CN112685570B - Multi-label graph-based power grid network frame knowledge graph construction method - Google Patents

Multi-label graph-based power grid network frame knowledge graph construction method Download PDF

Info

Publication number
CN112685570B
CN112685570B CN202011477778.3A CN202011477778A CN112685570B CN 112685570 B CN112685570 B CN 112685570B CN 202011477778 A CN202011477778 A CN 202011477778A CN 112685570 B CN112685570 B CN 112685570B
Authority
CN
China
Prior art keywords
label
grid
power grid
line
nodes
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN202011477778.3A
Other languages
Chinese (zh)
Other versions
CN112685570A (en
Inventor
解凯
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
NR Electric Co Ltd
NR Engineering Co Ltd
Original Assignee
NR Electric Co Ltd
NR Engineering Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by NR Electric Co Ltd, NR Engineering Co Ltd filed Critical NR Electric Co Ltd
Priority to CN202011477778.3A priority Critical patent/CN112685570B/en
Publication of CN112685570A publication Critical patent/CN112685570A/en
Application granted granted Critical
Publication of CN112685570B publication Critical patent/CN112685570B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
    • Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
    • Y04S10/00—Systems supporting electrical power generation, transmission or distribution
    • Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications

Landscapes

  • Supply And Distribution Of Alternating Current (AREA)

Abstract

The invention discloses a method for constructing a power grid network frame knowledge graph based on a multi-label graph, which formally defines the graph as the multi-label graph: PGKG = { V, E, Σ V, Σ E, L }, which is formed by connecting grid entities, grid entity attributes, and binary or multivariate relations among the grid entities, where V represents a node in the graph and is a set of grid facility entities such as a substation, a box transformer, a distribution room, and a branch box at each voltage level; e ⊆ V × V represents an edge in the diagram, which is a collection of power grid line entities such as power transmission lines, cables, low-voltage distribution cables and the like of each voltage class; the sigma V represents a node label, and is a set of static and dynamic attributes of the power grid facility entity; Σ E represents an edge label, which is a set of static and dynamic attributes of a grid line entity; l is a label mapping function for mapping labels to nodes and edges, the label mapping of a node as V → Σ V maps a grid facility attribute label to a grid facility, and the label mapping of an edge as E → Σ E maps a grid line attribute label to a grid line.

Description

Multi-label graph-based power grid network frame knowledge graph construction method
Technical Field
The invention relates to a power grid network frame knowledge graph construction method based on a multi-label graph, and belongs to the field of power grid basic knowledge framework construction.
Background
The automatic informatization systems of the power grid adopt IEC-61970 and IEC-61850 models, so that the information types, dimensions and quantity brought by the current power grid informatization and automation level rapid promotion are remarkably increased, and the requirement of rapid expansion modeling cannot be met.
With the development of artificial intelligence technology, the intelligent application of the power grid must be powerfully supported. As an important direction of artificial intelligence, the knowledge map is a mapping map of the professional knowledge field, and is a series of different graphs for displaying the relationship between the entity development process and the structure of the professional field, and the knowledge resources and the carriers thereof are described by using a visualization technology, and the knowledge and the mutual relation among the knowledge resources, the carriers, the knowledge resources, the analysis, the construction, the drawing and the display are mined, constructed, drawn and displayed. The knowledge graph is a modeling method of a general entity relationship, and the expansibility based on graph database modeling is very favorable for building a model for intelligent application of a power grid.
At present, a part of methods for applying the knowledge map exist in power grid informatization, but the method mainly applies a knowledge analysis means focusing on text information of documents such as power grid detection, regulation, alarm and the like, and has certain difference from being used as a real intelligent power grid application bottom model, which is mainly shown in that:
firstly, the current application is limited to text-based knowledge extraction and knowledge construction of natural language, and a mesh entity of a power grid is not constructed as relational mapping of a knowledge graph;
secondly, entities of the power grid are primary power facilities and lines, the knowledge graph takes the primary power facilities and lines as a static structure relation model, but a real power flow model needs to combine and express the power grid entities and a large amount of real-time power grid measurement information mapped to attributes of the knowledge graph entities;
thirdly, for mass equipment and attributes of the power grid, the existing knowledge graph construction method brings mass nodes and relationship edges according to entity one-by-one modeling, so that the operation performance of the knowledge graph is reduced, and an ontology and an instantiation must be defined reasonably;
and fourthly, various power grid informatization intelligent applications need to provide a general analysis method for power grid flow based on the combination of entity relations and entity attributes of a power grid network frame knowledge graph, such as electric island analysis of flow communication.
Disclosure of Invention
The technical problem to be solved by the invention is as follows: the method comprises the steps of taking a mesh entity of a power grid as relational mapping construction of the knowledge graph, and mapping real-time measurement information of the power grid to the label attribute of the entity in the knowledge graph.
The invention adopts the following technical scheme for solving the technical problems:
a power grid network frame knowledge graph construction method based on a multi-label graph defines the power grid network frame knowledge graph as a multi-label graph, and comprises the following steps: PGKG ═ V, E, Σ V, Σ E, L }, where: PG (PG)KG represents a grid network frame knowledge graph; v represents a node in the map and is a set of all power grid facility entities, wherein the power grid facility entities comprise all voltage-level transformer substations, box transformer substations, distribution rooms and branch boxes; e represents an edge in the graph and represents a side,
Figure BDA0002836165290000021
the method is a set of all power grid line entities, wherein the power grid line entities comprise power transmission lines, cables and low-voltage power distribution cables of all voltage classes; the sigma V represents a node label and is a set of static attributes and dynamic attributes of all power grid facility entities; Σ E represents an edge label, which is a set of static attributes and dynamic attributes of all grid line entities; l represents a label mapping function for mapping labels to nodes or edges, including label mapping of nodes and label mapping of edges, the label mapping of nodes is V → Σ V, i.e. mapping grid facility entity attribute labels to grid facility entities, and the label mapping of edges is E → Σ E, i.e. mapping grid line entity attribute labels to grid line entities.
As a preferred scheme of the method, the construction method comprises the step of defining the binary relation between a power grid facility entity and a power grid line entity in the power grid, namely constructing a direct power supply relation with one edge connecting two nodes in a power grid network frame knowledge graph, and formally defining the direct power supply relation as R2(V1，V2，E1-2) From two nodes V of binary relation1、V2And an edge E between nodes1-2Meaning that a binary relation is defined over the set V E, i.e.
Figure BDA0002836165290000022
As a preferable scheme of the method, the construction method comprises the step of defining the multivariate relation between a power grid facility entity and a power grid line entity in the power grid, wherein the multivariate relation is composed of a plurality of nodes and edges which are related to each other, namely the complex power supply relation with a plurality of edges connecting a plurality of nodes is constructed in a power grid structure knowledge graph, and the formal definition is Rnm(V1…n，E1…m，ΣnV，ΣmE) Byn number of nodes V1…nAnd an edge E between m nodes1…mRepresenting, by the label Σ of n nodes simultaneouslynLabel Σ for an edge between V and m nodesmAnd E auxiliary representation.
As a preferred solution of the method of the present invention, the label of the node is mapped as V → Σ V, that is, the grid facility entity attribute label is mapped to the grid facility entity, and the label domain of Σ V includes: the method comprises the following steps of carrying out expansion according to requirements on the name of a power grid facility, the type of the power grid facility, the number of transformers, the total capacity of the transformers, the highest voltage level, the total active sum, the total reactive sum, the load rate, the commissioning time and the position coordinate, wherein the total active sum, the total reactive sum and the load rate are real-time data mapped from a power grid real-time measurement database.
As a preferred solution of the method of the present invention, the label of the edge is mapped to E → Σ E, that is, the label domain of Σ E includes: the system comprises a line name, a line type, a line rated capacity, a line voltage grade, a line active state, a line reactive state, a line current, a line load rate, a commissioning time, a line moving position coordinate, a starting station bus, a starting station line switch state, an end station bus, an end station line switch state and expansion according to needs, wherein the line active state, the line reactive state, the line current, the line load rate, the starting station line switch state and the end station switch state are real-time data mapped from a power grid real-time measurement database.
The electric island is generated by calculation based on a power grid network frame knowledge graph, the power grid network frame knowledge graph is constructed based on the construction method of the power grid network frame knowledge graph based on the multi-label graph, and the electric island is characterized in that the electric island G '{ V', E ', Sigma V', Sigma E ', L' } for power grid flow communication is a subgraph of the power grid network frame knowledge graph, namely the subgraph
Figure BDA0002836165290000031
Figure BDA0002836165290000032
The generation method is that a plurality of direct power supply binary relations R are used as judgment conditions according to real-time measurement data of power grid facility entities and power grid line entities in the labels2Multiple relation R combined into complex power supplynmThe method comprises the following steps:
step 1, from the selected investigation node VxAt the beginning, for all the contained VxBinary relation R of nodesx 2Sequentially investigating, setting and selecting a relation RxyReach and VxV with direct binary relationyA node;
step 2, examine Vx→VySide E ofxyAccording to ExyTag Σ E ofxyIf the line current attribute of (2) is not zero, then V is representedxNode and VyThe nodes have a power grid flow communication relation;
step 3, mixing Vx，Vy，Exy，ΣVx，ΣVy，ΣExyAdding G';
step 4, examine VxNext binary relation R of nodesx 2Until V is adjustedxTraversing all binary relations of the nodes; marking of V in GxCompleting the investigation;
and 5, selecting nodes which are not marked and inspected in the G ', repeating the steps 1-4 until all the nodes in the G' are marked and inspected and stopped, wherein the G 'is the electric island with the communicated power flow of the power grid, V' represents the nodes in the electric island, E 'represents the edge in the electric island, Sigma V' represents a node label in the electric island, Sigma E 'represents an edge label in the electric island, and L' represents a label mapping function in the electric island.
Compared with the prior art, the invention adopting the technical scheme has the following technical effects:
the invention adopts a multi-label graph mode to construct a power grid network frame knowledge graph, takes a mesh entity of a power grid as the relational mapping construction of the knowledge graph, maps the real-time measurement information of the power grid to the label attribute of the entity in the knowledge graph, and provides a joint judgment algorithm of the entity relation and the entity real-time attribute of the power grid network frame knowledge graph, thereby providing a universal analysis method of the real-time power flow range of the power grid for various power grid information intelligent applications.
Drawings
Fig. 1 is the ontology of the grid network frame knowledge graph of the present invention.
Fig. 2 is a binary relationship of the grid framework knowledge graph of the present invention.
Fig. 3 is a multivariate relation of the grid network frame knowledge graph of the invention.
Fig. 4 is an instantiation method of the grid network frame knowledge map of the invention.
Fig. 5 is a flow chart of an electrical island for judging power flow communication of a power grid based on the power grid network frame knowledge map.
Fig. 6 shows that the power grid network frame knowledge graph expansion based on the invention is applied to an electric power guarantee informatization system.
Detailed Description
Reference will now be made in detail to embodiments of the present invention, examples of which are illustrated in the accompanying drawings. The embodiments described below with reference to the accompanying drawings are exemplary only for explaining the present invention and are not construed as limiting the present invention.
Power grid network frame knowledge graph body constructed based on multi-label graph
As shown in fig. 1, a Power Gird Knowledge Graph (Power Gird Knowledge Graph) is represented by a multi-label Graph, which can be formally defined as follows: PGKG ═ V, E, Σ V, Σ E, L }, and is a binary relationship R between grid entities, grid entity attributes, and grid entities2Or the multivariate relation RnmA connection formation, wherein: v represents nodes in the graph, and is a set of power grid facility entities such as transformer substations, box substations, distribution rooms, branch boxes and the like in each voltage class;
Figure BDA0002836165290000041
the edges in the representation are the set of power grid line entities such as transmission lines, cables, low-voltage distribution cables and the like of all voltage classes; the Sigma V represents a node label, and refers to a set of static and dynamic attributes of facilities such as transformer substations, box transformers, distribution rooms and branch boxes in all voltage classes; Σ E represents the edge label and,the method is a set of static and dynamic attributes of lines such as transmission lines, cables and low-voltage distribution cables of all voltage classes; l is a label mapping function for mapping labels to nodes and edges, the label mapping of a node as V → Σ V maps grid facility attribute labels to grid facilities (nodes), and the label mapping of an edge as E → Σ E maps grid line attribute labels to grid lines (edges).
As shown in fig. 2, the main construction method is to define a binary relationship between substation facilities and grid lines in the grid, that is, a power supply relationship between two (substation facility) nodes connected to a (line) edge is constructed in a grid rack knowledge graph, and may be formally defined as R2(V1，V2，E1-2) Represented by two nodes of a binary relationship (substation facilities) and a relationship edge (line) between the nodes, the binary relationship being defined on the set V E, i.e.
Figure BDA0002836165290000051
As shown in fig. 3, the second main construction method is to define the multivariate relationship between the substation facilities and the grid lines in the grid, and the multivariate relationship is composed of a plurality of nodes and edges which are related to each other, that is, a complex power supply relationship in which a plurality of (lines) are connected to a plurality of (substation facility) nodes is constructed in the grid rack knowledge graph, and can be formally defined as Rnm(V1.。n，E1.。m，ΣnV，ΣmE) The method is represented by relation edges (lines) among n nodes (substation facilities) and m nodes, and is represented by labels of the n nodes (substation facilities) and labels of the relation edges (lines) among the m nodes in an auxiliary mode.
As shown in fig. 1, the label mapping of the node is V → Σ V maps the grid facility attribute label to the facility node of the grid station, and the label domain of Σ V includes { the station installation name, the station installation type, the number of transformers, the total capacity of the transformer, the highest voltage level, the total active sum, the total reactive sum, the load factor, the commissioning time, the location coordinate }, and can be expanded as needed, wherein the total active sum, the total reactive sum, and the load factor are real-time data mapped from the grid real-time measurement database.
As shown in fig. 1, the label mapping of the edge is E → Σ E, the grid line attribute label is mapped to the edge of the grid line, and the label domain of Σ E includes { line name, line type, line rated capacity, line voltage class, line active, line reactive, line current, line load rate, commissioning time, line traveling position coordinate, bus of the origin station, line switch state of the origin station, bus of the end station, and line switch state of the end station }, and can be extended as needed, wherein the line active, line reactive, line current, line load rate, line switch state of the origin station, and line switch state of the end station are real-time data mapped from the grid real-time measurement database.
Instantiation of power grid network frame knowledge map
The instantiation of the power grid network frame knowledge graph is based on a power grid architecture and a service mode in the power supply field, and the uniform expression and management of the full-element multi-source heterogeneous knowledge of the power grid are realized. As shown in fig. 4, the connection with the real-scene objective things of the power grid is realized by using various field methods from structured, semi-structured and unstructured data and by using information such as power grid network frame, geographic space, power grid equipment ledger, state measurement, alarm, video, audio and picture.
1. Structured data-oriented power conservation knowledge extraction
The grid network frame knowledge is derived from data supporting a grid monitoring service system, such as transformer substations, lines, equipment asset information, association information and the like, and extraction of the knowledge from structured data such as a related database is an important method. The method is to realize the conversion from the structured data to the knowledge through an R2RML mapping language, and the language is a custom mapping language for expressing a data set from a relational database to a uniform Resource Description Framework (RDF). This mapping provides the ability to view existing relational data under the RDF data model, and may represent the original relational data based on user-defined structures and target vocabulary. Converting database data into instantiation data of an ontology, wherein the conversion process comprises generation of Uniform Resource Identifiers (URIs), definition of RDF classes and attributes, processing of empty nodes, expression of association relation among data and the like, and the basic conversion rule comprises the following steps:
(1) mapping tables in the database into RDF classes;
(2) mapping the column of the table in the database into RDF attribute;
(3) mapping each row in a database table into a resource/entity, and creating a URI of the resource/entity;
(4) mapping each cell value in a database table into a character value;
(5) if the value of the cell corresponds to a foreign key, it is replaced with the URI of the resource or entity to which the foreign key points.
2. Power-conserving knowledge extraction oriented to semi-structured data
And extracting relevant information and association thereof aiming at important semi-structured code data of the geographic space of a power grid architecture, a power grid equipment ledger, state measurement, alarm and the like to form an entity label.
The semi-structured power grid entity attribute data has better data quality and is an important data source for knowledge extraction, and although the semi-structured power grid entity attribute data does not conform to a relational database or other forms of data table form structures, labels or other marks are included to separate semantic elements and maintain the hierarchy of records and data fields. For semi-structured data in static and dynamic attributes of a power grid architecture, a D2R tool is used as a tool capable of converting contents in a relational database into RDF triples, data analysis is directly carried out according to semi-structured syntax, rules similar to structured data knowledge extraction are constructed, and related knowledge is extracted and mapped to tags of entities.
3. Non-structural data oriented power protection entity identification
The knowledge extraction is carried out on a large amount of unstructured grid data such as emergency plans, fault descriptions, operation rules, overhaul standards, work tickets and scheduling logs, so that the power grid knowledge map can be further enriched. The knowledge extraction comprises two steps of entity identification and relation extraction. The power grid network frame knowledge graph constructed by the invention can be used as a basic graph to assist entity recognition from related texts so as to obtain a better power grid field named entity recognition effect.
In the pre-training process of introduction and natural language processing, features obtained through deep learning are used as input of a sequence labeling learning model. The named entity recognition and knowledge graph construction are regarded as a bootstrap iteration BootStrap ping process, the entity recognition process is supervised and learned by utilizing the existing constructed power grid network frame knowledge graph to generate a training data set, and in the knowledge extraction process, the constructed knowledge graph is used again to perform iterative supervision to generate test data, so that the problem of constructing the data set in the power grid field is solved, and the result quality of named entity recognition is further improved.
Third, basic application of power grid network frame knowledge graph
The general analysis method for the power grid flow is carried out on the basis of the entity relationship and the entity attribute combination of the power grid network frame knowledge graph, and algorithm services such as electric island analysis and the like of flow communication can be provided for various power grid information-based intelligent applications. The specific algorithm shown in fig. 5 is as follows:
the electric island for judging the power flow communication of the power grid is a subgraph of a group of power grid stations connected by lines, i.e. the subgraph
Figure BDA0002836165290000071
Figure BDA0002836165290000072
The generation method is that a plurality of binary relations R are used as judgment conditions according to the measured real-time attribute of the power grid entity in the label2Are merged into a multivariate relation Rnm：
(1) From the selected survey node VxAt the beginning, according to all the contained VxBinary relation R of nodesx 2For sequential investigation, e.g. selecting a relation RxyReach and VxV with direct binary relationyA node;
(2) investigation of Vx→VySide E ofxyAccording to ExyTag Σ E ofxyLine current property ofIf not zero, then V is representedxNode and VyThe nodes have a power grid flow communication relation;
(3) will Vx，Vy，Exy，ΣVx，ΣVy，ΣExyAdding G';
(4) investigation of VxNext binary relation R of nodesx 2Until V is adjustedxTraversing all binary relations of the nodes; marking of V in GxCompleting the investigation;
(5) and (4) selecting nodes which are not marked and are inspected, and repeating the steps (1) - (4) until all the nodes in G 'are marked and inspected and stopped, wherein G' is an electric island set with power flow communication of a power grid.
The algorithm is a computing service of a power grid framework knowledge graph for various application basic electrified topological structures of a power grid, the knowledge graph is typical graph data in representation and conforms to graph data models such as an attribute graph or RDF (remote data format) triple group, the operation of the graph can be efficiently realized, and a storage management layer of the algorithm designs a special storage scheme for nodes, node attributes, edges, edge attributes and the like in the attribute graph structure, so that the access efficiency of the storage layer to the graph data is inherently superior to that of a relational database. By utilizing the calculation capability of the graph database of the knowledge graph, the calculation analysis of the power grid electric island can be rapidly provided.
Fourthly, expansion of power grid network frame knowledge map
The power grid network frame knowledge graph constructed by the invention can further expand the domain knowledge on the basis of the main entity and relationship of the network frame expressed by the power grid network frame knowledge graph to form a richer power grid domain knowledge graph, and the graphs can exist as subgraphs of the main graph.
In order to improve the query efficiency of various knowledge sub-fields, a general distributed extended storage and query method of graph data is adopted, and the basic idea is to introduce a graph segmentation and query decomposition strategy in the distributed storage and query processing of a large-scale knowledge graph. The graph is divided into a plurality of fragments, boundary nodes of each fragment are extended to n-hop (n-hop) neighbors, and the SPARQL query is divided into a plurality of subqueries for parallel evaluation. The nodes and the neighbors thereof are defined as node blocks, distributed storage is carried out on the node blocks by adopting heuristic rules, and meanwhile, inquiry is decomposed, so that the aims of increasing the parallelism and reducing the communication overhead are fulfilled. Thus, the core of the storage and query model is how the data set is partitioned. The processing method adopted here is to perform hash mapping on the data, and determine the storage partition of each data according to the corresponding hash value.
As shown in fig. 6, for example, a power grid company is deployed for power guarantee informatization application, and the extended application based on the power grid network frame knowledge graph includes:
(1) and (3) power protection information gathering and pushing: based on the knowledge graph, the power protection information is gathered and pushed; unified convergence representation of multi-source heterogeneous power protection information is achieved through a power protection knowledge graph, and query meeting specific requirements is achieved through a query interface (entity attribute relation) and a structured query language (sub-graph query). The method comprises entity information aggregation, situation information aggregation, information aggregation of a specific mode and the like.
(2) And (3) understanding and pushing a power-saving fault plan: and identifying the power grid fault brief report, and searching and pushing a power protection plan based on the understanding of the fault brief report. Matching the fault instance information with the fault in the concept knowledge graph, finding a plan and pushing the plan to the front end.
(3) And (4) reporting the power protection real-time condition: and (4) making a man-machine conversation model for the power protection system capability platform, combining the power protection knowledge graph and knowledge reasoning, completing semantic processing and automatically giving results to generate report contents within the service flow range.
(4) Knowledge search query service: and providing an active push service for searching and inquiring information according to the knowledge of the subject.
(5) Emergency plan recommendation service: and providing emergency plan recommendation service according to real-time power protection conditions.
The above embodiments are only for illustrating the technical idea of the present invention, and the technical idea of the present invention is not limited thereto, and any modifications made on the basis of the technical solution according to the technical idea of the present invention fall within the protective scope of the present invention.

Claims (6)

1. A construction method of a power grid network frame knowledge graph based on a multi-label graph is characterized in that the power grid network frame knowledge graph is defined as the multi-label graph, and the multi-label graph comprises the following steps: PGKG ═ V, E, Σ V, Σ E, L }, where: PGKG represents a power grid network frame knowledge graph; v represents a node in the map and is a set of all power grid facility entities, wherein the power grid facility entities comprise all voltage-level substations, box substations, distribution rooms and branch boxes; e represents an edge in the graph,
Figure FDA0002836165280000011
the method is a set of all power grid line entities, wherein the power grid line entities comprise power transmission lines, cables and low-voltage power distribution cables of all voltage classes; the Sigma V represents a node label, and is a set of static attributes and dynamic attributes of all power grid facility entities; the Σ E represents an edge label, which is a set of static attributes and dynamic attributes of all power grid line entities; l represents a label mapping function for mapping labels to nodes or edges, including label mapping of nodes and label mapping of edges, the label mapping of nodes is V → Σ V, i.e. mapping grid facility entity attribute labels to grid facility entities, and the label mapping of edges is E → Σ E, i.e. mapping grid line entity attribute labels to grid line entities.
2. The method for constructing the grid network frame knowledge graph based on the multi-label graph as claimed in claim 1, wherein the construction method comprises defining a binary relation between grid facility entities and grid line entities in a grid, namely constructing a direct power supply relation in which two nodes are connected by one edge in the grid network frame knowledge graph, and formally defining the direct power supply relation as R2(V1，V2，E1-2) From two nodes V of binary relation1、V2And an edge E between nodes1-2Meaning that a binary relation is defined over the set V E, i.e.
Figure FDA0002836165280000012
3. The method for constructing the grid network frame knowledge graph based on the multi-label graph as claimed in claim 1, wherein the construction method comprises defining a multivariate relationship between grid facility entities and grid line entities in a grid, and the multivariate relationship is composed of a plurality of nodes and edges which are related to each other, namely, a complex power supply relationship which is constructed in the grid network frame knowledge graph and formed by connecting a plurality of nodes by a plurality of edges is defined as Rnm(V1…n，E1…m，ΣnV，ΣmE) From n nodes V1…nAnd an edge E between m nodes1…mRepresenting, by the label Σ of n nodes simultaneouslynLabel Σ for an edge between V and m nodesmAnd E auxiliary representation.
4. The method for constructing the grid structure knowledge-graph based on the multi-label graph according to claim 1, wherein labels of the nodes are mapped to V → Σ V, that is, the grid facility entity attribute labels are mapped to the grid facility entities, and a label domain of Σ V includes: the method comprises the following steps of carrying out expansion according to requirements on the name of a power grid facility, the type of the power grid facility, the number of transformers, the total capacity of the transformers, the highest voltage level, the total active sum, the total reactive sum, the load rate, the commissioning time and the position coordinate, wherein the total active sum, the total reactive sum and the load rate are real-time data mapped from a power grid real-time measurement database.
5. The method for constructing the grid network frame knowledge graph based on the multi-label graph according to claim 1, wherein the label mapping of the edge is E → Σ E, that is, the grid line entity attribute label is mapped to the grid line entity, and the label domain of Σ E includes: the system comprises a line name, a line type, a line rated capacity, a line voltage grade, a line active state, a line reactive state, a line current, a line load rate, a commissioning time, a line moving position coordinate, a starting station bus, a starting station line switch state, an end station bus, an end station line switch state and expansion according to needs, wherein the line active state, the line reactive state, the line current, the line load rate, the starting station line switch state and the end station switch state are real-time data mapped from a power grid real-time measurement database.
6. The electrical island is generated by calculation based on a power grid network frame knowledge graph, the power grid network frame knowledge graph is constructed based on the construction method of the power grid network frame knowledge graph based on the multi-label graph as claimed in any one of claims 1 to 5, and the electrical island G '{ V', E ', Σ V', Σ E ', L' } is a subgraph of the power grid network frame knowledge graph, namely, a subgraph of the power grid network frame knowledge graph
Figure FDA0002836165280000021
Figure FDA0002836165280000022
The generation method is that a plurality of direct power supply binary relations R are used as judgment conditions according to real-time measurement data of power grid facility entities and power grid line entities in the labels2Multiple relation R combined into complex power supplynmThe method comprises the following steps:
step 1, from the selected investigation node VxAt the beginning, for all the contained VxBinary relation R of nodesx 2Sequentially investigating, setting and selecting a relation RxyReach and VxV with direct binary relationyA node;
step 2, examine Vx→VyEdge E ofxyAccording to ExyTag Σ E ofxyIf the line current attribute of (2) is not zero, then V is representedxNode and VyThe nodes have a power grid flow communication relation;
step 3, mixing Vx，Vy，Exy，ΣVx，ΣVy，ΣExyAdding G';
step 4, examine VxNext binary relation R of nodesx 2Until V is adjustedxTraversing all binary relations of the nodes; marking of V in GxCompletion of investigation；
And 5, selecting nodes which are not marked and inspected, repeating the steps 1-4 until all the nodes in G ' are marked and inspected, stopping inspection, wherein G ' is the electric island with the power grid flow communication, V ' represents the nodes in the electric island, E ' represents the edges in the electric island, Sigma V ' represents a node label in the electric island, Sigma E ' represents an edge label in the electric island, and L ' represents a label mapping function in the electric island.
CN202011477778.3A 2020-12-15 2020-12-15 Multi-label graph-based power grid network frame knowledge graph construction method Active CN112685570B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202011477778.3A CN112685570B (en) 2020-12-15 2020-12-15 Multi-label graph-based power grid network frame knowledge graph construction method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202011477778.3A CN112685570B (en) 2020-12-15 2020-12-15 Multi-label graph-based power grid network frame knowledge graph construction method

Publications (2)

Publication Number Publication Date
CN112685570A CN112685570A (en) 2021-04-20
CN112685570B true CN112685570B (en) 2022-07-22

Family

ID=75447959

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202011477778.3A Active CN112685570B (en) 2020-12-15 2020-12-15 Multi-label graph-based power grid network frame knowledge graph construction method

Country Status (1)

Country Link
CN (1) CN112685570B (en)

Families Citing this family (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113360641B (en) * 2021-05-07 2023-05-30 内蒙古电力（集团）有限责任公司乌兰察布电业局 Deep learning-based power grid fault handling plan semantic modeling system and method
CN113254669B (en) * 2021-06-15 2021-10-19 广东电网有限责任公司湛江供电局 Knowledge graph-based power distribution network CIM model information completion method and system
CN115048533B (en) * 2022-06-21 2023-06-27 四维创智（北京）科技发展有限公司 Knowledge graph construction method and device, electronic equipment and readable storage medium
CN115276003B (en) * 2022-09-19 2023-02-10 国网江西省电力有限公司电力科学研究院 Power distribution station area topological relation identification method based on multi-source data fusion knowledge graph

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109635127A (en) * 2019-02-20 2019-04-16 云南电网有限责任公司信息中心 A kind of power equipment portrait knowledge mapping construction method based on big data technology
CN111552813A (en) * 2020-03-18 2020-08-18 国网浙江省电力有限公司 Power knowledge graph construction method based on power grid full-service data

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109635127A (en) * 2019-02-20 2019-04-16 云南电网有限责任公司信息中心 A kind of power equipment portrait knowledge mapping construction method based on big data technology
CN111552813A (en) * 2020-03-18 2020-08-18 国网浙江省电力有限公司 Power knowledge graph construction method based on power grid full-service data

Also Published As

Publication number Publication date
CN112685570A (en) 2021-04-20

Similar Documents

Publication Publication Date Title
CN112685570B (en) Multi-label graph-based power grid network frame knowledge graph construction method
Tang et al. Enhancement of power equipment management using knowledge graph
WO2023115842A1 (en) Data-driven offline and online integrated simulation system and method for power distribution network
CN110674311A (en) Knowledge graph-based power asset heterogeneous data fusion method
US9798310B2 (en) Method for searching cross-regional power supply area based on CIM model and system thereof
CN109635127B (en) Power equipment portrait knowledge map construction method based on big data technology
Wang et al. Power system network topology identification based on knowledge graph and graph neural network
CN113849944A (en) Power grid topology service method, device and system based on graph data model
CN117273133A (en) Construction method of multi-source heterogeneous data knowledge graph of power distribution network
CN115757810A (en) Method for constructing standard ontology of knowledge graph
Tang et al. Automatic schema construction of electrical graph data platform based on multi-source relational data models
CN114117065A (en) Knowledge graph construction method and system based on power production statistical service
Jiang et al. Construction of substation engineering design knowledge graph based on “ontology seven-step method”
Chen et al. Application of knowledge graph in power system fault diagnosis and disposal: A critical review and perspectives
CN115937881A (en) Method for automatically identifying content of knowledge graph construction standard form
Si et al. Research on comprehensive evaluation of distribution network based on Knowledge Graphs
CN115599897A (en) Power distribution network equipment intelligent question-answering system construction method and system based on knowledge graph
CN115618007A (en) Monitoring data analysis system based on knowledge graph
CN112685608A (en) CYPHER-based power grid dispatching field knowledge attribute graph model construction method
Tang et al. Ontology based Knowledge Modeling and Extraction of Power Equipment Supply Chain
Wang et al. [Retracted] Design of an Underground Transmission Line Condition Fault Monitoring System for Power Grids Based on Data Analysis Algorithms
Xinjie et al. A Construction Method for the Knowledge Graph of Power Grid Supervision Business
CN117251582B (en) Substation knowledge graph construction and optimization method based on multi-view learning
Gu et al. Research on the Knowledge Graph Construction Technology and its Power Grid Applications
CN117934209B (en) Regional power system carbon emission big data analysis method based on knowledge graph

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant