Power grid risk intelligent early warning and analysis method, system and equipment
By building a multi-source data fusion model and risk knowledge database for power grids, the problems of strong manual dependence and insufficient data fusion in power grid maintenance are solved, and intelligent risk warning and analysis are achieved.
Patent Information
- Application Number
- CN202510809632.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-06-17
- Publication Date
- 2025-09-23
AI Technical Summary
Existing power grid maintenance risk warning technologies have problems such as strong manual dependence, insufficient data fusion and low intelligence level, making it difficult to proactively predict and identify risks before maintenance.
By acquiring multi-source heterogeneous data of the power grid, building a directed graph knowledge base and risk knowledge graph, using topological coupling coefficients and risk dependency coefficients to conduct risk analysis, set risk rules, and generate intelligent early warning sheets.
It achieves efficient and accurate risk warning, improves the completeness of risk identification and the ability to deal with complex scenarios, and can timely identify and generate effective prevention and control strategies.
Smart Images

Figure CN120688866A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of power system security early warning, and in particular to a method, system and equipment for intelligent early warning and analysis of power grid risks. Background Art
[0002] The power grid system is a network system with a complex and vast number of internal nodes. Grid maintenance and overhaul are crucial for ensuring safe and stable system operation, ensuring normal life, and maintaining the orderly flow of production activities. Due to the complexity and widespread distribution of the power grid, risk warning and assessment have become indispensable prerequisites for grid maintenance and overhaul. Effective grid risk warning measures require both the timely and accurate identification of risk indicators from real-time, multi-source data signals across the entire network and the precise analysis and handling of the corresponding incident types and locations, eliminating potential incidents within specific areas and within the permitted response timeframe.
[0003] At present, risk warning of power grids under maintenance mode mainly relies on the following technologies and methods:
[0004] The first is risk assessment based on expert experience. Dispatchers or operations and maintenance personnel manually analyze the risks associated with changes in grid operation based on historical experience, grid operating procedures, and maintenance plans. For example, they manually calculate the load transfer capacity under N-1 / N-2 fault conditions to determine whether there are equipment overloads or voltage limits. Risk assessment based on manual experience, a guiding technique found in foundational professional manuals such as the Grid Dispatching and Operation Risk Prevention and Control Manual, suffers from significant inefficiencies and struggles to cope with the complex and ever-changing operational scenarios of the entire grid. It can only serve as a safeguard in extreme situations.
[0005] The second is risk analysis based on real-time monitoring and early warning of SCADA / EMS systems and offline simulation software. Generally, these two methods are often used in combination. Real-time monitoring and early warning based on SCADA / EMS systems uses SCADA (Supervisory Control and Data Acquisition System) and EMS (Energy Management System) to monitor the operating status of the power grid in real time, and issues alarms in combination with a preset rule base (such as equipment load rate, voltage exceeding the limit, etc.); risk analysis based on offline simulation software uses power system simulation software such as PSCAD to perform flow calculations, short-circuit analysis, transient stability simulation, etc. on the power grid under maintenance mode to assess potential risks. This combination focuses on real-time monitoring, but because it is an offline simulation, it is difficult to conduct dynamic analysis and evaluation in combination with changes in power grid topology.
[0006] The third approach is predictive early warning based on big data analysis. This involves leveraging historical operational and meteorological data, combined with machine learning algorithms (such as random forests and LSTM), to predict risks that may arise during maintenance. Current application results suggest that big data analysis methods are primarily used for post-event analysis, making it difficult to proactively predict and identify risks and generate effective prevention and control strategies before maintenance. In summary, current power grid maintenance risk warning and analysis technologies and methods suffer from a high degree of manual dependency, insufficient data integration, and low intelligence. Summary of the Invention
[0007] In view of this, the purpose of the present invention is to provide a method, device and equipment for intelligent early warning and analysis of power grid risks to overcome the problems of strong manual dependence, insufficient data fusion and low intelligence level in current power grid system maintenance.
[0008] To achieve the above objectives, the present invention adopts the following technical solutions:
[0009] In a first aspect, the present application provides a method for intelligent early warning and analysis of power grid risks, comprising:
[0010] Acquire multi-source heterogeneous data of the power grid, and perform cleaning and gap filling processing on the multi-source heterogeneous data of the power grid;
[0011] For the processed data, association relationships and indexes are constructed through association keys, and a directed graph knowledge base is constructed based on the association relationships and indexes; wherein the association keys include device ID, timestamp, and substation or line code;
[0012] Determine a topological coupling coefficient between nodes based on directed path information and historical fault information between nodes in a directed graph knowledge base, wherein the topological coupling coefficient is used to quantify the coupling relationship between nodes;
[0013] Based on the directed graph knowledge base, a risk knowledge graph is constructed, wherein, in the risk knowledge graph, vertices represent entities, directed edges represent relationships between entities, and weights of directed edges represent risk dependency coefficients between entities; the entities include users, maintenance tasks, and equipment, and the risk dependency coefficients are determined based on the topological coupling coefficients and historical data to characterize the impact between entities;
[0014] Setting risk rules to provide early warning of risk events and risk levels when repairing target equipment; wherein the risk rules include entity status information, risk dependency relationship information between entities, and risk indicators, wherein the entity status information includes information of the target equipment to be repaired, the risk dependency relationship information includes a risk dependency coefficient, and the risk indicators include conditions that trigger corresponding risk events and risk levels;
[0015] For the maintenance of target equipment, a comprehensive risk characterization index is determined based on equipment risk, user impact and load transfer capability, and the risk warning level is determined and a warning order is generated based on the comprehensive risk characterization index.
[0016] Furthermore, in some embodiments of the present application, the multi-source heterogeneous data of the power grid includes: power grid topology data, maintenance plan data, user data, real-time monitoring data, and important user access mode data;
[0017] The types of nodes in the directed graph knowledge base include substation, line, load, user and maintenance status, and the types of edges include line connection, power supply relationship and maintenance impact.
[0018] Furthermore, in some embodiments of the present application, the following is further included:
[0019] Standardize the stored multi-source heterogeneous data of the power grid, including standardizing continuous data based on mean and standard deviation, and standardizing discrete data based on one-hot encoding;
[0020] The standardized data are aligned in time windows.
[0021] Furthermore, in some embodiments of the present application, the calculation formula of the topological coupling coefficient is:
[0022]
[0023] Where LS(u, v) represents the topological coupling coefficient between nodes u and v; Path(u, v) represents the set of directed paths between nodes u and v; CP(u, v) represents the total number of directed paths between nodes u and v; e represents the number of devices in route l in the set of directed paths; ρ e represents the failure rate of equipment e;
[0024] The calculation formula of the dependency coefficient is:
[0025] R i,j =αLS i,j +β·RF ij Among them, R i,j represents the risk dependence coefficient between entities i and j; α and β represent weight coefficients; RF ij represents the probability of risk spreading from entity i to j.
[0026] Furthermore, in some embodiments of the present application, the following is further included:
[0027] The state of the device is predicted based on multiple data sources and a state prediction dynamic adjustment formula. The state prediction dynamic adjustment formula is:
[0028]
[0029] in, represents the state prediction value of device i; represents the weight of data source k; represents the prediction of the state of device i by data source k; γ represents the attenuation coefficient; K represents the total number of data sources.
[0030] Furthermore, in some embodiments of the present application, the method further includes: constructing a user-grid coupling discrimination model, wherein the user-grid coupling discrimination model is used to calculate a user-grid coupling discrimination system, and the user-grid coupling discrimination coefficient is used to identify high-coupling users that require key protection during the maintenance of target equipment;
[0031] The user grid coupling discrimination model is expressed as:
[0032]
[0033] Among them, Coup j represents the user grid coupling discrimination coefficient of user j; g represents the upper power supply that supplies power to user j; G(j) represents the upper power supply set of user j; LS(g, j) represents the topological coupling coefficient between user j and power supply g; L j Represents the load level of user j; ∈ is a preset real number; where ∈ is used to prevent the denominator in the user-grid coupling discrimination model from being 0.
[0034] Furthermore, in some embodiments of the present application, the method further includes dynamically adjusting the weight of the directed edge based on the occurrence of random events, and the adjustment formula is:
[0035]
[0036] in, represents the weight of the adjusted directed edge; γ is the attenuation coefficient; e represents the random event in the path; ES e is the severity assessment coefficient of event e.
[0037] Furthermore, in some embodiments of the present application, the early warning sheet includes risk level information, risk description information, impact scope information, and control measures information;
[0038] Among them, the risk level information includes the risk warning level; the risk description information includes the relationship between entities in the risk knowledge graph library; the impact range information includes the list of associated devices or users determined by the user-grid coupling discrimination model; the control measure information is determined based on the information in the preset control measure library corresponding to the risk event and risk level.
[0039] In a second aspect, the present application provides a power grid risk intelligent early warning and analysis system, comprising:
[0040] A fusion module is configured to acquire multi-source heterogeneous data from the power grid, cleanse and fill gaps in the data, construct association relationships and indexes based on association keys for the processed data, and construct a directed graph knowledge base based on the association relationships and indexes; wherein the association keys include device IDs, timestamps, and substation or line codes; and determine topological coupling coefficients between nodes based on directed path information and historical fault information between nodes in the directed graph knowledge base, wherein the topological coupling coefficients are used to quantify the coupling relationships between nodes.
[0041] An early warning module is used to construct a risk knowledge graph based on the directed graph knowledge base, wherein, in the risk knowledge graph, vertices represent entities, directed edges represent relationships between entities, and weights of directed edges represent risk dependency coefficients between entities; the entities include users, maintenance tasks, and equipment, and the risk dependency coefficients are determined based on the topological coupling coefficients and historical data to characterize the impact between entities; risk rules are set to issue early warnings for risk events and risk levels when repairing target equipment; wherein the risk rules include entity status information, risk dependency relationship information between entities, and risk indicators, wherein the entity status information includes information on the target equipment to be repaired, the risk dependency relationship information includes the risk dependency coefficient, and the risk indicators include conditions for triggering corresponding risk events and risk levels;
[0042] The early warning order generation and intelligent analysis module is used for the maintenance of target equipment, determines the comprehensive risk characterization index based on equipment risk, user impact and load transfer capacity, and determines the risk warning level and generates an early warning order based on the comprehensive risk characterization index.
[0043] In a third aspect, the present application provides a power grid risk intelligent early warning and analysis device, including a processor and a memory, wherein the processor is connected to the memory:
[0044] The processor is configured to call and execute the program stored in the memory;
[0045] The memory is used to store the program, and the program is at least used to execute the above-mentioned power grid risk intelligent early warning and analysis method.
[0046] The present invention relates to the technical field of power system safety early warning, and in particular to a method, system and device for intelligent early warning and analysis of power grid risks, the method comprising: acquiring multi-source heterogeneous data of the power grid, and cleaning and supplementing the multi-source heterogeneous data of the power grid; constructing association relationships and indexes through association keys for the processed data, and constructing a directed graph knowledge base based on the association relationships and indexes; wherein the association keys include device IDs, timestamps and substation or line codes; determining the topological coupling coefficients between nodes based on directed path information and historical fault information between nodes in the directed graph knowledge base, the topological coupling coefficients being used to quantify the coupling relationships between nodes; and constructing a risk knowledge graph based on the directed graph knowledge base, wherein, in the risk knowledge graph, vertices represent entities and directed edges represent relationships between entities. The weights of directed edges represent the risk dependency coefficients between entities. Entities include users, maintenance tasks, and equipment. The risk dependency coefficients are determined based on the topological coupling coefficient and historical data and are used to characterize the impact between entities. Risk rules are set to provide early warnings of risk events and risk levels when repairing target equipment. Risk rules include entity status information, risk dependency relationship information between entities, and risk indicators. Entity status information includes information about the target equipment to be repaired, risk dependency information includes the risk dependency coefficient, and risk indicators include the conditions that trigger corresponding risk events and risk levels. For the maintenance of target equipment, a comprehensive risk characterization index is determined based on equipment risk, user impact, and load transfer capacity. Based on this comprehensive risk characterization index, the risk warning level is determined and an early warning notice is generated. In this way, by integrating multi-source data, an intelligent risk level determination model is established, enabling efficient, accurate, and intelligent risk warnings. BRIEF DESCRIPTION OF THE DRAWINGS
[0047] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.
[0048] Figure 1 This is a flow chart of a method for intelligent early warning and analysis of power grid risks provided by an embodiment of the present invention;
[0049] Figure 2 This is a schematic diagram of the structure of the power grid risk intelligent early warning and analysis system provided by an embodiment of the present invention;
[0050] Figure 3 It is a structural diagram of the power grid risk intelligent early warning and analysis device provided by an embodiment of the present invention. DETAILED DESCRIPTION
[0051] To make the objectives, technical solutions, and advantages of the present invention more apparent, the technical solutions of the present invention will be described in detail below. Obviously, the embodiments described are only some of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other implementations obtained by those of ordinary skill in the art without inventive effort are within the scope of protection of the present invention.
[0052] Figure 1 This is a flow chart of the intelligent early warning and analysis method for power grid risk provided by the embodiment of the present invention. Figure 1 , this embodiment may include the following steps:
[0053] S101. Acquire multi-source heterogeneous data of the power grid, and clean and fill in gaps in the multi-source heterogeneous data of the power grid.
[0054] As previously mentioned, previous methods for grid maintenance risk warning and analysis suffer from insufficient multi-source data fusion, negatively impacting the accuracy of risk identification and risk discrimination. Therefore, this application constructs a fusion model for grid multi-source data as a precursor to subsequent risk warning and identification. The data to be aggregated, i.e., the grid's multi-source heterogeneous data, specifically includes: grid topology data, maintenance plan data, user data, real-time monitoring data, and data on key user access methods.
[0055] Specifically, grid topology data is recorded or generated by production management systems, geographic information databases, and CIM / E-formatted grid model files. It primarily includes wiring diagrams, equipment connection relationships, and electrical parameters (such as impedance and capacity). Maintenance plan data is recorded or generated by scheduling plans and outage management systems. It primarily includes equipment to be repaired, maintenance schedules (start and end times), and the impact area (such as outage areas and load transfer plans). User data is recorded or generated by load management systems. It primarily includes user load curves, power supply information for key users (such as hospitals and data centers), and user power supply reliability requirements. Real-time monitoring data is recorded and monitored in real time by SCADA and EMS. It primarily includes real-time measurement data (such as voltage, frequency, and power), PMU dynamic data, and equipment status. Key user access method data is obtained through key user power supply contracts and site surveys. It primarily includes the number and serial number of power sources, backup power supply switchover time and status, and key user internal wiring diagrams. After acquisition, this data is cleaned and supplemented.
[0056] S102: Construct association relationships and indexes through association keys, and construct a directed graph knowledge base based on the association relationships and indexes.
[0057] The association key includes the device ID, timestamp, and substation or line code.
[0058] Table 1 below shows an exemplary index association method for constructing multi-source heterogeneous data of a power grid:
[0059] Table 1. Examples of index association methods for multi-source heterogeneous data in power grids
[0060] Data Category Associative Key Network topology Device ID (such as transformer number) Maintenance plan Device ID + timestamp User data Substation / Line Coding Real-time monitoring Device ID + timestamp
[0061] After the data is associated and indexed, a graph database is used to store the topological relationships between the data. Based on this graph database, the overall power grid data can be represented as a knowledge base based on a directed graph. Its edges and vertices (also known as nodes) are determined by the aforementioned data. Specifically, node types include substations, lines, loads, users, and maintenance status; edge types include line connections, power supply relationships, and maintenance impacts.
[0062] At this time, when you need to locate the location to be inspected or at risk, you can quickly query and locate the specific equipment, line or user that needs to be inspected and analyzed in the directed graph knowledge base by typing in the field.
[0063] S103 : Determine a topological coupling coefficient between nodes based on directed path information and historical fault information between nodes in a directed graph knowledge base.
[0064] Among them, the topological coupling coefficient is used to quantify the coupling relationship between nodes.
[0065] S104. Construct a risk knowledge graph based on the directed graph knowledge base.
[0066] In the risk knowledge graph, vertices represent entities, directed edges represent relationships between entities, and the weights of directed edges represent risk dependency coefficients between entities. Entities include users, maintenance tasks, and equipment. The risk dependency coefficient is determined based on the topological coupling coefficient and historical data, and is used to characterize the impact between entities.
[0067] S105. Risk rules are set to provide early warning of risk events and risk levels when the target equipment is being repaired.
[0068] Among them, the risk rules include entity status information, risk dependency information between entities and risk indicators. Among them, the entity status information includes the target equipment information to be repaired, the risk dependency information includes the risk dependency coefficient, and the risk indicators include the conditions for triggering corresponding risk events and risk levels.
[0069] S106. For the maintenance of target equipment, a comprehensive risk characterization index is determined based on equipment risk, user impact, and load transfer capability, and a risk warning level is determined based on the comprehensive risk characterization index and a warning order is generated.
[0070] Furthermore, in some embodiments of the present application, it also includes: standardizing the stored multi-source heterogeneous data of the power grid, including standardizing continuous data based on mean and standard deviation, and standardizing discrete data based on one-hot encoding; and aligning the time windows of the standardized data.
[0071] Specifically, in practical applications, since the above-mentioned multi-source heterogeneous data of the power grid includes multiple types of data, such as continuous data, discrete data, and state data (such as data whose state is determined by virtual variables), in some embodiments of the present application, it can also be standardized by the following formula:
[0072]
[0073] In the above formula, X norm represents the data after standardization of continuous data; μ represents the mean of this type of data; σ represents the standard deviation of this type of data; X code Indicates the data after discrete data is encoded; OneHot means the data is encoded with one hot encoding.
[0074] Furthermore, after completing the data normalization process, in order to ensure the temporal and spatial consistency of all data labels related to risk identification, in the embodiment of the present application, various types of data are also aligned in time windows to ensure synchronous analysis of topology, load, and maintenance status at the same time. Time alignment can be expressed by the following formula:
[0075]
[0076] In the above formula, represents the data sequence of the source data of each category after time window alignment; ∩(·) represents the intersection; K represents the total number of data sources; Indicates the start time of the kth data source; Represents the end time of the kth data source. The data slice after time window alignment can be expressed as:
[0077]
[0078] In the above formula, A time series representing the topological relationship of data in a data slice; represents the load time series in the data slice; Represents the data in the data slice that represents the maintenance status; other data types also include important user access methods, etc.
[0079] Furthermore, when performing risk identification, it is necessary to identify the key connection relationships between devices or nodes. To this end, the coupling relationship between devices must be quantified. Therefore, this application constructs a topology fusion process to calculate the topology coupling coefficient. The specific calculation formula is as follows:
[0080]
[0081] Where LS(u, v) represents the topological coupling coefficient between nodes u and v; Path(u, v) represents the set of directed paths between nodes u and v; CP(u, v) represents the total number of directed paths between nodes u and v; e represents the number of devices in route l in the set of directed paths; ρ e It represents the failure rate of device e based on historical failure fitting.
[0082] At the same time, considering that the data used for risk prediction in the actual maintenance and risk identification process comes from multi-dimensional prediction data (such as the samples used in load forecasting include both real-time data collected by SCADA and meteorological data, etc.), the dynamic adjustment mechanism added to the equipment status prediction in this application example is as follows:
[0083]
[0084] in, represents the state prediction value of device i; represents the weight of data source k; represents the prediction of the state of device i by data source k; γ represents the attenuation coefficient; K represents the total number of data sources.
[0085] In order to identify high-coupling users that need to be protected during maintenance, the following calculation model for user-grid coupling discrimination coefficient is constructed:
[0086]
[0087] Among them, Coup j represents the user grid coupling discrimination coefficient of user j; g represents the upper power supply that supplies power to user j; G(j) represents the upper power supply set of user j; LS(g, j) represents the topological coupling coefficient between user j and power supply g; L j Represents the load level of user j; ∈ is a preset real number (a very small real number generated to prevent the denominator from being 0).
[0088] In this way, user importance can be mapped to grid equipment risk, thereby introducing user power supply reliability requirements into the decision-making process.
[0089] The following is an analysis of a specific scenario to illustrate the role and significance of the above steps:
[0090] Scenario setting: Maintenance of a 220kV main transformer in a certain regional power grid.
[0091] The maintenance equipment is coded as [1, 0, 0] (transformer type) through formula (1); the SCADA overload alarm data within the maintenance period is filtered out through time alignment through formula (2); the LS of another main transformer and the key line is calculated to be 0.88 through formula (4) (in addition to the physical connection, the possible diffusion direction of the risk is determined by the topological relationship); the weight of the SCADA data source is calculated to be 0.9 through formula (5), and the remaining prediction data are eliminated, and the prediction of the SCADA data source is adopted (solving the problem of "who to trust" when there is a conflict in multi-source data); and the power supply guarantee of these three users should be given priority attention through formula (6).
[0092] Furthermore, after completing the fusion of multi-source data, a risk knowledge graph of the system is constructed to enable the system to automatically issue alerts, i.e., intelligent warnings, when conditions under risk rules are triggered.
[0093] Specifically, in the risk knowledge graph, users, maintenance tasks, and equipment are all set as entities. The relationship between them is used to determine the triggering rules of risk alerts. The risk rules can be exemplified by the following field codes:
[0094]
[0095] In the above example, risk_event(overload, Level) represents a risk event, the left side of the bracket represents the risk event type, which is load overload in this example; Level represents the risk level, which can be set according to the relevant safety regulations of the State Grid; :- means that when all the conditions on the right are met, the risk event alarm on the left is triggered; equipment(E, transformer) indicates that the type of equipment E is a transformer; maintainance(E, active) indicates that equipment E is in maintenance status; linked(E, Line, 0.8) indicates that the correlation between equipment E and the line is 0.8, and load_rate(Line) indicates the load rate of the line.
[0096] Referring to the example above, the risk rule includes the following elements:
[0097] Entity status information, specifically including entity category (such as user, equipment) and maintenance status.
[0098] Risk dependency information between entities, such as the risk correlation between transformers and lines, is used to measure the risk propagation monitoring path.
[0099] Risk indicators are specifically risk monitoring boundaries or judgment conditions set according to different risk event categories (such as the risk indicators that should be monitored in accordance with the "Guidelines for Safety and Stability of Power Systems" and other guidelines). For example, monitoring the relationship between two lines installed on the same pole is used to determine whether the "N-2 risk caused by maintenance of double circuits on the same pole" is triggered as a judgment condition; monitoring "whether the transformer oil temperature is greater than 90 degrees" is used as the boundary condition for the transformer oil temperature exceeding the limit.
[0100] In this way, according to the above elements, risk triggering rules and field settings, all entities in the system can be converted into a risk knowledge graph represented by a directed graph, whose vertices (i.e. nodes) represent entities, directed edges represent the relationships between entities, and the weights of directed edges represent the risk dependence coefficients between entities, i.e. risk dependence and causal relationships.
[0101] In some embodiments of the present application, the calculation formula of the above dependency coefficient is specifically:
[0102] R i,j =αLS i,j +β·RF ij (7)
[0103] Among them, R i,j represents the risk dependence coefficient between entities i and j; α and β represent weight coefficients; RF ij It represents the probability of risk spreading from entity i to j, which can be fitted by historical data.
[0104] It can be understood that the risk dependency coefficient in the above formula can be used as the edge weight in the risk knowledge graph. The role of the edge weight is to determine whether to trigger the inspection of related equipment. As a core element in the risk rule, this coefficient can be used as a judgment condition for the risk rule and a judgment tool for the direction of risk propagation. The specific triggering rules are described as follows:
[0105] Table 2. Relationship between the scope of boundary rights and early warning measures
[0106] Edge weight range Risk Communication Intensity Explanation Early warning impact 0.0~0.3 Weak correlation (negligible risk transfer) Do not trigger associated device check 0.3~0.7 Moderate correlation Needs to be checked but is not considered as a major risk source 0.7~1.0 Strong association (high risk transmission path) Immediate warning and activation of linkage control
[0107] In addition, as a complex system, the power grid may experience random events of different risk levels at any time. In order to dynamically respond to the impact of random events on risk warning, this application also dynamically adjusts the weights of directed edges based on the random events that occur. The specific adjustment formula is:
[0108]
[0109] in, represents the edge weight between entities after the random event triggering adjustment; γ represents the attenuation coefficient; e represents the random event that occurs in the path between i and j (such as a short circuit in line L1 between i and j); ESe represents the severity assessment coefficient of event e. Different event types can be evaluated by referring to the format and value range in Table 2. It should be noted that when edge weights are updated, the corresponding risk rules will also change. This allows for responses to dynamic disturbance events, making risk warnings and response measures more timely and accurate.
[0110] The following still uses a specific scenario to illustrate how the present invention realizes intelligent early warning, and the scenario is still the maintenance of a 220kV main transformer in the power grid in the above-mentioned region.
[0111] Enter main transformer maintenance in the risk knowledge library, and the field is maintenance(b1,active).
[0112] The graph reasoning process is as follows: First, the edge weight between main transformer b1 and line l5 is calculated to be 0.85, which triggers linked(b1, Line5, 0.85). This determines that the maintenance of main transformer b1 has a significant impact on line l5, that is, line l5 is a critical outgoing line. Then, the field load_rate(Line5) is detected to be 0.92, triggering the risk warning condition "transformer maintenance + critical line load greater than 90%", generating a Level-2 risk alarm.
[0113] On this basis, this application determines the comprehensive risk characterization index for the maintenance of target equipment from the perspective of equipment risk, user impact and load transfer capability, and determines the risk warning level and generates a warning order based on the comprehensive risk characterization index.
[0114] Specifically, the warning information generated in the above embodiment is given based on the risk rule trigger. On this basis, the present application also adds a comprehensive evaluation mechanism to conduct a comprehensive and integrated evaluation of the warning risk, including building a model from three dimensions: equipment risk, user impact, and load transfer capability.
[0115] For equipment risk, this application uses N-1 safety margin to characterize it, specifically
[0116]
[0117] In the above formula, D ke Represents the collection of key devices in the system; C i Indicates the rated core indicators of key equipment (such as rated power, rated capacity, etc.); Indicates the power load of key equipment i after maintenance.
[0118] As for the user impact, the calculation formula of the user-grid coupling coefficient in formula (6) can be directly used to express it.
[0119] The load transfer capacity can be expressed as follows:
[0120]
[0121] In the above formula, M represents the total number of spare lines; represents the remaining capacity of the backup line m; ΔP m Indicates the load that needs to be transferred to line m; P critical Indicates the critical load level that must be guaranteed.
[0122] Based on the above three variables, this application can construct the following systematic comprehensive risk characterization index model:
[0123]
[0124] In the above formula, w1, w2, and w3 represent the weights of the three indicators respectively; N ku Represents a collection of important users.
[0125] Based on the values calculated by the above system comprehensive risk characterization index model, the numerical range can be set to generate the following risk warning levels:
[0126] Risk level I: RI>0.8.
[0127] Risk Level II: 0.6 <RI≤0.8。
[0128] Risk Level III: RI≤0.6.
[0129] After completing the comprehensive risk assessment of the system, natural language generation technology can also be used to generate an early warning form with one click. The content of the early warning form includes risk level information, risk description information, impact range information and control measures information; among them, risk level information includes risk warning level; risk description information includes the relationship between entities in the risk knowledge library; impact range information includes the list of associated devices or users determined by the user-grid coupling discrimination model; control measures information is determined based on the information in the preset control measures library corresponding to the risk events and risk levels.
[0130] The content and form of the early warning sheet can be found in Table 3 below.
[0131] Table 3. Content and form of early warning notice
[0132] Field Generate Logic Data Source Risk Level Formula 11 calculation result + rule base matching Risk Assessment Model Risk Description Natural language template filling Relationships between entities in the risk knowledge graph Scope of impact List of associated devices / users Topological fusion result (Formula 6) Control measures Preset measures library + real-time optimization suggestions Historical disposal cases / expert knowledge base
[0133] Among them, for the risk knowledge graph that has been constructed, a control measures library can be pre-set in the system for each risk level and specific risk event category in accordance with corresponding safety specifications and expert experience. When an early warning form is generated, the control measures library can be retrieved according to the content of the early warning form by natural language to generate specific response measures suggestions.
[0134] The intelligent early warning and analysis method for power grid risk provided by the present application constructs a multi-source data fusion model for the power grid through processes such as data classification, cleaning and standardization, time window alignment, topology fusion, and user-grid coupling. By constructing a risk knowledge graph based on graph theory for the power grid, a risk trigger mechanism is given under corresponding rules. At the same time, based on the system comprehensive risk characterization coefficient and using natural language generation technology, early warning orders are intelligently generated. The intelligent and efficient multi-process of multi-source data fusion, risk assessment, intelligent early warning and early warning order generation is achieved. On this basis, the completeness of risk identification is improved through the dynamic fusion of multi-source heterogeneous data; the edge weight dynamic update mechanism of the risk knowledge graph provided by the present invention is used to improve the risk identification capability of the system in dealing with complex scenarios; by introducing the user-grid coupling coefficient, the user power supply security is introduced into the model, so that the power grid can more accurately judge the impact of maintenance on important users, thereby optimizing the corresponding security decisions.
[0135] Based on the same inventive concept, this application also provides a power grid risk intelligent early warning and analysis system, Figure 2 FIG. 1 is a schematic diagram of the structure of the intelligent early warning and analysis system for power grid risks provided by an embodiment of the present invention. Figure 2 As shown, the system at least includes:
[0136] The fusion module 11 is used to obtain multi-source heterogeneous data from the power grid, cleanse and fill gaps in the data, construct association relationships and indexes based on association keys for the processed data, and construct a directed graph knowledge base based on the association relationships and indexes. The association keys include device IDs, timestamps, and substation or line codes. The topological coupling coefficients between nodes are determined based on the directed path information and historical fault information between nodes in the directed graph knowledge base. The topological coupling coefficients are used to quantify the coupling relationships between nodes.
[0137] An early warning module 12 is configured to construct a risk knowledge graph based on a directed graph knowledge base, wherein, in the risk knowledge graph, vertices represent entities, directed edges represent relationships between entities, and weights of directed edges represent risk dependency coefficients between entities; entities include users, maintenance tasks, and equipment, and risk dependency coefficients are determined based on topological coupling coefficients and historical data to characterize the impact between entities; risk rules are set to provide early warnings of risk events and risk levels when repairing target equipment; wherein the risk rules include entity status information, risk dependency relationship information between entities, and risk indicators; wherein the entity status information includes information about the target equipment to be repaired, the risk dependency relationship information includes the risk dependency coefficient, and the risk indicators include conditions that trigger corresponding risk events and risk levels;
[0138] The early warning form generation and intelligent analysis module 13 is used to determine a comprehensive risk characterization index based on equipment risk, user impact and load transfer capability for the maintenance of target equipment, and to determine the risk warning level and generate an early warning form based on the comprehensive risk characterization index.
[0139] Regarding the system in the above embodiment, the specific manner in which each module performs operations has been described in detail in the embodiment of the method, and will not be elaborated here.
[0140] The present invention also provides a power grid risk intelligent early warning and analysis device for implementing the above method embodiment. Figure 3 Schematic diagram of the structure of the intelligent early warning and analysis device for power grid risk provided by the embodiment of the present invention. Figure 3 As shown, the intelligent early warning and analysis device for power grid risks in this embodiment includes a processor 21 and a memory 22, wherein the processor 21 is connected to the memory 22. The processor 21 is used to call and execute a program stored in the memory 22; the memory 22 is used to store a program, which is used to at least execute the method in the above embodiment.
[0141] The specific implementation plan of the power grid risk intelligent early warning and analysis device provided in the embodiment of this application can refer to the implementation plan of the power grid risk intelligent early warning and analysis method in any of the above embodiments, and will not be repeated here.
[0142] It can be understood that the same or similar parts of the above embodiments can be referenced to each other, and the contents not described in detail in some embodiments can refer to the same or similar contents in other embodiments.
[0143] It should be noted that, in the description of the present invention, the terms "first", "second", etc. are used for descriptive purposes only and should not be understood as indicating or implying relative importance. In addition, in the description of the present invention, unless otherwise specified, the meaning of "plurality" is at least two.
[0144] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, segment or portion of code comprising one or more executable instructions for implementing the steps of a specific logical function or process, and the scope of the preferred embodiments of the present invention includes alternative implementations in which functions may be performed out of the order shown or discussed, including performing functions in a substantially simultaneous manner or in the reverse order depending on the functions involved, which should be understood by those skilled in the art to which the embodiments of the present invention pertain.
[0145] It should be understood that various parts of the present invention can be implemented using hardware, software, firmware, or a combination thereof. In the above-described embodiments, multiple steps or methods can be implemented using software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented using hardware, as in another embodiment, any one of the following technologies known in the art or a combination thereof can be used: a discrete logic circuit having a logic gate circuit for implementing a logic function on a data signal, an application-specific integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.
[0146] Those skilled in the art will understand that all or part of the steps in the method of the above embodiment can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable storage medium. When the program is executed, it includes one or a combination of the steps of the method embodiment.
[0147] In addition, the functional units in the various embodiments of the present invention may be integrated into a single processing module, or each unit may exist physically separately, or two or more units may be integrated into a single module. The aforementioned integrated modules may be implemented in the form of hardware or in the form of software functional modules. If the integrated modules are implemented in the form of software functional modules and sold or used as independent products, they may also be stored in a computer-readable storage medium.
[0148] The storage medium mentioned above can be a read-only memory, a magnetic disk or an optical disk, etc.
[0149] Throughout this specification, reference to terms such as "one embodiment," "some embodiments," "examples," "specific examples," or "some examples" means that a specific feature, structure, material, or characteristic described in conjunction with that embodiment or example is included in at least one embodiment or example of the present invention. In this specification, schematic representations of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.
[0150] Although the embodiments of the present invention have been shown and described above, it will be understood that the above embodiments are illustrative and are not to be construed as limitations on the present invention. A person skilled in the art may change, modify, replace and modify the above embodiments within the scope of the present invention.
Claims
1. A method for intelligent early warning and analysis of power grid risks, characterized in that: include: Acquire multi-source heterogeneous data of the power grid, and perform cleaning and gap filling processing on the multi-source heterogeneous data of the power grid; For the processed data, association relationships and indexes are constructed through association keys, and a directed graph knowledge base is constructed based on the association relationships and indexes; wherein the association keys include device ID, timestamp, and substation or line code; Determine a topological coupling coefficient between nodes based on directed path information and historical fault information between nodes in a directed graph knowledge base, wherein the topological coupling coefficient is used to quantify the coupling relationship between nodes; Based on the directed graph knowledge base, a risk knowledge graph is constructed, wherein, in the risk knowledge graph, vertices represent entities, directed edges represent relationships between entities, and weights of directed edges represent risk dependency coefficients between entities; the entities include users, maintenance tasks, and equipment, and the risk dependency coefficients are determined based on the topological coupling coefficients and historical data to characterize the impact between entities; Setting risk rules to provide early warning of risk events and risk levels when repairing target equipment; wherein the risk rules include entity status information, risk dependency relationship information between entities, and risk indicators, wherein the entity status information includes information of the target equipment to be repaired, the risk dependency relationship information includes a risk dependency coefficient, and the risk indicators include conditions that trigger corresponding risk events and risk levels; For the maintenance of target equipment, a comprehensive risk characterization index is determined based on equipment risk, user impact and load transfer capability, and the risk warning level is determined and a warning order is generated based on the comprehensive risk characterization index.
2. The method for intelligent early warning and analysis of power grid risks according to claim 1 is characterized in that: The multi-source heterogeneous data of the power grid includes: power grid topology data, maintenance plan data, user data, real-time monitoring data and important user access mode data; The types of nodes in the directed graph knowledge base include substation, line, load, user and maintenance status, and the types of edges include line connection, power supply relationship and maintenance impact.
3. The method for intelligent early warning and analysis of power grid risks according to claim 1, characterized in that: Also includes: Standardize the stored multi-source heterogeneous data of the power grid, including standardizing continuous data based on mean and standard deviation, and standardizing discrete data based on one-hot encoding; The standardized data are aligned in time windows.
4. The method for intelligent early warning and analysis of power grid risks according to claim 1, characterized in that: The calculation formula of the topological coupling coefficient is: Where LS(u, v) represents the topological coupling coefficient between nodes u and v; Path(u, v) represents the set of directed paths between nodes u and v; CP(u, v) represents the total number of directed paths between nodes u and v; e represents the number of devices in route l in the set of directed paths; ρ e represents the failure rate of equipment e; The calculation formula of the dependency coefficient is: R i,j =αLS i,j +β·RF ij Among them, R i,j represents the risk dependence coefficient between entities i and j; α and β represent weight coefficients; RF ij represents the probability of risk spreading from entity i to j.
5. The method for intelligent early warning and analysis of power grid risks according to claim 1 is characterized in that: Also includes: The state of the device is predicted based on multiple data sources and a state prediction dynamic adjustment formula. The state prediction dynamic adjustment formula is: in, represents the state prediction value of device i; represents the weight of data source k; represents the prediction of the state of device i by data source k; γ represents the attenuation coefficient; K represents the total number of data sources.
6. The method for intelligent early warning and analysis of power grid risks according to claim 1, characterized in that: Also includes: Constructing a user-grid coupling discrimination model, wherein the user-grid coupling discrimination model is used to calculate a user-grid coupling discrimination system, and the user-grid coupling discrimination coefficient is used to identify high-coupling users that require key protection during target equipment maintenance; The user grid coupling discrimination model is expressed as: Among them, Coup j represents the user grid coupling discrimination coefficient of user j; g represents the upper power supply that supplies power to user j; G(j) represents the upper power supply set of user j; LS(g, j) represents the topological coupling coefficient between user j and power supply g; L j Represents the load level of user j; ∈ is a preset real number; where ∈ is used to prevent the denominator in the user-grid coupling discrimination model from being 0.
7. The method for intelligent early warning and analysis of power grid risks according to claim 1, characterized in that: Also includes: Based on the random events that occur, the weights of the directed edges are dynamically adjusted, and the adjustment formula is: in, represents the weight of the directed edge after adjustment; γ is the attenuation coefficient; e represents the random event in the pathway; ES e is the severity assessment coefficient of event e.
8. The method for intelligent early warning and analysis of power grid risks according to claim 1, characterized in that: The early warning sheet includes risk level information, risk description information, impact scope information and control measures information; Among them, the risk level information includes the risk warning level; the risk description information includes the relationship between entities in the risk knowledge graph library; the impact range information includes the list of associated devices or users determined by the user-grid coupling discrimination model; the control measure information is determined based on the information in the preset control measure library corresponding to the risk event and risk level.
9. An intelligent early warning and analysis system for power grid risks, characterized in that: include: A fusion module is used to obtain multi-source heterogeneous data of the power grid and clean and fill in the gaps of the multi-source heterogeneous data of the power grid; For the processed data, association relationships and indexes are constructed using association keys, and a directed graph knowledge base is constructed based on the association relationships and indexes; wherein the association keys include device IDs, timestamps, and substation or line codes; and based on directed path information and historical fault information between nodes in the directed graph knowledge base, a topological coupling coefficient between nodes is determined, wherein the topological coupling coefficient is used to quantify the coupling relationship between nodes; An early warning module is used to construct a risk knowledge graph based on the directed graph knowledge base, wherein, in the risk knowledge graph, vertices represent entities, directed edges represent relationships between entities, and weights of directed edges represent risk dependency coefficients between entities; the entities include users, maintenance tasks, and equipment, and the risk dependency coefficients are determined based on the topological coupling coefficients and historical data to characterize the impact between entities; risk rules are set to issue early warnings for risk events and risk levels when repairing target equipment; wherein the risk rules include entity status information, risk dependency relationship information between entities, and risk indicators, wherein the entity status information includes information on the target equipment to be repaired, the risk dependency relationship information includes the risk dependency coefficient, and the risk indicators include conditions for triggering corresponding risk events and risk levels; The early warning order generation and intelligent analysis module is used for the maintenance of target equipment, determines the comprehensive risk characterization index based on equipment risk, user impact and load transfer capacity, and determines the risk warning level and generates an early warning order based on the comprehensive risk characterization index.
10. An intelligent early warning and analysis device for power grid risks, characterized in that: The device comprises a processor and a memory, wherein the processor is connected to the memory: The processor is configured to call and execute the program stored in the memory; The memory is used to store the program, and the program is at least used to execute the power grid risk intelligent early warning and analysis method according to any one of claims 1-8.
Citation Information
Cited By
Laboratory risk assessment method and device and computer equipment
CN121434659A
Intelligent management method and system for power grid regulation and control service
CN121581679A
Power supply service digital graph linkage monitoring, control and early warning method
CN122048029A
A power service number map linkage monitoring, management and control early warning method
CN122048029B
