Automatic construction method and system for knowledge graph in PCBA (printed circuit board assembly) welding process field
By automatically building a PCBA welding process knowledge graph, the problem of low efficiency in PCBA welding process knowledge management in the existing technology is solved, timely update and accurate transfer of knowledge is achieved, and the efficiency and accuracy of process design is improved.
Patent Information
- Application Number
- CN202510357335.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-03-25
- Publication Date
- 2025-06-13
AI Technical Summary
The existing PCBA welding process knowledge management has problems such as incomplete information, lagging updates, and weak knowledge correlation, resulting in low process design efficiency.
The method of automatically constructing the PCBA welding process knowledge graph is adopted, and by sorting out multimodal data sources, building standard entity sets and relationship sets, the effective organization and management of knowledge is achieved.
It realizes the timely update and accurate transmission of PCBA welding process knowledge, improves data processing efficiency, promotes the unified organization and management of knowledge, and enhances the efficiency and accuracy of process design.
Smart Images
Figure CN120144786A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the field of PCBA (Printed Circuit Board Assembly) soldering process, and specifically relates to a method and system for automatically constructing a knowledge graph in the field of PCBA soldering process. Background Art
[0002] The input of PCBA soldering process knowledge is usually presented in multimodal forms such as text, pictures, and tables. The data volume is huge and the relationships are complex. The relevance between knowledge is weak, the unified organization and management of knowledge are difficult, and the process knowledge cannot be updated in a timely manner. There is a lack of a knowledge model with tight associations and an effective knowledge update method.
[0003] In the process of formulating traditional PCBA soldering processes, process knowledge is mostly accumulated, transmitted, and updated manually. There are often problems such as incomplete information and lagging knowledge updates. It relies on the personal knowledge and experience of process planners. After a large number of case retrievals and data consultations in relational databases or traditional knowledge bases, a relatively feasible PCBA soldering process is designed. This process preparation method requires multiple transfers between tables, and the information retrieval speed is slow, resulting in long working hours and low work efficiency for process designers. The incremental construction technology of knowledge graphs can effectively integrate multimodal data generated in production, automatically construct a knowledge graph of PCBA soldering process, and ensure the timely update and accurate transmission of process knowledge. This method can effectively extract and manage a large amount of process data. Summary of the Invention
[0004] Object of the Invention: The present invention provides a method and system for automatically constructing a knowledge graph in the field of PCBA soldering process, which can automatically extract PCBA soldering process data, construct a standard entity set and relationship set, realize the effective organization and management of multimodal process knowledge, and complete the construction of the PCBA soldering process knowledge graph.
[0005] Technical Solution: The automatic construction method of a knowledge graph in the field of PCBA soldering process according to the present invention specifically includes the following steps:
[0006] (1) Sort out the data sources of the knowledge graph for PCBA soldering process design. The data sources include process design guidelines, soldering process cards of examples, component specifications, equipment specifications, tooling data documents, process standard documents, and fault analysis reports; classify the obtained process design field knowledge into process route knowledge, component knowledge, and process parameter knowledge;
[0007] (2) Sort out the hierarchical structure among various concepts of the classified knowledge, perform knowledge fusion in the form of "entity - relationship - entity", and construct the ontology model and semantic framework of the PCBA soldering process knowledge graph;
[0008] (3) Construct the ontology model of the knowledge graph, extract data from multi-modal data through regular matching, OCR recognition and entity extraction technologies, and automatically construct the entity triple element information of the PCBA soldering process knowledge to form an entity set;
[0009] (4) Based on the semantic framework of the ontology model, associate the complex relationships between the process knowledge entities, construct a complex network relationship, and form a relationship set of entity - entity mapping;
[0010] (5) Read the entity set and relationship set after data cleaning to construct the PCBA soldering process knowledge graph.
[0011] Further, the process design guidelines described in step (1) include the general procedures of PCBA soldering processing, the main manufacturing unit, the types of work, a brief description of the process content, a brief description of the work step content, equipment, tooling and tools, and precautions information;
[0012] The process assembly flow card includes the product name, the whole product drawing number, the process flow of the product, the process name, a brief description of the process content, equipment, tooling and tools, the operating worker, the inspection tooling, the part number and the part name;
[0013] The component specification includes the material code of the component, the manufacturer, the quality grade, the process structure, the packaging form, the pin material, the pin plating and the housing material;
[0014] The equipment specification includes the use of the production line, the asset number, the equipment name, the equipment model, the equipment manufacturer, the current production line, the typical index range of the equipment and the adjustable parameters of the equipment;
[0015] The tooling data document includes the tooling number and the quantity of tooling;
[0016] The process standard documents include the requirements for electronic components, the requirements for the forming of electronic components and the requirements for PCBs;
[0017] The fault analysis report includes the fault phenomenon, the fault cause and the fault solution.
[0018] Further, the implementation process of classifying the obtained process design domain knowledge in step (1) is as follows:
[0019] According to the process design guide and process assembly flow card, taking the process flow as the main line, removing duplicate, outdated or information unrelated to the process route, sorting out the process name, brief description of process content, main manufacturing unit and data of brief description of process content, and classifying them as process route knowledge; According to the component specification, removing duplicate, outdated or information unrelated to the component, sorting out the basic information of the component, category code, package attribute, pin attribute, structure attribute, body code, data of working steps and processes used, and classifying them as component knowledge; According to the equipment specification, tooling data, process standard documents and fault analysis reports, removing duplicate, outdated or information unrelated to process parameters, sorting out the equipment type, equipment name, equipment type, equipment capacity scope, equipment adjustment parameter scope, tooling type, tooling name, tooling number, characteristic parameters applicable to the tooling, fault phenomenon, fault cause and fault solution, and classifying them as process parameter knowledge.
[0020] Further, the implementation process of step (2) is as follows:
[0021] Define the entity concepts in the process route knowledge, sort out the hierarchical relationships between various entities, and construct the semantic framework between the entities;
[0022] Define the entity concepts in the component knowledge, sort out the hierarchical relationships between various entities, and construct the semantic framework between the entities;
[0023] Define the entity concepts in the process parameter knowledge, sort out the hierarchical relationships between various entities, and construct the semantic framework between the entities;
[0024] Identify the entities with the same content, fuse the entities with consistent content for multiple times, so that the same entity is uniquely represented in the PCBA soldering knowledge graph ontology model, and express the process route knowledge, component knowledge and process parameter knowledge after knowledge fusion in a graphical way to form the PCBA soldering knowledge graph ontology model.
[0025] Further, the knowledge graph ontology model in step (3) includes three main nodes, namely the assembly process flow, component library and process parameter library; Among them, under the assembly process flow node, there are product name, process type and process nodes; Under the component knowledge library node, there are component, body attribute, package attribute and process attribute nodes; Under the process parameter library node, there are equipment library, tooling library, fault library and process rule library nodes.
[0026] Further, the implementation process of step (3) is as follows:
[0027] According to the PCBA soldering knowledge graph ontology model, for structured and semi-structured data, use regular matching to extract some entity elements of process route knowledge, component knowledge and process parameter knowledge;
[0028] For unstructured data, some entity elements of component knowledge and process parameter knowledge are recognized by OCR. First, the PyMuPDF library is used to extract the picture tables in the Pdf, and then the extracted picture tables are binarized, denoised, and enhanced. Finally, the entity elements in the picture tables are extracted through Tesseract OCR.
[0029] Define the type and unique code of each entity to form an entity set. The content of the entity set contains three columns. The first column is the class to which the entity belongs, the second column is the unique code of the entity, and the third column is the extracted content of the entity.
[0030] Further, the implementation process of step (4) is as follows:
[0031] According to the semantic framework in the entity set and the ontology model, each entity is associated. Based on the semantic framework of the ontology model, the data in the entity set is extracted to form a seven-tuple structure of "entity type - entity ID - entity name | relationship type | entity type - entity ID - entity name". The relationship set contains seven columns, where the first three columns are the head entities, the fourth column is the relationship between entities, and the last three columns are the tail entities.
[0032] Further, the implementation process of step (5) is as follows:
[0033] Embed the entity triple elements in the entity set, the triple and seven-tuple structured data in the relationship set into the Neo4j knowledge graph in turn, instantiate the PCBA soldering process knowledge graph ontology model, and realize the automatic construction of the PCBA soldering process knowledge graph.
[0034] A knowledge graph system in the field of PCBA soldering process described in the present invention includes:
[0035] A file import module that imports files related to the PCBA soldering process, including soldering process guides, component specifications, equipment operation manuals, and process rule documents.
[0036] An entity extraction module that automatically extracts key entities in the PCBA soldering process files, including process information, component information, equipment information, tooling information, and fault information, and constructs them into a standardized entity set.
[0037] A semantic network construction module that extracts entities in the process and automatically constructs relationships according to the semantic framework in the ontology model and the entity set, constructs a semantic network of process knowledge, and forms a standardized relationship set.
[0038] A knowledge graph construction module that automatically constructs a Neo4j knowledge graph according to the structures of the entity set and the relationship set.
[0039] The system adopts a BS architecture, with the Vue framework used for the front end and the Django framework for the back end. The front end displays a Web page, and the back end uses a Neo4j graph database to store and process process knowledge.
[0040] Beneficial effects: Compared with the prior art, the beneficial effects of the present invention are as follows: The present invention can automatically extract multi-modal data of the PCBA soldering process. The automatic extraction significantly reduces manual intervention, remarkably improves the data processing efficiency, helps in the management of standardized process data, ensures consistency, and facilitates later traceability and knowledge accumulation. The present invention uses a knowledge graph as a fusion model for multi-modal process data and complex relationships to achieve the unified organization and effective management of PCBA soldering process knowledge. The knowledge graph has the advantages of strong process knowledge relevance, accurate process expression, high operability, and strong scalability. Description of the Drawings
[0041] Figure 1 is a flowchart of an automated construction method for a knowledge graph in the field of PCBA soldering process;
[0042] Figure 2 is the ontology model of the knowledge graph in the field of PCBA soldering process in an embodiment of the present invention;
[0043] Figure 3 is a flowchart for extracting multi-modal data entities in the field of PCBA soldering process in an embodiment of the present invention;
[0044] Figure 4 is a schematic diagram of the structure and implementation of a knowledge graph system in the field of PCBA soldering process proposed by the present invention. Detailed Embodiments
[0045] The present invention will be further described in detail below with reference to the accompanying drawings.
[0046] As Figure 1 shown, an incremental construction method for a knowledge graph for intelligent generation of PCBA soldering process specifically includes the following steps:
[0047] Step 1: The first data source for the knowledge graph in the field of PCBA soldering process design is the process design guide, which contains information such as general procedures for PCBA soldering processing, main manufacturing units, job types, brief descriptions of process contents, brief descriptions of work steps, equipment, tooling, tools, and precautions.
[0048] The second data source is the process assembly flow card containing examples, which includes information such as product name, overall drawing number, product process flow, process name, brief description of process content, equipment, tooling, tools, operating workers, inspection tooling, part numbers, and part names.
[0049] The third data source is the component specification, which contains information such as the material code, manufacturer, quality grade, process structure, packaging form, pin material, pin plating, and housing material of the component.
[0050] The fourth data source is the equipment specification, which contains information such as the production line usage, asset number, equipment name, equipment model, equipment manufacturer, current production line, typical equipment index range, and adjustable equipment parameters.
[0051] The fifth data source is the tooling data, which contains information such as the tooling number and the quantity of tooling.
[0052] The sixth data source is the process standard documents, which contain information such as requirements for electronic components, forming requirements for electronic components, and requirements for PCBs.
[0053] The seventh data source is the fault analysis report, which contains information such as the fault phenomenon, fault cause, and fault solution.
[0054] Based on the process design guide and the process assembly flow card, with the process flow as the main line, remove duplicate, outdated, or information irrelevant to the process route, sort out the process name, brief description of the process content, main manufacturing unit, and data of the brief description of the process content, and classify them as process route knowledge; based on the component specification, remove duplicate, outdated, or information irrelevant to the component, sort out the basic information of the component, category code, packaging attribute, pin attribute, structure attribute, body code, and data of the working steps and processes used, and classify them as component knowledge; based on the equipment specification, tooling data, process standard documents, and fault analysis report, remove duplicate, outdated, or information irrelevant to the process parameters, sort out the equipment type, equipment name, equipment type, equipment capability scope, equipment adjustment parameter scope, tooling type, tooling name, tooling number, characteristic parameters applicable to the tooling, fault phenomenon, fault cause, and fault solution, and classify them as process parameter knowledge.
[0055] Step 2: For the classified knowledge, sort out the hierarchical structure among various concepts, perform knowledge fusion in the form of "entity - relationship - entity", and construct the ontology model and semantic framework of the PCBA soldering process knowledge graph.
[0056] Define the entity concepts in the process route knowledge, including entity concepts such as product name, whole product drawing number, product process flow, process, process name, brief description of the process content, equipment, tooling and tools, operation types, inspection tooling, part number, and part name, sort out the hierarchical relationships among each entity, and construct the semantic framework among the entities, such as: the semantic relationship between the process name class node and the brief description of the process content class node is "the brief description of the process content is".
[0057] Define the entity concepts in component knowledge, including entity concepts such as basic component information, category code, package attributes, pin attributes, structure attributes, body code, and working steps and processes. Sort out the hierarchical relationships among entities and construct the semantic framework between entities. For example, the relationship between the process attribute class node and the electrostatic level class node is "the electrostatic level is".
[0058] Define the entity concepts in process parameter knowledge, including equipment manuals, tooling data, process standard documents, and fault analysis reports. Sort out entities such as equipment type, equipment name, equipment type, equipment capability scope, equipment adjustment parameter scope, tooling type, tooling name, tooling number, characteristic parameters applicable to the tooling, fault phenomena, fault causes, and fault solutions. Sort out the hierarchical relationships among entities and construct the semantic framework between entities. For example, the relationship between the equipment class node and the equipment adjustable parameter class node is "the equipment adjustable parameter is".
[0059] Identify entities with the same content, fuse entities with consistent content, and make the same entity be uniquely represented in the PCBA soldering knowledge graph ontology model. Express the process route knowledge, component knowledge, and process parameter knowledge after knowledge fusion in a graphical way to form the PCBA soldering knowledge graph ontology model.
[0060] Step 3: According to the knowledge graph ontology model, extract data from multimodal data through regular matching, OCR recognition, and entity extraction technologies, and automatically construct the entity triple information of PCBA soldering process knowledge to form an entity set.
[0061] As Figure 2 shown, the PCBA soldering process domain knowledge graph ontology model describes the types of nodes and relationships and the hierarchical relationships among various concepts in the PCBA soldering process domain knowledge graph in the form of nodes and edges. In the knowledge graph ontology model, there are mainly three main nodes, namely the assembly process flow, component library, and process parameter library. Among them, under the assembly process flow node, there are nodes such as product name, process type, and process; under the component knowledge base node, there are nodes such as components, body attributes, package attributes, and process attributes; under the process parameter library node, there are nodes such as equipment library, tooling library, fault library, and process rule library.
[0062] As Figure 3 shown, the specific content of the PCBA soldering process domain multimodal data entity extraction flow chart includes: According to the PCBA soldering knowledge graph ontology model, for structured and semi-structured data, use regular matching to extract the entity elements of process route, components, and process parameters. For example, extract quality level data from the process guide document through the regular pattern "r'quality level.*? is " "[""]'".
[0063] According to the PCBA soldering knowledge graph ontology model, for unstructured data, some entity elements of component knowledge and process parameter knowledge are recognized by OCR. First, the PyMuPDF library is used to extract the picture tables in the Pdf, and then the extracted picture tables are binarized, denoised, and enhanced. Finally, the entity elements in the picture tables are extracted through Tesseract OCR.
[0064] Define the type and unique code of each entity to form an entity set. The content of the entity set contains three columns. The first column is the class to which the entity belongs, the second column is the unique code of the entity, and the third column is the extracted content of the entity. The structure of the entity set is like: "Process\tP001\tKit", "Process\tP002\tMake stencil", "Process\tP003\tPrepare", etc.
[0065] Step 4: According to the semantic framework in the entity set and the ontology model, associate each entity. Based on the semantic framework of the ontology model, extract the data in the entity set to form a seven-tuple structure of "entity type - entity ID - entity name|relationship type|entity type - entity ID - entity name"; the relationship set contains seven columns, where the first three columns are the head entities, the fourth column is the relationship between entities - entities, and the last three columns are the tail entities. The structure of the relationship set is like: "Step\tS003\tProcess step 1.2\tBrief description of the process step is\tStepContent\tC003\tKit the whole turnover box according to the production quantity", "Step\tS004\tProcess step 2.1\tBrief description of the process step is\tStepContent\tC004\tMake a stencil for printing solder paste, the stencil thickness is 0.12mm", etc.
[0066] Step 5: Embed the entity triple elements in the entity set, the triple and seven-tuple structured data in the relationship set into the Neo4j knowledge graph in turn, instantiate the PCBA soldering process knowledge graph ontology model, and realize the automatic construction of the PCBA soldering process knowledge graph.
[0067] The present invention also proposes a knowledge graph system in the field of PCBA soldering process, adopting a BS architecture, realizing the separation of the front and back ends through the Vue framework (front end) and the Django framework (back end). The front end displays a Web page, and the back end uses the Neo4j graph database to store and process process knowledge. Use the Neo4j graph database to store the entities, relationships, and their attribute data related to the PCBA soldering process, which is convenient for complex queries, dependency relationship modeling, and the display and reasoning of the knowledge graph. The specific four functional modules of the system are:
[0068] The file import module, whose main function is to import files related to the PCBA soldering process, including soldering process guides, component specifications, equipment operation manuals, process rule documents, etc.
[0069] The entity extraction module, whose main function is to automatically extract key entities from PCBA soldering process files, including process information, component information, equipment information, tooling information, fault information, etc., and construct a standardized entity set.
[0070] The semantic network construction module, whose main function is to extract entities in the process and automatically construct relationships based on the semantic framework and entity set in the ontology model, construct the semantic network of process knowledge, and form a standardized relationship set.
[0071] The knowledge graph construction module, whose main function is to automatically construct a Neo4j knowledge graph according to the structures of the entity set and the relationship set.
[0072] As Figure 4 shown, the knowledge graph system in the field of PCBA soldering process adopts a BS architecture, realizes the separation of the front and back ends through the Vue framework and the Django framework, displays the Web page at the front end, and stores and processes process knowledge using the Neo4j graph database at the back end. This system includes four functional modules: file import module, entity extraction module, semantic network construction module, and knowledge graph construction module.
[0073] On the Web page side, users upload files related to the PCBA soldering process, such as process guides, component specifications, equipment operation manuals, etc. through the file import module. The system automatically extracts key entities (such as process information, component information, equipment information, etc.) from the files through the entity extraction module and generates a standardized entity set. Then, the semantic network construction module extracts the entity relationships in the process according to the ontology model, constructs the semantic network of process knowledge, and forms a standardized relationship set. Finally, the knowledge graph construction module automatically constructs a knowledge graph in Neo4j according to the structures of the entity set and the relationship set to support querying, reasoning, and displaying.
[0074] The present invention has been described in detail above in conjunction with specific embodiments, but these descriptions should not be construed as limiting the present invention. Those skilled in the art understand that without departing from the spirit and scope of the present invention, various equivalent replacements, modifications, or improvements can be made to the technical solutions and their implementation manners of the present invention, and these all fall within the scope of the present invention. The protection scope of the present invention is subject to the appended claims.
Claims
1. A method for automatically constructing a knowledge graph in the field of PCBA assembly and welding process, characterized in that: The following steps are involved: (1) Sort out the data sources of the knowledge graph for PCBA assembly and soldering process design, including process design guidelines, example assembly and soldering process flow cards, component manuals, equipment manuals, tooling data documents, process standard documents, and failure analysis reports; classify the acquired process design domain knowledge into process route knowledge, component knowledge, and process parameter knowledge; (2) Sort out the hierarchical structure of various concepts in the classified knowledge, integrate the knowledge in the form of "entity-relationship-entity", and construct the PCBA assembly and welding process knowledge graph ontology model and semantic framework; (3) Construct a knowledge graph ontology model, extract multimodal data through regular matching, OCR recognition and entity extraction technology, and automatically construct the entity three-element information of PCBA assembly and welding process knowledge to form an entity set; (4) Based on the semantic framework of the ontology model, the complex relationships between process knowledge entities are associated, complex network relationships are constructed, and a relationship set of entity-entity mapping is formed; (5) Read the entity set and relationship set after data cleaning to build the PCBA assembly and soldering process knowledge graph.
2. The method for automatically constructing a knowledge graph in the field of PCBA welding process according to claim 1, characterized in that: The process design guide in step (1) includes the general process of PCBA assembly and welding, main manufacturing unit, type of work, brief description of process content, brief description of work step content, equipment, tooling and tools, and precautions; The process assembly flow card includes the product name, the whole piece drawing number, the product process flow, the process name, a brief description of the process content, equipment, tooling and tools, operating work, inspection tooling, component number and component name; The component specification includes the material code, manufacturer, quality grade, process structure, packaging form, pin material, pin plating and shell material of the component; The equipment manual includes the purpose of the production line, asset number, equipment name, equipment model, equipment manufacturer, current production line, typical equipment indicator range and equipment adjustable parameters; The tooling data document includes the tooling number and tooling quantity; The process standard documents include electronic component requirements, electronic component forming requirements and PCB requirements; The fault analysis report includes the fault phenomenon, fault cause and fault solution.
3. The method for automatically constructing a knowledge graph in the field of PCBA welding process according to claim 1, characterized in that: The process of classifying the acquired process design domain knowledge in step (1) is as follows: According to the process design guide and process assembly flow card, with the process flow as the main line, remove duplicate, outdated or irrelevant information, sort out the process name, process content brief description, main manufacturing unit and process content brief description data, and classify them as process route knowledge; according to the component manual, remove duplicate, outdated or irrelevant information, sort out the basic information of components, category code, package attributes, pin attributes, structure attributes, body code, and process step data, and classify them as component knowledge; Based on the equipment manual, tooling information, process standard documents and fault analysis reports, remove duplicate, outdated or information irrelevant to process parameters, sort out the equipment type, equipment name, equipment type, equipment capacity range, equipment adjustment parameter range, tooling type, tooling name, tooling number, tooling applicable characteristic parameters, fault phenomenon, fault cause and fault solution, and classify them as process parameter knowledge.
4. The method for automatically constructing a knowledge graph in the field of PCBA welding process according to claim 1, characterized in that: The implementation process of step (2) is as follows: Define entity concepts in process route knowledge, sort out the hierarchical relationships between entities, and build a semantic framework between entities; Define entity concepts in component knowledge, sort out the hierarchical relationships between entities, and build a semantic framework between entities; Define the entity concepts in process parameter knowledge, sort out the hierarchical relationships between entities, and build a semantic framework between entities; Determine entities with the same content, merge multiple entities with the same content, make the same entity uniquely represented in the PCBA assembly knowledge graph ontology model, express the process route knowledge, component knowledge and process parameter knowledge after knowledge fusion in a graphical way, and form a PCBA assembly knowledge graph ontology model.
5. The method for automatically constructing a knowledge graph in the field of PCBA welding process according to claim 1, characterized in that: The knowledge graph ontology model in step (3) includes three main nodes, namely, assembly process flow, component library and process parameter library; wherein, the assembly process flow node includes product name, process type and process node; the component knowledge base node includes component, ontology attribute, package attribute and process attribute node; the process parameter library node includes equipment library, tooling library, fault library and process rule library node.
6. The method for automatically constructing a knowledge graph in the field of PCBA welding process according to claim 1, characterized in that: The implementation process of step (3) is as follows: Based on the PCBA assembly knowledge graph ontology model, regular matching is used to extract some entity elements of process route knowledge, component knowledge and process parameter knowledge for structured and semi-structured data; For unstructured data, OCR is used to identify some entity elements of component knowledge and process parameter knowledge; first, the PyMuPDF library is used to extract the image table in the PDF, and then the extracted image table is binarized, denoised and enhanced, and finally the entity elements in the image table are extracted through Tesseract OCR; Define the type and unique code for each entity to form an entity set. The content of the entity set contains three columns, where the first column is the class to which the entity belongs, the second column is the unique code of the entity, and the third column is the extracted content of the entity.
7. The method for automatically constructing a knowledge graph in the field of PCBA welding process according to claim 1, characterized in that: The implementation process of step (4) is as follows: According to the semantic framework of the entity set and the ontology model, the entities are associated. Based on the semantic framework of the ontology model, the data in the entity set is extracted to form a seven-tuple structure of "entity type-entity ID-entity name|relationship type|entity type-entity ID-entity name"; the relationship set contains seven columns, of which the first three columns are head entities, the fourth column is the relationship between entities, and the last three columns are tail entities.
8. The method for automatically constructing a knowledge graph in the field of PCBA welding process according to claim 1, characterized in that: The implementation process of step (5) is as follows: The entity triples in the entity set and the triples and septuples in the relationship set are sequentially embedded into the Neo4j knowledge graph, and the PCBA assembly and welding process knowledge graph ontology model is instantiated to realize the automatic construction of the PCBA assembly and welding process knowledge graph.
9. A PCBA assembly and welding process domain knowledge graph system using the method according to any one of claims 1 to 8, characterized in that: include: The file import module imports files related to PCBA soldering process, including soldering process guide, component manual, equipment operation manual, and process rule documents; The entity extraction module automatically extracts key entities in the PCBA assembly and welding process files, including process information, component information, equipment information, tooling information, and fault information, and constructs them into a standardized entity set; The semantic network construction module extracts entities in the process and automatically constructs relationships based on the semantic framework and entity set in the ontology model, constructs a semantic network of process knowledge, and forms a standardized relationship set; The knowledge graph construction module automatically builds the Neo4j knowledge graph based on the structure of entity sets and relationship sets; The system adopts BS architecture, Vue framework is used in the front end, and Django framework is used in the back end. The front end displays web pages, and the back end adopts Neo4j graphic database to store and process process knowledge.