A knowledge graph-based complex equipment development demand structure information generation method

Abstract

The application discloses a kind of complex equipment development demand structure information generation method based on knowledge graph.There are following steps: first, based on BIGRU-CRF and CasRel model, construct field knowledge graph, and establish product function expression model according to the concept of extension base element;Second, based on product function expression model, field knowledge graph and correction cosine semantic similarity, the demand structure of complex equipment to be developed is solved, and the structure information is sought to mine the innovative structure in product design space;Finally, a node importance ranking method is proposed and used to sort the structure solution, and the leftmost demand structure information is generated.Finally, more accurate guidance is provided for the complex equipment development process in this field.

Description

Technical Field

[0001] This invention relates to the field of knowledge graph construction technology for complex equipment development, and specifically to a method for generating structural information on the development requirements of complex equipment based on knowledge graphs. Background Technology

[0002] In the development of complex equipment, engineers often encounter difficulties such as non-standard operation, lack of design knowledge, and incomplete consideration of problems, leading to low development efficiency and stagnation. Typically, to improve the standardization of complex equipment development, companies rely too heavily on the experience and knowledge of R&D personnel and previous design schemes. However, with the application of big data and artificial intelligence, a large number of innovative product structures exist on the internet. If the functional requirements of the product to be designed can be realized during the development of complex equipment, and these innovative structures can be collected and fully utilized through knowledge graph technology, even when facing the most complex equipment development solutions, it can help engineers improve their work efficiency and enhance the company's innovation vitality.

[0003] Traditional methods for developing complex equipment mainly include process-based, analogy-based, and knowledge-based approaches, but these methods have limitations. Process-based methods use behavior as an intermediary to connect structure and function, achieving a structure-to-function mapping. Then, through the relationships between functions indicated by meta-functions, the top-level function of the product is restored through functional merging operations, ultimately achieving an understanding of the development philosophy of complex equipment. However, this design method requires a high level of knowledge from designers. Analogy-based methods can enhance creativity by applying previous successful cases, but analogy can become overly reliant on similarity, requiring relatively high knowledge and potentially leading to design errors due to its limited applicability. Knowledge-based methods emphasize using domain-specific knowledge and experience to guide and manage each stage of complex equipment development. This method expresses the designer's ideas, principles, and guidelines with explicit knowledge, better reflecting product characteristics and meeting modern design needs compared to traditional process-based design processes. However, this method may neglect cross-domain innovation and consume more time. Therefore, proactive steps must be taken to improve the quality of design solutions, enhance objectivity, and increase efficiency. Summary of the Invention

[0004] This invention addresses the limitations of objectivity in solving product functions and the lack of effective auxiliary tools in the development of complex equipment. It proposes a knowledge graph-based method for generating requirement structure information for complex equipment development. First, a domain knowledge graph is constructed based on BIGRU-CRF and CasRel models, and a product function expression model is established according to the extension primitive concept. Second, based on the product function expression model, the domain knowledge graph, and modified cosine semantic similarity, the requirement structure of the product to be developed is solved. Finally, a node importance ranking method is proposed and used to rank the structural solutions, generating the product's requirement structure information. Ultimately, this provides more accurate guidance for the development process of complex equipment in this field.

[0005] To achieve the above objectives, this invention proposes a method for generating structural information on the development requirements of complex equipment based on knowledge graphs, comprising the following steps:

[0006] (1) Construct a dataset and build a domain knowledge graph using BIGRU-CRF and CasRel models.

[0007] 1.1) Select relevant patent texts, define the entity types to be extracted, and represent the entities using the "BMES"-based entity annotation method;

[0008] 1.2) Define the types of relations to be extracted, and use the sprite annotation assistant to annotate the relations;

[0009] 1.3) Use the BERT model to generate word vectors that fully incorporate information from the training text;

[0010] 1.4) Use the BIGRU-CRF model to complete the entity recognition task and the CasRel model to complete the relation extraction task, and construct a domain knowledge graph;

[0011] (2) Use the event-based model in extension primitive theory to represent the behavior and functional objects in the product function expression model.

[0012] (3) Based on this behavior, the modified cosine semantic similarity is used to match similar relation words in the knowledge graph, and all tail entities of the relation words are obtained to form a matrix.

[0013] (4) Extract the functional objects in the event element model and compare them with all elements in the candidate matrix. Use modified cosine semantic similarity for matching, collect the elements that exceed the similarity threshold, and inversely deduce all head entities through the element and its corresponding relation word to generate the structural solution set corresponding to the functional object. Iterate through all behaviors and functional objects in the product function expression model to reason out the overall structural solution set.

[0014] (5) Take the knowledge graph and the solution set of the requirement structure corresponding to the product function as input data and pass them into the node importance evaluation method.

[0015] (6) Using a node importance evaluation calculation method based on network analysis, sort all elements in the structural solution set according to the knowledge graph and node importance.

[0016] (7) Starting from the first functional object in the overall functional set, select the first element in the sorting result of the elements in its corresponding structural solution set, that is, the element with the highest node importance. Considering factors such as economy and convenience, determine whether it can serve as the final requirement structure information of the first functional object. If it can serve as the final result set, put it into the final result set. If it is not feasible, continue to take out the second element in the sorting result of the structural solution set and evaluate it again, and so on.

[0017] (8) Find the final requirement structure information of all functional objects in the overall functional set and form a complex equipment development plan.

[0018] In step (2), the product function expression model is constructed by gradually decomposing the structure from top to bottom, with each structure having a different functional representation. The "behavior" in the product function expression model describes how the structure achieves its function. This framework of "structure-behavior-function" or "system-behavior-function" simplifies the description of complex equipment development models, and this structured approach also facilitates computer processing. In this invention, the decomposition from the overall structure to subsystems and the further decomposition from subsystems to structures are both represented using the composite element model in extension theory; while the structure itself and its functions are represented using the event element model in extension theory, thus completing the construction process of the product function expression model. In the product function expression model, the most important elements are the "behavior" and "function" objects. Only when these two objects are accurately defined can the system seek the necessary structural information based on the knowledge graph.

[0019] In this invention, Γ(D) is used to denote the name of the complex equipment or its subsystem. i This indicates the i-th "behavior" of the complex equipment or its subsystem, while noun i Let i represent the i-th functional object of the complex equipment or its subsystem, then we have

[0020]

[0021] In step (3), a fixed set of verbs is defined. i and its corresponding noun i Let F be a set of product functional units to be solved. i , denoted as Fi =[verb i ,noun i Then the overall function set is M = {F1, F2, ..., F}. n In domain knowledge graphs, search and verbs i The names of relations with high modified cosine semantic similarity are selected, and the tail entities containing these relations are aggregated to form a matrix B consisting of "behavior" and "function" objects. i Then B i =[b1,b2,...,b n ]. Among them, b i Indicates the i-th verb. i Similar relation names, corresponding matrices This represents all tail entities corresponding to this relation term.

[0022] Then for the i-th group of product functional units F to be solved i Its matrix B i for:

[0023]

[0024] The calculation process for the modified cosine semantic similarity in step (3) is as follows:

[0025]

[0026] Here, Sim(D1,D2) represents the adjusted cosine semantic similarity between two entities D1 and D2. The semantic vector representing entity D1, The semantic vector representing entity D2. The semantic vector mean of entity D1, The semantic vector mean of entity D2.

[0027] In step (4), set B is... i Elements with a modified cosine semantic similarity of less than 0.8 After all removals, for the corrected candidate matrix, search for the corresponding triplet head entity. Product functional unit F i Candidate solution matrix C i Then the matrix is

[0028]

[0029] If C i If it is an empty set, then the user can define the default implementation functional unit F. i The structure.

[0030] Repeat the above steps until all functions in Γ(D) can find the candidate solution matrix C. i After summarizing, we obtain the overall structural solution set.

[0031] Step (6) proposes a node importance evaluation calculation method based on network analysis, and sorts all elements in the structural solution set according to the calculation results. The specific steps are as follows.

[0032] 6.1) Take a structural solution from the solution set of the requirement structure corresponding to the product function as node v. i Dijstra's algorithm is used to calculate the shortest path dis(v) between the current node and all other nodes. i ,v j ).

[0033] 6.2) Calculate the current node v i The proximity centrality C(i).

[0034] Define dis(v) i ,v j ) is node v i To node v j The shorter the path length, the greater the proximity to centrality. Node v i The proximity centrality can be represented as C(i).

[0035]

[0036] 6.3) Based on the constructed domain knowledge graph, solve for the current node v i neighborhood δ ki With node key domain O i .

[0037] Define node v i The neighborhood of a node is represented by the set of all nodes connected to it by an edge, and the neighborhood of a node is δ. ki Node v s neighborhood Indicates with v s The set of all nodes that are associated by an edge; for node v t neighborhood Indicates with v t The set of all nodes that are connected by an edge. Then node v i The node key field is represented as

[0038]

[0039] 6.4) Traverse the neighborhood δ ki Any two nodes v p and v q Solve for P(v)p ,v q ).

[0040] P(v p ,v q )={{v p ,v q}∪{v p ,v i ,v q}∪{v p ,v′ i ,v q |v′ i ∈F i ∩v′ i ≠v i}}

[0041] P(v p ,v q ) indicates that at node v i neighborhood δ ki Any two nodes v p and v q The set of shortest paths between.

[0042] 6.5) Calculate the value of the node v. i The number of shortest paths M(i) and the number of paths that do not pass through node v i The number of shortest paths N(i).

[0043] Defined in node v of degree k i neighborhood δ ki In the context, any pair of nodes can be connected via node v. i The number of shortest paths is M(i), which does not pass through node v. i The number of shortest paths is N(i). Assume node v s and v t The shortest path between them is w sj Based on the key domains of the nodes, M(i) and N(i) can be expressed as follows:

[0044] in

[0045] in

[0046] 6.6) Calculate node v i The criticality K(i).

[0047]

[0048] 6.7) In network analysis, due to differences in network structure, the number and location of nodes and edges affect the importance of each node within the entire network. In complex networks, there is a difference in importance between central and non-central nodes; whether nodes are connected also affects importance. Furthermore, the greater the number of shortest paths a node has, the more important it is. Therefore, the importance of a node v can be defined and calculated. i Importance D(i).

[0049]

[0050] A computer storage medium is provided, containing a computer program for executing a method for generating structural information on the development requirements of complex equipment based on knowledge graphs. When the program runs, it controls the device containing the computer storage medium to implement the method for generating structural information on the development requirements of complex equipment based on knowledge graphs.

[0051] A processor is proposed. The processor is used to run a program, wherein the program, when running, executes the aforementioned method for generating complex equipment development requirement structure information based on knowledge graphs.

[0052] The beneficial effects of this invention are as follows:

[0053] This paper proposes a knowledge graph-based method for generating structural information on the development requirements of complex equipment. The method first focuses on constructing a product functional expression model to effectively process, store, and express the deep-level knowledge of the designed product. After the model is built, the design knowledge stored within the knowledge graph is utilized to comprehensively explore structural solutions that meet functional requirements by mapping the product functional expression model to functional objects in the knowledge graph. This process essentially matches the structural framework in the product functional expression model with the functional objects in the knowledge graph to uncover innovative structures in the development space of complex equipment. After sorting, selecting, and evaluating the uncovered structural information, the most suitable design scheme is finally formed. This knowledge graph-based method for generating structural information on the development requirements of complex equipment enables designers to process key knowledge in design tasks more flexibly and accurately, thereby effectively improving design efficiency. Attached Figure Description

[0054] Figure 1 Overall flowchart of the method for generating structural information on the development requirements of complex equipment based on knowledge graphs.

[0055] Figure 2 Partial annotation of entity dataset.

[0056] Figure 3 Flowchart of a node importance evaluation method based on network analysis.

[0057] Figure 4Product design diagram of automated guided vehicles. Detailed Implementation

[0058] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The following description of at least one exemplary embodiment is merely illustrative and is in no way intended to limit the present invention or its application or use. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0059] The terminology used in this invention is intended to describe particular embodiments, not to set limitations. In this invention and related claims, singular terms such as “a,” “the,” and “the” also encompass their plural forms, unless the context clearly indicates otherwise. Such wording is intended to ensure that the form of the terminology does not restrict the understanding of the invention, allowing it to encompass a variety of possible implementations.

[0060] The specific embodiments of the present invention will now be described in detail and clearly with reference to the accompanying drawings. Figure 1 This paper presents a method for generating structural information on the development requirements of complex equipment based on knowledge graphs, including the following steps:

[0061] (1) Obtain relevant patents in the field of complex equipment, perform data annotation and processing, construct a dataset in the field of complex equipment, and then obtain the corresponding character vectors and word vectors, specifically:

[0062] 1.1) Taking the field of automated guided vehicles (AGVs) as an example, relevant patents in this field from 2020 onwards were obtained. The relevant entity types are categorized into five types according to the designers of complex equipment: AGV name, functional components, constituent parts, implemented functions, and subsystems. The annotation descriptions for entity recognition task data are shown in Table 1.

[0063] Table 1. Data annotation instructions for entity recognition task

[0064]

[0065]

[0066] Taking the example of "a patent mentions an automated guided vehicle that has laser sensors to detect obstacles in the environment. In addition, there is a navigation system to determine the vehicle's position," we will use the Genie Assistant for sequence labeling.

[0067] The labeled results are as follows Figure 2As shown in the figure. For the functional component entity of "laser sensor", the following annotation method is adopted: the character "激" is represented by B-FC, indicating that "激" is the first character of this type of entity of functional components; the characters "光", "传", "感", etc. are represented by M-FC, indicating that they are the middle characters of this type of entity of functional components; the character "器" is represented by E-FC, indicating that "器" is the last character of this type of entity of functional components. For the characters that do not want to be extracted and the punctuation marks that have no effect, the "S" notation is used. After annotation, the annotated dataset is finally obtained.

[0068] In the annotated dataset, the first column is each specific character that appears in the corpus, and the second column is one of "B", "M", "E", "S", and "O". When all the patent text data is annotated, more than 17,000 lines of entity recognition datasets in the field of automated guided vehicles are obtained;

[0069] 1.2) The construction of the text relation dataset is consistent with the source of the entity dataset. To meet the needs of knowledge mining and utilization in the process of complex equipment development, combined with the characteristics of entity annotation, the collected data is defined for relationships and data annotation.

[0070] The relationships between entities are defined as three categories: composition, realization, and unknown relationship. The composition relationship means that the head entity can be composed of the tail entity, the realization relationship means that the head entity can realize the function of the tail entity, and the unknown relationship means that the relationship between the two entities is not clear and有待挖掘 (to be explored). The relationship annotation instructions are shown in Table 2:

[0071] Table 2 Relationship annotation instructions

[0072]

[0073] The data format of the annotation example is shown in Table 3, including the head entity, the tail entity, the relationship between entities, and the sentence text. Similar to the division method of the entity recognition dataset, after sorting and summarizing, it is divided into two parts. One part is used for training the model, and the other part is used to verify the relationship recognition effect of the model. Similarly, to enable the model to be fully trained, the ratio of the training model dataset to the verification model dataset is 9:l.

[0074] Table 3 Partial annotation of the relationship dataset

[0075] <~

[0076]

[0077] 1.3) After the dataset construction is completed, the BERT model is used to obtain word vectors and character vectors, and the specific process is as follows;

[0078] a) It means that a knowledge text sequence to be extracted is segmented by character, and each independent character is used as input and fed into BERT

[0079] b) After encoding the text characters to be extracted using BERT, they can be converted into a 768-dimensional feature vector, i.e., h i ∈R 1 ×768 After linear transformation, the output of BERT can be mapped to a low-dimensional semantic feature vector.

[0080] 1.4) The semantic feature vectors are fed into the BIGRU-CRF model and the CasRel model for entity recognition and relation extraction. This yields a domain knowledge graph for automated guided vehicles.

[0081] (2) Using the model in extension primitive theory to represent the product function expression object, taking the field of automated guided vehicles as an example, the specific construction process is as follows:

[0082] Note that during model building, the first column requires a default structure provided by the user. This ensures that the result is not NULL if the final required structure fails to be solved.

[0083] a)

[0084] b)

[0085] c)

[0086] d)

[0087] e)J 0-DHMK =[Positioning achieved using magnetic sensors]

[0088] f)

[0089] g)

[0090] h)J 0-TXMK =[Using a wireless network for local area network transmission]

[0091] i)

[0092] j)

[0093] (3) Based on the "behavior-function" object in the product function expression model, in the driving module of the constructed automated guided vehicle product function expression model, "provide" is the first action, and "rotational motion power" is the corresponding first function object. This will be used as an example for illustration.

[0094] In the knowledge graph, searching for relation names with similar meanings to "provide" reveals three categories of relations with a modified cosine semantic similarity greater than or equal to 0.8: provide, realize, and through. All tail entities of these three relation categories are then collected and arranged into a matrix.

[0095] (4) Taking the first two relations "provide" and "realize" as examples, query all tail entities with a modified cosine semantic similarity greater than or equal to 0.8 with the tail entity named "rotational motion power", and search for its corresponding triple head entity to generate a partial requirement structure solution set corresponding to the product functional object, thus completing the functional requirement structure solution process of the automated guided vehicle drive module. As shown in Table 4, the partial requirement structure solution set of the target functional object "rotational motion power" in the drive module is displayed. After repeated iterations, the overall structure solution set can be inferred.

[0096] Table 4. Decomposition of Requirements Structure for the "Rotational Motion Power" Target Function Object in the Driving Module (Partial)

[0097]

[0098]

[0099] (5) The domain knowledge graph and the solution set of the requirement structure corresponding to the product function are used as input data and passed into the node importance evaluation method, such as... Figure 3 As shown.

[0100] (6) Taking the requirement structure solution set of the target functional object of "rotational motion power" of the automatic guided vehicle drive module as an example, the importance of the nodes is sorted and the top 10 items are selected for display. The following results can be obtained, which facilitates the final determination of the requirement structure information of the final complex equipment development scheme, as shown in Table 5.

[0101] Table 5 shows the importance ranking of the structural solution set of the "Rotational Motion Power" target functional requirement in the driving module (top 10 items).

[0102]

[0103] (7) Select the structure name that ranks first in importance and determine whether it can serve as the requirement structure of the final functional object. If it can serve as the requirement structure, put it directly into the final result set; if it cannot serve as the requirement structure, take out the second one in the ranking list and evaluate it again, and so on.

[0104] (8) Finally, the final requirement structure information for all functional objects in the overall functional set is obtained. See Table 6.

[0105] Table 6. Development Requirements Information for Complex Equipment in Automated Guided Vehicles

[0106]

[0107]

[0108] By comparing the final functional results of the complex equipment development with the initial functional results, in the drive module, the initial structure of the rotation and direction-changing functional object changed from a wheel to a Magnum wheel, indicating that the patent text has shown greater interest in the Magnum wheel structure in the past three years. Compared with traditional wheels, the Magnum wheel can achieve omnidirectional movement, translating in any direction within a plane while simultaneously rotating. In the navigation module, the initial structure of the positioning functional object changed from radar to SLAM laser navigation, demonstrating that laser technology is currently the mainstream in the AGV navigation field. Compared with traditional magnetic sensors, SLAM constructs maps with high accuracy, eliminates cumulative errors, and can directly... Used for positioning and navigation; in the carrying module, based on actual needs, specific structures suitable for transporting and holding goods can be selected to replace fixed pallets with cargo supports, improving the efficiency of goods transportation in specific scenarios; in the safety control module, a buffer structure is used instead of rubber strips, providing more robust protection for the automated guided vehicle and its carried goods; in the communication module, visible light wireless communication is used instead of traditional wireless networks to achieve signal stability; in the energy module, lithium batteries are used to extend cycle life, are environmentally friendly, and offer better safety; in the auxiliary module, the ammeter is replaced with a touchscreen for easier user reading. Observing Table 4-6 reveals that among the 15 functional objects, 8 innovative requirement structures were replaced based on knowledge graphs, resulting in an improved structure ratio of 53.33%; calculations show that the overall node importance of this solution increased by 0.164, and the average node importance per structure increased by 0.021. Finally, based on Table 4-6, further product structure selection can be made, and a product solution diagram of the automated guided vehicle can be drawn. Figure 4 In the diagram, circles represent the design structure, and boxes represent the implemented functional objects.

[0109] The above embodiments are merely illustrative of the design concepts and features described herein, and are intended to enable those skilled in the art to understand the essence of the invention and implement it accordingly. The scope of protection of this invention is not limited to the specific examples described above. Those skilled in the art can make appropriate modifications and changes as needed to adapt to different applications and implementation scenarios, and such modifications and changes are still included under the legal protection of this invention.

Claims

1. A method for generating structural information on the development requirements of complex equipment based on knowledge graphs, comprising the following steps: (1) Construct a dataset and use BIGRU-CRF and CasRel models to build a domain knowledge graph; 1.1) Select relevant patent texts, define the entity types to be extracted, and represent the entities using the "BMES"-based entity annotation method; 1.2) Define the types of relations to be extracted, and use the sprite annotation assistant to annotate the relations; 1.3) Use the BERT model to generate word vectors that fully incorporate information from the training text; 1.4) The BIGRU-CRF model is used to complete the entity recognition task, and the CasRel model is used to complete the relation extraction task, thus constructing a domain knowledge graph; (2) Use the event-based model in extension primitive theory to represent the behavior and functional objects in the product function expression model; (3) Based on this behavior, the modified cosine semantic similarity is used to match similar relation words in the knowledge graph, and all tail entities of the relation words are obtained to form a matrix; (4) Extract the functional objects in the event model and compare them with all elements in the candidate matrix. Use modified cosine semantic similarity to match them. Collect the elements that exceed the similarity threshold. Use the element and its corresponding relation word to inversely deduce all head entities and generate the structural solution set corresponding to the functional object. Iterate through all behaviors and functional objects in the product function expression model and reason to obtain the overall structural solution set. (5) Use the knowledge graph and the solution set of the requirement structure corresponding to the product function as input data and pass them into the node importance evaluation method; (6) Using a node importance evaluation calculation method based on network analysis, sort all elements in the structural solution set according to the knowledge graph and node importance; (7) Starting from the first functional object in the overall functional set, select the first element in the sorting result of the elements in its corresponding structural solution set, that is, the element with the highest node importance, and determine whether it can serve as the final requirement structure information of the first functional object. If it can serve as the final result set, put it into the final result set. If it is not feasible, continue to take out the second element in the sorting result of the structural solution set, evaluate it again, and so on. (8) Find the final requirement structure information of all functional objects in the overall functional set, and comprehensively form a complex equipment development plan; In step (6), the node importance evaluation calculation method based on network analysis is used. The specific process of sorting all elements in the structural solution set according to the knowledge graph and node importance is as follows: 6.1) Take a structural solution from the solution set of the requirement structure corresponding to the product function as node v. i Dijstra's algorithm is used to calculate the shortest path dis(v) between the current node and all other nodes. i ,v j ); 6.2) Calculate the current node v i The proximity centrality C(i); Define dis(v) i ,v j ) is node v i To node v j The shortest path length; the shorter the length, the greater the proximity to centrality; node v i The proximity centrality can be expressed as C(i); 6.3) Based on the constructed domain knowledge graph, solve for the current node v i neighborhood δ ki With node key domain O i ; Define node v i The neighborhood of a node is represented by the set of all nodes connected to it by an edge, and the neighborhood of a node is δ. ki Node v s neighborhood Indicates with v s The set of all nodes that are associated by an edge; for node v t neighborhood Indicates with v t The set of all nodes that are associated by an edge; then node v i The node key field is represented as 6.4) Traverse the neighborhood δ ki Any two nodes v p and v q Solve for P(v) p ,v q ); P(v p ,v q )={{v p ,v q }∪{v p ,v i ,v q }∪{v p ,v i ′,v q |v i ′∈F i ∩v i ′≠v i }}.; P(v p ,v q ) indicates that at node v i neighborhood δ ki Any two nodes v p and v q The set of shortest paths between; 6.5) Calculate the value of the node v. i The number of shortest paths M(i) and the number of paths that do not pass through node v i The number of shortest paths N(i); Defined in node v of degree k i neighborhood δ ki In the context, any pair of nodes can be connected via node v. i The number of shortest paths is M(i), which does not pass through node v. i The number of shortest paths is N(i); assuming node v s and v t The shortest path between them is w sj Based on the key domains of the nodes, M(i) and N(i) can be expressed as follows: in in 6.6) Calculate node v i The criticality K(i); 6.7) In network analysis, due to different network structures, the number and location of nodes and edges affect the importance of each node in the entire network; in complex networks, there is a difference in importance between central and non-central nodes; whether nodes are connected also affects importance; and the larger the number of shortest paths a node has, the more important the node is; therefore, the node v can be defined and calculated. i Importance D(i); Finally, all elements in the structural solution set are sorted according to the knowledge graph and node importance.

2. The method for generating structural information on the development requirements of complex equipment based on knowledge graphs according to claim 1, wherein the annotation and processing in step 1.1) specifically comprises: Obtain relevant patents from 2020 onwards in the field of complex equipment to be studied. Analyze the literature content and the field of complex equipment to be studied, and combine the suggestions of designers in this field to formulate the entity types to be identified in the knowledge graph. Use the Sprite Annotation Assistant and the BMES data annotation method to annotate the relevant text; BMES is a common data annotation method. In BMES, "B" represents the annotation mark at the beginning of an entity, "M" represents the annotation mark in the middle of an entity, and "E" represents the annotation mark at the end of an entity. "S" represents a non-entity mark. The entity's abbreviation must be followed by "B", "M" and "E". Finally, the annotation results are saved as a txt file.

3. The method for generating structural information on the development requirements of complex equipment based on knowledge graphs according to claim 1, wherein step 1.2) defines the type of relation to be extracted and uses a sprite annotation assistant to annotate the relation, specifically as follows: First, based on the text content and the technical experts in the relevant complex equipment field, the relationship type between the head entity and the tail entity in the text statement is defined; second, the annotation format {{"h":{"name":"Here is the name of the head entity"},"t":{"name":"Here is the name of the tail entity"},"relation":"Here is the relationship type","text":"Here is the content of the original text statement"} is adopted; finally, the relationship annotation is performed using the Sprite Annotation Assistant.

4. The method for generating structural information on the development requirements of complex equipment based on knowledge graphs according to claim 1, wherein step 1.3) uses the BERT model to generate word vectors that fully integrate training text information, specifically as follows: a) indicates that a sequence of knowledge text to be extracted is segmented into characters, with each independent character serving as input to BERT; b) After encoding the text characters to be extracted using BERT, they can be converted into a 768-dimensional feature vector, i.e., h i ∈R 1×768 After linear transformation, the output of BERT can be mapped to a low-dimensional semantic feature vector.

5. The method for generating structural information on the development requirements of complex equipment based on knowledge graphs according to claim 1, wherein the specific process of using the event-based model in extension primitive theory to represent the behavior and functional objects in the product function expression model in step (2) is as follows: Let Γ(D) represent the name of the complex equipment or its subsystem. i This indicates the i-th "behavior" of the complex equipment or its subsystem, while noun i Let i represent the i-th functional object of the complex equipment or its subsystem, then we have 6. The method for generating structural information on the development requirements of complex equipment based on knowledge graphs according to claim 1, wherein step (3) uses modified cosine semantic similarity to match similar relation words in the knowledge graph, obtains all tail entities of the relation words, and forms a matrix, the specific process is as follows: Define a fixed set of verbs i and its corresponding noun i Let F be a set of product functional units to be solved. i , denoted as F i =[verb i ,noun i Then the overall function set is M = {F1, F2, ..., F}. n In a domain knowledge graph, search and verb... i The names of relations with high semantic similarity based on modified cosine similarity are selected, and the tail entities containing these relations are aggregated to form a matrix B consisting of "behavior" and "function" objects. i Then B i =[b1,b2,...,b n ];in, b i Indicates the i-th verb. i Similar relation names, and their corresponding matrices This indicates all tail entities corresponding to this relation term; Then for the i-th group of product functional units F to be solved i Its matrix B i for: In step (3), the calculation process for the modified cosine semantic similarity is as follows: Where Sim(D1,D2) represents the adjusted cosine semantic similarity between two entities D1 and D2; where, The semantic vector representing entity D1, The semantic vector representing entity D2; The semantic vector mean of entity D1, The semantic vector mean of entity D2.

7. The method for generating structural information of complex equipment development requirements based on knowledge graphs according to claim 1, wherein the specific process of generating the structural solution set corresponding to the functional object in step (4) is as follows: Define set B i Elements with a modified cosine semantic similarity of less than 0.8 After all removals, for the corrected candidate matrix, search for the corresponding triplet head entity. Product functional unit F i Candidate solution matrix C i Then the matrix is If C i If it is an empty set, then the user can define the default implementation functional unit F. i The structure; Repeat the above steps until all functions in Γ(D) can find the candidate solution matrix C. i After summarizing, we obtain the overall structural solution set.

8. A computer storage medium, characterized in that, The storage medium contains a program that, when run, controls the device to execute the knowledge graph-based method for generating structural information on the development requirements of complex equipment as described in any one of claims 1 to 7.

9. A processor, characterized in that, The processor is used to execute a program, wherein the program executes the method for generating complex equipment development requirement structure information based on knowledge graphs as described in any one of claims 1 to 7 during runtime.

Images