Construction method of incidence relation graph model based on power basic resources and visualization system

Abstract

The invention relates to the technical field of electric power data visualization processing, and discloses a construction method and a visualization system of an incidence relation graph model based on electric power basic resources, and the system comprises an electric power data graph modeling module, an electric power data graph visualization construction module and an electric power data graph interaction module. According to the electric power associated machine account data, point-side dual-heterogeneous knowledge graph autonomous construction processing required by the electric power machine account data is carried out in combination with point-side dual-heterogeneous graph construction software, and standardized and unified modeling of the electric power machine account data is realized. And according to the electric power machine account data point-edge double-heterogeneous knowledge graph and the electric power machine account data node link graph visual layout feature information, in combination with point-edge double-heterogeneous graph construction software, intelligent customized construction output of the electric power machine account data node link graph is carried out, electric power machine account data visual construction is realized, and the electric power machine account data utilization quality is improved.

Description

Technical Field

[0001] This invention relates to the technical field of power data visualization processing, specifically to a method for constructing a relational graph model based on power infrastructure resources and a visualization system. Background Technology

[0002] The scale of the digital power infrastructure resource ledger is growing rapidly, and the traditional tabular and itemized content organization format faces several challenges, including: difficulty in reflecting the correlation between infrastructure resources, lack of a unified visualization view of infrastructure resources, and low efficiency in the retrieval and analysis of correlated infrastructure resources; the digital power infrastructure resources come from multiple data sources and are characterized by large quantity, diverse types, and complex relationships, leading to problems such as data ambiguity, redundant synonymous data, and sparse relationships, which limit the further use of digital power infrastructure resources, increase the burden on operation and maintenance personnel, and reduce the quality of operation and maintenance work; the existing power ledger data visualization construction cannot realize the intelligent processing of the graph modeling process, graph visualization process, and graph interaction process of ledger data, which reduces the standardization, unification, and visualization effect of the utilization of digital power infrastructure resources.

[0003] Chinese invention patent CN 117369813 B, published on June 21, 2024, discloses a method for visualizing an energy consumption monitoring indicator system based on a data platform. This method involves constructing a multi-dimensional energy consumption indicator system; tracing the source of electricity data according to the energy consumption indicator system to obtain energy consumption data; performing data governance based on the energy consumption data; selecting appropriate methods for data analysis and indicator calculation based on the governed data for different data types and energy consumption monitoring indicators; and integrating the calculated indicator data into a visualization engine based on graphical syntax design principles. After selecting the chart type, the visualization engine is used for rapid development, presenting the energy consumption indicator system according to different dimensions. However, the above technical solution cannot achieve intelligent construction of a knowledge graph of electricity data or interactive feedback of electricity data visualization. Summary of the Invention

[0004] (a) Technical problems to be solved

[0005] To address the shortcomings of existing power ledger data visualization methods, which fail to intelligently process the graph modeling, visualization, and interaction of ledger data, thus reducing the standardization, unification, and visualization of power digitization infrastructure resources, this paper aims to achieve the following: disambiguation and alignment preprocessing of power ledger data; scientific analysis of power ledger data source types; accurate construction of a dual heterogeneous knowledge graph of power ledger nodes and edges; intelligent matching of visual layout information of power ledger data node link graphs; precise visualization generation of power ledger data node link graphs; and efficient interaction with power ledger data node link graphs.

[0006] (II) Technical Solution

[0007] This invention is achieved through the following technical solution: a method for constructing a relational graph model based on power infrastructure resources, the method comprising the following steps:

[0008] S1. Collect real-time power ledger data, perform data disambiguation and alignment preprocessing on the power ledger data to generate power-related ledger data; analyze and process the source types of the power ledger data to obtain power ledger data source type analysis information; construct and process the point-edge dual heterogeneous knowledge graph required for the power ledger data to generate the power ledger data point-edge dual heterogeneous knowledge graph.

[0009] S2. Based on the analysis results of the source type of power ledger data, analyze the visual layout information of the node link diagram required for the visualization of power ledger data, and obtain the visual layout feature information of the node link diagram of power ledger data; perform node link diagram construction processing on the power ledger data, generate the power ledger data node link diagram and output it;

[0010] S3. Collect interactive feature information of the power ledger, perform interactive search processing on the node link graph of the historical power ledger data, and obtain the node link graph of the target interactive power ledger.

[0011] Preferably, the following steps are taken: First, real-time electricity ledger data is collected. Then, data disambiguation and alignment preprocessing are performed on the electricity ledger data to generate related electricity ledger data. Next, the source types of the electricity ledger data are analyzed to obtain source type analysis information. Finally, the required point-edge heterogeneous knowledge graph for the electricity ledger data is constructed.

[0012] S11. Collect power ledger information of the target power area online through the power data acquisition platform and obtain real-time power ledger data. The real-time power ledger data includes any one of power equipment characteristic information, operating status information and maintenance information. The power data acquisition platform includes any one of SG-CIM and Kunlun system.

[0013] S12. Based on the real-time power ledger data and the power-specific term matrix, perform power ledger data disambiguation preprocessing to obtain real-time power ledger disambiguation data and perform power ledger data association and alignment preprocessing to generate a power-related ledger dataset.

[0014] S13. Based on the power-related ledger dataset and the power data source feature keyword matrix, perform source type analysis processing on the power ledger data to obtain power ledger data source type analysis information;

[0015] S14. Based on the power-related ledger dataset, perform point-edge heterogeneous knowledge graph construction processing required for the power ledger data to generate a point-edge heterogeneous knowledge graph for the power ledger data.

[0016] Preferably, the steps for performing disambiguation preprocessing on the real-time power ledger data and the power-specific term matrix to obtain disambiguated real-time power ledger data, and then performing association and alignment preprocessing on the power ledger data to generate a power-related ledger dataset are as follows:

[0017] S121. Construct a matrix of terms specific to the power industry. ,in Indicates the first A set of specialized terms in the power industry, which represent technical terms in the general and standardized terminology of the power industry;

[0018] S122. The BERT language model algorithm is used to search for synonyms and near-synonyms in the text information of the real-time power ledger data, and the synonyms and near-synonyms searched in the real-time power ledger data are compared with the power-related term matrix. The terminology used in the power industry Semantic analysis was performed, and based on the results, text editing software was used to identify synonyms and near-synonyms in the real-time electricity ledger data, using semantically identical terms specific to the electricity industry. Text information replacement and reduplicated word deletion are performed to obtain real-time electricity ledger disambiguation data. The text editing software includes any one of WPS Office, Microsoft Word, Tencent Docs, and Shimo Docs.

[0019] S123. A clustering algorithm is used to perform semantic association alignment clustering on the real-time power ledger disambiguation data to obtain a power-related ledger dataset. ,in and They represent the first The and the first Data related to electricity.

[0020] Preferably, the steps for performing source type analysis on the power ledger data based on the power-related ledger dataset and the power data source feature keyword matrix to obtain power ledger data source type analysis information are as follows:

[0021] S131. Construct a keyword matrix for power data source characteristics. ,in Indicates the first This document defines power data source characteristic keywords corresponding to different power data source types. Power data source types include basic power attributes and identity information, power spatial location and topology relationships, power lifecycle and status information, and power operation and maintenance and related information. Basic power attributes and identity information includes the identification, technical parameters, and ownership relationships of power equipment. Power spatial location and topology relationships include the geographic spatial location, electrical topology location, and installation location of power equipment. Power lifecycle and status information includes commissioning information, maintenance and repair records, and defect and fault information. Power operation and maintenance and related information includes the responsible entity, document associations, and spare parts associations. The power data source characteristic keywords represent standard keywords set for different power data source types to describe the power data source.

[0022] S132, Transfer the power-related ledger dataset The power-related ledger data mentioned in the document to With the power data source feature keyword matrix Keywords describing the source characteristics of power data By performing text information matching, the dataset of ledgers associated with the electricity industry was retrieved. Matching keywords of the power data source features The corresponding power data source type information is used to generate power ledger data source type analysis information after data identification.

[0023] Preferably, the steps for constructing the point-edge heterogeneous knowledge graph required for the power ledger data based on the power-related ledger dataset are as follows:

[0024] S141. Obtain the power-related ledger dataset. ;

[0025] S142. Using point-edge dual heterogeneous graph construction software based on the aforementioned power association ledger dataset. The power-related ledger data mentioned in the document to The process involves constructing a point-edge heterogeneous knowledge graph of the power ledger data, and obtaining the point-edge heterogeneous knowledge graph of the power ledger data. The point-edge heterogeneous knowledge graph construction software includes any one of NetworkX, Cytoscape, Gephi, and Pajek.

[0026] Preferably, based on the analysis results of the power ledger data source type, the visual layout information of the node link graph required for power ledger data visualization is analyzed to obtain the visual layout feature information of the power ledger data node link graph; the operation steps for constructing and outputting the power ledger data node link graph are as follows:

[0027] S21. Based on the power ledger data source type analysis information and the power ledger data node link diagram standard visual layout information matrix, perform analysis and processing of the node link diagram visual layout information required for power ledger data visualization to obtain the power ledger data node link diagram visual layout feature information.

[0028] S22. Based on the dual heterogeneous knowledge graph of the power ledger data points and edges and the visual layout feature information of the power ledger data node link graph, perform node link graph construction processing on the power ledger data, generate the power ledger data node link graph and output it.

[0029] Preferably, the steps for analyzing and processing the visual layout information of the node link diagram required for power ledger data visualization based on the power ledger data source type analysis information and the standard visual layout information matrix of the power ledger data node link diagram to obtain the visual layout feature information of the power ledger data node link diagram are as follows:

[0030] S211. Construct a standard visual layout information matrix for the power ledger data node link diagram. ,in Indicates the first The standard visual layout information of the power ledger data node link diagram corresponding to different power data source types. The standard visual layout information of the power ledger data node link diagram represents the fill color information, border color information, label color information, layout shape information, edge type information, edge thickness information, edge size information, node type information, and node size information required for constructing the standard power ledger data node link diagram for different power data source types.

[0031] S212. Connect the power ledger data source type analysis information with the power ledger data node link diagram standard visual layout information matrix. The standard visual layout information of the power ledger data node link diagram described in the document Perform text information analysis on the source type of electricity data, and search for the standard visual layout information of the data node link diagram of the electricity ledger corresponding to the analysis information on the source type of electricity ledger data. The specific steps for constructing the visual layout feature information of the power ledger data node link graph are as follows:

[0032] S2121. Input text preprocessing: The power ledger data source type analysis information and the power ledger data node link diagram standard visual layout information matrix are processed. The standard visual layout information of the power ledger data node link diagram described in the document Chinese text data undergoes preprocessing including text size scaling to a fixed size and text normalization.

[0033] S2122, Convolutional Layer Processing: Analyze the source type information of the preprocessed power ledger data and the standard visual layout information of the power ledger data node link diagram. Convolution is used; the convolution formula is: ,in and These represent the first convolution adjustment factor and the second convolution adjustment factor, respectively. Represents a constant. The input variables for the convolution formula are the preprocessed data source type analysis information of the power ledger data and the standard visual layout information of the power ledger data node link diagram. ; express The convolution objective function, The base is The sum of the exponents is The exponential function; Represents the convolution objective function and exponential function In the interval The convolution value; after extracting the convolution, new data source type analysis information and standard visual layout information of the power ledger data node link diagram are formed. ;

[0034] S2123. Activation function layer processing: The convolutional layer outputs the source type analysis information of the power ledger data and the standard visual layout information of the power ledger data node link graph. Nonlinear transformations are performed using the ReLU activation function, the formula for which is: ,in This indicates the analysis information on the source type of the power ledger data and the standard visual layout information of the power ledger data node link diagram. The convolution value; Indicates the value Random numbers within a range;

[0035] S2124. Pooling layer processing: Analyze the source type information of the power ledger data after linear transformation and the standard visual layout information of the power ledger data node link diagram. Perform a downsampling operation;

[0036] S2125, Fully Connected Layer Processing: This involves analyzing the source type information of the sampled power ledger data and the standard visual layout information of the power ledger data node link diagram. Flattened into vectors, these vectors can be connected to fully connected layers for classification or regression.

[0037] S2126. Output Results: Output the power ledger data source type analysis information and the power ledger data node link graph standard visual layout information to the fully connected layer. Perform text information analysis on the source type of electricity data, and search for the standard visual layout information of the data node link diagram of the electricity ledger corresponding to the analysis information on the source type of electricity ledger data. The data is then used to generate a visual layout feature information for the power ledger data node link diagram.

[0038] Preferably, the steps for constructing and outputting the node link graph of the power ledger data based on the dual heterogeneous knowledge graph of the power ledger data points and edges, and the visual layout feature information of the node link graph of the power ledger data are as follows:

[0039] S221. Obtain the dual heterogeneous knowledge graph of the power ledger data points and edges and the visual layout feature information of the power ledger data node link graph;

[0040] S222. Using point-edge heterogeneous graph construction software, based on the visual layout information of the node link graph corresponding to the visual layout feature information of the node link graph of the power ledger data, the point-edge heterogeneous knowledge graph of the power ledger data is transformed into a node link graph through data visualization processing, and the power ledger data node link graph is generated after data identification.

[0041] S223. Output the generated power ledger data node link diagram through the display screen.

[0042] Preferably, the steps for collecting interactive feature information from the power ledger and performing interactive search processing on the node link graph of historical power ledger data to obtain the target interactive power ledger node link graph are as follows:

[0043] S31. Collect the feature information of the node link graph of the historical power ledger data required by the user online through the data input dialog box, and generate power ledger interactive feature information, which includes the name, time and item of the power ledger information;

[0044] S32. Using the SURF image search algorithm, based on the interactive feature information of the power ledger, the node link graph of historical power ledger data is searched in the power data acquisition platform, and a target interactive power ledger node link graph matrix is ​​generated after data identification. ,in Indicates the search term A diagram showing the interconnected nodes of the target-oriented power ledger.

[0045] A visualization system for constructing a relational graph model based on power infrastructure resources is provided to implement the method for constructing the relational graph model based on power infrastructure resources. The system includes a power data graph modeling module, a power data graph visualization construction module, and a power data graph interaction module.

[0046] The power data graph modeling module includes a power ledger data acquisition unit, a power ledger data disambiguation and alignment preprocessing unit, a power ledger data source type analysis unit, and a power ledger data point-edge dual heterogeneous knowledge graph construction unit.

[0047] The power ledger data acquisition unit collects real-time power ledger data through a power data acquisition platform. The power ledger data disambiguation and alignment preprocessing unit performs disambiguation preprocessing based on the real-time power ledger data and combines text editing software with power-related terms to obtain disambiguated power ledger data. It then performs association and alignment preprocessing to generate associated power ledger data. The power ledger data source type analysis unit analyzes the source type of the associated power ledger data and power data source characteristic keywords to obtain power ledger data source type analysis information. The power ledger data point-edge heterogeneous knowledge graph construction unit constructs the required point-edge heterogeneous knowledge graph based on the associated power ledger data and using point-edge heterogeneous knowledge graph construction software, generating a power ledger data point-edge heterogeneous knowledge graph.

[0048] The power data visualization construction module includes a power ledger data node link diagram visual layout analysis unit, a power ledger data node link diagram construction unit, and a power ledger data node link diagram output unit;

[0049] The power ledger data node link diagram visual layout analysis unit analyzes and processes the visual layout information of the node link diagram required for power ledger data visualization based on the power ledger data source type analysis information and the standard visual layout information of the power ledger data node link diagram, thereby obtaining the visual layout feature information of the power ledger data node link diagram; the power ledger data node link diagram construction unit constructs the node link diagram of the power ledger data based on the power ledger data point-edge heterogeneous knowledge graph, the visual layout feature information of the power ledger data node link diagram, and combined with point-edge heterogeneous graph construction software, thereby generating the power ledger data node link diagram; the power ledger data node link diagram output unit outputs the power ledger data node link diagram through a display screen.

[0050] The power data graph interaction module includes a power ledger interaction feature information collection unit and a power ledger data node link graph search unit.

[0051] The power ledger interactive feature information collection unit collects power ledger interactive feature information through a data input dialog box; the power ledger data node link graph search unit performs node link graph interactive search processing on historical power ledger data based on the power ledger interactive feature information and in conjunction with the power data collection platform to obtain the target interactive power ledger node link graph.

[0052] (III) Beneficial Effects

[0053] This invention provides a method for constructing a relational graph model based on power infrastructure resources and a visualization system therein. It offers the following advantages:

[0054] I. Accurately acquire real-time power ledger information through the power data acquisition platform to provide reliable data support for data disambiguation and alignment preprocessing of power ledger data; perform disambiguation and alignment preprocessing of power ledger data based on real-time power ledger data and combined with text editing software and power-related terminology to achieve intelligent standardization processing of data ambiguity, redundancy, and data silos in power ledger data, providing reliable data support for knowledge graph modeling of power ledger relationships; accurately analyze the source types of power ledger data based on power-related ledger data, combined with intelligent search algorithms and power data source feature keywords based on big data storage, to achieve refined classification of power ledger data; and autonomously construct and process the required point-edge-heterogeneous knowledge graph based on power-related ledger data and combined with point-edge-heterogeneous graph construction software to achieve standardized and unified modeling of power ledger data, improving the scientific nature of power ledger data utilization.

[0055] Second, by analyzing the source type of power ledger data, combining intelligent recognition algorithms with the standard visual layout information of the power ledger data node link graph stored in standard form, intelligent analysis of the visual layout information of the node link graph required for power ledger data visualization is performed, realizing customized layout of node link graphs based on power ledger data sources; based on the point-edge heterogeneous knowledge graph of power ledger data, the visual layout feature information of the power ledger data node link graph, and the point-edge heterogeneous graph construction software, intelligent customized construction and output of the power ledger data node link graph is performed, realizing the construction of power ledger data visualization and improving the quality of power ledger data utilization.

[0056] Third, by efficiently collecting interactive feature information of power ledger through the data input dialog box and combining it with the power data acquisition platform to perform efficient interactive search of the node link graph of historical power ledger data, the system achieves efficient and accurate interaction of power ledger data, improves the accuracy of power ledger data interaction retrieval, and reduces the workload and time of power operation and maintenance personnel. Attached Figure Description

[0057] Fig. 1 A schematic diagram of the modules of a visualization system for constructing a relational graph model based on power infrastructure resources, provided by the present invention;

[0058] Fig. 2 The flowchart illustrates a method for constructing a relational graph model based on power infrastructure resources, as provided by this invention. Detailed Implementation

[0059] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0060] An embodiment of the method for constructing a relational graph model based on power infrastructure resources and the visualization system thereof is as follows:

[0061] Example 1:

[0062] Please see Figs. 1-2 A method for constructing a relational graph model based on power infrastructure resources, comprising the following steps:

[0063] S1. Collect real-time power ledger data, perform data disambiguation and alignment preprocessing on the power ledger data to generate power-related ledger data; analyze and process the source types of the power ledger data to obtain power ledger data source type analysis information; construct and process the point-edge dual heterogeneous knowledge graph required for the power ledger data to generate the power ledger data point-edge dual heterogeneous knowledge graph.

[0064] S2. Based on the analysis results of the source type of power ledger data, analyze the visual layout information of the node link diagram required for the visualization of power ledger data, and obtain the visual layout feature information of the node link diagram of power ledger data; perform node link diagram construction processing on the power ledger data, generate the power ledger data node link diagram and output it;

[0065] S3. Collect interactive feature information of the power ledger, perform interactive search processing on the node link graph of the historical power ledger data, and obtain the node link graph of the target interactive power ledger.

[0066] For further details, please refer to Figs. 1-2 The process involves collecting real-time electricity ledger data, performing data disambiguation and alignment preprocessing on the electricity ledger data to generate electricity-related ledger data, analyzing the source types of the electricity ledger data to obtain source type analysis information, and constructing the required point-edge heterogeneous knowledge graph for the electricity ledger data. The steps for generating the point-edge heterogeneous knowledge graph for the electricity ledger data are as follows:

[0067] S11. Collect power ledger information of the target power area online through the power data acquisition platform and obtain real-time power ledger data. The real-time power ledger data includes any one of power equipment characteristic information, operating status information and maintenance information. The power data acquisition platform includes any one of SG-CIM and Kunlun system.

[0068] S12. Based on the real-time power ledger data and the power-specific term matrix, perform power ledger data disambiguation preprocessing to obtain real-time power ledger disambiguation data and perform power ledger data association and alignment preprocessing to generate a power-related ledger dataset.

[0069] S13. Based on the power-related ledger dataset and the power data source feature keyword matrix, perform source type analysis and processing of power ledger data to obtain power ledger data source type analysis information;

[0070] S14. Based on the power-related ledger dataset, construct the point-edge heterogeneous knowledge graph required for the power ledger data, and generate the point-edge heterogeneous knowledge graph of the power ledger data.

[0071] The steps for disambiguating and preprocessing power ledger data based on real-time power ledger data and a power-specific term matrix to generate a power-related ledger dataset are as follows:

[0072] S121. Construct a matrix of terms specific to the power industry. ,in Indicates the first These are specialized terms in the power industry, representing technical terms that are generally standardized in the power sector.

[0073] S122. Use the BERT language model algorithm to search for synonyms and near-synonyms in the text information of the real-time power ledger data, and compare the searched synonyms and near-synonyms with the power industry-specific term matrix. Specialized terms in the Chinese power sector Semantic analysis was performed, and based on the results, text editing software was used to identify synonyms and near-synonyms in the real-time power ledger data, using semantically identical power-specific terms. Perform text information replacement and reduplicated word removal processing to obtain real-time electricity ledger disambiguation data. Text editing software includes any one of WPS Office, Microsoft Word, Tencent Docs, and Shimo Docs.

[0074] S123. A clustering algorithm is used to perform semantic association alignment clustering on the disambiguation data of the real-time power ledger, resulting in a power-related ledger dataset. ,in and They represent the first The and the first Data related to electricity.

[0075] The steps for analyzing the source types of power ledger data based on the power-related ledger dataset and the power data source feature keyword matrix are as follows:

[0076] S131. Construct a keyword matrix for power data source characteristics. ,in Indicates the first This document defines the characteristic keywords of power data sources corresponding to different power data source types. The power data source types include basic power attributes and identity information, power spatial location and topology relationships, power lifecycle and status information, and power operation and maintenance and related information. Basic power attributes and identity information includes the identification, technical parameters, and ownership relationships of power equipment. Power spatial location and topology relationships include the geographic spatial location, electrical topology location, and installation location of power equipment. Power lifecycle and status information includes commissioning information, maintenance and repair records, and defect and fault information. Power operation and maintenance and related information includes the responsible entity, document associations, and spare parts associations. The power data source characteristic keywords represent standard keywords used to describe the power data source for different power data source types.

[0077] S132, Transfer the power-related ledger dataset China Power Related Ledger Data to Keyword matrix of power data source features Keywords of China Power Data Source Characteristics By performing text information matching, a dataset of ledgers related to electricity was retrieved. Matching power data source characteristics keywords The corresponding power data source type information is used to generate power ledger data source type analysis information after data identification.

[0078] The steps for constructing a heterogeneous knowledge graph of power ledger data based on the power-related ledger dataset are as follows:

[0079] S141. Obtain the power-related ledger dataset. ;

[0080] S142. Using point-edge dual heterogeneous graph construction software based on power association ledger dataset. China Power Related Ledger Data to The process involves constructing a point-edge heterogeneous knowledge graph from the electricity ledger data, and obtaining the point-edge heterogeneous knowledge graph from the electricity ledger data. The software used for constructing the point-edge heterogeneous knowledge graph includes any one of NetworkX, Cytoscape, Gephi, and Pajek.

[0081] The power ledger data acquisition unit accurately acquires real-time power ledger information using a power data acquisition platform, providing reliable data support for data disambiguation and alignment preprocessing. The power ledger data disambiguation and alignment preprocessing unit, based on real-time power ledger data and combined with text editing software and power-related terminology, performs intelligent standardization processing of data ambiguity, redundancy, and data silos, providing reliable data support for power ledger relationship knowledge graph modeling. The power ledger data source type analysis unit accurately analyzes the source types of power ledger data based on power-related ledger data, combined with intelligent search algorithms and power data source feature keywords based on big data storage, achieving refined classification of power ledger data. The power ledger data point-edge heterogeneous knowledge graph construction unit autonomously constructs the required point-edge heterogeneous knowledge graph based on power-related ledger data and point-edge heterogeneous graph construction software, achieving standardized and unified modeling of power ledger data and improving the scientific nature of power ledger data utilization.

[0082] For further details, please refer to Figs. 1-2 Based on the analysis results of the power ledger data source types, the visual layout information of the node link graph required for power ledger data visualization is analyzed, and the visual layout feature information of the power ledger data node link graph is obtained. The operation steps for constructing and outputting the power ledger data node link graph are as follows:

[0083] S21. Based on the analysis information of the source type of the power ledger data and the standard visual layout information matrix of the power ledger data node link diagram, perform analysis and processing of the visual layout information of the node link diagram required for the visualization of the power ledger data to obtain the visual layout feature information of the power ledger data node link diagram.

[0084] S22. Based on the dual heterogeneous knowledge graph of power ledger data points and edges and the visual layout feature information of power ledger data node link graph, construct the node link graph of power ledger data, generate the power ledger data node link graph and output it.

[0085] The steps for analyzing and processing the visual layout information of the node link diagram required for power ledger data visualization, based on the analysis information of the power ledger data source type and the standard visual layout information matrix of the power ledger data node link diagram, are as follows:

[0086] S211. Construct a standard visual layout information matrix for the power ledger data node link diagram. ,in Indicates the first The standard visual layout information of the power ledger data node link diagram corresponding to different power data source types. The standard visual layout information of the power ledger data node link diagram represents the fill color information, border color information, label color information, layout shape information, edge type information, edge thickness information, edge size information, node type information, and node size information required for constructing the standard power ledger data node link diagram for different power data source types.

[0087] S212. Integrate the power ledger data source type analysis information with the standard visual layout information matrix of the power ledger data node link diagram. Standard visual layout information for data node link diagram of China Power Ledger Perform text information analysis on the source types of electricity data, and search for the standard visual layout information of the electricity ledger data node link diagram corresponding to the analysis information on the source types of electricity ledger data. The specific steps for constructing the visual layout feature information of the power ledger data node link diagram are as follows:

[0088] S2121. Input text preprocessing: This involves analyzing the source type information of the power ledger data and creating a standard visual layout matrix of the power ledger data node link diagram. Standard visual layout information for data node link diagram of China Power Ledger Chinese text data undergoes preprocessing including text size scaling to a fixed size and text normalization.

[0089] S2122, Convolutional Layer Processing: Analyzes the source type information and standard visual layout information of the power ledger data node link diagram after preprocessing. Convolution is used; the convolution formula is: ,in and These represent the first convolution adjustment factor and the second convolution adjustment factor, respectively. Represents a constant. The input variables for the convolution formula are the preprocessed power ledger data source type analysis information and the standard visual layout information of the power ledger data node link diagram. ; express The convolution objective function, The base is The sum of the exponents is The exponential function; Represents the convolution objective function and exponential function In the interval The convolution value; after extracting the convolution, new power ledger data source type analysis information and power ledger data node link diagram standard visual layout information are generated. ;

[0090] S2123, Activation Function Layer Processing: This process analyzes the source type of the electricity ledger data and provides standard visual layout information for the electricity ledger data node link diagram output by the convolutional layer. Nonlinear transformations are performed using the ReLU activation function, the formula for which is: ,in This section displays the analysis information on the source types of electricity ledger data and the standard visual layout information of the electricity ledger data node link diagram. The convolution value; Indicates the value Random numbers within a range;

[0091] S2124. Pooling Layer Processing: Analysis of source type information and standard visual layout information of power ledger data node link diagram after linear transformation. Perform a downsampling operation;

[0092] S2125, Fully Connected Layer Processing: This process integrates the source type analysis information of the sampled power ledger data with the standard visual layout information of the power ledger data node link diagram. Flattened into vectors, these vectors can be connected to fully connected layers for classification or regression.

[0093] S2126. Output Results: Outputs of the fully connected layer include analysis information on the source types of power ledger data and standard visual layout information for the power ledger data node link diagram. Perform text information analysis on the source types of electricity data, and search for the standard visual layout information of the electricity ledger data node link diagram corresponding to the analysis information on the source types of electricity ledger data. The data is then used to generate a visual layout feature information for the power ledger data node link diagram.

[0094] The steps for constructing and outputting the node link graph of the power ledger data based on the heterogeneous knowledge graph of the data points and edges and the visual layout feature information of the node link graph of the power ledger data are as follows:

[0095] S221. Obtain the visual layout feature information of the dual heterogeneous knowledge graph of power ledger data points and edges and the link graph of power ledger data nodes.

[0096] S222. Using point-edge heterogeneous graph construction software, based on the visual layout information of the node link graph corresponding to the visual layout feature information of the node link graph of the power ledger data, the point-edge heterogeneous knowledge graph of the power ledger data is transformed into a node link graph through data visualization processing, and then the power ledger data node link graph is generated after data identification.

[0097] S223. Output the generated power ledger data node link diagram on the display screen.

[0098] The power ledger data node link diagram visual layout analysis unit analyzes the visual layout information of the node link diagram required for power ledger data visualization based on the power ledger data source type analysis information, combined with intelligent recognition algorithms and standard visual layout information of the power ledger data node link diagram stored in a standard manner. This enables customized layout of the node link diagram based on the power ledger data source. The power ledger data node link diagram construction unit and the power ledger data node link diagram output unit work together to intelligently and customarily construct and output the power ledger data node link diagram based on the power ledger data point-edge heterogeneous knowledge graph, the visual layout feature information of the power ledger data node link diagram, and the point-edge heterogeneous graph construction software. This enables the construction of power ledger data visualization and improves the quality of power ledger data utilization.

[0099] For further details, please refer to Figs. 1-2 The steps for collecting interactive feature information from electricity ledgers and performing interactive search processing on the node link graph of historical electricity ledger data to obtain the node link graph of the target interactive electricity ledger are as follows:

[0100] S31. Collect the feature information of the node link graph of the historical power ledger data required by the user online through the data input dialog box, and generate the power ledger interactive feature information, which includes the name, time and item of the power ledger information;

[0101] S32. Using the SURF image search algorithm, based on the interactive feature information of the power ledger, the node link graph of historical power ledger data is searched in the power data acquisition platform, and a target interactive power ledger node link graph matrix is ​​generated after data identification. ,in Indicates the search term A diagram showing the interconnected nodes of the target-oriented power ledger.

[0102] By combining the power ledger interactive feature information collection unit and the power ledger data node link graph search unit, the system efficiently collects power ledger interactive feature information using a data input dialog box and performs efficient node link graph interactive search of historical power ledger data using the power data acquisition platform. This achieves efficient and accurate power ledger data interaction, improves the accuracy of power ledger data interaction retrieval, and reduces the workload and time of power operation and maintenance personnel.

[0103] Example 2:

[0104] Please see Figs. 1-2 A visualization system for constructing a relational graph model based on power infrastructure resources is provided to implement a method for constructing a relational graph model based on power infrastructure resources. The system includes a power data graph modeling module, a power data graph visualization construction module, and a power data graph interaction module.

[0105] The power data graph modeling module includes a power ledger data acquisition unit, a power ledger data disambiguation and alignment preprocessing unit, a power ledger data source type analysis unit, and a power ledger data point-edge dual heterogeneous knowledge graph construction unit.

[0106] The system comprises the following components: a power ledger data acquisition unit, which collects real-time power ledger data through a power data acquisition platform; a power ledger data disambiguation and alignment preprocessing unit, which performs disambiguation preprocessing on real-time power ledger data using text editing software and power-related terminology to obtain disambiguated real-time power ledger data, and performs association and alignment preprocessing to generate associated power ledger data; a power ledger data source type analysis unit, which analyzes the source type of power ledger data based on associated power ledger data and power data source characteristic keywords to obtain power ledger data source type analysis information; and a power ledger data point-edge heterogeneous knowledge graph construction unit, which constructs the required point-edge heterogeneous knowledge graph based on associated power ledger data using point-edge heterogeneous graph construction software, generating a power ledger data point-edge heterogeneous knowledge graph.

[0107] The power data visualization construction module includes a power ledger data node link diagram visual layout analysis unit, a power ledger data node link diagram construction unit, and a power ledger data node link diagram output unit;

[0108] The power ledger data node link diagram visual layout analysis unit analyzes and processes the visual layout information of the node link diagram required for power ledger data visualization based on the power ledger data source type analysis information and the standard visual layout information of the power ledger data node link diagram, obtaining the visual layout feature information of the power ledger data node link diagram; the power ledger data node link diagram construction unit constructs the node link diagram of the power ledger data based on the power ledger data point-edge heterogeneous knowledge graph, the visual layout feature information of the power ledger data node link diagram, and combined with point-edge heterogeneous graph construction software, generating the power ledger data node link diagram; the power ledger data node link diagram output unit outputs the power ledger data node link diagram through a display screen.

[0109] The power data graph interaction module includes a power ledger interaction feature information collection unit and a power ledger data node link graph search unit;

[0110] The power ledger interactive feature information collection unit collects power ledger interactive feature information through a data input dialog box; the power ledger data node link graph search unit performs interactive search processing on the node link graph of historical power ledger data based on the power ledger interactive feature information and in conjunction with the power data collection platform to obtain the target interactive power ledger node link graph.

[0111] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for constructing a relational graph model based on basic power resources, characterized in that, The method includes the following steps: S1. Collect real-time power ledger data, perform data disambiguation and alignment preprocessing on the power ledger data to generate power-related ledger data; analyze and process the source types of the power ledger data to obtain power ledger data source type analysis information; construct and process the point-edge dual heterogeneous knowledge graph required for the power ledger data to generate the power ledger data point-edge dual heterogeneous knowledge graph. S2. Based on the analysis results of the source type of power ledger data, analyze the visual layout information of the node link diagram required for the visualization of power ledger data, and obtain the visual layout feature information of the node link diagram of power ledger data. Perform node link graph construction on the power ledger data, generate and output the power ledger data node link graph; S3. Collect interactive feature information of the power ledger, perform interactive search processing on the node link graph of the historical power ledger data, and obtain the node link graph of the target interactive power ledger.

2. The method for constructing a relational graph model based on power infrastructure resources according to claim 1, characterized in that: The operation steps of S1 are as follows: S11. Collect power ledger information of the target power area online through the power data acquisition platform and obtain real-time power ledger data; S12. Based on the real-time power ledger data and the power-specific term matrix, perform power ledger data disambiguation preprocessing to obtain real-time power ledger disambiguation data and perform power ledger data association and alignment preprocessing to generate a power-related ledger dataset. S13. Based on the power-related ledger dataset and the power data source feature keyword matrix, perform source type analysis processing on the power ledger data to obtain power ledger data source type analysis information; S14. Based on the power-related ledger dataset, perform point-edge heterogeneous knowledge graph construction processing required for the power ledger data to generate a point-edge heterogeneous knowledge graph for the power ledger data.

3. The method for constructing a relational graph model based on power infrastructure resources according to claim 2, characterized in that: The operation steps of S12 are as follows: S121. Construct a matrix of terms specific to the power industry. The include ;in Indicates the first A term specific to the power industry; S122. The BERT language model algorithm is used to search for synonyms and near-synonyms in the text information of the real-time power ledger data, and the synonyms and near-synonyms found in the real-time power ledger data are compared with the... The above Semantic analysis is performed, and based on the results, text editing software is used to identify synonyms and near-synonyms in the real-time electricity ledger data, using words with the same semantic meaning. Text information replacement and reduplicated word deletion are performed to obtain disambiguation data for real-time electricity ledgers; S123. A clustering algorithm is used to perform semantic association alignment clustering on the real-time power ledger disambiguation data to obtain a power-related ledger dataset. The include and ;in and They represent the first The and the first Data related to electricity.

4. The method for constructing a relational graph model based on power infrastructure resources according to claim 3, characterized in that: The operation steps of S13 are as follows: S131. Construct a keyword matrix for power data source characteristics. The include ;in Indicates the first Keywords representing the characteristics of different power data sources; S132, the above The above to With the The above Perform text information matching to search for results matching the given information. The matching The corresponding power data source type information is used to generate power ledger data source type analysis information after data identification.

5. The method for constructing a relational graph model based on power infrastructure resources according to claim 4, characterized in that: The operation steps of S14 are as follows: S141, Obtain the ; S142, Using point-edge dual heterogeneous map construction software based on the above The above to A point-edge heterogeneous knowledge graph was constructed from the power ledger data, resulting in the power ledger data point-edge heterogeneous knowledge graph.

6. The method for constructing a relational graph model based on power infrastructure resources according to claim 5, characterized in that: The operation steps of S2 are as follows: S21. Based on the power ledger data source type analysis information and the power ledger data node link diagram standard visual layout information matrix, perform analysis and processing of the node link diagram visual layout information required for power ledger data visualization to obtain the power ledger data node link diagram visual layout feature information. S22. Based on the dual heterogeneous knowledge graph of the power ledger data points and edges and the visual layout feature information of the power ledger data node link graph, perform node link graph construction processing on the power ledger data, generate the power ledger data node link graph and output it.

7. The method for constructing a relational graph model based on power infrastructure resources according to claim 6, characterized in that: The operation steps of S21 are as follows: S211. Construct a standard visual layout information matrix for the power ledger data node link diagram. The include ;in Indicates the first Standard visual layout information for the data node link diagram of the power ledger corresponding to various power data source types; S212, Combine the power ledger data source type analysis information with the... The above Perform text information analysis on the source type of electricity data, and search for the corresponding information in the electricity ledger data source type analysis. The specific steps for constructing the visual layout feature information of the power ledger data node link graph are as follows: S2121. Input text preprocessing: The power ledger data source type analysis information and the... The above Chinese text data undergoes preprocessing including text size scaling to a fixed size and text normalization. S2122, Convolutional Layer Processing: Analyze the source type information of the preprocessed power ledger data and the... Convolution processing is used; S2123, Activation function layer processing: The convolutional layer outputs the power ledger data source type analysis information and the... Nonlinear transformation is performed using the ReLU activation function; S2124, Pooling Layer Processing: Analyze the source type information of the power ledger data after linear transformation and the... Perform downsampling operation; S2125, Fully Connected Layer Processing: Analyze the source type information of the sampled power ledger data and the... Flattened into vectors, these vectors can be connected to fully connected layers for classification or regression. S2126. Output Results: The output of the fully connected layer includes the analysis information on the source type of the power ledger data and the... Perform text information analysis on the source type of electricity data, and search for the corresponding information in the electricity ledger data source type analysis. The data is then used to generate a visual layout feature information for the power ledger data node link diagram.

8. The method for constructing a relational graph model based on power infrastructure resources according to claim 7, characterized in that: The operation steps of S22 are as follows: S221. Obtain the dual heterogeneous knowledge graph of the power ledger data points and edges and the visual layout feature information of the power ledger data node link graph; S222. Using point-edge heterogeneous graph construction software, based on the visual layout information of the node link graph corresponding to the visual layout feature information of the node link graph of the power ledger data, the point-edge heterogeneous knowledge graph of the power ledger data is transformed into a node link graph through data visualization processing, and the power ledger data node link graph is generated after data identification. S223. Output the generated power ledger data node link diagram through the display screen.

9. The method for constructing a relational graph model based on power infrastructure resources according to claim 8, characterized in that: The operation steps of S3 are as follows: S31. Collect the feature information of the node link graph of the historical power ledger data required by the user online through the data input dialog box, and generate interactive feature information of the power ledger. S32. Using the SURF image search algorithm, based on the interactive feature information of the power ledger, the node link graph of historical power ledger data is searched in the power data acquisition platform, and a target interactive power ledger node link graph matrix is ​​generated after data identification. The include ;in Indicates the search term A diagram showing the interconnected nodes of the target-oriented power ledger.

10. A visualization system for constructing a graph model of relationships based on power infrastructure resources, used to implement the method for constructing a graph model of relationships based on power infrastructure resources as described in any one of claims 1-9, characterized in that: The system includes a power data graph modeling module, a power data graph visualization construction module, and a power data graph interaction module.

Images