Risk monitoring method, device and equipment for power grid and storage medium

Abstract

The invention discloses a risk monitoring method and device for a power grid, equipment and a storage medium. The method comprises the following steps: acquiring initial safety production data of the power grid acquired by a plurality of different system platforms in real time; performing data cleaning on each piece of initial safety production data; constructing a data view according to the initial safety production data after data cleaning; performing structured processing on the data view to obtain a knowledge structure graph; the knowledge structure graph comprises each node and each edge; the operation state of the power grid is monitored in real time, and when abnormity is monitored, problem tracing is carried out based on the knowledge structure graph so as to determine abnormal nodes and edges; and establishing an abnormal data link according to the abnormal nodes and edges, and analyzing the abnormal data link to obtain a risk monitoring result of the power grid. According to the method, the multi-source data is acquired, so that the risk monitoring is more sufficient, the data association degree can be enhanced by constructing the knowledge structure graph, the risk monitoring is comprehensively and quickly performed, and the normal operation of a power grid is facilitated.

Description

Technical Field

[0001] The present application relates to the technical field of power grid risk monitoring, and in particular to a method, device, equipment and storage medium for power grid risk monitoring. Background Art

[0002] With the rapid development of the power industry and the deepening of globalization, the challenges faced by power safety production are becoming increasingly complex. These challenges not only come from traditional factors such as equipment aging, human errors, and natural disasters, but may also be exacerbated by the application of new technologies, management omissions, etc. As the cornerstone of the operation of modern society, the stability and reliability of the power grid are directly related to the healthy development of the national economy and the harmony and stability of society. Once a safety accident occurs, it will cause huge economic losses. Therefore, in-depth research on risk management of power safety production, especially risk situation perception and pre-control, is of great significance to improving the safety and reliability of the power grid.

[0003] Traditional risk monitoring methods are relatively limited, usually only conducting single-dimensional analysis on a single data source. Therefore, risk identification is not comprehensive, affecting the accuracy of monitoring and being detrimental to the normal operation of the power grid. Summary of the invention

[0004] In view of this, the present application provides a risk monitoring method, device, equipment and storage medium for a power grid, which is used to solve the problem that traditional risk monitoring methods are relatively limited and usually only perform a single-dimensional analysis on a single data source. Therefore, risk identification is not comprehensive, affecting the accuracy of monitoring and being detrimental to the normal operation of the power grid.

[0005] To achieve the above objectives, the proposed solution is as follows:

[0006] In a first aspect, a risk monitoring method for a power grid includes:

[0007] Real-time acquisition of various initial safety production data collected by multiple different system platforms on the power grid;

[0008] Carry out data cleaning on the initial safety production data mentioned in each article respectively;

[0009] Construct data views from initial safety production data after each piece of data is cleaned;

[0010] Structuring the data view to obtain a knowledge structure graph; the knowledge structure graph includes each node and each edge;

[0011] Monitor the operation status of the power grid in real time. When the operation status of the power grid is detected to be abnormal, trace the source of the problem based on the knowledge structure graph to determine the abnormal nodes and edges.

[0012] An abnormal data chain is established by the abnormal nodes and edges, and the abnormal data chain is analyzed to obtain a risk monitoring result of the power grid.

[0013] Preferably, each piece of the initial production safety data is cleaned to obtain each piece of first production safety data, including:

[0014] For each piece of the initial production safety data, determining the missing sub-data in the initial production safety data that meets the missing condition;

[0015] Obtaining each parameter type corresponding to the missing sub-data;

[0016] Selecting a corresponding filling strategy from a preset filling strategy set according to each of the parameter types;

[0017] The missing sub-data is filled with data using the filling strategy.

[0018] Preferably, data cleaning is performed on each of the initial safety production data, including:

[0019] For each piece of the initial production safety data, determining each piece of equipment operation status sub-data and operating environment sub-data in the initial production safety data;

[0020] Use the pre-trained multi-dimensional anomaly detection model to process each piece of equipment operation status sub-data and operating environment sub-data to obtain the anomaly value of the initial safety production data;

[0021] If the abnormal value is not less than a preset first threshold, the piece of initial security data is deleted;

[0022] If the abnormal value is not less than the preset second threshold value and is less than the first threshold value, the initial safety data is corrected.

[0023] Preferably, the multi-dimensional anomaly detection model includes a data splitting module, a first feature extraction module, a second feature extraction module, a first abnormal data determination module, a second abnormal data determination module, a misjudgment detection module, and an abnormality assessment module;

[0024] Wherein, the input end of the data splitting module is used as the input end of the multi-dimensional anomaly detection model;

[0025] The input end of the first feature extraction module and the input end of the second feature extraction module are both connected to the output end of the data splitting module, the output end of the first feature extraction module is connected to the input end of the first abnormal data determination module, the output end of the second feature extraction module is connected to the input end of the second abnormal data determination module, the output end of the first abnormal data determination module and the output end of the second abnormal data determination module are both connected to the input end of the misjudgment detection module, and the output end of the misjudgment detection module is connected to the input end of the abnormality assessment module;

[0026] The output end of the anomaly assessment module serves as the output end of the multi-dimensional anomaly detection model.

[0027] Preferably, the data view is constructed from the initial safety production data after cleaning each piece of data, including:

[0028] The initial safety production data after data cleaning is used as the first safety production data, and the data type, storage format and physical storage location of each piece of the first safety production data are determined respectively;

[0029] Establishing a corresponding relationship between each piece of the first production safety data according to the data type, storage format and physical storage location;

[0030] Establishing a unified naming method and data format, and performing data standardization processing on each of the first production safety data according to the naming method and data format to obtain each of the second production safety data;

[0031] Constructing a logical data model, and associating each of the second production safety data in the logical data model according to the corresponding relationship;

[0032] Based on the association rule of the logical data model, the associated pieces of second production safety data are fused through the virtualization layer to form a data view.

[0033] Preferably, the step of performing structural processing on the data view to obtain a knowledge structure graph includes:

[0034] According to the preset entity type, determine each core entity data in the data view, and determine the attribute of each core entity data, and the corresponding relationship between each core entity data;

[0035] Taking each of the core entity data as each node;

[0036] Connecting the nodes according to the corresponding relationship to form edges;

[0037] By using the attributes, each node and each edge is labeled to obtain a knowledge structure graph.

[0038] Preferably, it also includes:

[0039] Establishing solutions based on the risk monitoring results, applying the solutions to the power grid, and performing real-time verification to determine verification results;

[0040] If the verification result is abnormal, determining each key execution node in the abnormal data chain;

[0041] Determine, from the knowledge structure graph, a characteristic data chain formed by each of the key execution nodes;

[0042] Determine whether the characteristic data chain is abnormal;

[0043] If so, the solution is rectified according to the characteristic data chain.

[0044] In a second aspect, a risk monitoring device for a power grid includes:

[0045] The safety production data acquisition module is used to obtain in real time various initial safety production data collected by multiple different system platforms on the power grid;

[0046] A data cleaning module, used to clean the initial safety production data of each item;

[0047] A data view building module is used to build a data view from the initial safety production data after each piece of data is cleaned;

[0048] A structured processing module, used for performing structured processing on the data view to obtain a knowledge structure graph; the knowledge structure graph includes various nodes and various edges;

[0049] A problem tracing module is used to monitor the operation status of the power grid in real time. When the operation status of the power grid is detected to be abnormal, the problem is traced based on the knowledge structure graph to determine the abnormal nodes and edges;

[0050] The analysis module is used to establish an abnormal data chain based on the abnormal nodes and edges, and analyze the abnormal data chain to obtain a risk monitoring result of the power grid.

[0051] In a third aspect, a risk monitoring device for a power grid includes a memory and a processor;

[0052] The memory is used to store programs;

[0053] The processor is used to execute the program to implement each step of the risk monitoring method for the power grid as described in any one of the first aspects.

[0054] In a fourth aspect, a storage medium stores a computer program thereon, wherein when the computer program is executed by a processor, each step of the risk monitoring method for a power grid as described in any one of the first aspects is implemented.

[0055] It can be seen from the above technical scheme that the present application obtains in real time various initial production safety data collected by multiple different system platforms for the power grid; performs data cleaning on each of the initial production safety data respectively; constructs a data view from the initial production safety data after data cleaning; performs structured processing on the data view to obtain a knowledge structure graph; the knowledge structure graph contains various nodes and edges; monitors the operating status of the power grid in real time, and when it is detected that the operating status of the power grid is abnormal, traces the problem based on the knowledge structure graph to determine the abnormal nodes and edges; establishes an abnormal data chain from the abnormal nodes and edges, and analyzes the abnormal data chain to obtain the risk monitoring result of the power grid. This application obtains the initial safe production data of the power grid collected by multiple different system platforms. Multi-source data can make risk monitoring more comprehensive, and data cleaning of the initial safe production data can make the data more accurate. When analyzing the data, a data view is constructed from the data, and a structured representation is performed to obtain a knowledge structure graph, which can enhance the correlation between the data. The data correspondence can be analyzed more efficiently based on the connection or position of the nodes and edges. At the same time, the operating status of the power grid is monitored in real time. The location of the anomaly can be determined based on the knowledge structure graph, and then an abnormal data chain can be established. The risk monitoring results can be analyzed comprehensively and quickly, and the power grid maintenance can be carried out based on the risk monitoring results, which is conducive to the normal operation of the power grid. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0057] Figure 1 An optional flow chart of a risk monitoring method for a power grid provided in an embodiment of the present application;

[0058] Figure 2 A schematic diagram of the structure of a multi-dimensional anomaly detection model provided in an embodiment of the present application;

[0059] Figure 3 An optional flow chart of another risk monitoring method for a power grid provided in an embodiment of the present application;

[0060] Figure 4A schematic diagram of the structure of a risk monitoring device for a power grid provided in an embodiment of the present application;

[0061] Figure 5 A schematic diagram of the structure of a risk monitoring device for a power grid provided in an embodiment of the present application. DETAILED DESCRIPTION

[0062] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0063] In recent years, domestic and foreign scholars and institutions have carried out extensive research in the field of power production safety risk management, striving to improve the safety management level of power production through advanced theories and methods. In addition, domestic research also pays special attention to the combination of theory and practice. Through a large number of empirical studies and case analyses, valuable experience has been accumulated, providing a solid theoretical foundation and technical support for the safety management and risk control of power companies.

[0064] The embodiment of the present invention provides a method for monitoring the risk of a power grid. The method can be applied to various computer terminals or smart terminals. The execution subject can be a processor or a server of the computer terminal or the smart terminal. The method flow chart of the method is as follows: Figure 1 As shown, specifically including:

[0065] S1: Real-time acquisition of various initial safety production data collected from the power grid by multiple different system platforms.

[0066] In order to ensure the comprehensiveness and adequacy of data and overcome the defect of single data in existing monitoring methods, this application obtains data from multiple data sources and aggregates the power grid safety production data collected by multiple different system platforms. This can ensure that certain uncommon data will not be missed and improve the accuracy of risk monitoring.

[0067] The system platform can be a power grid management platform, a smart safety monitoring system, a SCADA system, a DMS (distribution management system), an AMI system (advanced metering infrastructure), etc. Although these platforms can obtain the safe production data of the power grid and there will be duplicate data, some data cannot be collected by certain system platforms. For example, the SCADA system can collect real-time operation data of the power grid, such as the voltage, current, power, frequency, etc. of the substation, but it cannot collect data on the user side, such as detailed electricity consumption data and real-time electricity consumption behavior of the user terminal. The AMI system can accurately collect data on the user side, making the data more comprehensive.

[0068] Initial production safety data may include equipment ledgers, equipment operating status, operation plans, operation processes, personnel qualifications, equipment defect records, power outage data, project progress and settlement, operation violation records, maintenance records, accident reports and other data in multiple dimensions. These data have the characteristics of multiple data sources, large quantity, wide coverage and dispersion, and can all be subjected to secondary analysis.

[0069] One or more pieces of initial safety production data can be collected from each system platform.

[0070] S2: Clean the initial production safety data mentioned in each item respectively.

[0071] Big data processing technology can be used to comprehensively clean all initial safety production data, including but not limited to solving data missing, duplication, anomaly and other problems, which can significantly improve the accuracy, completeness and consistency of data. Cleaning data before actually entering risk monitoring can avoid subsequent rework or correction due to data quality issues, saving time and resources.

[0072] S3: Build a data view from the initial safety production data after each piece of data is cleaned.

[0073] The cleaned initial production safety data can be divided according to core elements, such as equipment, operation risks, etc., and detailed definitions and classifications can be carried out to clarify the attributes and associations of each core element, such as equipment model, installation location, maintenance records, operation risk level, operator qualifications, operating environment conditions, etc. Then, the power grid can be associated with the equipment through the equipment ID, the equipment and the operation through the operation ID, and the operation and the risk through the risk level, thus forming a data view with clear logic and consistent data, which can provide clear data support for the risk monitoring of the power grid.

[0074] You can choose to use SQL or visualization tools to define the logic of the data view. The data view can display the real-time status of the equipment, allowing staff to quickly locate problems. This not only simplifies the data access process, but also improves data security and availability.

[0075] S4: Structural processing is performed on the data view to obtain a knowledge structure graph; the knowledge structure graph includes each node and each edge.

[0076] Structural graph is a graphical tool used to describe and represent the relationship between elements in a complex system. It visualizes the structural information in the system in the form of nodes (representing entities or concepts) and edges (representing relationships or connections), making it easier for machines to understand and process and improve monitoring efficiency.

[0077] The nodes in the knowledge structure map can include multiple types: topological nodes, which represent nodes of electrical connection relationships, mainly including various types of equipment; physical nodes, which represent real entities in the physical world, such as main transformers, tools, equipment, operators, etc.; logical nodes, which represent other logical entities other than real entities, such as work teams, work qualifications, work plans, etc.; time attribute nodes, including summary nodes, detail nodes and business process nodes. Summary nodes represent the summary of records of process entities and display the current status, such as tool inspection summary, equipment power outage summary, etc. Detail nodes represent the storage of record details of process entities, such as tool inspection records, equipment power outage application forms, fault power outage records, etc. Business process nodes represent nodes formed by separating some attributes from physical nodes and logical nodes according to business processes, such as defect discovery, defect grading, and defect elimination arrangements.

[0078] The edges between nodes may include ordinary edges, which directly reflect the clear relationship between two nodes, such as the relationship between equipment and planned operations, and the relationship between a team and a work plan; and data inspection calculation edges, which represent data anomaly relationships dynamically generated through data recommendation algorithms or rules.

[0079] S5: Monitor the operation status of the power grid in real time. When the operation status of the power grid is detected to be abnormal, trace the problem based on the knowledge structure graph to determine the abnormal nodes and edges.

[0080] By constructing a knowledge structure graph, bidirectional coupling of the data chain and the business chain can be achieved. In other words, the knowledge structure graph establishes the association or logical relationship between the data of each node at the data level, and at the business level, the corresponding evaluation index system can be designed according to the actual process or operating status of the power grid production and operation, and covers multiple dimensions such as equipment health, operational compliance, and personnel skills. By mapping these indicators to the nodes and edges on the knowledge structure graph, the data chain and the business chain can be linked together to achieve comprehensive risk monitoring and evaluation of the power grid's safe production status at different time points.

[0081] When a problem occurs at a certain node, the source of the problem can be traced based on the knowledge structure map, which is both convenient and efficient.

[0082] S6: Establish an abnormal data chain based on the abnormal nodes and edges, and analyze the abnormal data chain to obtain a risk monitoring result of the power grid.

[0083] After determining the abnormal nodes and edges, an abnormal data chain can be established, so that the scope of abnormal risk can be evaluated, the root cause of abnormal risk can be identified, and the risk points and propagation paths in the power grid can be systematically monitored, which can improve the accuracy of risk monitoring. With the help of machine learning algorithms, the risk monitoring results can be deeply analyzed to discover potential safety hazards, predict future trends, and provide scientific decision-making basis for staff.

[0084] In the above scheme, the present application obtains the initial safe production data of the power grid collected by multiple different system platforms. Multi-source data can make risk monitoring more comprehensive, and data cleaning of the initial safe production data can make the data more accurate. When analyzing the data, a data view is constructed from the data, and a structured representation is performed to obtain a knowledge structure graph, which can enhance the correlation between the data. The data correspondence can be analyzed more efficiently based on the connection or position of the nodes and edges. At the same time, the operating status of the power grid is monitored in real time. The location of the anomaly can be determined based on the knowledge structure graph, and then an abnormal data chain can be established. The risk monitoring results can be comprehensively and quickly analyzed, and the power grid maintenance can be carried out based on the risk monitoring results, which is conducive to the normal operation of the power grid.

[0085] This application is applied to the user interaction interface to facilitate staff to conduct intuitive and real-time risk monitoring.

[0086] In the method provided in the embodiment of the present invention, each piece of the initial safety production data is cleansed separately to obtain a process of each piece of first safety production data, which may include multiple types of data cleaning, such as filling missing data, processing abnormal data, deleting duplicate data, etc., which are described in detail below.

[0087] 1) Fill in missing data.

[0088] Select a data filling algorithm to fill in missing values. For example, if a sub-data is missing in an initial safety production data, the position of the missing sub-data can be determined according to the order or position of the sub-data in the initial safety production data, and the known value of the previous sub-data can be used to fill it; if multiple consecutive missing sub-data are encountered, fill them in the same way as above.

[0089] Optionally, a dynamic filling method can also be used, that is, for each piece of the initial production safety data, determine the missing sub-data in the initial production safety data that meets the missing conditions; obtain each parameter type corresponding to the missing sub-data; select a corresponding filling strategy from a pre-set filling strategy set according to each parameter type; and use the filling strategy to fill the missing sub-data.

[0090] This can be more flexible and intelligent, and different filling strategies can be selected according to different situations. The filling strategy set is pre-constructed and optimized according to the characteristics of different system platforms and the parameter types corresponding to different data. For example, for data of equipment defect record type, when filling the corresponding missing sub-data, the filling strategy can be matched with parameter types such as equipment model and installation location, such as mean filling, nearest neighbor filling or model-based predictive filling. This can make the filling more accurate and overcome the problem of the relatively single existing filling technology.

[0091] 2) Handle abnormal data.

[0092] Anomaly detection algorithms can be used to detect anomalies in each piece of initial production safety data, mark the abnormal data, and process the abnormal data based on the results of anomaly detection, such as deletion, correction, replacement, etc.

[0093] However, in order to make anomaly detection more accurate, and considering that some data may be identified as abnormal on the surface, but in fact due to some special reasons, it is not truly abnormal data, it is necessary to identify it to prevent misjudgment and affect the overall data. Specifically:

[0094] For each piece of the initial production safety data, determining each piece of equipment operation status sub-data and operating environment sub-data in the initial production safety data;

[0095] Use the pre-trained multi-dimensional anomaly detection model to process each piece of equipment operation status sub-data and operating environment sub-data to obtain the anomaly value of the initial safety production data;

[0096] If the abnormal value is not less than a preset first threshold, the piece of initial security data is deleted;

[0097] If the abnormal value is not less than the preset second threshold value and is less than the first threshold value, the initial safety data is corrected.

[0098] Specifically, the equipment operation status sub-data can directly reflect the health status and working efficiency of the production equipment in the power grid, which is directly related to the failure and risk of the equipment. In addition, the equipment operation status sub-data is usually numerical data, which is easy to analyze and process. The multi-dimensional anomaly detection model can efficiently identify the anomalies therein; the operating environment sub-data can reflect the overall safety status of the power grid production environment, has multi-dimensional characteristics, and it can complement the equipment operation status sub-data, and can be combined for analysis to improve the accuracy of anomaly detection and reduce misjudgment of anomalies.

[0099] 3) Delete duplicate data.

[0100] By using unique identifiers to remove duplicate data, a unique identifier, such as UUID or primary key, can be assigned to each seed data in each piece of initial production safety data. This identifier remains unchanged during subsequent processing, and duplicate data can be identified and removed by checking the identifier.

[0101] Among them, the structure of the multi-dimensional anomaly detection model is as follows Figure 2 As shown, the model includes a data splitting module, a first feature extraction module, a second feature extraction module, a first abnormal data determination module, a second abnormal data determination module, a misjudgment detection module, and an abnormality assessment module;

[0102] Wherein, the input end of the data splitting module is used as the input end of the multi-dimensional anomaly detection model;

[0103] The input end of the first feature extraction module and the input end of the second feature extraction module are both connected to the output end of the data splitting module, the output end of the first feature extraction module is connected to the input end of the first abnormal data determination module, the output end of the second feature extraction module is connected to the input end of the second abnormal data determination module, the output end of the first abnormal data determination module and the output end of the second abnormal data determination module are both connected to the input end of the misjudgment detection module, and the output end of the misjudgment detection module is connected to the input end of the abnormality assessment module;

[0104] The output end of the anomaly assessment module serves as the output end of the multi-dimensional anomaly detection model.

[0105] Specifically, the first abnormal data determination module and the second abnormal data determination module can adopt a fusion of multiple algorithms, including the isolation forest algorithm, the LOF local isolation factor algorithm, the deep learning model, etc., so as to fully utilize the advantages of different algorithms and improve the accuracy and robustness of abnormal data determination.

[0106] The data splitting module can split each piece of initial production safety data into equipment operation status sub-data and operating environment sub-data;

[0107] The first feature extraction module can extract data features of the equipment operation status sub-data and transmit them to the first abnormal data determination module; the second feature extraction module can extract data features of the operating environment sub-data and transmit them to the second abnormal data determination module.

[0108] The first abnormal data determination module can determine the abnormal data therein according to the data characteristics of the equipment operation status sub-data as abnormal status data; the second abnormal data determination module can determine the abnormal data therein according to the data characteristics of the operating environment sub-data as abnormal environment data;

[0109] The misjudgment detection module can combine abnormal status data with non-abnormal operating environment sub-data, and combine abnormal environment data with non-abnormal equipment operation status sub-data to perform misjudgment detection, and output truly abnormal data to the abnormality assessment module. For example, the number of power outage data in the equipment operation status sub-data is abnormal status data, but by analyzing the corresponding non-abnormal equipment operation status sub-data, such as historical power outage patterns and environmental factors (such as weather, equipment load), it is determined whether there is a misjudgment; the abnormality assessment module is used to evaluate the abnormal values ​​of truly abnormal data.

[0110] Optionally, for the multi-dimensional anomaly detection model, a modular design can be set up to facilitate future expansion and maintenance, and a unified standardized interface can be defined to ensure smooth data exchange between different types of data.

[0111] The following is a detailed explanation of the steps in this application of constructing a data view from the initial production safety data after cleaning each piece of data.

[0112] The initial safety production data after data cleaning is used as the first safety production data, and the data type, storage format and physical storage location of each piece of the first safety production data are determined respectively;

[0113] Establishing a corresponding relationship between each piece of the first production safety data according to the data type, storage format and physical storage location;

[0114] Establishing a unified naming method and data format, and performing data standardization processing on each of the first production safety data according to the naming method and data format to obtain each of the second production safety data;

[0115] Constructing a logical data model, and associating each of the second production safety data in the logical data model according to the corresponding relationship;

[0116] Based on the association rule of the logical data model, the associated pieces of second production safety data are fused through the virtualization layer to form a data view.

[0117] Specifically, the data view finally formed by the above process is to integrate the initial safety production data after all data cleaning, and different data have different formats, characteristics, etc. Therefore, in order to facilitate integration, realize seamless connection between different data sources, and for stability after integration, unified standardization is required, in which the corresponding relationship needs to be determined, such as the corresponding relationship between equipment and operation, the corresponding relationship between personnel and qualifications, and the corresponding matching is performed according to the data type, storage format and physical storage location. The reasoning ability can also be used to automatically discover the implicit association in the data (such as the potential relationship between a certain type of equipment defect and a specific operation violation), further enhancing the analysis value of the data; to establish a unified naming method and data format, it is necessary to analyze the data format of each first safety production data, such as relational database, non-relational database, CSV format, etc., and analyze the structure, such as tables, documents, etc. When standardizing, type conversion and unit standardization can be performed through ETL tools or scripts, such as converting all data into a string format represented by UTF-8 encoding: $$ S{UTF-8}=S{original} ltimes F_{UTF-8} $$, which can ensure that all data maintain a unified framework.

[0118] A logical data model can be selected (such as a star model or a snowflake model). After the second production safety data is associated with the logical data model, data fusion can be performed through the virtualization layer according to the data access method (such as API, database connection, etc.) and permission requirements set by the logical data model. The set standardized data interface specifications (such as RESTful API, GraphQL) and data exchange protocols (such as JSON, XML) can also be followed to ensure that data collected by different system platforms can interact through a unified interface and protocol, thereby achieving real-time data integration and building a unified view without moving data. Standardized formats (such as RDF, OWL) can also be defined, and view building tools such as Neo4j and Apache Jena can be used to form a comprehensive, consistent, interconnected, and dynamically updated data view, thereby improving data availability and providing strong data support for power grid management, equipment maintenance, and risk control.

[0119] The data view constructed by data realizes the intelligent collection of data sources and the accurate output of data chains. For different business scenarios, it can realize the accurate output mechanism of data chains, which can quickly retrieve relevant data and assist staff in responding.

[0120] Optionally, the process of performing structured processing on the data view to obtain a knowledge structure graph includes:

[0121] According to the preset entity type, determine each core entity data in the data view, and determine the attribute of each core entity data, and the corresponding relationship between each core entity data;

[0122] Taking each of the core entity data as each node;

[0123] Connecting the nodes according to the corresponding relationship to form edges;

[0124] By using the attributes, each node and each edge is labeled to obtain a knowledge structure graph.

[0125] Specifically, it can make data retrieval and application more efficient, quickly locate and retrieve relevant information, and support more advanced data analysis and mining tasks. For example, by querying specific types of equipment failure cases through attribute label combinations, or analyzing the probability of human errors under specific operating conditions, potential safety hazards can be discovered in advance, improving the accuracy of risk perception and problem location.

[0126] For example, a topological node, which is a device type, can be connected to other nodes such as equipment ledgers, maintenance records, and fault reports. A physical node, which is a worker type, can be connected to nodes such as work plans, personnel qualifications, and on-site environments. Each node and edge can be labeled for subsequent screening and hierarchical display, including but not limited to node type, business field, importance level, etc., such as "high-voltage equipment", "regular inspections", and "special operations". These labels can not only make data retrieval and application more efficient, but also quickly locate and retrieve relevant information, and support more advanced data analysis and mining tasks, such as querying specific types of equipment failure cases through attribute label combinations, or analyzing the probability of human errors under specific operating conditions, so as to discover potential safety hazards in advance and improve the accuracy of risk perception and problem location. In addition, a complete knowledge structure map update and maintenance mechanism can be established to ensure the timeliness and accuracy of the data and reflect the latest safety production conditions and management requirements.

[0127] For the knowledge structure map, data verification can also be performed regularly to ensure data consistency and integrity and avoid data redundancy and errors.

[0128] Specifically, this method can perform data reverse deduction and realize reverse tracing from risk status to management problems. For example, when monitoring the operating status of the power grid in real time, when the power grid load in a certain area is detected to be abnormal, the equipment information, historical fault records and maintenance status corresponding to the relevant nodes and edges in the area can be automatically retrieved according to the knowledge structure graph to determine the risk monitoring results; data recording and analysis can also be performed. For example, if equipment failures frequently occur in a certain area of ​​the power grid, the equipment maintenance records, operation logs, environmental conditions and other data corresponding to the relevant nodes and edges can be analyzed according to the knowledge structure graph to find out the root cause of the failure (equipment aging, improper maintenance or operational errors, etc.) as the risk monitoring result. It can not only reveal the direct cause of the risk, but also discover potential indirect factors, providing a basis for comprehensive problem solving.

[0129] Furthermore, the method further comprises:

[0130] Establishing solutions based on the risk monitoring results, applying the solutions to the power grid, and performing real-time verification to determine verification results;

[0131] If the verification result is abnormal, determining each key execution node in the abnormal data chain;

[0132] Determine, from the knowledge structure graph, a characteristic data chain formed by each of the key execution nodes;

[0133] Determine whether the characteristic data chain is abnormal;

[0134] If so, the solution is rectified according to the characteristic data chain.

[0135] Specifically, the risk monitoring results will be used to establish solutions, which can be a hierarchical and graded risk prevention and control measures list, to provide accurate decision-making support for staff at different levels, to ensure that each level can obtain suggestions that match their responsibilities. For example, for senior managers, the solutions will focus on the adjustment of the overall strategy and the optimal allocation of resources. For middle-level managers, specific management process improvements and team training suggestions will be provided, while for front-line operators, detailed operating instructions and safe operating procedures will be provided. In order to improve the effectiveness and feasibility of the solutions, we can learn from the successful experience of similar typical cases in history and use them as a reference for the formulation of current solutions to ensure that the solutions are fully verified and optimized.

[0136] A closed-loop monitoring framework can be established to track the implementation and effectiveness of the solutions to ensure the effective implementation of risk prevention measures. Once deviations or poor results are found during the implementation of measures, early warnings can be issued in a timely manner, and adjustment suggestions can be provided to ensure that risk prevention measures can be continuously improved and optimized. For example, for equipment maintenance in the solution, key nodes such as equipment downtime, maintenance personnel presence, and maintenance tool preparation can be set, and data from these nodes can be collected in real time through sensors or information systems to ensure that each step is executed according to the process specified in the solution.

[0137] After rectification, the implementation can continue, and a comparison process can be set up to evaluate the effect of rectification by comparing the changes in risk indicators of the solution measures after rectification. For example, indicators such as equipment failure rate, number of violations of operation rules, and qualified rate of personnel safety training can be set up. Through comparative analysis of historical data and real-time data, it can be judged whether the solution measures of rectification have achieved the expected improvement effect. Furthermore, big data and artificial intelligence technologies can be used to realize online supervision of the implementation of rectification measures and online evaluation of their effectiveness. For example, for an equipment maintenance task, a continuous data chain can be established by the corresponding nodes from task issuance, personnel dispatch, tool preparation, on-site operation to task completion. By monitoring the data changes of these nodes, the progress of the implementation of rectification measures can be understood in real time, and deviations can be discovered and corrected in time; at the same time, based on these real-time data, the changes in various risk indicators will be automatically calculated, and an evaluation report will be generated to provide decision support for staff. In order to ensure the application and optimization of the closed-loop monitoring framework, a feedback optimization loop can be established, that is, after each rectification measure is completed, the data and effectiveness evaluation results of the execution process are automatically collected, and a comprehensive analysis is conducted to extract successful experiences and existing problems. In one example, if the equipment failure rate drops significantly after a corrective measure is implemented, it can be marked as "effective", otherwise it can be marked as "needs improvement". The effectiveness evaluation results can also be fed back into the knowledge structure map.

[0138] The risk monitoring results can be used to analyze some management situations, covering multiple scenarios, and can involve nodes, edges or labels corresponding to multiple data, such as:

[0139] The arrangement of maintenance operation mode is unreasonable: power outage application form, equipment ledger, customer complaint work order, fault ticket; equipment operation and maintenance management is not in place: risk assessment data (importance, health), equipment fault ticket, operation plan (regular inspection plan, maintenance plan, pre-test plan, maintenance plan, defect elimination plan, acceptance plan, inspection plan), equipment ledger; equipment defect management is not in place: spare parts, power outage application form, fault ticket, outsourced unit, equipment defects; operation process management is not in place: saturation, supervision plan, inspection personnel, in-place control records; operation plan management is not in place: operation plan (annual plan, monthly plan, weekly plan, daily plan), operation documents (work ticket, operation ticket, construction operation ticket), fault repair order; operation violation management is not in place: violations, operating personnel, work team, outsourced unit.

[0140] Among them, the solutions for the shortage of spare parts are: analyze the demand report of spare parts to ensure that the demand is reported in time; for defect management, the solutions can be: if defect handling is often delayed due to the lack of power outage windows, arrange inspection and maintenance plans in advance according to the power outage windows to avoid delays caused by insufficient power outage windows.

[0141] In an example, a general process could be as follows Figure 3 shown.

[0142] and Figure 1 Corresponding to the method described above, the embodiment of the present invention also provides a risk monitoring device for a power grid, which is used to monitor Figure 1 In the specific implementation of the method, the risk monitoring device for the power grid provided by the embodiment of the present invention can be used in a computer terminal or various mobile devices, combined with Figure 4 , introduce the risk monitoring device of the power grid, such as Figure 4 As shown, the device may include:

[0143] The safety production data acquisition module 10 is used to acquire in real time various pieces of initial safety production data collected from the power grid by multiple different system platforms;

[0144] A data cleaning module 20, used for cleaning each piece of the initial production safety data;

[0145] A data view building module 30, used to build a data view from the initial safety production data after each piece of data is cleaned;

[0146] A structured processing module 40 is used to perform structured processing on the data view to obtain a knowledge structure graph; the knowledge structure graph includes various nodes and edges;

[0147] A problem tracing module 50 is used to monitor the operation status of the power grid in real time. When the operation status of the power grid is detected to be abnormal, the problem is traced based on the knowledge structure graph to determine the abnormal nodes and edges;

[0148] The analysis module 60 is used to establish an abnormal data chain based on the abnormal nodes and edges, and analyze the abnormal data chain to obtain a risk monitoring result of the power grid.

[0149] It can be seen from the above technical scheme that the present application obtains the initial safe production data of the power grid collected by multiple different system platforms. Multi-source data can make risk monitoring more comprehensive, and data cleaning of the initial safe production data can make the data more accurate. When analyzing the data, a data view is constructed from the data, and a structured representation is performed to obtain a knowledge structure graph, which can enhance the correlation between the data. The data correspondence can be analyzed more efficiently according to the connection or position of the nodes and edges. At the same time, the operating status of the power grid can be monitored in real time, and the location of the anomaly can be determined based on the knowledge structure graph, and then an abnormal data chain can be established, which can comprehensively and quickly analyze the risk monitoring results. Subsequently, the power grid can be maintained according to the risk monitoring results, which is conducive to the normal operation of the power grid.

[0150] Furthermore, an embodiment of the present application provides a risk monitoring device for a power grid. Optionally, Figure 5 The hardware structure diagram of the risk monitoring equipment of the power grid is shown in FIG. Figure 5 The hardware structure of the risk monitoring device for a power grid may include: at least one processor 01 , at least one communication interface 02 , at least one memory 03 and at least one communication bus 04 .

[0151] In the embodiment of the present application, the number of the processor 01 , the communication interface 02 , the memory 03 , and the communication bus 04 is at least one, and the processor 01 , the communication interface 02 , and the memory 03 communicate with each other through the communication bus 04 .

[0152] The processor 01 may be a central processing unit (CPU), or an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present invention.

[0153] The memory 03 may include a high-speed RAM memory, and may also include a non-volatile memory (non-volatile memory), etc., such as at least one disk memory.

[0154] The memory stores a program, and the processor can call the program stored in the memory, and the program is used to execute the following power grid risk monitoring method, including:

[0155] Real-time acquisition of various initial safety production data collected by multiple different system platforms on the power grid;

[0156] Carry out data cleaning on the initial safety production data mentioned in each article respectively;

[0157] Construct data views from initial safety production data after each piece of data is cleaned;

[0158] Structuring the data view to obtain a knowledge structure graph; the knowledge structure graph includes each node and each edge;

[0159] Monitor the operation status of the power grid in real time. When the operation status of the power grid is detected to be abnormal, trace the source of the problem based on the knowledge structure graph to determine the abnormal nodes and edges.

[0160] An abnormal data chain is established by the abnormal nodes and edges, and the abnormal data chain is analyzed to obtain a risk monitoring result of the power grid.

[0161] Optionally, the detailed functions and extended functions of the program may refer to the description of the risk monitoring method for the power grid in the method embodiment.

[0162] The embodiment of the present application further provides a storage medium, which may store a program suitable for execution by a processor, and when the program is executed, controls the device where the storage medium is located to execute the following power grid risk monitoring method, including:

[0163] Real-time acquisition of various initial safety production data collected by multiple different system platforms on the power grid;

[0164] Carry out data cleaning on the initial safety production data mentioned in each article respectively;

[0165] Construct data views from initial safety production data after each piece of data is cleaned;

[0166] Structuring the data view to obtain a knowledge structure graph; the knowledge structure graph includes each node and each edge;

[0167] Monitor the operation status of the power grid in real time. When the operation status of the power grid is detected to be abnormal, trace the source of the problem based on the knowledge structure graph to determine the abnormal nodes and edges.

[0168] An abnormal data chain is established by the abnormal nodes and edges, and the abnormal data chain is analyzed to obtain a risk monitoring result of the power grid.

[0169] Specifically, the storage medium may be a computer-readable storage medium, and the computer-readable storage medium may be an electronic memory such as a flash memory, an EEPROM (Electrically Erasable Programmable Read-Only Memory), an EPROM, a hard disk or a ROM.

[0170] Optionally, the detailed functions and extended functions of the program may refer to the description of the risk monitoring method for the power grid in the method embodiment.

[0171] In addition, the functional modules in the various embodiments of the present disclosure can be integrated together to form an independent part, or each module can exist separately, or two or more modules can be integrated to form an independent part. If the function is implemented in the form of a software functional module and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present disclosure is essentially or the part that contributes to the prior art or the part of the technical solution can be embodied in the form of a software product, which is stored in a storage medium and includes a number of instructions for a computer device (which can be a personal computer, a live broadcast device, or a network device, etc.) to perform all or part of the steps of the methods of the various embodiments of the present disclosure.

[0172] Finally, it should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the presence of other identical elements in the process, method, article or device including the elements.

[0173] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0174] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for monitoring risk of a power grid, characterized in that: include: Real-time acquisition of various initial safety production data collected by multiple different system platforms on the power grid; Carry out data cleaning on the initial safety production data mentioned in each article respectively; Construct data views from initial safety production data after each piece of data is cleaned; Structuring the data view to obtain a knowledge structure graph; the knowledge structure graph includes each node and each edge; Monitor the operation status of the power grid in real time. When the operation status of the power grid is detected to be abnormal, trace the source of the problem based on the knowledge structure graph to determine the abnormal nodes and edges. An abnormal data chain is established by the abnormal nodes and edges, and the abnormal data chain is analyzed to obtain a risk monitoring result of the power grid.

2. The method according to claim 1, characterized in that: Each of the initial safety production data is cleaned to obtain each first safety production data, including: For each piece of the initial production safety data, determining the missing sub-data in the initial production safety data that meets the missing condition; Obtaining each parameter type corresponding to the missing sub-data; Selecting a corresponding filling strategy from a preset filling strategy set according to each of the parameter types; The missing sub-data is filled with data using the filling strategy.

3. The method according to claim 1, characterized in that The initial safety production data mentioned in each article shall be cleaned separately, including: For each piece of the initial production safety data, determining each piece of equipment operation status sub-data and operating environment sub-data in the initial production safety data; Use the pre-trained multi-dimensional anomaly detection model to process each piece of equipment operation status sub-data and operating environment sub-data to obtain the anomaly value of the initial safety production data; If the abnormal value is not less than a preset first threshold, the piece of initial security data is deleted; If the abnormal value is not less than the preset second threshold value and is less than the first threshold value, the initial safety data is corrected.

4. The method according to claim 3, characterized in that The multi-dimensional anomaly detection model includes a data splitting module, a first feature extraction module, a second feature extraction module, a first abnormal data determination module, a second abnormal data determination module, a misjudgment detection module, and an anomaly assessment module; Wherein, the input end of the data splitting module is used as the input end of the multi-dimensional anomaly detection model; The input end of the first feature extraction module and the input end of the second feature extraction module are both connected to the output end of the data splitting module, the output end of the first feature extraction module is connected to the input end of the first abnormal data determination module, the output end of the second feature extraction module is connected to the input end of the second abnormal data determination module, the output end of the first abnormal data determination module and the output end of the second abnormal data determination module are both connected to the input end of the misjudgment detection module, and the output end of the misjudgment detection module is connected to the input end of the abnormality assessment module; The output end of the anomaly assessment module serves as the output end of the multi-dimensional anomaly detection model.

5. The method according to claim 1, characterized in that: The data view is constructed from the initial safety production data after cleaning each piece of data, including: The initial safety production data after data cleaning is used as the first safety production data, and the data type, storage format and physical storage location of each piece of the first safety production data are determined respectively; Establishing a corresponding relationship between each piece of the first production safety data according to the data type, storage format and physical storage location; Establishing a unified naming method and data format, and performing data standardization processing on each of the first production safety data according to the naming method and data format to obtain each of the second production safety data; Constructing a logical data model, and associating each of the second production safety data in the logical data model according to the corresponding relationship; Based on the association rule of the logical data model, the associated pieces of second production safety data are fused through the virtualization layer to form a data view.

6. The method according to claim 1, characterized in that The structural processing of the data view to obtain a knowledge structure graph includes: According to the preset entity type, determine each core entity data in the data view, and determine the attribute of each core entity data, and the corresponding relationship between each core entity data; Taking each of the core entity data as each node; Connecting the nodes according to the corresponding relationship to form edges; By using the attributes, each node and each edge is labeled to obtain a knowledge structure graph.

7. The method according to any one of claims 1 to 6, characterized in that: Also includes: Establishing solutions based on the risk monitoring results, applying the solutions to the power grid, and performing real-time verification to determine verification results; If the verification result is abnormal, determining each key execution node in the abnormal data chain; Determine, from the knowledge structure graph, a characteristic data chain formed by each of the key execution nodes; Determine whether the characteristic data chain is abnormal; If so, the solution is rectified according to the characteristic data chain.

8. A risk monitoring device for a power grid, characterized in that: include: The safety production data acquisition module is used to obtain in real time various initial safety production data collected by multiple different system platforms on the power grid; A data cleaning module, used to clean the initial safety production data of each item; A data view building module is used to build a data view from the initial safety production data after each piece of data is cleaned; A structured processing module, used for performing structured processing on the data view to obtain a knowledge structure graph; the knowledge structure graph includes various nodes and various edges; A problem tracing module is used to monitor the operation status of the power grid in real time. When the operation status of the power grid is detected to be abnormal, the problem is traced based on the knowledge structure graph to determine the abnormal nodes and edges; The analysis module is used to establish an abnormal data chain based on the abnormal nodes and edges, and analyze the abnormal data chain to obtain a risk monitoring result of the power grid.

9. A risk monitoring device for a power grid, characterized in that: including memory and processor; The memory is used to store programs; The processor is used to execute the program to implement each step of the risk monitoring method for a power grid as described in any one of claims 1-7.

10. A storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, each step of the risk monitoring method for a power grid as described in any one of claims 1 to 7 is implemented.

Images