Industrial neural network graph mode query method and device based on SIPyOC-RC model

Abstract

The invention provides an industrial neural network graph mode query method and device based on an SIPyOC-RC model. The method comprises the following steps: acquiring a pre-constructed industrial neural network graph based on the SIPyOC-RC model; wherein the industrial neural network graph comprises a brain graph and an instance graph, the brain graph comprises brain graph nodes and a connection relation, the instance graph comprises instance graph nodes and a connection relation, and the instance graph nodes are instantiation results of the brain graph nodes; obtaining a user query request, and analyzing the user query request into an intention format based on the SIPyOC-RC model; and performing mode query on the industrial neural network diagram according to the intention format to obtain a mode query result. According to the industrial neural network graph mode query method and device based on the SIPyOC-RC model provided by the invention, mode query of the industrial neural network graph based on the SIPyOC-RC model is realized, rapid, efficient and accurate general query of one graph can be realized, and real-time accurate query of enterprise information is assisted.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to an industrial neural network graph pattern query method and device based on a SIPyOC-RC model. Background Art

[0002] In the general industrial manufacturing field, the information systems of various departments of enterprises are independent of each other, and the channels for data communication have many defects, such as the inability to synchronize data in real time, data errors during data synchronization, etc. An effective way to solve these problems is to construct an industrial neural network graph and establish a query method for the industrial neural network graph.

[0003] The SIPyOC-RC (Source Input Process yield Output Customer - Resource Control) model is an improved model of SIPOC (Supplier; Input; Process; Output; Customer), which can be used to clearly describe the business process. However, the prior art has not yet provided an industrial neural network graph and its pattern query method based on the SIPyOC-RC model. Summary of the invention

[0004] The present invention provides an industrial neural network graph pattern query method and device based on the SIPyOC-RC model, so as to solve the defect that an industrial neural network graph and its pattern query method based on the SIPyOC-RC model have not been provided in the prior art, and realizes providing an industrial neural network graph and its pattern query method and device based on the SIPyOC-RC model.

[0005] The present invention provides an industrial neural network graph pattern query method based on a SIPyOC-RC model, comprising the following steps: obtaining a pre-constructed industrial neural network graph based on the SIPyOC-RC model; wherein the industrial neural network graph comprises a mind map and an instance map, the mind map comprises mind map nodes and connection relationships, the instance map comprises instance map nodes and connection relationships, and the instance map nodes are the instantiation results of the mind map nodes; obtaining a user query request, parsing the user query request into an intent format based on the SIPyOC-RC model; performing a pattern query on the industrial neural network graph according to the intent format to obtain a pattern query result.

[0006] According to a method for querying an industrial neural network graph pattern based on a SIPyOC-RC model provided by the present invention, before obtaining a pre-constructed industrial neural network graph based on a SIPyOC-RC model, the method further includes: performing task decomposition on an industrial production process to construct a mind map activity node; connecting each of the mind map activity nodes according to the pre- and post-relationships and inclusion relationships of each of the mind map activity nodes; parsing the resources, control conditions, input objects and output objects of each of the mind map activity nodes, respectively establishing corresponding resource nodes, control condition nodes, input object nodes and output object nodes, and establishing connections between the resource nodes, the control condition nodes, the input object nodes and the output object nodes and the mind map activity nodes, thereby obtaining a mind map; in the actual production process, instantiating the mind map nodes in the mind map to obtain instance nodes and the connection relationships between the instance nodes, thereby obtaining an instance graph.

[0007] According to a method for querying an industrial neural network graph pattern based on a SIPyOC-RC model provided by the present invention, the query pattern includes a single-node query; corresponding to the single-node query, the intent format includes information on the query pattern, the query node, the node attributes and the query intent; performing a pattern query on the industrial neural network graph according to the intent format includes: performing query pattern routing according to the query pattern, routing to a single-node query pattern, and querying the node attributes of the query node according to the query intent.

[0008] According to a method for querying an industrial neural network graph pattern based on a SIPyOC-RC model provided by the present invention, the query pattern includes a two-node query and a three-node or more query. Corresponding to the two-node query and the three-node or more query, the intent format includes query pattern, query node, node attribute, query intent and information about the relationship between nodes; performing a pattern query on the industrial neural network graph according to the intent format includes: performing query pattern routing according to the query pattern, routing to a corresponding two-node query pattern or a three-node or more query pattern, and querying the node attributes of the query node according to the query intent and the relationship between the nodes.

[0009] According to an industrial neural network graph pattern query method based on the SIPyOC-RC model provided by the present invention, the two-node query includes querying the second instance graph active node according to the first instance graph active node; accordingly, the node attribute of the query node is queried according to the query intent and the relationship between the nodes, including: locating the first instance graph active node; taking the first instance graph active node as the starting node, and according to the structure of the instance graph and the relationship between the nodes, screening the endpoints that meet the relationship between the nodes in the instance graph, and using the endpoints as the second instance graph active nodes; acquiring and returning the node attribute information of the second instance graph active node according to the query intent.

[0010] According to an industrial neural network graph pattern query method based on the SIPyOC-RC model provided by the present invention, the two-node query includes querying the first mind map activity node according to the third instance graph activity node; accordingly, the querying the node attribute of the query node according to the query intent and the relationship between the nodes includes: locating the third instance graph activity node; obtaining the second mind map activity node corresponding to the third instance graph activity node according to the structure of the industrial neural network graph; taking the second mind map activity node as the starting node, according to the structure of the mind map and the relationship between the nodes, screening the endpoints that meet the relationship between the nodes in the mind map, and taking the endpoints as the first mind map activity nodes; obtaining and returning the node attribute information of the first mind map activity node according to the query intent.

[0011] According to an industrial neural network graph pattern query method based on the SIPyOC-RC model provided by the present invention, the two-node query includes querying the fourth mind map activity node according to the third mind map activity node; accordingly, the node attribute of the query node is queried according to the query intent and the relationship between the nodes, including: locating the third mind map activity node; taking the third mind map activity node as the starting node, and according to the structure of the mind map and the relationship between the nodes, screening the endpoints that meet the relationship between the nodes in the mind map, and using the endpoints as the fourth mind map activity node; acquiring and returning the node attribute information of the fourth mind map activity node according to the query intent.

[0012] According to a method for querying an industrial neural network graph pattern based on a SIPyOC-RC model provided by the present invention, the two-node query includes querying a fifth mind map activity node according to a first instance map node; wherein the first instance map node includes an input object node, an output object node, a resource node or a control condition node; accordingly, querying the node attribute of the query node according to the query intent and the relationship between the nodes includes: locating the first instance map node; searching for a fourth instance map activity node connected to the first instance map node in the instance map; searching for a fifth mind map activity node corresponding to the fourth instance map activity node according to the structure of the industrial neural network; obtaining a second node set consisting of the topmost nodes in the first node set consisting of the fifth mind map activity node; obtaining a third node set consisting of instantiated nodes of the second node set; searching the instance map for a terminal point from the fourth instance map activity node to the third node set via the instance map activity node, obtaining the node attribute of the terminal point according to the query intent, and returning the node attribute as the node attribute of the fifth mind map activity node.

[0013] According to a method for querying an industrial neural network graph pattern based on a SIPyOC-RC model provided by the present invention, the two-node query includes querying a first mind map node according to a second instance map node; wherein the second instance map node and the first mind map node include an input object node, an output object node, a resource node or a control condition node; accordingly, querying the node attribute of the query node according to the query intent and the relationship between the nodes includes: locating the second instance map node; searching for a fifth instance map activity node connected to the second instance map node in the instance map; searching for a sixth mind map activity node corresponding to the fifth instance map activity node according to the structure of the industrial neural network; obtaining a fifth node set consisting of the topmost nodes in a fourth node set consisting of the sixth mind map activity node; obtaining a sixth node set consisting of instantiated nodes of the fifth node set; searching for a terminal point from the fifth instance map activity node to the sixth node set via the instance map activity node in the instance map; and obtaining and returning information on the node attribute of the node connected to the terminal point and having the same type as the first mind map node according to the query intent.

[0014] According to an industrial neural network graph pattern query method based on the SIPyOC-RC model provided by the present invention, the query intent includes node attribute description information query intent and node attribute statistical information query intent.

[0015] The present invention also provides an industrial neural network graph pattern query device based on the SIPyOC-RC model, comprising the following modules: a construction module, used to obtain a pre-constructed industrial neural network graph based on the SIPyOC-RC model; wherein the industrial neural network graph includes a mind map and an instance map, the mind map includes mind map nodes and connection relationships, the instance map includes instance map nodes and connection relationships, and the instance map nodes are the results of instantiation of the mind map nodes; a parsing module, used to obtain a user query request, and parse the user query request into an intent format based on the SIPyOC-RC model; a query module, used to perform a pattern query on the industrial neural network graph according to the intent format to obtain a pattern query result.

[0016] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, an industrial neural network graph pattern query method based on the SIPyOC-RC model as described above is implemented.

[0017] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements an industrial neural network graph pattern query method based on the SIPyOC-RC model as described in any one of the above.

[0018] The present invention also provides a computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the industrial neural network graph pattern query method based on the SIPyOC-RC model as described in any one of the above is implemented.

[0019] The industrial neural network graph pattern query method and device based on the SIPyOC-RC model provided by the present invention obtain a pre-built industrial neural network graph based on the SIPyOC-RC model, obtain a user query request, parse the user query request into an intent format based on the SIPyOC-RC model, perform a pattern query on the industrial neural network graph according to the intent format, and obtain a pattern query result, thereby realizing a pattern query of the industrial neural network graph based on the SIPyOC-RC model, and can realize fast, efficient and accurate one-graph universal query, thereby facilitating real-time and accurate query of enterprise information. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0021] Figure 1 It is a structural diagram of the SIPyOC-RC model.

[0022] Figure 2 It is one of the flow charts of the industrial neural network graph pattern query method based on the SIPyOC-RC model provided by the present invention.

[0023] Figure 3 It is a structural diagram of the industrial neural network graph based on the SIPyOC-RC model.

[0024] Figure 4 This is the second flow chart of the industrial neural network graph pattern query method based on the SIPyOC-RC model provided by the present invention.

[0025] Figure 5 It is a structural schematic diagram of an industrial neural network graph pattern query device based on a SIPyOC-RC model provided by the present invention.

[0026] Figure 6 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0027] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0028] Figure 1 The following is a schematic diagram of the structure of the SIPyOC-RC model. The SIPyOC-RC model contains seven types of nodes and seven types of relationships, among which: S stands for Supplier, which is the provider of input.

[0029] I stands for Input, which is the input of an activity or process.

[0030] P stands for Process, which is an activity or process.

[0031] O stands for Output, which is the output of the activity.

[0032] The C to which the output points represents the Consumer, which is the recipient of the output.

[0033] R stands for Resource, which is the resource required for the activity.

[0034] The C pointing to the activity stands for Control, which is the control condition of the activity.

[0035] Figure 2 This is one of the flow charts of the industrial neural network graph pattern query method based on the SIPyOC-RC model provided by the present invention, such as Figure 2 As shown, the method includes: Step S1, obtaining a pre-built industrial neural network diagram based on the SIPyOC-RC model; wherein the industrial neural network diagram includes a mind map and an instance map, the mind map includes mind map nodes and connection relationships, the instance map includes instance map nodes and connection relationships, and the instance map nodes are the results of instantiation of the mind map nodes.

[0036] The industrial neural network graph pattern query method based on the SIPyOC-RC model provided by the present invention can be applied to the server, and the method can be applied to large aircraft manufacturing or other fields.

[0037] Industrial neural network graph data storage should generally be stored in distributed high-availability graph databases, such as Neo4j, Galaxy Base, NebulaGraph, TigerGraph and other distributed graph databases. The industrial neural network graph has established a graph construction specification based on the SIPyOC-RC model, including the mind map and the instance map. The mind map is the planned part of the industrial neural network graph, and the instance map is the instance part of the industrial neural network graph. The instance map is the instantiation result after adding real attributes to the mind map.

[0038] Figure 3 This is a schematic diagram of the structure of an industrial neural network diagram based on the SIPyOC-RC model. A process is a set of interrelated or interacting activities that use specific inputs to produce desired outputs by performing certain operations or transformation mechanisms.

[0039] An activity (P) usually refers to a series of operations, tasks, actions or work units performed to achieve a specific business goal or process. It constitutes the basic component of a process and is the specific step performed in the process. An activity (P) includes a process. Figure 3 In the example, "gateway" belongs to the active node.

[0040] Input (I) refers to all raw materials, information, documents or other necessities that must be provided in order for the process to operate smoothly.

[0041] Output (O) refers to the result produced at the end of a process, which can be a product, service, report, or other form of value deliverable.

[0042] Resources (R), especially non-consumable resources, such as personnel, equipment, tooling, etc.

[0043] Control conditions (C) refer to standards, specifications, manuals, requirement documents, etc.

[0044] Step S2: Obtain a user query request, and parse the user query request into an intent format based on the SIPyOC-RC model.

[0045] The server can receive the user query request sent by the client, and parse the user query request into an intent format based on the SIPyOC-RC model, wherein the intent format includes the preset necessary information required for responding to the user query request to perform an industrial neural network graph mode query based on the SIPyOC-RC model.

[0046] Step S3: perform a pattern query on the industrial neural network diagram according to the intent format to obtain a pattern query result.

[0047] According to the intent format based on the SIPyOC-RC model obtained through parsing, a pattern query is performed on the industrial neural network graph to obtain the pattern query result.

[0048] The industrial neural network graph pattern query method based on the SIPyOC-RC model provided by the present invention obtains a pre-built industrial neural network graph based on the SIPyOC-RC model, obtains a user query request, parses the user query request into an intent format based on the SIPyOC-RC model, performs a pattern query on the industrial neural network graph according to the intent format, and obtains a pattern query result, thereby realizing a pattern query of the industrial neural network graph based on the SIPyOC-RC model, and can realize fast, efficient and accurate one-graph universal query, thereby facilitating real-time and accurate query of enterprise information.

[0049] According to a method for querying an industrial neural network graph pattern based on a SIPyOC-RC model provided by the present invention, before obtaining a pre-constructed industrial neural network graph based on a SIPyOC-RC model, the method further includes: performing task decomposition on an industrial production process to construct a mind map activity node; connecting each of the mind map activity nodes according to the pre- and post-relationships and inclusion relationships of each of the mind map activity nodes; parsing the resources, control conditions, input objects and output objects of each of the mind map activity nodes, respectively establishing corresponding resource nodes, control condition nodes, input object nodes and output object nodes, and establishing connections between the resource nodes, the control condition nodes, the input object nodes and the output object nodes and the mind map activity nodes, thereby obtaining a mind map; in the actual production process, instantiating the mind map nodes in the mind map to obtain instance nodes and the connection relationships between the instance nodes, thereby obtaining an instance graph.

[0050] Before obtaining the pre-built industrial neural network graph based on the SIPyOC-RC model, the method further includes building an industrial neural network graph based on the SIPyOC-RC model. The following example illustrates the process of building an industrial neural network graph based on the SIPyOC-RC model.

[0051] The industrial neural network diagram is closely related to the design, manufacturing, procurement, logistics and other processes of manufacturing industrial enterprises, and describes the various activities of the enterprise. The industrial neural network diagram starts with the construction of the mind map. First, the task is disassembled according to the process activities of each link of the enterprise such as design and manufacturing, and the mind map P activity node is constructed. For example, the wing design and tail design activities of the aircraft are linked to each mind map P activity according to the business's front-end and back-end relationship and inclusion relationship into the framework of the mind map of the industrial neural network diagram. Further, the mind map resources (such as wing designers, design software, etc.) required for each mind map P activity, the input, output, control conditions, etc. of the mind map P are analyzed, and the corresponding resource node R, control condition node C, input node I and output node O are constructed, and the connection with the mind map P activity is established, so as to obtain the mind map of the industrial neural network diagram.

[0052] After sorting out the mind map of industrial enterprises, the mind map is instantiated in the actual design process, manufacturing process, and logistics process to obtain instance nodes and the connection relationship between instance nodes, which are stored in the instance map. Each mind map node can be instantiated multiple times into multiple instance nodes.

[0053] The industrial neural network graph pattern query method based on the SIPyOC-RC model provided by the present invention decomposes tasks in the industrial production process, constructs mind map activity nodes, connects each mind map activity node according to the front-end and back-end relationship and the inclusion relationship of each mind map activity node, analyzes the resources, control conditions, input objects and output objects of each mind map activity node, respectively establishes corresponding resource nodes, control condition nodes, input object nodes and output object nodes, and establishes connections between the resource nodes, control condition nodes, input object nodes and output object nodes and the mind map activity nodes, thereby obtaining a mind map. In the actual production process, the mind map nodes in the mind map are instantiated to obtain instance nodes and the connection relationship between the instance nodes, thereby obtaining an instance graph, and realizing the construction of the industrial neural network graph.

[0054] According to a method for querying an industrial neural network graph pattern based on a SIPyOC-RC model provided by the present invention, the query pattern includes a single-node query; corresponding to the single-node query, the intent format includes information on the query pattern, the query node, the node attributes and the query intent; performing a pattern query on the industrial neural network graph according to the intent format includes: performing query pattern routing according to the query pattern, routing to a single-node query pattern, and querying the node attributes of the query node according to the query intent.

[0055] The query mode includes a single-node query, and the single-node query includes only one query node. For a single-node query, the intent format includes information about the query mode, query node, node attributes, and query intent. Among them, the query mode is a single-node query, which can be reflected by the presence of only one node in the intent format. When the query mode is a single-node query, a pattern query is performed on the industrial neural network graph according to the intent format, including: query mode routing according to the query mode, routing to the single-node query mode, and querying the node attributes of the query node according to the query intent.

[0056] The industrial neural network graph pattern query method based on the SIPyOC-RC model provided by the present invention improves the accuracy of single node mode query by setting the intent format including the query mode, query node, node attributes and information of the query intent, and querying the node attributes of the query node according to the query intent when the query mode is a single node query.

[0057] According to a method for querying an industrial neural network graph pattern based on a SIPyOC-RC model provided by the present invention, the query pattern includes a two-node query and a three-node or more query. Corresponding to the two-node query and the three-node or more query, the intent format includes query pattern, query node, node attribute, query intent and information about the relationship between nodes; performing a pattern query on the industrial neural network graph according to the intent format includes: performing query pattern routing according to the query pattern, routing to a corresponding two-node query pattern or a three-node or more query pattern, and querying the node attributes of the query node according to the query intent and the relationship between the nodes.

[0058] The query mode also includes two-node query and three-node query, in which case the query mode involves two or more nodes. When the query mode is two-node query and three-node query, the intent format includes the query mode, query node, node attribute, query intent and information about the relationship between nodes.

[0059] For example, the intent format can be as follows: [{node 1, attributes and query intent}, [node relationship information], {node 2, attributes and query intent}, ...].

[0060] It is understandable that if there is no need to obtain the node attribute information of a certain node, the corresponding node may not include the node attribute information. The query intent can use a set symbol or string to indicate what kind of query to make, such as using "?" to indicate obtaining specific node attribute information. A set of node relationship information can be established in advance, and the correspondence between the node relationship information and the path information can be set. When querying the pattern, the path information is determined and queried based on the node relationship information. In the intent format, the mind map nodes and instance map nodes can be distinguished by preset symbols. For example, P represents the instance map activity node, and P′ represents the mind map activity node.

[0061] For example, what devices are needed to query the activity with ID 1732?

[0062] After natural language understanding and processing, the query intent can be converted into the following intent format: [{Activity P, ID: '1732'}, [Required], {Device R, Name: '?'}]

[0063] According to the query intent, the pattern query routing is performed, and the query pattern type to which it belongs is analyzed, and then classified and analyzed. Among them, the two-node query pattern means that based on the information of a node in the graph, the information of another node associated with the node through several paths is queried. According to the structure of the industrial neural network graph, it is divided into brain graph and instance graph nodes. Under this model, the two-node query pattern can be shown in Table 1: Table 1

[0064] Among them, P represents the instance graph activity node, I represents the instance graph input object node, O represents the instance graph output object node, R represents the instance graph resource node, and C represents the instance graph control condition node. P′ represents the mind map activity node, I′ represents the mind map input object node, O′ represents the mind map output object node, R′ represents the mind map resource node, and C′ represents the mind map control condition node. The common two-node query mode is checked in Table 1 with "√".

[0065] The industrial neural network graph pattern query method based on the SIPyOC-RC model provided by the present invention improves the accuracy of two-node query and three-node query by setting the intention format including the query mode, query node, node attribute, query intention and information about the relationship between nodes, and querying the node attributes of the query node according to the query intention and the relationship between nodes when the query mode is two-node query and three-node query or more.

[0066] According to an industrial neural network graph pattern query method based on the SIPyOC-RC model provided by the present invention, the two-node query includes querying the second instance graph active node according to the first instance graph active node; accordingly, the node attribute of the query node is queried according to the query intent and the relationship between the nodes, including: locating the first instance graph active node; taking the first instance graph active node as the starting node, and according to the structure of the instance graph and the relationship between the nodes, screening the endpoints that meet the relationship between the nodes in the instance graph, and using the endpoints as the second instance graph active nodes; acquiring and returning the node attribute information of the second instance graph active node according to the query intent.

[0067] Among them, the first instance graph activity node can be represented as a P1 node, and the second instance graph activity node can be represented as a P2 node.

[0068] When querying the instance graph P2 node based on the instance graph P1 node, first locate the instance graph P1 node, and based on the industrial neural network diagram of the specific business and the path information obtained based on the relationship between nodes (the relationship between nodes can be extracted through natural language processing according to the user's query request) (such as path requirements such as path, pre- and post-path, and number of jumps), select the end point of the specified path as P2 node, and return the required node attribute information on the P2 node. That is, when querying the instance graph P2 node based on the instance graph P1 node, search the query path in the instance graph based on the instance graph P1 node and the instance graph structure (search the instance graph path), and find all instance P2 nodes that meet the query intent.

[0069] Example: User query: What is the post-plan for plan with ID 123?

[0070] Intent format: [{Plan P1, ID: '123'}, [Postfix], {Plan P2, Name: '?'}].

[0071] Execution process: First locate the plan node P with ID '123', search according to the [post-position] path, retrieve all plan P2 nodes that meet the starting point ID '123' and the path [post-position], and return the names of all P2 nodes that meet the conditions.

[0072] The industrial neural network graph pattern query method based on the SIPyOC-RC model provided by the present invention locates the first instance graph active node, takes the first instance graph active node as the starting node, screens the endpoints that satisfy the node-to-node relationship in the instance graph according to the structure of the instance graph and the node-to-node relationship, takes the endpoints as the second instance graph active node, obtains and returns the node attribute information of the second instance graph active node according to the query intent, and realizes querying the instance graph active node according to the instance graph active node.

[0073] According to an industrial neural network graph pattern query method based on the SIPyOC-RC model provided by the present invention, the two-node query includes querying the first mind map activity node according to the third instance graph activity node; accordingly, the querying the node attribute of the query node according to the query intent and the relationship between the nodes includes: locating the third instance graph activity node; obtaining the second mind map activity node corresponding to the third instance graph activity node according to the structure of the industrial neural network graph; taking the second mind map activity node as the starting node, according to the structure of the mind map and the relationship between the nodes, screening the endpoints that meet the relationship between the nodes in the mind map, and taking the endpoints as the first mind map activity nodes; obtaining and returning the node attribute information of the first mind map activity node according to the query intent.

[0074] The third instance map activity node can be represented as P3, and the first mind map activity node can be represented as P1′.

[0075] When querying the mind map node P1′ based on the instance map node P3, first locate the instance map node P3, and find the source mind map node P3′ of the instance map node P3 based on the structure of the business industry neural network graph (find the mind map upward, that is, find the instance map node P3 as the end point, and the starting point of the path with "instance" as the edge is the source mind map node), and then find the path that meets the user's requirements in the mind map (such as the type of path and the number of jumps, etc.), find the end points of these paths as the mind map P1′ nodes, and return the information of these nodes. That is, find the corresponding mind map node P3′ from the instance P3 node (find the mind map upward), and find all P1′ nodes that meet the query intent based on the direction and graph structure.

[0076] Example: User query: What are the categories of secondary sub-plans of plan with ID 123?

[0077] Intent format: [{Plan P3, ID: '123'}, [Contains, Level 2], {Plan P1′, Type: '?'}].

[0078] Execution process: First locate the instance P3 node, find the source mind map node P3′ of the instance P3 node according to the structure of the business industry neural network graph, find all paths with the mind map P3′ as the starting point, the path is 'included', and the number of jumps is two, obtain the end point P1′ of the path, and return the plan type in S for the set S composed of all such P1′ nodes. For example, there are 3 A plans, 1 C plan, and 2 D plans in S. Then the secondary sub-plans have three types, namely A, C, and D.

[0079] The industrial neural network graph pattern query method based on the SIPyOC-RC model provided by the present invention locates the third instance graph activity node, obtains the second mind map activity node corresponding to the third instance graph activity node according to the structure of the industrial neural network graph, takes the second mind map activity node as the starting node, and screens the end points that meet the relationship between nodes in the mind map according to the structure of the mind map and the relationship between nodes, takes the end points as the first mind map activity node, obtains and returns the node attribute information of the first mind map activity node according to the query intention, and realizes the query of the mind map activity node according to the instance graph activity node.

[0080] According to an industrial neural network graph pattern query method based on the SIPyOC-RC model provided by the present invention, the two-node query includes querying the fourth mind map activity node according to the third mind map activity node; accordingly, the node attribute of the query node is queried according to the query intent and the relationship between the nodes, including: locating the third mind map activity node; taking the third mind map activity node as the starting node, and according to the structure of the mind map and the relationship between the nodes, screening the endpoints that meet the relationship between the nodes in the mind map, and using the endpoints as the fourth mind map activity node; acquiring and returning the node attribute information of the fourth mind map activity node according to the query intent.

[0081] The third mind map activity node can be represented as P3′, and the fourth mind map activity node can be represented as P4′.

[0082] When querying the mind map node P4′ based on the mind map node P3′, first locate the mind map node P3′, and according to the relationship information between nodes given by the user's query request, find the paths that meet the user's requirements (such as the type of path and the number of jumps, etc.) in the mind map, find the end point of these paths, the mind map node P4′, and return the information of these nodes.

[0083] Example: User query: What are the types of secondary sub-plans of type A plan?

[0084] Intent format: [{Plan P3′, Type: A}, [Contains, Level 2], {Plan P4′, Type: “?”}].

[0085] Execution process: The first step is to locate the mind map node P3′ of the A type plan, find all paths with the mind map P3′ as the starting point, the path as 'included', and the number of jumps as two, obtain the end point P4′ of the path, and return all such P4′. That is, according to the mind map P3′ node and the mind map structure, search for the query path in the graph (mind map finds the path), and find all the mind map P4′ nodes that meet the query intent.

[0086] The industrial neural network graph pattern query method based on the SIPyOC-RC model provided by the present invention locates the third mind map activity node, takes the third mind map activity node as the starting node, screens the end points that meet the relationship between nodes in the mind map according to the structure of the mind map and the relationship between nodes, takes the end points as the fourth mind map activity node, obtains and returns the node attribute information of the fourth mind map activity node according to the query intention, and realizes the query of the mind map activity node according to the mind map activity node.

[0087] According to a method for querying an industrial neural network graph pattern based on a SIPyOC-RC model provided by the present invention, the two-node query includes querying a fifth mind map activity node according to a first instance map node; wherein the first instance map node includes an input object node, an output object node, a resource node or a control condition node; accordingly, querying the node attribute of the query node according to the query intent and the relationship between the nodes includes: locating the first instance map node; searching for a fourth instance map activity node connected to the first instance map node in the instance map; searching for a fifth mind map activity node corresponding to the fourth instance map activity node according to the structure of the industrial neural network; obtaining a second node set consisting of the topmost nodes in the first node set consisting of the fifth mind map activity node; obtaining a third node set consisting of instantiated nodes of the second node set; searching the instance map for a terminal point from the fourth instance map activity node to the third node set via the instance map activity node, obtaining the node attribute of the terminal point according to the query intent, and returning the node attribute as the node attribute of the fifth mind map activity node.

[0088] The first instance graph node may be an input object node I, an output object node O, a resource node R or a control condition node C. The first instance graph node may be represented as an instance graph I / O / R / C node. The fifth mind map activity node may be represented as P5'.

[0089] When querying the mind map P5′ node based on the instance graph I / O / R / C node, first find the instance I / O / R / C node, find the active node P1 connected to the instance I / O / R / C node, search the set S1 consisting of all mind map nodes P1′ that can be instantiated as P1 based on the industrial neural network graph structure, query the set S2 consisting of the top-level nodes in the set S1, and form the set S3 of the instantiated nodes of all points in S2, search the endpoints of all reachable paths from P1 to S3 in the instance graph (the path can only pass through P type nodes), and use the node attributes of the endpoint as the node attributes of the P5′ node and return them.

[0090] For example, a user queries: What are the categories of plans to which the output O1 numbered MI-0248 belongs?

[0091] Intended format: [{Output O1, Number: MI-0248}, [Belongs to], {Plan P, Type: "?"}].

[0092] Execution process: First find the node of output O1 numbered MI-0248, find the active node P1 that generates O1, search for the set S1 consisting of all the mind map nodes P1′ that can be instantiated as P1 according to the industrial neural network graph structure, query the set S2 consisting of the top nodes in the set S1, and form the set S3 of the instantiated nodes of all points of S2, and search for the final active node of all reachable paths from P1 to S3 in the instance graph (the path can only pass through the P type node) and return it. For example, there are 3 A plans, 1 C plan, and 2 D plans in the returned active nodes. Then the plans to which the output O1 numbered MI-0248 belongs are of three types, namely A, C, and D.

[0093] For another example, the user queries: What are the categories of plans that the output O1 with the number MI-0248 belongs to? The query is also performed according to the above execution process. No further details will be given here.

[0094] The industrial neural network graph pattern query method based on the SIPyOC-RC model provided by the present invention locates the first instance graph node, searches for the fourth instance graph active node connected to the first instance graph node in the instance graph, searches for the fifth mind map active node corresponding to the fourth instance graph active node according to the structure of the industrial neural network, obtains the second node set composed of the top nodes in the first node set composed of the fifth mind map active node, obtains the third node set composed of the instantiation nodes of the second node set, searches for the fourth instance graph active node through the instance graph active node to the third node set in the instance graph, obtains the node attributes of the endpoint according to the query intent, and returns the node attributes as the node attributes of the fifth mind map active node, thereby realizing the query of the mind map active node according to the instance graph I / O / R / C node.

[0095] According to a method for querying an industrial neural network graph pattern based on a SIPyOC-RC model provided by the present invention, the two-node query includes querying a first mind map node according to a second instance map node; wherein the second instance map node and the first mind map node include an input object node, an output object node, a resource node or a control condition node; accordingly, querying the node attribute of the query node according to the query intent and the relationship between the nodes includes: locating the second instance map node; searching for a fifth instance map activity node connected to the second instance map node in the instance map; searching for a sixth mind map activity node corresponding to the fifth instance map activity node according to the structure of the industrial neural network; obtaining a fifth node set consisting of the topmost nodes in a fourth node set consisting of the sixth mind map activity node; obtaining a sixth node set consisting of instantiated nodes of the fifth node set; searching for a terminal point from the fifth instance map activity node to the sixth node set via the instance map activity node in the instance map; and obtaining and returning information on the node attribute of the node connected to the terminal point and having the same type as the first mind map node according to the query intent.

[0096] The second instance graph node may be an input object node, an output object node, a resource node or a control condition node. The first mind map node may be an input object node, an output object node, a resource node or a control condition node. The second instance graph node may be represented as an instance graph I / O / R / C node, and the first mind map node may be represented as a mind map I′ / O′ / R′ / C′ node.

[0097] According to the instance graph I / O / R / C node, query the mind graph I′ / O′ / R′ / C′ node, first find the instance graph I / O / R / C node, find the active node P1 connecting the instance graph I / O / R / C node, search the set S4 composed of all mind graph nodes P1′ that can be instantiated as P1 according to the industrial neural network graph structure, query the set S5 composed of the top nodes in the set S4, and form the set S6 of the instantiated nodes of all points in S5, search the final active node set S7 of all reachable paths from P1 to S6 in the instance graph (the path can only pass through P type nodes), and return the information of the required I / O / R / C nodes connected to the S7 set.

[0098] For example, a user queries: When generating output O1 numbered MI-0248, which types of equipment are required to operate?

[0099] Intent format: [{output O1, number: MI-0248}, [need to run], {device R, type: "?"}].

[0100] Execution process: First find the output O1 node numbered MI-0248, find the active node P1 that generates O1, search for the set S4 consisting of all mind map nodes P1′ that can be instantiated as P1 based on the industrial neural network graph structure, query the set S5 consisting of the top-level nodes in the set S4, and form the set S6 of the instantiated nodes of all points in S5, search for the final active node set S7 consisting of all reachable paths from P1 to S6 in the instance graph (the path can only pass through P-type nodes), and return the device type information of all R nodes connected to the S7 set.

[0101] The industrial neural network graph pattern query method based on the SIPyOC-RC model provided by the present invention locates the second instance graph node, searches for the fifth instance graph active node connected to the second instance graph node in the instance graph, searches for the sixth mind map active node corresponding to the fifth instance graph active node according to the structure of the industrial neural network, obtains the fifth node set consisting of the top nodes in the fourth node set consisting of the sixth mind map active node, obtains the sixth node set consisting of the instantiated nodes of the fifth node set, searches for the endpoint of the fifth instance graph active node through the instance graph active node to the sixth node set in the instance graph, obtains and returns the node attribute information of the node connected to the endpoint and consistent with the first mind map node type according to the query intent, and realizes the query of the mind map I' / O' / R' / C' node according to the instance graph I / O / R / C node.

[0102] According to an industrial neural network graph pattern query method based on the SIPyOC-RC model provided by the present invention, the query intent includes node attribute description information query intent and node attribute statistical information query intent.

[0103] The query intent includes the node attribute description information query intent and the node attribute statistical information query intent. The node attribute description information query intent is used to query what the specific node attributes are. The node attribute statistical information query intent is used to query the statistical indicators of the node attributes.

[0104] Determine the query intent. If you want to query a specific attribute, you can use (specific attribute: "?") to mark the query intent. If you want to query statistical indicators, such as the number, average, maximum and minimum values, you can use (specific attribute: "count" / "average" / "max" / "min") to mark the query intent.

[0105] The industrial neural network graph pattern query method based on the SIPyOC-RC model provided by the present invention improves the flexibility of pattern query by including node attribute description information query intention and node attribute statistical information query intention in the query intention.

[0106] Figure 4 This is the second flow chart of the industrial neural network graph pattern query method based on the SIPyOC-RC model provided by the present invention. Figure 4 As shown in the figure, given an industrial neural network diagram based on the SIPyOC-RC model, the requirement is to query the information of the P' node of the mind map based on the instance activity node P. The implementation process is as follows: first, obtain the user query request, further identify the query pattern requirement, route the query pattern to the two-node query pattern, further identify it as the "instance P query mind map P'" pattern, execute the query pattern instance, and return the output result.

[0107] The industrial neural network graph pattern query method based on the SIPyOC-RC model provided by the present invention relates to a pattern query method for complex information in the field of industrial manufacturing. For the industrial neural network graph constructed based on the SIPOC improved model, pattern query technology is used to extract information. On the basis of the industrial neural network graph based on the SIPOC improved model, a general query mode that fully utilizes the graph structure is proposed, which can realize the general query of the industrial neural network graph. This pattern query method fully utilizes the structure of the SIPyOC-RC model, realizes the precise positioning of information, lays an important foundation for industrial enterprises to establish an accurate information query and analysis system, and promotes the development of industrial intelligent manufacturing.

[0108] The industrial neural network graph pattern query method based on the SIPyOC-RC model provided by the present invention has the following advantages: (1) The present invention designs a variety of query modes using graph structures for industrial neural network graphs based on the SIPOC improved model. The query has a high degree of freedom, good versatility and scalability, and is suitable for most intelligent query scenarios in industrial fields. (2) The present invention has good operability and can locally adjust the query mode according to the actual conditions in the industrial scene; (3) The present invention adopts an improved SIPOC model, couples the industrial neural network graph, considers the influence of various factors on the graph structure, and improves the intelligent query and analysis effect; (4) The present invention can not only be used for intelligent query, but also can integrate multiple query results to form result problem analysis.

[0109] The industrial neural network graph pattern query device based on the SIPyOC-RC model provided by the present invention is described below. The industrial neural network graph pattern query device based on the SIPyOC-RC model described below and the industrial neural network graph pattern query method based on the SIPyOC-RC model described above can be referenced to each other.

[0110] Figure 5Schematic diagram of the structure of the industrial neural network graph pattern query device based on the SIPyOC-RC model provided by the present invention. Figure 5 As shown, the device includes a construction module 10, a parsing module 20 and a query module 30, wherein: the construction module 10 is used to obtain a pre-constructed industrial neural network diagram based on the SIPyOC-RC model; wherein the industrial neural network diagram includes a mind map and an instance map, the mind map includes mind map nodes and connection relationships, the instance map includes instance map nodes and connection relationships, and the instance map nodes are the results of instantiation of the mind map nodes; the parsing module 20 is used to obtain a user query request, and parse the user query request into an intent format based on the SIPyOC-RC model; the query module 30 is used to perform a pattern query on the industrial neural network diagram according to the intent format to obtain a pattern query result.

[0111] The industrial neural network graph pattern query device based on the SIPyOC-RC model provided by the present invention obtains a pre-built industrial neural network graph based on the SIPyOC-RC model, obtains a user query request, parses the user query request into an intent format based on the SIPyOC-RC model, performs a pattern query on the industrial neural network graph according to the intent format, and obtains a pattern query result, thereby realizing a pattern query of the industrial neural network graph based on the SIPyOC-RC model, and can realize fast, efficient and accurate one-graph universal query, thereby facilitating real-time and accurate query of enterprise information.

[0112] Figure 6 An example of a physical structure diagram of an electronic device is shown in FIG. Figure 6 As shown, the electronic device may include: a processor 610, a communication interface 620, a memory 630 and a communication bus 640, wherein the processor 610, the communication interface 620 and the memory 630 communicate with each other through the communication bus 640. The processor 610 may call the logic instructions in the memory 630 to execute the industrial neural network graph pattern query method based on the SIPyOC-RC model, the method comprising: obtaining a pre-built industrial neural network graph based on the SIPyOC-RC model; wherein the industrial neural network graph comprises a mind map and an instance map, the mind map comprises mind map nodes and connection relationships, the instance map comprises instance map nodes and connection relationships, and the instance map nodes are the instantiation results of the mind map nodes; obtaining a user query request, parsing the user query request into an intent format based on the SIPyOC-RC model; performing a pattern query on the industrial neural network graph according to the intent format to obtain a pattern query result.

[0113] In addition, the logic instructions in the above-mentioned memory 630 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0114] On the other hand, the present invention also provides a computer program product, which includes a computer program, and the computer program can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the industrial neural network graph pattern query method based on the SIPyOC-RC model provided by the above methods, and the method includes: obtaining a pre-built industrial neural network graph based on the SIPyOC-RC model; wherein the industrial neural network graph includes a mind map and an instance map, the mind map includes mind map nodes and connection relationships, the instance map includes instance map nodes and connection relationships, and the instance map nodes are the results of instantiation of the mind map nodes; obtaining a user query request, parsing the user query request into an intent format based on the SIPyOC-RC model; performing a pattern query on the industrial neural network graph according to the intent format to obtain a pattern query result.

[0115] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to execute the industrial neural network graph pattern query method based on the SIPyOC-RC model provided by the above-mentioned methods, the method comprising: obtaining a pre-built industrial neural network graph based on the SIPyOC-RC model; wherein the industrial neural network graph comprises a mind map and an instance map, the mind map comprises mind map nodes and connection relationships, the instance map comprises instance map nodes and connection relationships, and the instance map nodes are the results of instantiation of the mind map nodes; obtaining a user query request, parsing the user query request into an intent format based on the SIPyOC-RC model; performing a pattern query on the industrial neural network graph according to the intent format to obtain a pattern query result.

[0116] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0117] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0118] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An industrial neural network graph pattern query method based on the SIPyOC-RC model, characterized in that: include: Obtain a pre-built industrial neural network diagram based on the SIPyOC-RC model; wherein the industrial neural network diagram includes a mind map and an instance map, the mind map includes mind map nodes and connection relationships, the instance map includes instance map nodes and connection relationships, and the instance map nodes are the instantiation results of the mind map nodes; Obtaining a user query request, and parsing the user query request into an intent format based on a SIPyOC-RC model; A pattern query is performed on the industrial neural network graph according to the intent format to obtain a pattern query result.

2. The industrial neural network graph pattern query method based on the SIPyOC-RC model according to claim 1 is characterized in that: Before obtaining the pre-built industrial neural network diagram based on the SIPyOC-RC model, the method further includes: Decompose the industrial production process into tasks and construct activity nodes in the mind map; Connecting each of the mind map activity nodes according to the preceding and following relationships and the inclusion relationship of each of the mind map activity nodes; Analyzing the resources, control conditions, input objects and output objects of each of the mind map activity nodes, respectively establishing corresponding resource nodes, control condition nodes, input object nodes and output object nodes, and establishing connections between the resource nodes, the control condition nodes, the input object nodes and the output object nodes and the mind map activity nodes, thereby obtaining a mind map; In the actual production process, the mind map nodes in the mind map are instantiated to obtain instance nodes and connection relationships between the instance nodes, thereby obtaining an instance map.

3. The industrial neural network graph pattern query method based on the SIPyOC-RC model according to claim 1 is characterized in that: The query mode includes a single node query; corresponding to the single node query, the intent format includes information on the query mode, query node, node attributes and query intent; The performing pattern query on the industrial neural network diagram according to the intent format includes: performing query pattern routing according to the query pattern, routing to a single-node query pattern, and querying the node attributes of the query node according to the query intent.

4. The industrial neural network graph pattern query method based on the SIPyOC-RC model according to claim 1 is characterized in that: The query mode includes two-node query and three-node or more query. Corresponding to the two-node query and three-node or more query, the intent format includes information on the query mode, query nodes, node attributes, query intent, and relationships between nodes; The performing a pattern query on the industrial neural network graph according to the intent format includes: Query mode routing is performed according to the query mode, routed to a corresponding two-node query mode or a three-node or more query mode, and the node attributes of the query node are queried according to the query intent and the relationship between the nodes.

5. The industrial neural network graph pattern query method based on the SIPyOC-RC model according to claim 4 is characterized in that: The two-node query includes querying the second instance graph activity node according to the first instance graph activity node; Correspondingly, querying the node attribute of the query node according to the query intent and the relationship between the nodes includes: Locating the first instance graph activity node; Taking the first instance graph activity node as the starting node, based on the structure of the instance graph and the relationship between the nodes, screening the endpoints that satisfy the relationship between the nodes in the instance graph, and taking the endpoints as the second instance graph activity nodes; The node attribute information of the active node of the second instance graph is acquired and returned according to the query intent.

6. The industrial neural network graph pattern query method based on the SIPyOC-RC model according to claim 4 is characterized in that: The two-node query includes querying the first mind map activity node according to the third instance map activity node; Correspondingly, querying the node attribute of the query node according to the query intent and the relationship between the nodes includes: Locating the third instance graph activity node; Acquire a second mind map activity node corresponding to the third instance map activity node according to the structure of the industrial neural network map; Taking the second mind map activity node as the starting node, and based on the structure of the mind map and the relationship between the nodes, screening the end point in the mind map that satisfies the relationship between the nodes, and taking the end point as the first mind map activity node; The node attribute information of the first mind map activity node is acquired and returned according to the query intention.

7. The industrial neural network graph pattern query method based on the SIPyOC-RC model according to claim 4 is characterized in that: The two-node query includes querying the fourth mind map activity node according to the third mind map activity node; Correspondingly, querying the node attribute of the query node according to the query intent and the relationship between the nodes includes: Locating the third mind map activity node; Taking the third mind map activity node as the starting node, and based on the structure of the mind map and the relationship between the nodes, selecting the end point in the mind map that satisfies the relationship between the nodes, and taking the end point as the fourth mind map activity node; The node attribute information of the fourth mind map activity node is acquired and returned according to the query intention.

8. The industrial neural network graph pattern query method based on the SIPyOC-RC model according to claim 4 is characterized in that: The two-node query includes querying the fifth mind map activity node according to the first instance graph node; wherein the first instance graph node includes an input object node, an output object node, a resource node or a control condition node; Correspondingly, querying the node attribute of the query node according to the query intent and the relationship between the nodes includes: Locating the first instance graph node; Searching in the instance graph for a fourth instance graph active node connected to the first instance graph node; Searching for a fifth mind map activity node corresponding to the fourth instance map activity node according to the structure of the industrial neural network; Acquire a second node set consisting of the topmost nodes in the first node set consisting of the active nodes of the fifth mind map; Acquire a third node set consisting of instantiated nodes of the second node set; Search the instance graph for the endpoint of the fourth instance graph activity node through the instance graph activity node to the third node set, obtain the node attribute of the endpoint according to the query intent, and return the node attribute as the node attribute of the fifth mind map activity node.

9. The industrial neural network graph pattern query method based on the SIPyOC-RC model according to claim 4 is characterized in that: The two-node query includes querying the first mind map node according to the second instance map node; wherein the second instance map node and the first mind map node include input object nodes, output object nodes, resource nodes or control condition nodes; Correspondingly, querying the node attribute of the query node according to the query intent and the relationship between the nodes includes: Locating the second instance graph node; Searching in the instance graph for a fifth instance graph active node connected to the second instance graph node; Searching for a sixth mind map activity node corresponding to the fifth instance map activity node according to the structure of the industrial neural network; Acquire a fifth node set consisting of the topmost nodes in the fourth node set consisting of the sixth mind map activity nodes; Acquire a sixth node set consisting of instantiated nodes of the fifth node set; Searching the instance graph for a terminal point from the fifth instance graph activity node to the sixth node set via the instance graph activity node; According to the query intent, information on the node attributes of the node connected to the endpoint and having the same node type as the first mind map node is obtained and returned.

10. The industrial neural network graph pattern query method based on the SIPyOC-RC model according to claim 1 is characterized in that: The query intent includes node attribute description information query intent and node attribute statistical information query intent.

11. An industrial neural network graph pattern query device based on the SIPyOC-RC model, characterized in that: include: A construction module is used to: obtain a pre-constructed industrial neural network diagram based on the SIPyOC-RC model; wherein the industrial neural network diagram includes a mind map and an instance map, the mind map includes mind map nodes and connection relationships, the instance map includes instance map nodes and connection relationships, and the instance map nodes are the instantiation results of the mind map nodes; A parsing module, used to: obtain a user query request, and parse the user query request into an intent format based on a SIPyOC-RC model; The query module is used to perform a pattern query on the industrial neural network diagram according to the intent format to obtain a pattern query result.

Images