A method for constructing a process design resource model based on knowledge graph

Through knowledge graph technology, the process design resource model is constructed and the part feature similarity is calculated, which solves the problems of low efficiency and insufficient resource sharing in traditional process design, realizes the reuse of process knowledge and unified management of resources, and improves design efficiency and sharing Spend.

CN114626190BActive Publication Date: 2024-11-05TIANJIN UNIV
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202111500782.1
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2021-12-09
Publication Date
2024-11-05
Estimated Expiration
2041-12-09

AI Technical Summary

Technical Problem

Traditional process design has low efficiency, inconsistency, unstable quality, and dispersed design data, resulting in insufficient cooperation and resource information sharing among enterprises, making it difficult to achieve reuse of process knowledge and unified management of resources.

Method used

Using a multi-level modeling method based on knowledge graph, the process design resource knowledge graph is constructed through process route splitting, part feature modeling and processing equipment resource modeling, and the process design resource reuse and resource cross-border by calculating part feature similarity is realized. Departmental and cross-organizational sharing.

Benefits of technology

It improves the reuse rate of process design knowledge and the sharing of resources among enterprises, solves the problems of fragmentation of process design knowledge and resource dispersion, and improves design efficiency and consistency.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114626190B_ABST
    Figure CN114626190B_ABST
Patent Text Reader

Abstract

The present invention relates to a method for constructing a process design resource model based on a knowledge graph, comprising the following steps: a first step, modeling process design resources for a process route: modeling process design resources based on the process route, splitting the process route into a series of processes, and splitting each process into a series of steps, taking the step as the basic unit of a process design instance, and establishing a step meta-knowledge model; classifying the step meta-knowledge model according to processing methods, with each category being in a parallel relationship, and each category can be further classified according to processing accuracy; a second step, modeling part features; a third step, modeling processing equipment resources; a fourth step, using knowledge graph technology to extract the relationship between nodes; a fifth step, constructing a process design resource knowledge graph; and a sixth step, recommending processing steps and processing equipment.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the technical field of knowledge management in manufacturing industry, and specifically is a method for constructing a process design resource model based on a knowledge graph. Background Art

[0002] Under the requirements of intelligent manufacturing, higher requirements are placed on the design resource organization and information sharing forms of enterprises. Traditional process design has problems such as low process design efficiency, poor consistency, unstable quality, discrete design data, and difficulty in achieving optimization. This leads to insufficient cooperation and resource information sharing between enterprises. The various departments within the enterprise do not have a comprehensive grasp of process design resource information and do not share it in a timely manner. The process design of customized products is cumbersome, and it is difficult to utilize existing process routes and the reuse rate of process knowledge.

[0003] In response to the above problems, this patent designs a multi-level modeling method for process design resources based on knowledge graphs, and achieves the purpose of reusing process design knowledge by calculating the similarity of part features. This patent aims to explore a unified modeling method for design resources between enterprises, and construct a process design resource knowledge graph based on the modeling method to improve the sharing rate and efficiency of enterprise process design. Summary of the invention

[0004] The purpose of the present invention is to provide a method for constructing a process design resource model. By adopting the method of the present invention, process design knowledge can be converted into structured data, and a unified modeling method is developed for the heterogeneity, multi-source and knowledge reusability of process design resources, and the purpose of process knowledge reuse is achieved by finding similar part features. The technical solution is as follows:

[0005] A method for constructing a process design resource model based on a knowledge graph, wherein the process design process includes: determining the production program for parts processing, describing the specifications and output of the product; formulating the production process of the parts, describing the production process flow, the specifications and quantity of the main production equipment and auxiliary equipment; calculating the quantity of raw materials and semi-finished products required for part production. The construction of the process design resource model includes the following steps:

[0006] The first step is to model the process design resources for the process route.

[0007] The process design resource modeling is carried out based on the process route. The process route is divided into a series of processes, and each process is divided into a series of steps. The step is used as the basic unit of the process design instance, and the step meta-knowledge model is established.

[0008] The process step meta-knowledge model is classified according to the processing method. Each category is in parallel relationship. Each category can be further classified according to the processing accuracy.

[0009] The process-step meta-knowledge is mainly divided into the basic concepts, attributes and other process equipment of process design knowledge. The process-step meta-SKU is represented by the following formula:

[0010] SK U ={SK C ,SK A ,PE}

[0011] SK C The basic concept of the process step element, SK A Indicates the parameters of the process step element. PE represents the processing equipment used in the process step and sets its main properties:

[0012] (1) Basic concepts include processing features, parts, and processing methods. Processing features refer to the basic features of the parts processed in this process step; processing methods refer to the processing methods used in this process step; and parts refer to the parts processed in this process step.

[0013] (2) Regarding the process step parameters, they include process step ID, cutting speed, spindle speed, cutting depth, feed rate, and working time. The process step ID is the unique identifier of the process step and is used to number the process step. The cutting speed, spindle speed, cutting depth, and feed rate are the basic parameters in the machining process. The working time is the time consumed by the process step, which is the sum of the mobile working time and the auxiliary working time.

[0014] (3) Regarding processing equipment, it is divided into cutting equipment, process equipment, testing equipment, and heat treatment equipment;

[0015] The second step is part feature modeling

[0016] The part features are described by the following formula:

[0017] F={F Man ,F Mat ,F Sha}

[0018] F Man Represents management features, which mainly include relevant attribute information of parts, including part drawing number, part name, part number, and part type;

[0019] F Mat Represents material characteristics, including relevant property information of the material, including material type, material grade, and material performance;

[0020] F Sha Represents shape features, which mainly reflect the shape structure of parts. This feature includes design feature information and manufacturing feature information;

[0021] Using the already constructed part feature model, construct the part feature vector:

[0022] F ea =[F c ,F s ,R,A cc ,M]

[0023] F c Represents the part feature type, F s Represents the dimensional parameters of the part features, R represents the roughness requirement, A cc represents the accuracy requirement, and M represents the material parameter;

[0024] Part feature parameter F s There are different property settings for different part types;

[0025] Material parameters, as the basic characteristics of the material that makes up a part, are represented by the following vectors:

[0026] M=[M T ,M g ,M EM ,M YS ]

[0027] M T Represents the material type, M g Represents the material grade, M EM Represents the elastic modulus of the material, M YS Represents the yield strength of the material;

[0028] The third step is to model the processing equipment resources

[0029] For the processing equipment of the work step meta-knowledge model, since it plays a relatively important role in processing and manufacturing, the processing equipment resources are modeled in combination with the work step meta-knowledge model; processing equipment resources mainly refer to the mechanical equipment required to process a part, which is the manufacturing resource facing the bottom layer of the manufacturing system, and it mainly includes machine tools, cutting tools, fixtures, measuring tools, etc.

[0030] A more detailed classification of processing equipment:

[0031] (1) For cutting equipment, it is further divided into lathes, milling machines, planers, grinders, boring machines, drilling machines, CNC machine tools, machining centers and others;

[0032] (2) For process equipment, it is further divided into tools, fixtures, gauges and molds;

[0033] (3) For testing equipment, it is further divided into displacement testing equipment, speed testing equipment, temperature testing equipment, humidity testing equipment, pressure testing equipment and other testing equipment;

[0034] (4) Heat treatment equipment is further divided into carburizing furnace, vacuum furnace, tempering furnace, annealing furnace, quenching furnace and other heat treatment equipment;

[0035] The processing equipment resources are formally described from four attributes: basic attributes, state attributes, functional attributes, and service attributes. The basic attributes are used to describe the basic situation of a resource, including resource ID, resource name, resource unit, location, contact information, and resource nature.

[0036] The status attribute is used to describe the current status of the resource, including idle, not fully loaded, fully loaded, and under maintenance;

[0037] Functional attributes are used to describe the functional conditions of the services provided by a resource, including service objects, processing accuracy, processing size, processing shape, processing method, input and output;

[0038] Service attributes are used to describe the specific conditions of the services provided by a resource, including the service mode and resource quantity;

[0039] The fourth step is to use knowledge graph technology to extract the relationship between nodes;

[0040] (1) In the process-step knowledge model, the nodes include the process-step concept, the part itself, and the materials that make up the part; in the processing resource model, the nodes include the processing equipment, the workshop to which the processing equipment belongs, the company, and the region;

[0041] (2) There is a "use" relationship between the process step element concept and the processing equipment; there is a "processing" relationship between the process step element concept and the parts; there is a "location" relationship between the processing equipment and the workshop, company and region to which it belongs; there is a "composition" relationship between the parts and the materials of the parts;

[0042] Step 5: Build a knowledge graph of process design resources

[0043] Using the constructed process design resource model and the extracted nodes and relationships, the existing process design knowledge is processed and represented as a data structure that can be recognized by computers, and a process design resource knowledge graph is constructed to facilitate subsequent storage and retrieval;

[0044] The sixth step is to use the constructed model and process design knowledge graph, and calculate the similarity between the new part features and the existing part features to recommend processing steps and processing equipment.

[0045] Among them, the part feature similarity calculation method can be:

[0046] For the calculation of part feature similarity, since the part feature vector contains non-numeric parameters and numerical parameters,

[0047] Similarity calculation method for non-numeric parameters in part feature types: Set the basic part feature set FT. If feature x and feature y do not belong to the set, or the two features do not belong to the same category at the basic feature level, then their similarity is 0. If both belong to the same category, build a Chinese word segmentation main dictionary and calculate its semantic similarity. The semantic similarity range is [0,1]. The closer to 1, the more similar the two are.

[0048] For numerical parameters, the cosine similarity calculation method is used for calculation. The cosine value of the angle between two vectors in the vector space is used as a measure of the size of the difference between two individuals. The closer the calculation result is to 1, the closer the angle is to 0 degrees, which means that the two vectors are more similar.

[0049] The total similarity between feature vector X and feature vector Y is the weighted sum of the similarities of the feature parameters of each part.

[0050] The present invention constructs a process design resource model to make the characterization of process design resources more accurate, and on this basis constructs a cross-departmental and cross-organizational process design resource knowledge graph to solve the problems of process design knowledge fragmentation and process design resource dispersion, greatly improving the reuse rate of process design knowledge and the sharing of process design resources among enterprises. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 Modeling roadmaps for process design resources;

[0052] Figure 2 It is the process route hierarchy diagram;

[0053] Figure 3 Design knowledge concept diagram for process

[0054] Figure 4 It is the resource attribute diagram of processing equipment;

[0055] Figure 5 This is an example of a machining process card;

[0056] Figure 6 For some Java programs;

[0057] Figure 7 Some nodes of Neo4j graph database. DETAILED DESCRIPTION

[0058] The technical solution of the present invention will be described clearly and completely below in conjunction with the accompanying drawings of the present invention. Obviously, the described embodiments are only 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.

[0059] 1. The present invention provides a process design resource modeling method based on knowledge graph.

[0060] The process design process mainly includes: determining the production program of parts processing, explaining the specifications and output of the products; formulating the production process of parts, explaining the production process flow, specifications and quantity of main production equipment and auxiliary equipment; calculating the quantity of raw materials and semi-finished products required for part production.

[0061] Therefore, the most important thing in the process design is the formulation of the process route. Figure 2 A hierarchical diagram of the process route is given.

[0062] 2. Modeling of process design resources based on process routes.

[0063] The process design resource modeling is carried out based on the process route, which is divided into a series of processes, and each process is divided into a series of steps. The step is the basic unit of process knowledge and is an inseparable knowledge unit in the process of process design knowledge expression. In order to facilitate the analysis and description of more complex processes, better organize production and calculate working hours, the step is used as the basic unit of the process design instance, and the step meta-knowledge model is established.

[0064] The process step meta-knowledge model is classified according to the processing method, and is divided into drilling, boring, turning, milling, etc., and each category is in parallel. Each category can be further classified according to the processing accuracy. For example, the turning category can be divided into finishing turning, semi-finishing turning, rough turning, etc., and the same is true for other categories.

[0065] According to the process manual and expert knowledge, combined with the basic characteristics of the process steps in the processing route, the process step meta-knowledge is mainly divided into the basic concepts of process design knowledge, attributes and other process equipment. Therefore, the process step meta-SKU can be represented by the following formula, and the relationship between the concepts is shown in the figure below: Figure 3 As shown,

[0066] SK U ={SK C ,SK A ,PE}

[0067] SK C The basic concept of the process step element, SK AIndicates the parameters of the process step element. PE represents the processing equipment used in the process step and sets its main properties:

[0068] (1) Regarding basic concepts, they mainly include processing features, parts, and processing methods. Processing features are the basic features of the parts processed in this process step, such as holes, planes, grooves, etc.; processing methods are the processing methods used in this process step, such as turning, milling, etc.; parts are the parts processed in this process step.

[0069] (2) Regarding the process step parameters, they mainly include process step ID, cutting speed, spindle speed, cutting depth, feed rate, and working time. The process step ID is the unique identifier of the process step, which is used to number the process step; cutting speed, spindle speed, cutting depth, and feed rate are the basic parameters in the machining process; working time is the time consumed by the process step, which is the sum of mobile working time and auxiliary working time.

[0070] (3) Regarding processing equipment, it is divided into cutting equipment, process equipment, testing equipment, and heat treatment equipment.

[0071] 3. Part feature modeling

[0072] Model and analyze the part entity in step 2. The part features can be described and expressed by the following formula:

[0073] F={F Man ,F Mat ,F Sha}

[0074] F Man Represents management features, which mainly include relevant attribute information of parts, such as part drawing number, part name, part number, and part type.

[0075] F Mat Represents material characteristics, mainly including relevant property information of the material, such as material type, material grade, material performance, etc.

[0076] F Sha It represents shape feature and mainly reflects the shape structure of the part. This feature has design feature information (geometric shape and related information, etc.) and manufacturing feature information (precision features, surface roughness, etc.).

[0077] Using the already constructed part feature model, construct the part feature vector:

[0078] F ea =[F c ,F s ,R,A cc ,M]

[0079] F c Represents the part feature type, Fs Represents the dimensional parameters of the part features, R represents the roughness requirement, A cc represents the accuracy requirement and M represents the material parameter.

[0080] Part feature parameter F s Different part types have different property settings, which can be expressed as:

[0081] F s ={F s1 ,F s2 ,...,F sn}

[0082] Material parameters, as the basic characteristics of the material that makes up a part, can be expressed by the following vectors:

[0083] M=[M T ,M g ,M EM ,M YS ]

[0084] M T Represents the material type, M g Represents the material grade, M EM Represents the elastic modulus of the material, M YS Represents the yield strength of the material.

[0085] 4. Modeling of processing equipment resources

[0086] For the processing equipment in the work step meta-knowledge model in step 2, since it plays a relatively important role in processing and manufacturing, the processing equipment resources are modeled in combination with the work step meta-knowledge model. Processing equipment resources mainly refer to the mechanical equipment required to process a part. They are manufacturing resources facing the bottom layer of the manufacturing system. They mainly include machine tools, cutting tools, fixtures, measuring tools, etc.

[0087] A more detailed classification of processing equipment:

[0088] (1) For cutting equipment, it is further divided into lathes, milling machines, planers, grinders, boring machines, drilling machines, CNC machine tools, machining centers and others;

[0089] (2) For process equipment, it is further divided into tools, fixtures, gauges and molds;

[0090] (3) For testing equipment, it is further divided into displacement testing equipment, speed testing equipment, temperature testing equipment, humidity testing equipment, pressure testing equipment and other testing equipment;

[0091] (4) Heat treatment equipment is further divided into carburizing furnace, vacuum furnace, tempering furnace, annealing furnace, quenching furnace and other heat treatment equipment.

[0092] The processing equipment resources are formally described from four attributes: basic attributes, state attributes, functional attributes, and service attributes. The processing equipment attribute MS can be described by the following formula:

[0093] MS={MS B ,MS Sta ,MS F ,MS Ser}

[0094] Among them, MS B Represents basic attributes, MS Sta Represents the state attribute, MS F Represents functional attributes, MS Ser Represents service attributes.

[0095] Specific property settings are as follows: Figure 4 As shown, the explanation is as follows:

[0096] (1) Basic attributes are used to describe the basic situation of a resource, including resource ID, resource name, resource unit, location, contact information, resource nature, etc.

[0097] (2) The state attribute is used to describe the current state of the resource, including idle, not fully loaded, fully loaded, and under maintenance;

[0098] (3) Functional attributes are used to describe the functional conditions of the services provided by a resource, including service objects, processing accuracy, processing size, processing shape, processing method, input and output, etc.

[0099] (4) Service attributes are used to describe the specific circumstances of the services provided by a resource, including service methods and resource quantities.

[0100] 5. Use knowledge graph technology to extract the relationships between nodes.

[0101] A knowledge graph is a knowledge base based on a semantic network, that is, a knowledge base with a directed graph structure, where the nodes of the graph represent entities or concepts, and the edges of the graph represent various semantic relationships between entities / concepts. Therefore, constructing a knowledge graph requires extracting nodes and relationships in the process design resource model.

[0102] (1) In the process-step knowledge model, the nodes include the process-step concept, the part itself, and the materials that make up the part; in the processing resource model, the nodes include the processing equipment, the workshop to which the processing equipment belongs, the company, and the region.

[0103] (2) There is a "use" relationship between the process step element concept and processing equipment; there is a "processing" relationship between the process step element concept and parts; there is a "location" relationship between processing equipment and its workshop, company and region; there is a "composition" relationship between parts and part materials.

[0104] 6. Build a knowledge graph of process design resources.

[0105] The existing process design knowledge is processed by utilizing the constructed process design resource model and the extracted nodes and relationships, and represented as a data structure that can be recognized by computers. Java is used as the programming language and Neo4j as the database to construct a process design resource knowledge graph for subsequent storage and retrieval.

[0106] Will be like Figure 5 The process card shown is expressed as structured data:

[0107] (1) According to the work step meta-knowledge model, the work step id is set to "workStep_1", the name of the processed part is sleeve, the part material is 65Mn, the processing method is drilling, the processing equipment is vertical milling machine Z550, the process equipment is straight shank twist drill, high-speed steel motorized tap, special fixture, three-purpose vernier caliper, and the cutting speed, spindle speed, cutting depth, feed rate, and working time are 15m / min, 735r / min, 3.4mm, 0.13mm / r / , and 0.52min respectively;

[0108] (2) According to the work step equipment resource model, the workshop ID of the processing equipment is set to "workshop_1", its company ID is set to "company_1", and the company's location is "Tianjin".

[0109] Use Java as the programming language and Neo4j as the database to build a knowledge graph. Java is an object-oriented programming language that is powerful and easy to use. It can be used to write desktop applications, web applications, distributed systems, and embedded system applications. Neo4j is a high-performance NOSQL graph database that stores structured data on the network instead of in tables. It is a high-performance graph database engine.

[0110] The WorkStep class is written in Java language to represent the work step entity. According to the previous work step entity modeling results, the member variables of the WorkStep class are shown in Table 1; and the Equipment class is written to represent the processing equipment entity, and the Part class represents the part entity.

[0111] Table 1 WorkStep class member variable table

[0112]

[0113] Use Java language to write the addWorkStep method to add the work step meta-entity to the graph database Neo4j and store the relevant parameters in the work step meta-entity; write the addWorkStepRelation method to create the relationship between the work step meta-entity and other entities (such as process equipment, parts) in the graph database Neo4j.

[0114] Write a test class WorkStepTest and use the Juint unit test framework to test the correctness of the program. First, create the WorkStep class entity workStep, set various parameters for workStep, use addWorkStep to add the workStep entity to the Neo4j graph database, and use the addWorkStepRelation method to add the relationship between the workStep entity and other entities.

[0115] The final result is the generation of WorkStep entities in the graph database Neo4j, and the generation of relationships between various entities. Figure 7 As shown, the process step entity has a relationship with the straight shank twist drill high speed, steel motor tap special fixture, and three-purpose vernier caliper process equipment entities; it has a relationship with the processing equipment entity vertical milling machine Z550, and the milling machine Z550 belongs to workshop_1, and workshop_1 belongs to company 1; the processed part is a sleeve, and the blank material of the sleeve is blank_1.

[0116] 7. Part feature similarity calculation

[0117] For parts with high similarity, the similarity of their processing steps is also high. Therefore, for new part features, by screening the existing part features in the process knowledge graph that have high similarity with the new part features, their processing steps and processing equipment are automatically recommended, so as to achieve the purpose of automatic generation of processes and reuse of process design knowledge.

[0118] Regarding the calculation of part feature similarity, since the part feature vector contains non-numeric parameters and numerical parameters, the calculation methods are different.

[0119] The calculation method for non-numeric parameters such as part feature types is shown in the following formula. Set the basic part feature set FT, which includes basic feature types such as holes, planes, and slots. If feature x and feature y do not belong to the set, or the two features do not belong to the same category at the basic feature level, then their similarity is 0. If both belong to the same category, build a Chinese word segmentation main dictionary and calculate their semantic similarity. The semantic similarity range is [0,1]. The closer to 1, the more similar the two are.

[0120]

[0121] For numerical parameters, the cosine similarity calculation method is used for calculation. The cosine value of the angle between two vectors in the vector space is used as a measure of the size of the difference between two individuals. The formula is shown below. The closer the calculation result is to 1, the closer the angle is to 0 degrees, which means that the two vectors are more similar.

[0122]

[0123] Therefore, the total similarity between feature vector X and feature vector Y is as follows, where w i is the weight sum of each part, sim i (x,y) is the similarity of each part.

[0124] SIM(X,Y)=w i sim i (x,y)

[0125] The hole processing method is selected to calculate its similarity. Some characteristic parameters are shown in the following table.

[0126] Table 2 Process design example data information

[0127]

[0128] According to the above formula, F ea1 The similarities with other features are 0.596, 0.828, 0.858, 0.985, 0.998, 0.833, and 0.978, respectively. Therefore, we can refer to Fea4, Fea5, and F in the process knowledge graph. ea8 Based on the process design knowledge, we can design the process steps and select the processing equipment.

[0129] In summary, the modeling approach can be summarized as follows:

[0130] (1) Divide the process route into a series of processes, and divide each process into a series of steps.

[0131] (2) In order to facilitate the analysis and description of more complex processes and better organize production and calculate working hours, the process step is taken as the basic unit of the process design instance and a process step meta-knowledge model is established.

[0132] (3) Perform feature modeling on the part entities in the process design example and construct the part feature vector.

[0133] (4) Combined with the already constructed process step meta-knowledge model, model the material resources in the process design resources and model the processing equipment resources.

[0134] (5) Map the concepts and entities in the model to nodes in the knowledge graph and extract the relationships between the nodes.

[0135] (6) Construct a knowledge graph of process design resources. Using the completed process step meta-knowledge model and processing equipment resource model, the existing process design knowledge is processed and represented as a data structure that can be recognized by computers. Java is used as the programming language and Neo4j as the database to construct a knowledge graph of process design resources for subsequent storage and retrieval.

[0136] (7) By utilizing the constructed model and process design knowledge graph and calculating the similarity between the new part features and the existing part features, the processing steps and processing equipment are recommended to achieve the purpose of reusing process design knowledge.

[0137] (8) Optimize the model to obtain a process design resource model with a high degree of structure and sharing.

Claims

1. A method for constructing a process design resource model based on a knowledge graph, wherein the process design process include: Determine the production program for parts processing and explain the product specifications and output; Formulate the production process of parts, explain the production process, specifications and quantity of production equipment and auxiliary equipment; Calculate the quantity of raw materials and semi-finished products required for part generation. The construction of the process design resource model includes the following steps: The first step is to model the process design resources for the process route. The process design resource modeling is carried out based on the process route. The process route is divided into a series of processes, and each process is divided into a series of steps. The step is used as the basic unit of the process design instance, and the step meta-knowledge model is established. The process step meta-knowledge model is classified according to the processing method, each category is in parallel relationship, and each category is further classified according to the processing accuracy; The process-step meta-knowledge is divided into the basic concepts, attributes and other process equipment of process design knowledge. The process-step meta-SKU is represented by the following formula: EN U ={SK C ,SK A ,PE} SK C The basic concept of the process step element, SK A Indicates the parameters of the process step element. PE represents the processing equipment used in the process step and sets its properties: (1) Basic concepts include processing features, parts, and processing methods. Processing features refer to the basic features of the parts processed in this process step; processing methods refer to the processing methods used in this process step; and parts refer to the parts processed in this process step. (2) Regarding the process step parameters, they include process step ID, cutting speed, spindle speed, cutting depth, feed rate, and working time. The process step ID is the unique identifier of the process step and is used to number the process step. The cutting speed, spindle speed, cutting depth, and feed rate are the basic parameters in the machining process. The working time is the time consumed by the process step, which is the sum of the mobile working time and the auxiliary working time. (3) Regarding processing equipment, it is divided into cutting equipment, process equipment, testing equipment, and heat treatment equipment; The second step is part feature modeling The part features are described by the following formula: F={F Man ,F Mat ,F Sha } F Man Represents management features, which contain relevant attribute information of parts, including part drawing number, part name, part number, and part type; F Mat Represents material characteristics, including relevant property information of the material, including material type, material grade, and material performance; F Sha Represents shape features, reflecting the shape structure of parts. This feature includes design feature information and manufacturing feature information; Using the already constructed part feature model, construct the part feature vector: F ea =[F c ,F s ,R,A cc ,M] F c Represents the part feature type, F s Represents the dimensional parameters of the part features, R represents the roughness requirement, A cc represents the accuracy requirement, and M represents the material parameter; Part feature parameter F s There are different property settings for different part types; Material parameters, as the basic characteristics of the material that makes up a part, are represented by the following vectors: M=[M T ,M g ,M EM ,M YS ] M T Represents the material type, M g Represents the material grade, M EM Represents the elastic modulus of the material, M YS Represents the yield strength of the material; The third step is to model the processing equipment resources. For the processing equipment of the work step meta-knowledge model, since it plays a relatively important role in processing and manufacturing, the processing equipment resources are modeled in combination with the work step meta-knowledge model; processing equipment resources refer to the mechanical equipment required to process a part, which is the manufacturing resource facing the bottom layer of the manufacturing system, including machine tools, cutting tools, fixtures, and measuring tools; A more detailed classification of processing equipment: (1) For cutting equipment, it is further divided into lathes, milling machines, planers, grinders, boring machines, drilling machines, CNC machine tools, machining centers and others; (2) For process equipment, it is further divided into tools, fixtures, gauges and molds; (3) For testing equipment, it is further divided into displacement testing equipment, speed testing equipment, temperature testing equipment, humidity testing equipment, pressure testing equipment and other testing equipment; (4) Heat treatment equipment is further divided into carburizing furnace, vacuum furnace, tempering furnace, annealing furnace, quenching furnace and other heat treatment equipment; The processing equipment resources are formally described from four attributes: basic attributes, state attributes, functional attributes, and service attributes. The basic attributes are used to describe the basic situation of a resource, including resource ID, resource name, resource unit, location, contact information, and resource nature. The status attribute is used to describe the current status of the resource, including idle, not fully loaded, fully loaded, and under maintenance; Functional attributes are used to describe the functional conditions of the services provided by a resource, including service objects, processing accuracy, processing size, processing shape, processing method, input and output; Service attributes are used to describe the specific conditions of the services provided by a resource, including the service mode and resource quantity; The fourth step is to use knowledge graph technology to extract the relationship between nodes; (1) In the process-step knowledge model, the nodes include the process-step concept, the part itself, and the materials that make up the part; in the processing resource model, the nodes include the processing equipment, the workshop to which the processing equipment belongs, the company, and the region; (2) There is a "use" relationship between the process step element concept and the processing equipment; there is a "processing" relationship between the process step element concept and the parts; there is a "location" relationship between the processing equipment and the workshop, company and region to which it belongs; there is a "composition" relationship between the parts and the materials of the parts; Step 5: Build a knowledge graph of process design resources Using the constructed process design resource model and the extracted nodes and relationships, the existing process design knowledge is processed and represented as a data structure that can be recognized by computers, and a process design resource knowledge graph is constructed to facilitate subsequent storage and retrieval; The sixth step is to use the constructed model and process design knowledge graph, and calculate the similarity between the new part features and the existing part features to recommend processing steps and processing equipment.

2. The model building method according to claim 1, It is characterized in that The calculation method of part feature similarity is: For the calculation of part feature similarity, since the part feature vector contains non-numeric parameters and numerical parameters, Similarity calculation method for non-numeric parameters in part feature types: Set the basic part feature set FT. If feature x and feature y do not belong to the set, or the two features do not belong to the same category at the basic feature level, then their similarity is 0. If both belong to the same category, build a Chinese word segmentation main dictionary and calculate its semantic similarity. The semantic similarity range is [0,1]. The closer to 1, the more similar the two are. For numerical parameters, the cosine similarity calculation method is used for calculation. The cosine value of the angle between two vectors in the vector space is used as a measure of the size of the difference between two individuals. The closer the calculation result is to 1, the closer the angle is to 0 degrees, which means that the two vectors are more similar. The total similarity between feature vector X and feature vector Y is the weighted sum of the similarities of the feature parameters of each part.