the corresponding branch to 0, thus forming the node admittance matrix after the fault. .

[0065] Among them, the anticipated fault set is a set of potential power grid fault scenarios that are pre-set and need to be simulated and analyzed one by one. It usually includes the interruption of single components such as critical lines and transformers. Critical equipment refers to components that are pre-selected based on their capacity, power supply range and importance in the network topology. Their failure will have a significant impact on system security or load supply. Specifically, it includes: transformers with a rated capacity of ≥100MVA, equipment with a voltage level of 220kV and above, main power supply lines that bear more than 80% of the load in the area, and tie lines connecting different regional power grids.

[0066] Based on the node admittance matrix after the fault The node voltage and branch power flow after the fault are obtained by iterative calculation using the Newton-Raphson method, forming a fault state analysis result that includes the bus voltage and branch power flow after the fault. Based on the fault state analysis results and steady state analysis results, after traversing the expected fault set, the complete steady state and fault state analysis results are output. Step 2.2: Based on the steady-state and fault-state analysis results, combined with the equipment's rated parameters and cross-sectional power transmission limits, identify and quantify the risk levels of equipment overload and cross-sectional limit exceedance, and generate a structured risk list.

[0067] Specifically, for equipment (lines and transformers), the active and reactive power of each branch are obtained based on the steady-state and fault-state analysis results. The rated capacity of the equipment is then introduced, and the equipment load rate is calculated through ratio calculation. :

[0068] in, For actual apparent power, satisfy , The active power of the branch circuit. The reactive power of the branch circuit; The rated capacity of the equipment; Equipment load rate; For each cross section, the active power is obtained based on the steady-state and fault-state analysis results. A transmission power limit for the cross section is introduced, and the load factor of the cross section is calculated through the ratio. :

[0069] in, The active power of the cross section. This represents the cross-sectional power transmission limit; The cross-sectional load factor; a cross-section refers to a set of electrically closely related lines or transformers in a power grid that share a specific power transmission task, and its total power transmission capacity has an upper limit.

[0070] Based on device load rate and cross-sectional load factor By introducing preset risk level thresholds (normal ≤80%, caution 80%-90%, warning 90%-100%, emergency >100%), the risk status of equipment and cross-sections is divided into multiple levels: That is, when the equipment load rate When the load rate is less than or equal to the normal threshold (e.g., 80%), the device is considered to be in normal condition; when the device load rate... When the load rate is greater than the normal threshold (e.g., 80%) and less than or equal to the warning threshold (e.g., 90%), the device is determined to be in a state of alert; when the device load rate... When the load rate is greater than the warning threshold (e.g., 90%) and less than or equal to the emergency threshold (e.g., 100%), the device is considered to be in a warning state; when the device load rate... If the percentage exceeds the emergency threshold (e.g., 100%), the device is considered to be in an emergency. Similarly, when the cross-sectional load factor When the load rate is less than or equal to the normal threshold (e.g., 80%), the section is considered to be in a normal state; when the section load rate... When the load rate is greater than the normal threshold (e.g., 80%) and less than or equal to the warning threshold (e.g., 90%), the section is determined to be in a state of alert; when the section load rate... When the load rate is greater than the warning threshold (e.g., 90%) and less than or equal to the emergency threshold (e.g., 100%), the section is determined to be in a warning state; when the section load rate... If the value exceeds the emergency threshold (e.g., 100%), the section is determined to be in an emergency state.

[0071] Iterate through all equipment and sections, determine their risk level, and record any abnormal conditions as risk events, generating a structured risk list. The risk list includes the risk equipment number, risk type (overload or section exceeding limit), risk level (caution, warning, emergency), and occurrence time.

[0072] Step 2.3: Use the load forecasting model to predict future load changes, associate the risk list with the forecast information with the corresponding grid equipment in the full voltage level grid topology map, and generate an enhanced topology map with risk level, type and trend.

[0073] Specifically, historical load data is acquired, including the active power of each bus node at multiple time points and date-type data. The historical load data is then subjected to max-min normalization to obtain the historical load sequence. ;

[0074] in, For historical load sequences, This represents the load value (active power) at time t. The time step is set to 0; the historical load data is subjected to max-min normalization to make it fall within the [0,1] interval, so as to improve the model training efficiency and convergence stability.

[0075] In a preferred embodiment of the present invention, a long short-term memory network is used as the load forecasting model, and its input is a historical load sequence. The output is the load forecast value of each bus node in the future time period T (e.g., the next hour). .

[0076] The Long Short-Term Memory (LSTM) network models long-term dependencies through gating units including input gates, forget gates, output gates, and memory units. The load forecasting model specifically includes an LSTM layer with 64 such gating units, followed by a fully connected output layer. The model is trained using the Adam optimizer (learning rate 0.001), with mean squared error (MSE) as the loss function, a batch size of 32, and one-hot encoding of date features as an auxiliary input to ensure the accuracy of load forecasting.

[0077] Load forecast Inverse normalization to actual power prediction .

[0078] Based on the risk list, the corresponding equipment node is located in the full voltage level power grid topology map according to the risk equipment number, and the actual power prediction value is calculated. Rated capacity of the corresponding device node The ratio of the two values ​​is used to obtain the predicted load factor. ;

[0079] in, To predict load factor; Based on predicted load rate Based on the established risk level thresholds, the predicted risk level is determined. The predicted risk level, risk type, actual power prediction value, and predicted load rate are overlaid as layers onto the full voltage level power grid topology map to form an enhanced full topology map with risk level, type, and trend.

[0080] Specifically, in the enhanced topology map, different colors or icons represent different risk levels (e.g., green for normal, yellow for caution, orange for warning, and red for emergency), and the trend arrows or values ​​of the actual power prediction are marked to help operators intuitively identify the future risk distribution and evolution trend.

[0081] Step 3: Based on the enhanced topology full graph, the fault point is determined by the dynamic adaptive fault credibility assessment model, and a fault handling plan is generated based on the knowledge graph reasoning method. The knowledge graph reasoning method is based on a preset power grid knowledge graph. The graph nodes include equipment nodes (such as lines, transformers, and switches), fault nodes (such as overload and short circuit), and load nodes. The edges include equipment-fault association edges, electrical connection edges, and action constraint edges. The system generates a contingency plan that includes fault isolation and load transfer by combining preset reasoning rules (such as "if the line is overloaded, disconnect its upstream and downstream isolating switches and then select an alternative power supply path") with real-time topology.

[0082] Specifically, in this embodiment, real-time operating data after a fault occurs is acquired, including current values, voltage values, and switch status signals at each monitoring point. At the same time, a fault information matrix is ​​constructed by combining the grid equipment connection relationships and path impedances in the enhanced topology full map. Based on the fault information matrix, a breadth-first search algorithm is used to start from the initial node of the fault report, traverse its electrical connection path, and filter out electrical connection paths that may contain faulty equipment as candidate fault paths. Then, the voltage drop and current surge at each node on the candidate fault path are detected level by level, and the voltage deviation at each node is calculated. With current deviation :

[0083]

[0084] in, The node's rated voltage, This is the reference current value when the node is operating normally; These are the voltage values ​​at each monitoring point after the fault occurred. The current values ​​at each monitoring point after the fault occurred; The voltage deviation at the node. This represents the current deviation at the node.

[0085] Based on the voltage deviation of each node With current deviation In addition to path impedance, a comprehensive failure index is calculated for each candidate fault path. : First, to balance the dimensions and numerical range, each physical quantity is normalized to a per-unit value:

[0086]

[0087]

[0088] in, The system's rated voltage. The rated current of the system. Reference impedance ( ); The total path impedance is given by the path impedance of each line segment in the enhanced topology diagram. The summation is obtained; This is the per-unit value of the voltage deviation. This is the per-unit value of the current deviation. This represents the per-unit value of the path impedance; Furthermore, this invention adopts a dynamic adaptive fault credibility assessment model to replace the traditional static weighted comprehensive fault index. By integrating topological influence factors, operational risk trend factors and equipment health decay factors, it achieves multi-dimensional dynamic assessment of fault candidate paths.

[0089] (a) Topological Influence Factor

[0090] Topological impact factors are used to quantify the positional importance of node i in the overall network topology. When a node located on a critical tie line or at a key substation fails, its impact is far greater than that of a terminal node. Based on the constructed full-voltage-level power grid topology map, graph theory indices are used to quantify the topological importance of nodes:

[0091] in, For nodes The normalized value of the betweenness centrality reflects the frequency with which the node is the intermediate node of the shortest path in the network. The higher the value, the more the node is in a critical channel position in the entire network. For nodes The normalized value of electrical coupling reflects the degree of electrical distance between this node and other nodes; and For the corresponding weight coefficients, satisfying .

[0092] Specifically, node betweenness centrality The calculation formula is:

[0093] in, For nodes To the node The total number of shortest paths, For the nodes The number of shortest paths.

[0094] Node electrical coupling Based on nodes Calculation of equivalent electrical impedance between the node and other nodes:

[0095] in, For nodes With nodes The equivalent electrical impedance between them This represents the total number of nodes in the entire network. and Perform maximum-min normalization respectively to obtain and This makes its value range [0,1].

[0096] (b) Operational risk trend factor

[0097] Operational risk trend factors are used to link fault location with system operation risk early warning information. If a fault occurs on equipment or sections that have previously been warned as "attention" or "warning", the reliability of the fault in this alarm should be increased accordingly. Risk level mapping values ​​taken from the risk list:

[0098] That is, if node If the corresponding equipment has been marked as having an abnormal risk level in the risk list, then its risk level is mapped to the corresponding trend factor value, which serves as prior information for fault judgment.

[0099] (c) Equipment health degradation factor

[0100] The equipment health degradation factor is used to incorporate the sub-health state of equipment reflected by online monitoring data (such as transformer oil temperature, partial discharge, cable joint temperature, etc.) into the fault diagnosis model. When equipment is in a sub-healthy state, its prior probability of failure is higher.

[0101]

[0102] in, For equipment The number of online monitoring indicators; For the first The normalized deviation value of a monitoring indicator is defined as the normalized result of the difference between the actual measured value and the normal operation reference value of the indicator, and the value range is [0,1]. The weighting coefficients for each monitoring indicator reflect the contribution of different monitoring indicators to the equipment's health status, satisfying... .

[0103] Specifically, when all monitoring indicators of the equipment are within the normal range ( When ≈0), ≈0 indicates that the equipment is in good health and does not add any additional gain to the reliability of the fault; when one or more monitoring indicators show abnormal deviations, The corresponding increase in the value reflects an increase in the sub-health level of the equipment.

[0104] (d) Dynamic integrated fault index

[0105] This invention employs a nonlinear fusion function to integrate the aforementioned factors with the basic electrical quantity deviations, replacing the traditional linear weighted sum model. Specifically, it uses a per-unit value for voltage deviation... Per-unit value of current deviation Path impedance per unit value Topological Influence Factor Operational risk trend factors and equipment health decay factor Composition of feature vectors:

[0106] The dynamic comprehensive fault index is calculated using a nonlinear fusion function trained based on historical fault data.

[0107] in, This is the weight matrix. For bias vectors, The sigmoid activation function is defined as follows:

[0108] W and b are obtained through training on historical fault data, making the model output closer to the distribution of real faults. During training, the feature vectors of each candidate path in the historical fault records are used as input, and the label (0 or 1) indicating whether the path is an actual fault path is used as output. The binary cross-entropy loss function is used for optimization.

[0109] The higher the value, the greater the likelihood that the candidate failure path will fail.

[0110] Select comprehensive fault indicators The largest candidate fault path is taken as the most likely fault path. The voltage deviation gradient is obtained by calculating the ratio of the difference in voltage deviation between adjacent nodes on the most likely fault path to the path impedance. The segment with the largest gradient value is the location of the fault point.

[0111] Where Gradient represents the voltage deviation gradient. node Voltage deviation, For nodes Voltage deviation, node and Adjacent nodes; For nodes and The line impedance between them.

[0112] Furthermore, based on the location of the fault point, in the enhanced topology map, with the device at the location of the fault point as the center, a breadth-first search algorithm is used to traverse its electrical connection path to identify all power outage areas and critical load points affected by the fault, while marking the upstream power source and downstream load nodes of the fault point. Based on the marked upstream power supply points and downstream load nodes of the fault point, the location of operable disconnect switches is identified in the enhanced topology map, and the upstream and downstream switches closest to the fault point are opened to form electrical isolation; for ring network structures, it is ensured that no new islands are created after isolation.

[0113] In the enhanced topology map, starting from the power outage area, a breadth-first search algorithm is used to traverse its electrical connection paths, filtering for paths with load rates lower than a preset load rate threshold. (like A set of load transfer schemes is formed by identifying feasible paths (90% or more) to achieve load transfer efficiency. in, The load factor threshold for the transfer path is a conservative margin standard set during the initial screening of transfer paths. It is used to quickly eliminate paths where critical equipment is close to full load and transmission margin is insufficient, ensuring that candidate paths have basic load transfer capabilities.

[0114] In this embodiment, the load transfer scheme refers to the specific operational plan for restoring power supply to the power outage area by switching the backup power path. Its core is to achieve safe load transfer through topology reconstruction after isolating the fault. The load transfer scheme set is a collection of all backup power paths that meet the preset load rate threshold and have safe power supply capability, selected based on the enhanced topology full map.

[0115] Based on the load transfer scheme set, for each load transfer scheme, if the load rate of each branch after transfer does not exceed the equipment's rated load rate threshold... (like =95%), node voltage deviation is within the allowable voltage deviation threshold. (like If the load transfer scheme passes the verification test, the load transfer rate is within the range of ±10% and no new equipment overload or cross-sectional overload occurs. The verified load transfer scheme will be output as a fault handling plan.

[0116] Example 2 like Figure 2 As shown, this invention also proposes an online security analysis, early warning, and auxiliary decision-making system for power distribution networks, including a topology full-map construction unit, a security analysis and early warning unit, and a collaborative scheduling unit: The topology full map construction unit collects multi-source graph model data and real-time operation data from the main grid, distribution network and microgrid, identifies the topological connection relationship of grid equipment based on graph theory algorithm, and dynamically maintains the topological relationship through incremental update algorithm to construct a full topology map of the power grid at all voltage levels.

[0117] The topology map construction unit includes a topology relationship identification module and a dynamic maintenance module. The topology relationship identification module, based on multi-source graph model data and real-time operation data, traverses the power grid equipment nodes and connection edges using a depth-first search algorithm to identify and establish the electrical connection relationships of the entire network, forming an initial topology structure. The dynamic maintenance module, based on the initial topology structure and combined with switch change signals extracted from real-time operation data, uses an incremental update algorithm to locally reconstruct and update the parts of the topology structure that have changed, forming a full voltage level power grid topology map.

[0118] The safety analysis and early warning unit performs ground-state power flow calculation and static N-1 safety analysis on the full voltage level power grid topology map. It identifies the risks of cross-section overload and equipment overload, and uses the load prediction model to generate an enhanced topology map with risk level, type and trend.

[0119] The safety analysis and early warning unit includes a safety analysis module, an operational risk assessment module, and a risk map generation module. The safety analysis module, based on a full-voltage-level power grid topology map, uses the Newton-Raphson method for ground-state power flow calculation and employs the static N-1 breaking analysis method to simulate the anticipated fault set, outputting steady-state and fault-state analysis results including bus voltage and branch power flow. The operational risk assessment module, based on the steady-state and fault-state analysis results and combined with equipment rated parameters and cross-sectional power transmission limits, identifies and quantifies the risk levels of equipment overload and cross-sectional limit exceedances, generating a structured risk list. The risk map generation module uses a load forecasting model to predict future load changes, associates the risk list and forecast information with the corresponding power grid equipment in the full-voltage-level power grid topology map, and generates an enhanced topology map with risk levels, types, and trends.

[0120] The collaborative scheduling unit is based on an enhanced topology graph, determines the fault point through a fault location algorithm, and generates a fault handling plan based on a knowledge graph reasoning method.

[0121] Example 3 The present invention also proposes a terminal, including a processor and a storage medium: The storage medium is used to store instructions; The processor is used to perform the steps of the above method according to the instructions.

[0122] Example 4 The present invention also proposes a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the above-described method.

[0123] This disclosure can be a system, method, and / or computer program product. A computer program product may include a computer-readable storage medium having computer-readable program instructions loaded thereon for causing a processor to implement various aspects of this disclosure.

[0124] Computer-readable storage media can be tangible devices capable of holding and storing instructions for use by an instruction execution device. Computer-readable storage media can be, for example—but not limited to—electrical storage devices, magnetic storage devices, optical storage devices, electromagnetic storage devices, semiconductor storage devices, or any suitable combination of the foregoing. More specific examples (a non-exhaustive list) of computer-readable storage media include: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), static random access memory (SRAM), portable compact disc read-only memory (CD-ROM), digital multifunction disc (DVD), memory sticks, floppy disks, mechanical encoding devices, such as punch cards or recessed protrusions storing instructions thereon, and any suitable combination of the foregoing. The computer-readable storage media used herein are not to be construed as transient signals themselves, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through waveguides or other transmission media (e.g., light pulses through fiber optic cables), or electrical signals transmitted through wires.

[0125] The computer-readable program instructions described herein can be downloaded from computer-readable storage media to various computing / processing devices, or downloaded via a network, such as the Internet, local area network, wide area network, and / or wireless network, to an external computer or external storage device. The network may include copper transmission cables, fiber optic transmission, wireless transmission, routers, firewalls, switches, gateway computers, and / or edge servers. A network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards them to the computer-readable storage media in the respective computing / processing device.

[0126] Computer program instructions used to perform the operations of this disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, status setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Smalltalk, C++, etc., and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The computer-readable program instructions may execute entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or may be connected to an external computer (e.g., via the Internet using an Internet service provider). In some embodiments, electronic circuitry, such as programmable logic circuitry, field-programmable gate arrays (FPGAs), or programmable logic arrays (PLAs), is personalized by utilizing the status information of the computer-readable program instructions to implement various aspects of this disclosure.

[0127] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the protection scope of the claims of the present invention.

Claims

1. A method for online safety analysis, early warning, and auxiliary decision-making in power distribution networks, characterized in that, include: Step 1: Collect multi-source graph model data and real-time operation data of the main grid, distribution network and microgrid, identify and dynamically update the topological connection relationship of grid equipment, and construct a full topological graph of the grid at all voltage levels; Step 2: Perform ground-state power flow calculation and static N-1 security analysis on the full voltage level power grid topology map to identify cross-section overload and equipment overload risks, and use the load forecasting model to generate an enhanced topology map with risk level, type and trend. Step 3: Based on the enhanced topology full graph, the fault point is determined by the dynamic adaptive fault credibility assessment model, and a fault handling plan is generated based on the knowledge graph reasoning method. The dynamic adaptive fault credibility assessment model integrates topology influence factors, operational risk trend factors and equipment health decay factors, and calculates the dynamic comprehensive fault index of the fault candidate path through a nonlinear fusion function, and determines the fault point based on the dynamic comprehensive fault index.

2. The method for online safety analysis, early warning, and auxiliary decision-making in a power distribution network according to claim 1, characterized in that: Step 1 involves collecting multi-source graph model data and real-time operational data from the main grid, distribution grid, and microgrids; identifying and dynamically updating the topological connections of grid equipment; and constructing a full-scale grid topology map across all voltage levels. Specific implementation methods include: Step 1.1: Based on multi-source graph model data and real-time operation data, traverse the power grid equipment nodes and connection edges, identify and establish the electrical connection relationship of the entire network, and form the initial topology; Step 1.2: Based on the initial topology, and combined with the switch change signals extracted from real-time operating data, the parts of the topology that have changed are locally reconstructed and updated to form a complete topology map of the power grid at all voltage levels.

3. The method for online safety analysis, early warning, and auxiliary decision-making in a power distribution network according to claim 1, characterized in that: Step 2 involves performing ground-state power flow calculations and static N-1 security analysis on the full voltage level power grid topology map to identify cross-section overload and equipment overload risks. An enhanced topology map with risk levels, types, and trends is then generated using a load forecasting model. Specific implementation methods include: Step 2.1: Perform ground-state power flow calculation based on the full voltage level grid topology map, and use the static N-1 interruption analysis method to simulate the expected fault set, outputting steady-state and fault-state analysis results including bus voltage and branch power flow; Step 2.2: Based on the steady-state and fault-state analysis results, combined with the equipment's rated parameters and cross-sectional power transmission limits, the risk levels of equipment overload and cross-sectional limit exceedance are obtained, and a structured risk list is generated. Step 2.3: Use the load forecasting model to obtain forecast information, associate the risk list with the forecast information with the corresponding grid equipment in the full voltage level grid topology map, and generate an enhanced topology map with risk level, type and trend.

4. The method for online safety analysis, early warning, and auxiliary decision-making in a power distribution network according to claim 3, characterized in that: In step 2.2, the risk levels of equipment overload and cross-sectional exceedance are obtained, and a structured risk list is generated. Specific steps include: For the equipment, the active power and reactive power of each branch are obtained based on the steady-state and fault-state analysis results. The rated capacity of the equipment is introduced, and the equipment load rate is calculated by ratio. For the cross section, the active power of the cross section is obtained based on the steady-state and fault-state analysis results. A cross section transmission power limit is introduced, and the cross section load rate is calculated by the ratio. Based on equipment load rate and cross-section load rate, a preset risk level threshold is introduced to classify the risk status of equipment and cross-section into multiple levels. Iterate through all equipment and cross-sections, determine the risk level, and record any non-normal status as a risk event, generating a structured risk list. The risk list includes the risk equipment number, risk type, risk level, and occurrence time.

5. The method for online safety analysis, early warning, and auxiliary decision-making in a power distribution network according to claim 1, characterized in that: In step 3, the implementation method of the dynamic adaptive fault reliability assessment model includes: Based on the connection relationships and path impedances of power grid equipment in the enhanced topology graph, a fault information matrix is ​​constructed; and a breadth-first search algorithm is used to select candidate fault paths starting from the fault reporting origin node. Calculate the basic electrical quantity deviations, including voltage deviation per unit value, for each node on the candidate fault path. Per-unit value of current deviation and path impedance per unit value ; based on , , Calculate the topology impact factor, operational risk trend factor, and equipment health degradation factor for each node: A nonlinear fusion function is used to fuse topology influence factors, operational risk trend factors, equipment health degradation factors, and basic electrical quantity deviations to obtain a dynamic comprehensive fault index.

6. The method for online safety analysis, early warning, and auxiliary decision-making in a power distribution network according to claim 5, characterized in that: For any node, the topological influence factor is calculated as follows: The topology influence factor is obtained by weighted summing of the normalized values ​​of the betweenness centrality and electrical coupling of the nodes. Betweenness centrality is the frequency at which the node is the intermediate node of the shortest path; electrical coupling is calculated based on the equivalent electrical impedance between the node and other nodes.

7. The method for online safety analysis, early warning, and auxiliary decision-making in a power distribution network according to claim 5, characterized in that: The calculation method for the operating risk trend factor is as follows: The operational risk trend factor is obtained based on the risk level mapping.

8. The method for online safety analysis, early warning, and auxiliary decision-making in a power distribution network according to claim 5, characterized in that: The calculation method for the equipment health degradation factor is as follows: The normalized result of the difference between the actual measured value and the normal operating reference value of each equipment indicator is calculated, and the weighted normalized result is subtracted to obtain the difference result. Multiply the interpolation results corresponding to all devices to obtain the product result, and subtract the product result to obtain the device health decay factor.

9. The method for online safety analysis, early warning, and auxiliary decision-making in a power distribution network according to claim 5, characterized in that: A nonlinear fusion function is used to fuse topology influence factors, operational risk trend factors, equipment health degradation factors, and basic electrical quantity deviations to obtain a dynamic comprehensive fault index. The specific steps include: The feature vector is composed of voltage deviation per unit value, current deviation per unit value, path impedance per unit value, topology influence factor, operational risk trend factor, and equipment health degradation factor. The dynamic comprehensive fault index is calculated by inputting the feature vector into the nonlinear fusion function obtained by training based on historical fault data.

10. The method for online safety analysis, early warning, and auxiliary decision-making in a power distribution network according to claim 1, characterized in that: In step 3, the fault point is determined based on the dynamic comprehensive fault index, and a fault handling plan is generated based on the knowledge graph reasoning method. The specific methods include: The candidate fault path with the largest dynamic comprehensive fault index is selected as the most likely fault path, and the fault location is determined by calculating the voltage deviation gradient between adjacent nodes on the most likely fault path. Based on the location of the fault point, the outage area and key load points affected by the fault are identified in the enhanced topology map. The upstream power supply point and downstream load node of the fault point are marked, the location of the disconnecting switch to be operated is identified to form electrical isolation, feasible power transfer paths that meet the load rate threshold are screened, and a fault handling plan is generated after verification.

11. A distribution network online security analysis, early warning, and auxiliary decision-making system utilizing the method of any one of claims 1-10, comprising a topology full-map construction unit, a security analysis and early warning unit, and a collaborative scheduling unit, characterized in that: The topology map construction unit is used to collect multi-source map model data and real-time operation data of the main grid, distribution network and microgrid, identify and update the topology connection relationship of grid equipment, and construct a full topology map of the grid at all voltage levels; The safety analysis and early warning unit is used to perform basic state power flow calculation and static N-1 safety analysis on the full voltage level power grid topology map, identify the risk of cross-section overload and equipment overload, and generate an enhanced full topology map with risk level, type and trend using the load forecasting model; The collaborative scheduling unit is used to determine fault points based on the enhanced topology full graph, through a dynamic adaptive fault credibility assessment model, and to generate fault handling plans based on knowledge graph reasoning methods.

12. The system according to claim 11, characterized in that, The topology map construction unit includes a topology relationship identification module and a dynamic maintenance module; the security analysis and early warning unit includes a security analysis module, an operational risk assessment module, and a risk map generation module.

13. The system according to claim 11, characterized in that, The dynamic adaptive fault reliability assessment model in the collaborative scheduling unit integrates the topology influence factor, operational risk trend factor and equipment health decay factor of each node, and comprehensively evaluates the candidate fault path through a nonlinear fusion function trained based on historical fault data to determine the fault location.

14. The system according to claim 11, characterized in that, The topology influence factor quantifies the importance of a node's position in the overall network topology based on node betweenness centrality and electrical coupling; the operational risk trend factor is determined based on the risk level mapping of the risk list in the enhanced topology graph. The equipment health decay factor is determined based on online monitoring data of the equipment.

15. The system according to claim 10, characterized in that, The collaborative scheduling unit identifies the power outage area based on the location of the fault point, and generates a fault handling plan through electrical isolation and load transfer path screening and verification.

16. The system according to claim 11, characterized in that, The safety analysis module uses the Newton-Raphson method to calculate the ground-state power flow and the static N-1 interruption analysis method to simulate the expected fault set.

17. The system according to claim 11, characterized in that, The risk map generation module uses a long short-term memory network as a load forecasting model to predict future load changes and associates the risk list with the forecast information to generate an enhanced topology map.

18. A terminal, comprising a processor and a storage medium; characterized in that: The storage medium is used to store instructions; The processor is configured to operate according to the instructions to perform the steps of the method according to any one of claims 1-10.

19. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the program implements the steps of the method according to any one of claims 1-10.

Images