Knowledge-based mechanical structure intelligent design method, system and device and medium

Abstract

The invention relates to the field of mechanical intelligent design, in particular to a knowledge-based mechanical structure intelligent design method, system and device and a medium, and the key points of the technical scheme are that the method comprises the following steps: obtaining design demand information; converting the design demand information into a plurality of knowledge points; decomposing each knowledge point into a plurality of knowledge types, wherein each knowledge type corresponds to a plurality of pieces of knowledge content; combining the knowledge type and the knowledge content of each knowledge point to finally form a plurality of knowledge points meeting design requirements; and combining the plurality of knowledge points meeting the design requirements to form a design result, and outputting the design result. Starting from knowledge calling in the actual mechanical structure design process, the mechanical structure intelligent design scheme based on knowledge is constructed by mining design knowledge, decomposing knowledge points and designing knowledge combination, the logicality and interpretability of the design result are high, and the design quality and efficiency are remarkably improved.

Description

Technical Field

[0001] The present application relates to the field of mechanical design, and in particular to a knowledge-based intelligent design method, system, device and medium for mechanical structures. Background Art

[0002] Mechanical design is the premise of mechanical manufacturing. The country's technological promotion in intelligent manufacturing and the accelerated iteration of manufacturing product upgrades require rapid development of mechanical design. With the rapid development of artificial intelligence and big data technology and the implementation of industry applications, intelligent mechanical structure design methods or technologies have become an urgent need in the current mechanical design field.

[0003] Among the relevant intelligent mechanical structure design schemes, some are to build an intelligent design system based on a multi-label mechanical mechanism intelligent design algorithm model according to multi-category data labels, import the design requirements into the intelligent design system and then output the mechanical mechanism design scheme; some are to obtain the recommended digital model features by extracting the features of the design requirement data, and then determine the recommended category corresponding to the recommended features, and determine the mechanical design scheme from the historical mechanical design schemes corresponding to the recommended category; others first obtain the design requirement information, including digital model information and process information, and target work scenario category, and then establish an intelligent algorithm model to recommend the corresponding target design scheme based on the obtained requirement information.

[0004] The above-mentioned related solutions are all based on the mechanical design demand data and the digital model information of the design results, to establish and train intelligent algorithms to achieve intelligent recommendation of mechanical structure design solutions. However, the final decision or result given by using the digital model information of historical design results to train the intelligent algorithm model is basically obtained through the "experience" obtained from the analysis of a large amount of historical design result data, which is mostly statistically significant and only the result is obtained, which makes the logic and interpretability of the recommended results of the current mechanical structure intelligent design poor, and has a limited scope in actual mechanical design applications. Summary of the invention

[0005] In order to improve the problem of poor logic and interpretability of the results recommended by intelligent design of mechanical structures, the present application provides a knowledge-based intelligent design method, system, device and medium for mechanical structures.

[0006] In a first aspect, the present application provides a knowledge-based intelligent design method for mechanical structures, which adopts the following technical solutions: A knowledge-based intelligent design method for mechanical structures, comprising: Obtain design requirement information; Convert the design requirement information into a number of knowledge points; Decomposing each of the knowledge points into a plurality of knowledge types and a plurality of knowledge contents corresponding to each of the knowledge types; combining the knowledge type and the knowledge contents of each of the knowledge points to finally form a plurality of knowledge points that meet the design requirements; Recommend design results and the knowledge content involved based on multiple knowledge points that meet design requirements; Output the design results and the knowledge content involved.

[0007] In a specific feasible implementation scheme, the design requirement information includes several minimum requirement information units; the minimum requirement information units are obtained based on data imported by the user, and / or obtained by semantic extraction based on information input by the user, and / or formed based on several options selected by the user from preset minimum requirement options.

[0008] In a specific possible implementation scheme, the step of converting the design requirement information into a plurality of knowledge points specifically includes: The mapping relationship between historical design requirement information and knowledge points corresponding to the mechanical structure is used to train the model to convert the current design requirement information into a plurality of knowledge points.

[0009] In a specific implementation scheme, the knowledge type of each of the knowledge points includes at least one of basic theoretical knowledge, experience-based knowledge, rule-based knowledge, demand-constrained knowledge and other categories; After the step of decomposing each of the knowledge points into a plurality of knowledge types, the method further includes: The knowledge content and quantity corresponding to each of the knowledge types of each of the knowledge points are marked, and the group of marking information is stored in a database.

[0010] In a specific possible implementation scheme, the step of combining the knowledge type and knowledge content of each of the knowledge points to finally form a plurality of knowledge points that meet the design requirements specifically includes: The knowledge type and knowledge content of each current knowledge point are combined using the mapping relationship training model of the knowledge type and corresponding knowledge content combination of the historical knowledge points to obtain a plurality of knowledge points that meet the design requirements.

[0011] In a specific possible implementation scheme, the step of recommending design results and the knowledge contents involved therein based on a plurality of knowledge points that meet the design requirements specifically includes: The mapping relationship between multiple knowledge points that meet historical design requirements and historical design results is used to train the model, and the design results and the knowledge contents involved are recommended according to the current multiple knowledge points that meet the design requirements.

[0012] In the second aspect, the present application provides a knowledge-based mechanical structure intelligent design system, which adopts the following technical solutions: A knowledge-based mechanical structure intelligent design system includes a demand acquisition module, a composite neural network model and an output module; the demand acquisition module is used to acquire design demand information; the output module is used to output recommended design results; the composite neural network model is used to implement the following process: converting the design demand information into a plurality of knowledge points; decomposing each of the knowledge points into a plurality of knowledge types, wherein each of the knowledge types corresponds to a plurality of knowledge contents; combining the knowledge type and the knowledge contents of each of the knowledge points to finally form a plurality of knowledge points that meet the design demand; and recommending the design results and the knowledge contents involved based on the plurality of knowledge points that meet the design demand.

[0013] In a third aspect, the present application provides a knowledge-based intelligent design device for mechanical structures, which adopts the following technical solutions: A knowledge-based mechanical structure intelligent design device comprises a memory and a processor, wherein the memory stores at least one instruction, at least one program, code set or instruction set, and the at least one instruction, at least one program, code set or instruction set is loaded and executed by the processor to implement the knowledge-based mechanical structure intelligent design method as described above.

[0014] In a fourth aspect, the present application provides a computer-readable storage medium, which adopts the following technical solution: A computer-readable storage medium stores at least one instruction, at least one program, code set or instruction set, and the at least one instruction, at least one program, code set or instruction set is loaded and executed by a processor to implement the knowledge-based intelligent design method for mechanical structures as described above.

[0015] In summary, the present application includes at least one of the following beneficial technical effects: 1. From the perspective of knowledge call in the design process of design engineers, a knowledge-based intelligent design scheme for mechanical structures was constructed. After obtaining the relevant demand information, the recommended design results and the knowledge content involved were finally output through four processes: knowledge point conversion, knowledge decomposition, knowledge combination and result recommendation. This improved the poor logic and interpretability of the design results, significantly improved the design quality, and thus significantly improved the design efficiency; 2. The recommendation process of design results is based on knowledge logic, and the recommended design results are composed of the knowledge types and knowledge contents involved in the design, which can explain the logic of intelligent design. Therefore, the recommended design results are more reliable and more conducive to promoting the actual industrial application of intelligent design; 3. Compared with the traditional methods in related technologies, the knowledge link designed in this application is divided into four parts: knowledge point conversion, knowledge decomposition, knowledge combination and result recommendation. It can reduce the overall data volume in the training process of a single algorithm model to a certain extent, so as to reduce memory consumption and lower training costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 It is a flow chart of the knowledge-based intelligent design method of mechanical structure according to an embodiment of the present application.

[0017] Figure 2 It is a schematic diagram of the fixture column layout design process for parts welding processing according to an embodiment of the present application.

[0018] Figure 3 It is a structural schematic diagram of the position relationship between parts and columns in the fixture column layout design for parts welding processing in an embodiment of the present application.

[0019] Figure 4 It is a schematic diagram of the fixture connection structure design process for parts welding processing according to an embodiment of the present application.

[0020] Figure 5 It is a structural schematic diagram of the positional relationship among parts, columns, and fixture connection structures in the fixture connection structure design for parts welding processing in an embodiment of the present application.

[0021] Figure 6 It is a structural block diagram of the knowledge-based mechanical structure intelligent design system of an embodiment of the present application.

[0022] Explanation of the reference numerals: 1. Part; 2. Column; 3. Fixture connection structure; 101. Acquisition module; 102. Composite neural network model; 103. Output module. DETAILED DESCRIPTION

[0023] In order to make the objectives, technical solutions and advantages of the present application clearer, the implementation methods of the present application will be further described in detail below with reference to the accompanying drawings.

[0024] First, the terms involved in the embodiments of the present application are introduced.

[0025] 3D digital model refers to a product model made using three-dimensional software, such as UG, SolidWorks, etc.

[0026] RPS point information, translated into English as Reference Point System, is called positioning reference point system. Each positioning reference point is called RPS point. RPS point is also called design reference point, which is often used in the field of automobile manufacturing. The precise position and dimensional accuracy of parts are the key to achieving good controllability and fit. The means to meet these requirements is RPS. RPS requires that the positioning position and size are the same at every stage from research and development to product manufacturing, from single parts to assembly to the completion of the whole vehicle, and the reference conversion is avoided as much as possible during the process. Through the use of RPS, product quality and process accuracy can be improved.

[0027] Composite neural network model. Neural network is an important technology of deep learning in artificial intelligence. In the field of mechanical design, it learns mechanical structure design data. After training with a large amount of mechanical structure design data, a mapping relationship training model is formed. New design requirements are input into the model to output recommended design results. The composite neural network model is different from the neural network model. It has a corresponding basic structure, that is, it synthesizes multiple neural network models. When learning, it needs to learn the basic structure and the composite method at the same time. It can reason, interfere and diagnose the system. In addition, the composite model also has many ideal theoretical properties, and performs very well in interpretability and sample generation, which can diagnose errors more conveniently.

[0028] The implementation of the knowledge-based intelligent design method for mechanical structures of the present application is described in detail below in conjunction with the accompanying drawings of the specification.

[0029] Reference Figure 1 , an embodiment of the present application provides a knowledge-based intelligent design method for a mechanical structure, comprising: S100, obtaining design requirement information.

[0030] The design requirement information includes several minimum requirement information units, among which the minimum requirement information unit is the descriptive information of the function, performance, appearance, etc. of the structure. This information constitutes the data set of the design requirement information, which is used as the basis for the next step of analysis and processing.

[0031] For example, refer to Figure 2 In the fixture column layout design for part welding processing, the design requirement information includes: (11) part 3D digital model and RPS point; (12) process information: loading the part in the middle position on the left; (13) control mode: pneumatic; (14) structure type: column; (15) appearance: uniform and neat. Among them, the fixture design reference point (RPS point) information for part welding processing is shown in Table 1 below: name X Y Z 3Fx 171 569.998 509.2 103fx 171 449.498 270.2 1Hyz 176 449.498 421 104fx 171 -0.002 521 105fx 190.25 -0.002 279 2Hz 130.13027 -245.002 388.56083 5Fx 176 -500.002 350.9 4Fx 176 -500.002 506.9 4Fc 87.020262 -261 177.62278 6Fc 162.01997 -301 169.74 5Fc 164.02099 -213 169.52968 Table 1 Information on the fixture design reference point (RPS point) for parts welding processing In one embodiment of the present application, the minimum requirement information unit is obtained based on the data imported by the user. For example, if the data imported by the user includes part 3D digital model, RPS point, process information: part in the middle position on the left, and structure type: column, then the three minimum requirement information units of "part 3D digital model and RPS point", "process information: part in the middle position on the left", and "structure type: column" will be obtained.

[0032] In another embodiment of the present application, the minimum required information unit can also be obtained by semantic extraction based on the information input by the user, and the result output after semantic extraction, that is, the output minimum required information unit, is in a specific format for better recognition in the future. For example, if the information input by the user is "adopting pneumatic control", the result output after semantic recognition is "control mode: pneumatic".

[0033] In another embodiment of the present application, the minimum requirement information unit can also be formed based on several options selected by the user from preset minimum requirement options. For example, when selecting a control mode, the user is given options such as pneumatic, hydraulic, and electric. After the user selects the pneumatic option and confirms it, the minimum requirement information unit formed is "Control mode: pneumatic."

[0034] S200, converting the design requirement information into a number of knowledge points.

[0035] Specifically, the mapping relationship between historical design requirement information and knowledge points corresponding to the mechanical structure is used to train the model to convert the current design requirement information into several knowledge points. Figure 2 and Figure 3 , take the fixture column layout design for part welding processing as an example. When welding part 1, it needs to be fixed with a fixture first, and the fixture is installed on column 2. Now we need to design the fixture column 2 layout based on the structure of part 1. For example, the current design requirement information is "(11) part 3D digital model and RPS point; (12) process information: loading part in the middle position on the left; (13) control mode: pneumatic; (14) structure type: column; (15) appearance: uniform and neat", if the corresponding knowledge points involved are obtained through the mapping relationship training model between the historical design requirement information and the knowledge points corresponding to the mechanical structure, the corresponding knowledge points involved are "(21) feature analysis of the part to be processed; (22) the number of fixture units; (23) column placement", then the current design requirement information is converted into the following knowledge points: (21) feature analysis of the part to be processed; (22) the number of fixture units; (23) column placement.

[0036] It should be noted that the trained neural network algorithm model, i.e., the above-mentioned mapping relationship training model, is included in a composite neural network model, and the composite neural network model includes a total of 4 parts, namely, the corresponding implementation of the four processes of knowledge point conversion such as S200, knowledge decomposition such as S300, knowledge combination such as S400, and result recommendation such as S500. In other words, the mapping relationship training models referred to in the following S300, S400 and S500 are also included in the composite neural network model, which will not be repeated here. Each part of the composite neural network model includes an input layer, a hidden layer, and an output layer, and the input layer and output layer of each part can simultaneously serve as the output layer or input layer of other parts. Among them, the training of the composite neural network model is a prior art, and the purpose of this application is to construct a knowledge-based intelligent design method for mechanical structures to improve the poor logic and interpretability of the design results of the relevant technology, rather than the generation of the composite neural network model, so it will not be repeated here.

[0037] Taking the fixture column layout design for the above-mentioned parts welding processing as an example, in the implementation, after the part 3D digital model and RPS point information are imported into the composite neural network model, combined with the process information and the distribution of weld points of the part 3D digital model, the structural and shape characteristics of the parts to be processed and the distribution of RPS points are analyzed, and then the number of fixture units and the column positions for placing the fixture units are determined according to the part structure, shape and RPS points. Among them, the part 3D digital model, RPS points and process information in the design requirement information are mainly used to determine the number and layout of the fixture units, the control method and structure type are used to determine the structural form of the fixture unit, and the appearance is mainly to limit the uniformity and neatness of the column placement, that is: the knowledge points involved are the feature analysis of the parts to be processed, the number of fixture units and the column placement position.

[0038] S300, decomposing each knowledge point into a number of knowledge types and a number of knowledge contents corresponding to each knowledge type.

[0039] Here, the mapping relationship between historical knowledge points and knowledge types and their knowledge contents is used to train the model to decompose each current knowledge point into several knowledge types and several knowledge contents corresponding to each knowledge type. Among them, the knowledge type of each knowledge point includes at least one of the five categories of basic theoretical knowledge, experience-based knowledge, rule-based knowledge, demand-constrained knowledge and other categories. Each knowledge type of each knowledge point corresponds to corresponding knowledge content. Taking the layout design of the fixture column for parts welding processing as an example, refer to Figure 2, if the knowledge points involved in the current design requirement information are "(21) Feature analysis of the parts to be processed; (22) The number of fixture units; (23) The placement of the columns", then the knowledge types and contents involved in "(21) Feature analysis of the parts to be processed" are "(211) Empirical type: part structure, shape, RPS point distribution, welding point distribution; (212) Demand constraint type: loading position and direction", and the knowledge types and contents involved in "(22) The number of fixture units" are "(221) Empirical type: RPS point merging; (222) Rule type: In principle, a single column should not exceed 3 RPS points; (223) Demand constraint type: The number of columns should be as small as possible to meet the requirements", and the knowledge types and contents involved in "(23) The placement of the columns" are "(231) Empirical type: The number of unit RPS points merged, interference (welding gun, too close), connection structure size constraints, merged RPS point flatness; (232) Demand constraint type: welding gun position; (233) Basic theory type: reliable load transfer".

[0040] S400, combining the knowledge type and knowledge content of each knowledge point, and finally forming a plurality of knowledge points that meet the design requirements.

[0041] Specifically, the knowledge type and knowledge content of each current knowledge point are combined to obtain multiple knowledge points that meet the design requirements by using the mapping relationship training model of the knowledge type and its corresponding knowledge content combination of the historical knowledge points. Taking the fixture column layout design for parts welding processing as an example, for each knowledge point involved, several solutions that meet the design requirements, that is, meet all the knowledge content requirements of the current knowledge point, will be determined based on all its knowledge types and contents. Such solutions are knowledge points that meet the design requirements. Taking the fixture column layout design for part welding processing as an example, the knowledge point that meets the design requirements based on the knowledge type and content corresponding to "(22) The number of fixture units" is "the number of fixture units is ac"; the knowledge point that meets the design requirements based on the knowledge type and content corresponding to "(23) Column placement" is "the set of column placement positions b"; the knowledge point that meets the design requirements based on the knowledge type and content involved in "(21) Feature analysis of the part to be processed" is "the analysis content of the part structure features (such as the flatness and three-dimensionality of the part structure, the local stiffness of the part structure, etc.)", which is combined with the column placement and the number of fixture units to realize the design of the column placement position and the design of the number of fixture units.

[0042] S500, recommending design results and the knowledge contents involved based on multiple knowledge points that meet the design requirements.

[0043] Specifically, the mapping relationship between multiple knowledge points that meet historical design needs and historical design results is used to train the model to recommend design results based on multiple current knowledge points that meet design needs. That is, the design result is a solution that simultaneously meets the content requirements of multiple knowledge points, and the solution is presented in the form of a three-dimensional design model.

[0044] The recommended result includes two parts: (1) the design result, i.e., the three-dimensional design model, and (2) the corresponding design knowledge content; the design knowledge content is the knowledge content of multiple knowledge points that meet the design requirements formed by knowledge combination in S400. It should be noted that after S300 and S400, the knowledge content and quantity corresponding to each knowledge type of each knowledge point will be marked, and the corresponding marking information will be stored in the database for reference, so that the corresponding knowledge content can be attached when the design result is recommended in S500.

[0045] S600, output the recommended design results and the knowledge content involved.

[0046] In order to make the implementation process of the scheme of this embodiment clearer, the above-mentioned knowledge-based mechanical structure intelligent design method is further introduced below by taking the fixture connection structure design of parts welding processing as an example: Reference Figure 4 and Figure 5 First, we need to obtain the design requirement information. The design requirement information for the fixture connection structure 3 of the welding process of part 1 is as follows: (11) fixture layout design model, (12) part structure: part 3D digital model, (13) RPS point information, (14) appearance: uniform and neat. Among them, the fixture layout design model is shown in Figure 3 , RPS point information is shown in Table 1. Then the design requirement information is imported into the trained composite neural network model, and after knowledge point transformation, knowledge decomposition and knowledge combination, the design results and the knowledge content involved are recommended, and the recommended design results and the knowledge content involved are output.

[0047] Specifically, the design of the fixture connection structure for parts welding processing in this embodiment involves the following knowledge points: (21) the shape of the connection structure, (22) the size of the connection structure, and (23) the interference check between the connection structure and other parts. Among them, the knowledge type and knowledge content involved in the knowledge point "(21) Shape of the connection structure" are: (211) Empirical type: single L or double L shape, easy to process, (212) Rule type: in principle not more than 3L, (213) Basic theory type: load size and transfer stability; the knowledge type and knowledge content involved in the knowledge point "(22) Size of the connection structure" are: (221) Empirical type: the size range of the connection structure is 65-90mm, and the width and thickness are 20mm; the knowledge type and knowledge content involved in the knowledge point "(23) Interference check between the connection structure and other components" are: (231) Empirical type: interference (between the connection structure, between the connection structure and the column, between the connection structure and the parts, between the connection structure and the welding gun), (232) Rule type: the distance between the parts is greater than 10mm, and verification is required if it is between 5-10mm.

[0048] Subsequently, the model is trained using the mapping relationship between the knowledge type of the historical knowledge point and its corresponding knowledge content combination in the trained composite neural network model, and the knowledge type and knowledge content of each current knowledge point are combined to obtain three knowledge points that meet the design requirements. Finally, the model is trained using the mapping relationship between the part of the composite neural network model corresponding to the result recommendation, that is, multiple knowledge points that meet the historical design requirements and the historical design results, and the design results and the knowledge content involved are recommended and output based on the current three knowledge points that meet the design requirements.

[0049] Based on the same inventive concept as above, the present application also discloses a knowledge-based intelligent design system for mechanical structures. Figure 6 The system includes a demand acquisition module 101, a composite neural network model 102 and an output module 103. The demand acquisition module 101 is connected to the composite neural network model 102, and is used to obtain design demand information and input it into the composite neural network model 102. The composite neural network model 102 includes four parts, which respectively correspond to the knowledge point conversion, knowledge decomposition, knowledge combination and result recommendation as described in the embodiment of the knowledge-based mechanical structure intelligent design method, and are used to implement the following process: converting the design demand information into several knowledge points; decomposing each knowledge point into several knowledge types, wherein each knowledge type corresponds to several knowledge contents; combining the knowledge type and knowledge contents of each knowledge point to finally form multiple knowledge points that meet the design requirements; and recommending the design results and the knowledge contents involved according to the multiple knowledge points that meet the design requirements. The output module 103 is also connected to the composite neural network model 102, and is used to output the recommended design results and the knowledge contents involved.

[0050] Based on the same inventive concept mentioned above, an embodiment of the present application also discloses a knowledge-based mechanical structure intelligent design device, including a memory and a processor, wherein the memory stores at least one instruction, at least one program, code set or instruction set, and the at least one instruction, at least one program, code set or instruction set is loaded and executed by the processor to implement the knowledge-based mechanical structure intelligent design method provided by the above method embodiment.

[0051] Based on the same inventive concept mentioned above, an embodiment of the present application also discloses a computer-readable storage medium, which stores at least one instruction, at least one program, code set or instruction set, and the at least one instruction, at least one program, code set or instruction set can be loaded and executed by a processor to implement the knowledge-based mechanical structure intelligent design method provided by the above method embodiment.

[0052] It should be understood that the "plurality" mentioned in this article refers to two or more. "And / or" describes the association relationship of the associated objects, indicating that there can be three relationships. For example, A and / or B can mean: A exists alone, A and B exist at the same time, and B exists alone. The character " / " generally indicates that the associated objects are in an "or" relationship.

[0053] Those skilled in the art will appreciate that all or part of the steps to implement the above embodiments may be accomplished by hardware or by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium. The storage medium mentioned above may include, for example, various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0054] The above description is only an optional embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.

Claims

1. A knowledge-based intelligent design method for mechanical structures, characterized in that: include: Obtain design requirement information; Convert the design requirement information into a number of knowledge points; Decomposing each of the knowledge points into a plurality of knowledge types and a plurality of knowledge contents corresponding to each of the knowledge types; The knowledge type and knowledge content of each of the knowledge points are combined to finally form a plurality of knowledge points that meet the design requirements; Recommend design results and the knowledge content involved based on multiple knowledge points that meet design requirements; Output the design results and the knowledge content involved.

2. The knowledge-based intelligent design method for mechanical structures according to claim 1, characterized in that: The design requirement information includes several minimum requirement information units; the minimum requirement information units are obtained based on data imported by the user, and / or obtained by semantic extraction based on information input by the user, and / or formed based on several options selected by the user from preset minimum requirement options.

3. The knowledge-based intelligent design method for mechanical structures according to claim 1, characterized in that: The step of converting the design requirement information into a plurality of knowledge points specifically includes: The mapping relationship between historical design requirement information and knowledge points corresponding to the mechanical structure is used to train the model to convert the current design requirement information into a plurality of knowledge points.

4. The knowledge-based intelligent design method for mechanical structures according to claim 1, characterized in that: The knowledge type of each of the above-mentioned knowledge points includes at least one of the following five types: basic theoretical knowledge, experience-based knowledge, rule-based knowledge, demand-constrained knowledge, and other categories; After the step of decomposing each of the knowledge points into a plurality of knowledge types, the method further includes: The knowledge content and quantity corresponding to each of the knowledge types of each of the knowledge points are marked, and the group of marking information is stored in a database.

5. The knowledge-based intelligent design method for mechanical structures according to claim 1, characterized in that: The step of combining the knowledge type and the knowledge content of each of the knowledge points to finally form a plurality of knowledge points that meet the design requirements specifically includes: The knowledge type and knowledge content of each current knowledge point are combined using the mapping relationship training model of the knowledge type and corresponding knowledge content combination of the historical knowledge points to obtain a plurality of knowledge points that meet the design requirements.

6. The knowledge-based intelligent design method for mechanical structures according to claim 1, characterized in that: The step of recommending design results and the knowledge contents involved therein according to a plurality of knowledge points meeting the design requirements specifically includes: The model is trained by utilizing the mapping relationship between multiple knowledge points that meet historical design requirements and historical design results to recommend design results and the knowledge involved based on multiple current knowledge points that meet design requirements.

7. A knowledge-based intelligent design system for mechanical structures, characterized in that: It includes a demand acquisition module, a composite neural network model and an output module; the demand acquisition module is used to acquire design demand information; the output module is used to output the recommended design results; The composite neural network model is used to implement the following process: converting the design requirement information into a number of knowledge points; decomposing each of the knowledge points into a number of knowledge types, wherein each of the knowledge types corresponds to a number of knowledge contents; combining the knowledge type and the knowledge contents of each of the knowledge points to finally form a plurality of knowledge points that meet the design requirements; and recommending design results and the knowledge contents involved based on the plurality of knowledge points that meet the design requirements.

8. A knowledge-based intelligent design device for mechanical structures, characterized in that: It comprises a memory and a processor, wherein the memory stores at least one instruction, at least one program, code set or instruction set, and the at least one instruction, at least one program, code set or instruction set is loaded and executed by the processor to implement the knowledge-based mechanical structure intelligent design method as described in any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that: The readable storage medium stores at least one instruction, at least one program, code set or instruction set, and the at least one instruction, at least one program, code set or instruction set is loaded and executed by a processor to implement the knowledge-based mechanical structure intelligent design method as described in any one of claims 1 to 6.

Images