Power grid regulation cross-level knowledge graph-oriented sharing method and device, and electronic equipment

Abstract

The application discloses a kind of sharing method, device and electronic equipment for power grid regulation and control cross-level knowledge graph, the method includes step 1: data demand side power grid sends query request to open platform;Step 2: in the open platform, in the consistent name description of power grid registration information is inquired, corresponding power grid is regarded as target power grid;Step 3: in the open platform, the graph mode layer information of query request and target power grid are matched, whether according to matching success executes step 4 or 5;Step 4: data is extracted from the graph data layer of target power grid;Step 5: according to the graph mode layer information in query request, query graph data from target power grid.The above technical scheme is used, and the target power grid company is selected by the consistency of name description, data is extracted in the target power grid company, and is fused into the graph data of demand side, to avoid the influence of cross-level power grid company knowledge graph difference on cross-level knowledge graph data sharing.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of knowledge graph, and particularly relates to a sharing method and device for power grid regulation and control cross-level knowledge graph and electronic equipment. BACKGROUND

[0002] Power grid dispatching is one of the most important and complex scenarios in a power grid system. Power grid dispatching institutions in China are divided into five levels, namely, a national dispatching center, a sub-center dispatching center, a provincial dispatching center, a regional dispatching center, and a county-level dispatching center. The data sharing problem of the knowledge graph of power grid companies at all levels needs to be solved in power grid dispatching.

[0003] The power grid regulation and control cross-level knowledge graph sharing system involves two or more graphs. According to actual investigation, the construction of the knowledge graph of power grid companies at all levels has certain differences. The differences are business differences and graph design differences, which are specifically manifested in the same knowledge graph mode layer and different data layers, and different knowledge graph mode layers and different data layers.

[0004] In the process of cross-level knowledge graph sharing of power grid companies, cross-level data is graph structure data, and has the property of the knowledge graph constructed by the power grid company of the source. In order to realize the sharing and data fusion of the knowledge graph of cross-level power grid companies, the differences of the knowledge graph of power grid companies at all levels and the influence of the differences on the cross-level knowledge graph data sharing of power grid companies need to be solved. There is no ideal technical solution in the prior art to solve the above problems. SUMMARY

[0005] The purpose of the present application is to provide a sharing method and device for power grid regulation and control cross-level knowledge graph and electronic equipment. The method selects a target power grid company from numerous power grid companies by consistency comparison of the graph name and business description between cross-level power grid companies, extracts corresponding data in the target power grid company according to the graph mode layer information of the demand side, and fuses the data into the graph data of the demand side, thereby avoiding the influence of the differences of the knowledge graph of cross-level power grid companies on the cross-level knowledge graph data sharing. Further, a virtual graph is generated on a local graph platform based on the extracted graph data, so as to ensure that the property of the knowledge graph after the cross-level knowledge fusion is unchanged.

[0006] Technical scheme: The application provides a sharing method for power grid regulation and control cross-level knowledge graph, comprising: an open platform accepts a query request sent by a data demand side power grid; wherein the registration information of each level power grid on the open platform includes name description and graph pattern layer information; the open platform queries in the power grid registration information, obtains the name description consistent with the name description in the query request, and takes the power grid corresponding to the name description consistent with the name description as a target power grid; the open platform matches the graph pattern layer information of the target power grid according to the graph pattern layer information in the query request, if the matching is successful, extracts data from the graph data layer of the target power grid according to the data relationship and data described by the matching consistent graph pattern layer information, and sends the data to the data demand side power grid, on the local graph platform of the data demand side power grid, according to the data relationship of the graph data obtained from the target power grid and the graph data of the data demand side power grid itself, a corresponding virtual graph is generated, and the graph data source of the corresponding business is corrected as the virtual graph; if the matching is not successful, the graph data is queried from the target power grid according to the graph pattern layer information in the query request, the queried graph data is converted into relational data, and the data is returned to the data demand side power grid, on the local graph platform of the data demand side power grid, according to the data relationship of the graph data obtained from the target power grid and the graph data of the data demand side power grid itself, a corresponding virtual graph is generated, and the graph data source of the corresponding business is corrected as the virtual graph.

[0007] Specifically, the name description in the registration information includes service name, graph name and business description.

[0008] Specifically, the local graph platform accepts the query request sent by the data demand side power grid, and after querying, the local graph platform sends the queried data to the data demand side power grid, if the sent data cannot meet the business demand of the data demand side power grid, the open platform accepts the query request sent by the data demand side power grid.

[0009] Specifically, the open platform queries the graph name in the power grid registration information, and the text similarity of the graph name in the query request, if the text similarity is not lower than the text standard threshold, it is determined that the graph name in the power grid registration information is consistent with the graph name in the query request, and the power grid corresponding to the graph name consistent with the graph name is taken as the target power grid.

[0010] Specifically, if the text similarity is lower than the text standard threshold, the open platform queries the business description in the power grid registration information, and the text similarity of the business description in the query request, the business description in the power grid registration information with the highest text similarity is identified as consistent with the business description in the query request, and the power grid corresponding to the consistent business description is regarded as the target power grid.

[0011] Specifically, the open platform performs character cyclic traversal comparison on the names of each node of the graph schema layer information, and for nodes with the same name, the center of the feature vector is also compared, if the center of the feature vector is the same, the graph schema layer information in the query request is matched successfully with the graph schema layer information of the target power grid.

[0012] Specifically, the open platform extracts the data in the triple structure from the graph data layer of the target power grid, and the data demand side power grid performs knowledge fusion with the local data in the graph data layer after obtaining the data, forms a new virtual node in the graph data layer, and sends the data after knowledge fusion to the target power grid.

[0013] Specifically, the data demand side power grid interprets the graph schema layer information of itself, converts the relational data returned by the target power grid into data in the triple structure, and performs knowledge fusion with the local data in the graph data layer to form a new virtual node in the graph data layer.

[0014] The application also provides a sharing system for power grid regulation and control cross-level knowledge graph, comprising a data demand side power grid, an open platform, a target power grid and a data demand side local graph platform, wherein: the data demand side power grid is used to send a query request to the open platform; the registration information of each power grid on the open platform comprises name description and graph schema layer information; the open platform is used to query in the power grid registration information, obtain the name description consistent with the name description in the query request, and take the corresponding power grid as the target power grid; match the graph schema layer information in the query request with the graph schema layer information of the target power grid, if the matching is successful, extract the data from the graph data layer of the target power grid according to the data relationship and data described by the matched graph schema layer information, and return the data to the data demand side power grid; if the matching fails, query the graph data from the target power grid according to the graph schema layer information in the query request, convert the queried graph data into relational data, and return the data to the data demand side power grid; the data demand side local graph platform generates the corresponding virtual graph according to the data relationship of the graph data obtained from the target power grid and the graph data of the data demand side power grid itself, and corrects the graph data source of the corresponding business to the virtual graph.

[0015] The application further provides an electronic device comprising a memory and a processor, wherein the memory stores a computer program executable by the processor, and when the computer program is executed by the processor, the above-mentioned method for sharing a cross-level knowledge graph for power grid regulation and control is executed.

[0016] Advantages: Compared with the prior art, the application has the following remarkable advantages: by comparing the consistency of the graph names and business descriptions between cross-level power grid companies, a target power grid company is selected from a plurality of power grid companies, corresponding data is extracted in the target power grid company according to the graph mode layer information of the demand side, and the data is fused into the graph data of the demand side, the graph mode layer information is a graph comprising nodes and relationships between the nodes, thereby avoiding the influence of the difference between cross-level power grid knowledge graphs on cross-level knowledge graph data sharing; further, based on the extracted graph data, a virtual graph is generated on the local graph platform, which can ensure that the property of the knowledge graph after data cross-level knowledge fusion is unchanged and effective business calling can be performed. BRIEF DESCRIPTION OF DRAWINGS

[0017] Figure 1 A flowchart of a method for sharing a cross-level knowledge graph for power grid regulation and control is provided.

[0018] Figure 2 A structural diagram of a capability development platform is provided.

[0019] Figure 3 A diagram of a graph mode layer is provided.

[0020] Figure 4 A diagram of a knowledge fusion mode of data with the same graph mode layer is provided.

[0021] Figure 5 A diagram of a knowledge fusion mode of data with different graph mode layers is provided. DETAILED DESCRIPTION

[0022] The technical solutions of the application will be further described below with reference to the drawings.

[0023] Reference is made to Figure 1 A flowchart of a method for sharing a cross-level knowledge graph for power grid regulation and control is provided.

[0024] In the embodiment of the application, the data demand side power grid sends a query request to the local graph platform, the local graph platform returns the data obtained by the query after the query, and if the returned data cannot meet the business demand of the data demand side power grid, step 1 is performed.

[0025] In a specific implementation, the local graph platform (local graph service platform) is a local platform of a data demand side power grid, stores knowledge graph data related to power grid business, and can provide the data demand side power grid company with knowledge graph data related to business. When the data demand side power grid sends a query request to the local graph platform, the local graph platform can query related data in the local graph platform according to the service name, graph name, business description, and graph mode layer information in the query request, the query manner can refer to steps 2 to 5, and the data is returned to the data demand side power grid. If the business requirements of the data demand side power grid can be met, the query process can end, and if not, step 1 is performed.

[0026] Referring to Figure 2 which is a structural schematic diagram of the capability development platform provided by the application.

[0027] Step 1: The data demand side power grid sends a query request to the open platform.

[0028] In a specific implementation, the power grid company sends a query request to the open platform (capability open platform) according to specific regulation and control business and actual conditions, and the query request can include the service name, graph name, business description, and graph mode layer information related to regulation and control business.

[0029] Step 2: In the open platform, the registration information of the power grid is queried to obtain the name description consistent with the name description in the query request, the corresponding power grid is taken as a target power grid, and step 3 is performed.

[0030] In the embodiment of the application, the registration information of each level of power grid on the open platform includes name description and graph mode layer information.

[0031] In a specific implementation, the name description includes the name and business description of the regulation and control business, and the registration information of each level and each local power grid company on the open platform also includes the business and capability that they can provide, specifically including the related name description. Through the consistency query of the name description of the query request and the registration information, other level power grid companies that can provide or meet the regulation and control business required by the data demand side power grid can be determined, and they are taken as target power grid companies.

[0032] In the embodiment of the application, the name description in the registration information includes the service name, graph name, and business description.

[0033] In the embodiment of the present application, on the open platform, the text similarity of the graph name in the power grid registration information and the graph name in the query request is queried, if the text similarity is not lower than the text standard threshold, it is determined that the graph name in the power grid registration information is consistent with the graph name in the query request, and then step 3 is executed.

[0034] In the embodiment of the present application, if the text similarity is lower than the text standard threshold, the text similarity of the service description in the power grid registration information and the service description in the query request is queried, and the service description in the power grid registration information with the highest text similarity is determined to be consistent with the service description in the query request, and then step 3 is executed.

[0035] In specific implementation, the name description in the registration information includes a service name, a graph name and a service description, and the consistent name description in the power grid registration information in step 2 can be subdivided into graph name consistency query and service description consistency query.

[0036] In specific implementation, the graph constructed by the power grid control service has great business similarity in use scenarios, for example, the graph about the physical structure of the power grid is generally named as “power grid static topology graph”, and the graph related to fault estimation is generally named as “fault plan graph”. According to the actual situation of the power grid control, the graph name matching can calculate the text similarity according to the text edit distance, if the text similarity is not lower than the text standard threshold (which can be set according to the actual application scenario), it is determined that the graph name in the power grid registration information is consistent with the graph name in the query request, in the case of consistency, the target power grid company can be determined, then the graph mode layer information matching in step 3 can be directly executed, in the case of inconsistency, it indicates that there is no similar graph name in the registration information of the open platform, and further, the target power grid company needs to be determined through the consistency matching of the service description.

[0037] In specific implementation, the text similarity of the service description in the power grid registration information and the service description in the query request is queried, the main technical means adopts NLP (natural language processing) technology, the page description text is converted into a vector, the service description of the service provided by the capability development platform is also converted into a vector, then the vector similarity algorithm is adopted for comparison, the service description with the highest text similarity in the registration information is obtained, the service provided is determined as the target graph service, and the corresponding power grid company is determined as the target power grid company. The vector similarity can be obtained according to the static pre-training model word2vec or the dynamic pre-training model bert.

[0038] Step 3, on the open platform, the graph pattern layer information in the query request is matched with the graph pattern layer information of the target power grid, if the matching is successful, step 4 is executed, if the matching fails, step 5 is executed.

[0039] Referring to Figure 3 It is a schematic diagram of the graph pattern layer provided by the application.

[0040] In the embodiment of the application, the graph pattern layer information is a graph including nodes and the relationship between the nodes.

[0041] In the embodiment of the application, in the matching process, while the names of the nodes of the graph pattern layer information are compared by character circular traversal, the feature vector centers of the nodes with the same name are also compared, if the feature vector centers are the same, the graph pattern layer information in the query request is matched with the graph pattern layer information of the target power grid successfully.

[0042] In the specific implementation, the graph pattern layer information is used to describe the data attributes and data relationships of the data in the graph data layer, the required data can be effectively obtained through the graph pattern layer information, and based on the data attributes and data relationships, the graph pattern layer information can be directly fused into the local graph data layer and graph pattern layer and directly used for the application of the regulation and control business. In the knowledge graph of the power grid company, the graph pattern layer information is embodied in the form of a graph, which includes multiple nodes and the relationship between the nodes, the nodes can represent different functional organizations or functional units in the power grid company, and the relationship between the nodes can represent the working relationship or control mode between the two.

[0043] In the specific implementation, since the matching of the pattern layer is the matching of the graph, whether the two graphs are the same is determined not only by the character circular traversal comparison, but also by the synchronous comparison of the feature vector centrality of the graph, only when the node name is the same and the feature vector centrality of the nodes with the same name is the same, it can be determined that the two graphs are the same, and the graph pattern layer matching is successful. The feature vector centrality is used to represent the features of the node itself and the node relationship features of the node and other nodes.

[0044] Referring to Figure 4 It is a schematic diagram of the knowledge fusion mode of the data when the graph pattern layers are the same.

[0045] Step 4, the data relationship and data described by the matched graph pattern layer information are used to extract data from the graph data layer of the target power grid, and the data is returned to the data demand side power grid, and step 6 is entered.

[0046] In the embodiment of the present application, the data in the form of triple structure is extracted from the graph data layer of the target power grid, the data demander power grid performs knowledge fusion with the local data in the graph data layer after obtaining the data, new virtual nodes are formed in the graph data layer, and the data after knowledge fusion is sent to the target power grid.

[0047] In a specific implementation, after the mode layer is matched successfully, the graph data layer data of the target power grid is obtained according to the mode layer, Figure 4 In the embodiment, Schema is the mode layer, label is the node type of the graph database Neo4j, and subClassOf represents a relationship description. InStance is the data layer of the graph.

[0048] In a specific implementation, the mode layers are the same, and the power grid knowledge graph cross-level performance is shown on the demand of the data demander power grid company on the graph data layer of the target power grid company, for example, Figure 4 As shown in the lower part, the data demander power grid obtains the data layer data (triple structure) from the target power grid company according to the mode layer through the service bus. After the data is obtained, new virtual nodes are formed in the graph data layer through knowledge fusion, the newly obtained data and the original data are fused to form new data relationships, and the new data relationships are returned to the target power grid, and then step 6 is entered.

[0049] Referring to Figure 5 The present application provides a schematic diagram of the knowledge fusion mode of different graph mode layers.

[0050] Step 5: According to the graph mode layer information in the query request, the graph data of the target power grid is queried, the queried graph data is converted into relational data, and the data is returned to the data demander power grid, and step 6 is entered.

[0051] In the embodiment of the present application, the data demander power grid interprets the graph mode layer information of itself, converts the relational data returned by the target power grid into data in the form of triple structure, and performs knowledge fusion with the local data in the graph data layer to form new virtual nodes in the graph data layer.

[0052] In a specific implementation, when the schema layers do not match, data cannot be directly called to the target power grid, and data cannot be directly fused based on data relationships. The data demander power grid queries data from the target power grid graph data layer according to the service bus based on the graph schema layer name information, and queries related data in the target power grid graph data layer according to data attributes and data relationships. The data can be queried according to business similarity and the like. Then, the target power grid converts the queried graph data into relational data, and returns the relational data structure to the data demander power grid. The proxy service of the data demander power grid converts the relational data returned by the target power grid into triple data again, and compares the triple data with the existing data attributes and data relationships of the graph according to the similarity of the data attributes and data relationships, and then merges the triple data into the graph of the data demander power grid. Then, step 6 is performed.

[0053] In step 6, a corresponding virtual graph is generated on the local graph platform of the data demander power grid according to the data relationships of the graph data obtained from the target power grid and the graph data of the data demander power grid, and the graph data source of the corresponding business is corrected to the virtual graph.

[0054] In a specific implementation, the virtual graph is a graph structure, that is, a data relationship based on data attributes, but the graph data is not stored. The graph data exists temporarily according to the business demand, and disappears when the data flow ends. The virtual graph retains the data properties of the knowledge graph. Because of the characteristics of the related algorithms of the knowledge graph, the virtual graph is completely retained.

[0055] In a specific implementation, the knowledge graph of the cross-level power grid is shared. The data demander power grid generally cannot save the data of the target power grid company. The reason is that the data volume is very large, and there is a large storage pressure. At the same time, the data is also a valuable asset of each power grid company. Therefore, after the corresponding triple data is obtained through step 4 or step 5, the triple data is combined with the business demand, the graph data of the data demander power grid, and the triple data obtained by the remote service to generate a new graph that meets the business demand. The graph exists only in the cache. Then, the local graph service is called again, the graph data source is corrected to the virtual graph, the processing result is returned by combining the corresponding business demand algorithm.

[0056] The application also provides a sharing system for grid regulation and control cross-level knowledge graph, comprising: a data demand side grid, an open platform, a target grid and a data demand side local graph platform, wherein: the data demand side grid is configured to send a query request to the open platform; registration information of each grid on the open platform comprises name description and graph schema layer information; the open platform is configured to query the grid registration information to obtain name description consistent with the name description in the query request, and take the corresponding grid as the target grid; match the graph schema layer information of the target grid with the graph schema layer information in the query request, if the matching is successful, extract data from the graph data layer of the target grid according to the data relationship and data described by the matching consistent graph schema layer information, and return the data to the data demand side grid; if the matching fails, query graph data from the target grid according to the graph schema layer information in the query request, convert the obtained graph data into relational data, and return the data to the data demand side grid; and the data demand side local graph platform generates corresponding virtual graph according to the data relationship of the graph data obtained from the target grid and the graph data of the data demand side grid itself, and corrects the graph data source of the corresponding business to the virtual graph.

[0057] In the embodiment of the application, the name description in the registration information in the open platform comprises service name, graph name and business description.

[0058] In the embodiment of the application, the data demand side grid sends a query request to the local graph platform, and the local graph platform returns the obtained data after the query, if the returned data cannot meet the business demand of the data demand side grid, the data demand side grid sends a query request to the open platform.

[0059] In the embodiment of the application, the open platform is configured to query the text similarity of the graph name in the grid registration information and the graph name in the query request, if the text similarity is not lower than the text standard threshold, it is determined that the graph name in the grid registration information is consistent with the graph name in the query request, and then the graph schema layer information matching is performed.

[0060] In the embodiment of the application, the open platform is configured to query the text similarity of the business description in the grid registration information and the business description in the query request if the text similarity is lower than the text standard threshold, determine that the business description in the grid registration information with the highest text similarity is consistent with the business description in the query request, and then perform the graph schema layer information matching.

[0061] In the embodiment of the present application, in the matching process, while the names of each node of the graph schema layer information are compared by character circular traversal, for the nodes with the same name, the comparison of the feature vector centers is also performed, if the feature vector centers are the same, the graph schema layer information in the query request is matched successfully with the graph schema layer information of the target power grid; the graph schema layer information is a graph including nodes and the relationship between nodes.

[0062] In the embodiment of the present application, the data demander power grid is used to extract the data of the triple structure from the graph data layer of the target power grid, and after the data is acquired, the data demander power grid performs knowledge fusion with the local data in the graph data layer, forms a new virtual node in the graph data layer, and sends the data after knowledge fusion to the target power grid.

[0063] In the embodiment of the present application, the data demander power grid is used to interpret the graph schema layer information of itself, converts the relational data returned by the target power grid into the data of the triple structure, performs knowledge fusion with the local data in the graph data layer, and forms a new virtual node in the graph data layer.

[0064] The present application also provides an electronic device comprising a memory and a processor, wherein the memory stores a computer program executable by the processor, and when the computer program is executed by the processor, the sharing method for the power grid control cross-level knowledge graph is executed.

[0065] The present application also provides a computer readable storage medium, wherein the computer readable storage medium stores a computer program, and when the computer program is executed, the sharing method for the power grid control cross-level knowledge graph provided by any feasible implementation manner described above can be realized.

[0066] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can adopt a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can adopt the form of a computer program product implemented on one or more computer usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer usable program codes.

[0067] The computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart Figure 1 one or more functions specified in the flowchart or multiple flows and / or blocks. Figure 1 one or more functions specified in the flowchart or multiple flows and / or blocks.

[0068] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instructions which implement the Figure 1 one or more functions specified in the flowchart or multiple flows and / or blocks. Figure 1 one or more functions specified in the flowchart or multiple flows and / or blocks.

[0069] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart Figure 1 one or more functions specified in the flowchart or multiple flows and / or blocks. Figure 1 one or more functions specified in the flowchart or multiple flows and / or blocks.

Claims

1. A method for sharing a grid-oriented cross-level knowledge graph, characterized in that, The method comprises the following steps: An open platform accepts a query request sent by a data demand side power grid; Wherein, the registration information of each level power grid on the open platform comprises name description and graph schema layer information; The open platform queries the registration information of the power grid, obtains the name description consistent with the name description in the query request, and takes the power grid corresponding to the name description consistent with the name description as a target power grid; The open platform matches the graph schema layer information of the target power grid according to the graph schema layer information in the query request, extracts data from the graph data layer of the target power grid according to the data relationship and data described by the matching consistent graph schema layer information, and sends the data to the data demand side power grid, so that a corresponding virtual graph is generated on the local graph platform of the data demand side power grid according to the data relationship of the graph data obtained from the target power grid and the graph data of the data demand side power grid itself, and the graph data source of the corresponding business is corrected to the virtual graph; If the matching is unsuccessful, the graph data of the target power grid is queried according to the graph schema layer information in the query request, the queried graph data is converted into relational data, and the data is returned to the data demand side power grid, so that a corresponding virtual graph is generated on the local graph platform of the data demand side power grid according to the data relationship of the graph data obtained from the target power grid and the graph data of the data demand side power grid itself, and the graph data source of the corresponding business is corrected to the virtual graph.

2. The method of claim 1, wherein, The name description in the registration information comprises a service name, a graph name and a business description.

3. The method of claim 2, wherein, Further comprising: A local graph platform accepts a query request sent by the data demand side power grid, and sends the data obtained by querying to the data demand side power grid, wherein, if the sent data cannot meet the business demand of the data demand side power grid, the open platform accepts a query request sent by the data demand side power grid.

4. The method of claim 2, wherein, The open platform queries the graph name in the registration information of the power grid, and the text similarity of the graph name in the query request, if the text similarity is not lower than a text standard threshold, it is determined that the graph name in the registration information of the power grid is consistent with the graph name in the query request, and the power grid corresponding to the graph name consistent with the graph name is taken as a target power grid. If the text similarity is lower than the text standard threshold, the open platform queries the business description in the registration information of the power grid, and the text similarity of the business description in the query request, the business description in the registration information of the power grid with the highest text similarity is determined to be consistent with the business description in the query request, and the power grid corresponding to the business description consistent with the business description is taken as a target power grid.

5. The method of claim 4, wherein, The open platform matches the graph schema layer information of the target power grid according to the graph schema layer information in the query request, comprising: ​ 6. The method of claim 1, wherein, ​ The open platform performs character cyclic traversal comparison on the names of each node of the graph pattern layer information, and for nodes with the same name, performs comparison on the feature vector centers, if the feature vector centers are the same, the graph pattern layer information in the query request is successfully matched with the graph pattern layer information of the target power grid; the graph pattern layer information is a graph including nodes and relationships between nodes.

7. The method of claim 1, wherein, The data extraction from the graph data layer of the target power grid and the data sending to the data demander power grid include: The open platform extracts the data in the form of triple structure from the graph data layer of the target power grid, and the data demander power grid performs knowledge fusion with local data in the graph data layer after obtaining the data, forms a new virtual node in the graph data layer, and sends the data after knowledge fusion to the target power grid.

8. The method of claim 7, wherein, The conversion of the graph data obtained by the query into relational data and the data return to the data demander power grid include: The data demander power grid interprets the graph pattern layer information of itself, converts the relational data returned by the target power grid into data in the form of triple structure, and performs knowledge fusion with local data in the graph data layer, forming a new virtual node in the graph data layer.

9. A grid-oriented sharing system for cross-level knowledge graphs, characterized in that, It includes: a data demander power grid, an open platform, a target power grid, and a data demander local graph platform, wherein: The data demander power grid is configured to send a query request to the open platform; the registration information of each level of power grid on the open platform includes name description and graph pattern layer information; The open platform is configured to query in the registration information of the power grid, obtain a name description consistent with the name description in the query request, and take a power grid corresponding to the name description as a target power grid; match the graph pattern layer information in the query request with the graph pattern layer information of the target power grid, if the matching is successful, extract data from the graph data layer of the target power grid according to the data relationship and data described by the matching graph pattern layer information, and return the data to the data demander power grid; if the matching fails, query graph data from the target power grid according to the graph pattern layer information in the query request, convert the graph data obtained by the query into relational data, and return the data to the data demander power grid; The data demander local graph platform generates a corresponding virtual graph according to the data relationship of the graph data obtained from the target power grid and the graph data of the data demander power grid itself, and corrects the graph data source of the corresponding business to the virtual graph.

10. An electronic device, comprising: It includes a memory and a processor, and a computer program executable by the processor is stored in the memory, when the computer program is executed by the processor, the method in any one of claims 1 to 8 is executed.

Images