An intelligent interaction-based structural design analysis and optimization integrated method and system

Abstract

The application discloses a kind of based on intelligent interaction's structure design analysis optimization integrated method and system, the method includes following steps: obtaining input content and converting into text information, obtains semantic information based on semantic recognition, compare and extract the key word of structure design, the key parameter of structure design and the important information of structure analysis with corpus, generate optimization program, match corresponding optimization program and fill in parameter data in semantic information respectively, generate executable program;Carry out CAD system modeling, carry out mesh division, and adopt equal geometric proportion boundary finite element to carry out structure performance analysis, call the optimization program of design target and carry out topological optimization, when satisfying topological optimization convergence condition, smooth processing is carried out to topological structure and the topological structure after optimization is output.The application realizes the integrated design of structure modeling, analysis and optimization, with the characteristics of simple operation, short processing time and high intelligent degree.

Description

Technical Field

[0001] This invention relates to the field of structural design optimization technology, specifically to an integrated method and system for structural design analysis and optimization based on intelligent interaction. Background Technology

[0002] With the continuous development of intelligent manufacturing, traditional structural design methods can no longer meet the design needs of high-end manufacturing. Advanced product structural design methods with short design cycles and high automation are crucial for achieving intelligent manufacturing, interconnectivity, and improving product structural performance. However, existing structural design technologies often employ design processes such as CAD modeling, CAE analysis, structural optimization, and model reconstruction. CAD models cannot be directly analyzed using CAE, requiring cumbersome model conversion, and designers must be involved throughout the entire design process, resulting in high labor and time costs. Therefore, there is an urgent need for a technical solution that enables intelligent interaction in areas such as intelligent product manufacturing, achieving intelligent and efficient integrated structural design, analysis, and optimization to meet the current design needs of high-end manufacturing. Summary of the Invention

[0003] To overcome the shortcomings and deficiencies of existing technologies, this invention provides an integrated method and system for structural design, analysis, and optimization based on intelligent interaction. This invention utilizes NLP algorithms to extract keywords from the input information provided by designers and supplements and improves necessary information through human-computer interaction. By calling a program command library, the above information is imported into the integrated design system to generate an optimization program, thereby carrying out integrated design of structural modeling, analysis, and optimization. The geometrical scale boundary finite element method based on NURBS-enhanced quadtrees / octrees can be used for structural analysis and optimization. This invention solves the problems of high professional requirements for designers and low system intelligence in traditional structural design processes. Furthermore, this invention features simple operation, short processing time, and high intelligence.

[0004] To achieve the above objectives, the present invention adopts the following technical solution:

[0005] This invention provides an integrated method for structural design, analysis, and optimization based on intelligent interaction, comprising the following steps:

[0006] Obtain the designer's voice or text input;

[0007] Based on speech recognition, voice input is converted into text information;

[0008] Semantic information is obtained by performing semantic recognition on text information based on NLP algorithms;

[0009] We construct a keyword command library for structural design, a key parameter command library for structural design, and an important information command library for structural analysis as a corpus for training NLP algorithms.

[0010] By comparing semantic information with a corpus, keywords for structural design, key parameters for structural design, and important information for structural analysis can be extracted from the semantic information.

[0011] Based on the preset parameters required for structural design analysis and optimization, it is determined whether the extracted semantic information is missing or whether the parameters required for structural design analysis and optimization are modified. If it is determined that the parameters are missing or modified, the currently identified information is stored, and the missing information is returned or the identified information is modified. If it is determined that the parameters are not missing or modified, the optimization program is generated.

[0012] The optimization procedure includes: determining the optimization method, design domain, design objective, constraints, and parameters required for structural design analysis and optimization; constructing a topology optimization model; performing numerical solutions based on the finite element method; conducting sensitivity analysis and suppressing numerical instability based on a filtering method; iteratively updating the design variables until the convergence conditions are met, and then outputting the optimal material distribution scheme.

[0013] Based on the optimization method, design goal, and constraints obtained from the semantic information, the corresponding optimization program is matched, and the parameter data from the semantic information is filled into the optimization program to generate an executable program.

[0014] The system performs automated modeling in a CAD system, meshes the structure, and uses equal-geometric-scale boundary finite element method for structural performance analysis. It then calls the optimization program corresponding to the design objective to perform topology optimization. When the topology optimization convergence condition is met, the topology is smoothed and the optimized topology is output.

[0015] As a preferred technical solution, the keyword command library for structural design includes: design objectives, load conditions, and constraint conditions;

[0016] The key parameter command library for structural design includes: material elastic modulus or Young's modulus, Poisson's ratio, density, heat transfer coefficient, thermoelastic matrix, fluid viscosity, and stress;

[0017] The essential command library for structural analysis includes static analysis, dynamic analysis, thermodynamic analysis, heat dissipation analysis, heat conduction analysis, and fluid analysis.

[0018] As a preferred technical solution, the design objectives include: minimum flexibility, maximum stiffness, structural flexibility, minimum heat dissipation weakness, etc.; the boundary conditions include: the upper / lower / left / right boundaries are completely fixed, the upper / lower / left / right boundaries are fixed only in the x-direction, y-direction, or z-direction, constraining an endpoint, constraining a feature point, and insulating boundaries;

[0019] The load conditions include: force, pressure, load, thermal load, heat source, and flow velocity.

[0020] The constraints include: volume fraction, stress constraint, vibration frequency constraint, and displacement constraint.

[0021] As a preferred technical solution, the process of returning and supplementing the missing information or modifying the identified information involves supplementing the missing information by calling default values ​​and NLP semantic recognition.

[0022] In CAD modeling, determine whether the model is fully defined; in isogeometric statics analysis, determine whether all material properties, boundary conditions, and load conditions are given; in isogeometric topology optimization, determine whether the design domain, objective function, and constraints are given.

[0023] As a preferred technical solution, the optimization methods include: topology optimization, shape optimization, and size optimization;

[0024] The design objectives include: minimum structural flexibility, lightest mass, smallest volume, minimum heat dissipation, minimum energy consumption, and multi-objective coupling;

[0025] The constraints include: volume constraints, displacement constraints, stress constraints, temperature constraints, and multiple constraints.

[0026] As a preferred technical solution, the automated modeling of the CAD system uses isogeometric splines to construct spline boundary representation models, wherein the isogeometric splines include NURBS, T-splines, and B-splines;

[0027] The meshing is performed using spline basis functions or quadtree / octree meshes.

[0028] As a preferred technical solution, automated modeling is performed using a CAD system, mesh generation is carried out, and structural performance analysis is conducted using a finite element method with equal geometric scale. Topology optimization is then performed by calling the optimization program corresponding to the design objectives, specifically including:

[0029] The structural model is built using NURBS, and the mesh is generated based on quadtree decomposition technology. The stiffness matrix of the elements inside the design domain is pre-stored. For the elements on the NURBS boundary, the node vector corresponding to the intersection point is obtained by inverse calculation of the points based on the coordinates of the intersection point of the NURBS curve and the quadtree mesh through the physical point information. The new NURBS curve is obtained by inserting the nodes. The control points of the reconstructed NURBS curve are distributed at the intersection point of the NURBS curve and the quadtree mesh.

[0030] A scaled boundary coordinate system is established based on the scaled boundary finite element method. The boundary elements are reduced in dimension and discretized into NURBS line elements and ordinary line elements. The displacement fields of the two are constructed by NURBS shape functions and traditional Lagrange shape functions, respectively. Numerical solutions are performed in the circumferential direction and analytical solutions are performed in the radial direction.

[0031] For three-dimensional problems, meshing is performed based on octree decomposition. The internal hexahedral elements are pre-stored using stiffness matrices. For elements on the boundary, the NURBS surface information is reconstructed by inverse point calculation and node insertion based on the intersection of the NURBS surface and the octree mesh. The boundary elements are then discretized into NURBS surfaces and ordinary planes using proportional boundary finite element method, and analyzed and solved separately to complete the topology optimization of the equal geometric scale boundary finite element method.

[0032] As a preferred technical solution, the executable program includes:

[0033] Establish a CAD model, generate NURBS surfaces, decompose and divide the design domain using a quadtree and construct NURBS elements, and obtain a CAE model based on constraints and load conditions.

[0034] The CAE model is used to call the finite element topology optimization program of equal geometric scale boundary based on control point density to obtain the optimization results, which are then smoothed.

[0035] This invention also provides an integrated system for structural design, analysis and optimization based on intelligent interaction, comprising: an input content acquisition module, an input content conversion module, a semantic recognition module, a corpus construction module, an information extraction module, a semantic information analysis module, an optimization program generation module, an executable program generation module, a CAD modeling module, a mesh generation module, a structural performance analysis module, a topology optimization module, and a result output module;

[0036] The input content acquisition module is used to acquire the voice input or text input content of the designer;

[0037] The input content conversion module is used to convert voice input content into text information based on speech recognition;

[0038] The semantic recognition module is used to perform semantic recognition on text information based on NLP algorithms to obtain semantic information;

[0039] The corpus construction module is used to construct a keyword command library for structural design, a key parameter command library for structural design, and an important information command library for structural analysis as a corpus for NLP algorithm training.

[0040] The information extraction module is used to compare semantic information with the corpus and extract keywords for structural design, key parameters for structural design, and important information for structural analysis from the semantic information.

[0041] The semantic information analysis module is used to determine whether the extracted semantic information is missing or whether the parameters required for structural design analysis and optimization are modified based on preset parameters for structural design analysis and optimization. If it is determined that the parameters are missing or modified, the currently identified information is stored, and the missing information is returned or the identified information is modified. If it is determined that the parameters are not missing or modified, the optimization program is generated.

[0042] The optimization program generation module is used to generate an optimization program, which includes: determining the optimization method, design domain, design objective, constraints and parameters required for structural design analysis and optimization; constructing a topology optimization model; performing numerical solution based on the finite element method; performing sensitivity analysis and suppressing numerical instability based on a filtering method; iteratively updating the design variables until the convergence condition is met, and then outputting the optimal material distribution scheme.

[0043] The executable program generation module is used to match the corresponding optimization program according to the optimization method, design goal and constraint conditions in the acquired semantic information, and at the same time fill the parameter data in the semantic information into the optimization program to generate an executable program.

[0044] The CAD modeling module is used for automated modeling of the CAD system.

[0045] The grid division module is used to perform grid division;

[0046] The structural performance analysis module is used to perform structural performance analysis using the constant geometric scale boundary finite element method.

[0047] The topology optimization module is used to call the optimization program corresponding to the design target to perform topology optimization until the topology optimization convergence condition is met.

[0048] The result output module is used to smooth the topology and output the optimized topology.

[0049] As a preferred technical solution, the mesh generation module is used for mesh generation, and the structural performance analysis module is used for structural performance analysis using the constant geometric scale boundary finite element method, specifically including:

[0050] The structural model is built using NURBS, and the mesh is generated based on quadtree decomposition technology. The stiffness matrix of the elements inside the design domain is pre-stored. For the elements on the NURBS boundary, the node vector corresponding to the intersection point is obtained by inverse calculation of the points based on the coordinates of the intersection point of the NURBS curve and the quadtree mesh through the physical point information. The new NURBS curve is obtained by inserting the nodes. The control points of the reconstructed NURBS curve are distributed at the intersection point of the NURBS curve and the quadtree mesh.

[0051] A scaled boundary coordinate system is established based on the scaled boundary finite element method. The boundary elements are reduced in dimension and discretized into NURBS line elements and ordinary line elements. The displacement fields of the two are constructed by NURBS shape functions and traditional Lagrange shape functions, respectively. Numerical solutions are performed in the circumferential direction and analytical solutions are performed in the radial direction.

[0052] For three-dimensional problems, meshing is performed based on octree decomposition. The internal hexahedral elements are pre-stored using stiffness matrices. For elements on the boundary, the NURBS surface information is reconstructed by inverse point calculation and node insertion based on the intersection of the NURBS surface and the octree mesh. The boundary elements are then discretized into NURBS surfaces and ordinary planes using proportional boundary finite element method, and analyzed and solved separately to complete the topology optimization of the equal geometric scale boundary finite element method.

[0053] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0054] This invention employs an integrated technical solution combining natural language processing, human-computer interaction, and structural design analysis and optimization. It can guide structural design through voice / text input, eliminating the need for cumbersome manual operation procedures and requiring no specialized training for designers. Furthermore, the application of quadtree / octree decomposition technology significantly reduces mesh generation time, while the use of NURBS curves combined with proportional boundary finite element method further reduces the scale of mesh generation. This provides an efficient and accurate solution, solving the problem of low intelligence and automation in traditional structural design processes. It achieves the technical effect of convenient, intuitive, intelligent, interactive, and low-barrier-to-entry structural design. Attached Figure Description

[0055] Figure 1 This is a flowchart illustrating the integrated method for structural design, analysis, and optimization based on intelligent interaction according to the present invention.

[0056] Figure 2 This is a schematic diagram of the user interface of the integrated structural design, analysis and optimization method based on intelligent interaction of the present invention;

[0057] Figure 3 This is a schematic diagram of the initial CAD model and boundary conditions of the present invention;

[0058] Figure 4 This is a schematic diagram of the topology optimization process of the present invention;

[0059] Figure 5 This is a schematic diagram of the internal unit form after the quadtree decomposition of the present invention;

[0060] Figure 6 This is a schematic diagram of the boundary elements and NURBS reconstruction of the present invention;

[0061] Figure 7 This is a schematic diagram of the topological configuration resulting from the structural design optimization of this invention;

[0062] Figure 8 This is a schematic diagram of an editable CAD model showing the structural design optimization results of this invention. Detailed Implementation

[0063] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0064] Example 1

[0065] like Figure 1 As shown, this embodiment provides an integrated method for structural design, analysis, and optimization based on intelligent interaction. This embodiment uses a classic cantilever beam as an example. The user wants to obtain a cantilever beam of a specified weight. The optimization task is completed based on finite element topology optimization with constant geometric scale boundaries. Of course, this embodiment is not limited to cantilever beam structures, nor is it limited to finite element topology optimization with constant geometric scale boundaries. Specifically, it includes the following steps:

[0066] Step S1: Obtain the designer's voice input or text input;

[0067] Step S2: Use computer-aided automatic speech recognition (APR) technology to process the speech information and convert it into text information;

[0068] like Figure 2 As shown, the text information is extracted as follows: "Perform topology optimization design on a rectangular design domain of 80mm×40mm. The material is aluminum alloy. The goal is structural flexibility. The volume is the constraint. Constrain the left boundary and apply a downward force to the lower right corner." Based on the prompt, the information is supplemented and improved as follows: "Apply a downward force of 10 Newtons to the lower right corner and find the structure after the volume is reduced by 50%."

[0069] Step S3: Based on the text information, use NLP algorithms for semantic recognition, and correct the typos and grammatical errors in the user input based on statistical models;

[0070] In step S3, NLP algorithms are used for semantic recognition. A large amount of speech / text content is collected through input devices, and the semantic material library is denoised to establish a semantic material library for the learning and training of the algorithm. At the same time, the semantic information input by the designers each time is also saved to enrich the semantic material library, which can continuously improve the recognition accuracy of the NLP algorithm.

[0071] Step S4: By comparing with the corpus, extract relevant parameter information, including keywords, key parameters and important information in the text, such as: optimization design goals, constraints, boundary conditions, external loads, material parameters, etc.

[0072] In step S4, keywords, key parameters, and important information are extracted from the text. First, a keyword command library, a key parameter command library, and an important information command library are constructed as the corpus for training the NLP algorithm. The main contents are as follows:

[0073] The keyword command library includes: design objectives, load conditions, and constraint conditions. Design objectives include: minimum flexibility, maximum stiffness, structural flexibility, and minimum heat dissipation weakness. Boundary conditions include: completely fixed upper / lower / left / right boundaries, upper / lower / left / right boundaries fixed only in the x, y, or z directions, constraining an endpoint, constraining a feature point, and insulating boundaries. Load conditions include: force, pressure, load, thermal load, heat source, and flow velocity. Constraint conditions include: volume fraction, stress constraint, vibration frequency constraint, and displacement constraint.

[0074] The key parameter command library includes: material elastic modulus / Young's modulus, Poisson's ratio, density, heat transfer coefficient, thermoelastic matrix, fluid viscosity, stress, etc.

[0075] The important information command library includes: static analysis, dynamic analysis, thermodynamic analysis, heat dissipation analysis, heat conduction analysis, fluid analysis, etc.

[0076] In this embodiment, the CAD model is rectangular in shape, with dimensions of 80mm in length and 40mm in height; the CAE model is derived from the aforementioned rectangle, with aluminum alloy as the filling material, a degree-of-freedom constraint type, and a constraint direction covering all degrees of freedom, applied to the left side of the rectangle; the load type is a concentrated load, with a magnitude of 10N, directed downwards, applied to the lower right corner, as shown below. Figure 3 As shown, the design domain for topology optimization is the aforementioned rectangle, the optimization objective is structural flexibility, and the constraint is a 50% volume constraint.

[0077] Step S5: Based on the parameters necessary for structural design analysis and optimization, such as materials, loads, and constraints, determine whether the extracted text information is missing or needs modification. If no information is missing or needs modification, proceed to the next step. If information is missing or needs modification, store the currently identified text information and return to step S1. A computer dialog box will prompt the designer to supplement the missing information or modify the identified information, thus achieving human-computer interaction. If no information is missing, proceed to step S6.

[0078] In this embodiment, by calling default values ​​and using NLP semantic recognition, necessary but unmentioned conditions in step S4 are automatically supplemented. For example: in the CAD model, the default rectangular modeling position is the origin of the coordinate system, and the rotation angle around each coordinate axis is 0, thus inferring that the two diagonal points of the rectangle are (0,0) and (80mm,40mm); in the CAE model, the default element size is 1mm, inferring that constraints are applied to the line segment from (0,0) to (0,40mm), and concentrated loads are applied at (80mm,0) in the negative y-axis direction; in the topology optimization model, the default optimization convergence condition is that the volume change is less than 1% or the number of iterations is greater than 500. Combining the above conditions, in CAD modeling, it is determined whether the model is fully defined, including position and size; in isogeometric static analysis, it is determined whether material properties, boundary conditions, and load conditions are all given; in isogeometric topology optimization, it is determined whether the design domain, objective function, and constraint conditions are given.

[0079] The human-computer interaction in step S5 allows for the modification and improvement of input information. In this step, the computer system will provide timely feedback on the recognition information for observation and verification. In addition, the recognition information can be modified and improved through voice / text input until the input information meets the designer's requirements and complies with the operating conditions of the integrated program.

[0080] Step S6: Once the text information identified through the corpus meets the relevant optimization settings, the required program command library will be called to embed various necessary information into the optimization program and generate an executable optimization program.

[0081] The program command library in step S6 consists of programs already configured in the system. Each optimization method, design goal, and constraint corresponds to a different optimization program. The corresponding optimization program is invoked based on the optimization method, design goal, and constraints obtained from the acquired text information. Simultaneously, the parameter data from the text information is filled into the program to generate an executable program.

[0082] like Figure 4As shown, taking topology optimization as an example, the optimization procedure includes: 1. Preliminary preparation work, determining the optimization method, design domain, design objective, constraints, and parameters necessary for optimization; 2. Determining the modeling scheme, establishing corresponding mathematical models using optimization methods such as variable density method, level set method, progressive structural optimization method, and moving deformable component method; 3. Numerical solution scheme, using finite element method, equal geometric scale boundary finite element method, etc. for numerical solution; 4. Sensitivity analysis, using density filtering, sensitivity filtering, and other filtering methods to suppress possible numerical instability phenomena, such as gray-scale elements, local extrema, mesh dependence, and checkerboard pattern; 5. Iterative update of design variables, using optimization criterion method, moving asymptote method, sequential convex programming method, etc. for iterative update.

[0083] The optimization methods include topology optimization, shape optimization, and size optimization; the design objectives include minimizing structural flexibility, weight, volume, heat dissipation weakness, and energy consumption, as well as multi-objective coupling. Constraints include volume constraints, displacement constraints, stress constraints, temperature constraints, and multiple constraints.

[0084] In this embodiment, based on the parameters obtained from the above steps and the inferred modules to be executed, a complete executable program is generated. This program includes CAD system automated modeling, constant geometric scale boundary finite element static analysis, constant geometric scale boundary finite element topology optimization based on control point density, and post-processing of optimization results.

[0085] Step S7: Execute the integrated program. First, perform automated modeling of the CAD system according to the program commands, and then perform structural performance analysis on the above model.

[0086] Step S7's automated CAD system modeling primarily utilizes boundary representation models constructed from geometric splines such as NURBS, T-splines, and B-splines. Structural performance analysis can be performed accurately using spline basis functions for mesh generation, or adaptive quadtree / octree mesh generation can be employed, followed by accurate analysis using constant geometric scale boundary finite element method. Using constant geometric scale boundary finite element method for performance analysis allows direct use of the spline boundary representation model generated by the CAD system, eliminating the need for interpolation approximation of the precise CAD boundaries. Structural optimization involves calling the optimization program corresponding to the design objectives for topology optimization.

[0087] Step S7, the integration process, can be achieved using the finite element analysis method with equal geometric scale boundaries. Due to the consistency of mathematical expression, NURBS is used to establish the structural model. Then, adaptive, fast, and efficient mesh generation is performed using a quadtree decomposition technique that satisfies the 2:1 rule. Since there are only 16 element types within the design domain, such as... Figure 5As shown, six of these elements are illustrated. The remaining elements can be derived by rotating the following elements. The problem of suspended nodes can be solved using scaled boundary finite element methods. Therefore, the stiffness matrices of the aforementioned elements can be pre-calculated and stored in memory for quick retrieval during the solution process. For example... Figure 6 As shown, for elements on the NURBS boundary (i.e., elements where the CAD geometric model boundary is located), the intersection coordinates of the NURBS curve and the quadtree mesh are used to inversely calculate the points using physical point information, thereby obtaining the node vectors corresponding to the intersection points. New NURBS curves are then obtained through node insertion. At this point, control points of the reconstructed NURBS curve are distributed at the intersection points of the curve and the quadtree mesh, facilitating subsequent accurate solutions. Then, a scaled boundary coordinate system is established using the scaled boundary finite element method, and the boundary elements are dimensionally reduced and discretized into NURBS line elements and ordinary line elements. Their displacement fields are constructed using NURBS shape functions and traditional Lagrange shape functions, respectively. Numerical solutions are performed in the circumferential direction, and analytical solutions are performed in the radial direction. Since the NURBS curve can accurately describe the boundary of the geometric model, there is no need to provide a large number of seed points for detailed domain boundary subdivision, which undoubtedly reduces meshing costs and further improves solution efficiency. Similarly, for three-dimensional problems, octree decomposition can be used for meshing, and the internal hexahedral elements still use a method of pre-storing the stiffness matrix for fast retrieval. For boundary elements, the NURBS surface information is reconstructed using point inverse calculation and node insertion based on the intersection of the NURBS surface and the octree mesh, facilitating subsequent accurate solutions. Finally, the boundary elements are discretized into NURBS surfaces and ordinary planes using scaled boundary finite element analysis, and analyzed and solved separately. Based on this, equal-geometric-scale boundary finite element topology optimization is performed, achieving integrated structural modeling, optimization, and analysis, avoiding the cumbersome model conversion problems caused by traditional CAE analysis using Lagrange interpolation.

[0088] The executable program generated in step S6 is executed automatically. This executable program performs the following functions: (1) it establishes a rectangular model with diagonal points located at (0, 0) and (80mm, 40mm) respectively, generating a NURBS surface; (2) it decomposes the design domain using a quadtree and constructs NURBS elements, applying the aforementioned constraints and load conditions to obtain a CAE model; (3) based on this CAE model, it calls a finite element topology optimization program based on control point density with equal geometric scale boundaries, where the optimization objective, constraints, and design domain are consistent with those in step S4, such as... Figure 7 As shown, the optimization results are obtained; (4) After obtaining the optimization results, as shown Figure 8 As shown, the results are post-processed to transform the analysis model into an editable CAD model, meeting subsequent design and manufacturing requirements.

[0089] Step S8: Real-time feedback from the intelligent interactive system during the design process allows designers to determine whether optimization requires improving specific structural performance through voice / text input. Finally, the structural optimization program is automatically executed to generate the optimized design structure.

[0090] In step S8, during the integration process, the intelligent interactive system provides real-time feedback on the design flow. Designers can modify the structural performance according to design requirements via voice / text input. The structural optimization system then calls the optimization program corresponding to the design objective to perform topology optimization. When the topology optimization convergence condition is met, the model is smoothed, and the optimized structure is exported. Figure 7 As shown, the optimized model has jagged boundaries, which does not meet manufacturing requirements, so it needs to be smoothed.

[0091] Step S9: The designer judges whether the result of the automatically optimized structure is satisfactory. If satisfied, the design result model is output; if not satisfied, the designer inputs the information to be improved, returns to step S7, and performs structural optimization again until the designer is satisfied, and then outputs the design result model.

[0092] Example 2

[0093] Except for the following technical contents, the technical contents of this embodiment are the same as those of Embodiment 1;

[0094] This embodiment provides an integrated system for structural design, analysis and optimization based on intelligent interaction, including: an input content acquisition module, an input content conversion module, a semantic recognition module, a corpus construction module, an information extraction module, a semantic information analysis module, an optimization program generation module, an executable program generation module, a CAD modeling module, a mesh generation module, a structural performance analysis module, a topology optimization module, and a result output module;

[0095] In this embodiment, the input content acquisition module is used to acquire the designer's voice input content or text input content;

[0096] In this embodiment, the input content conversion module is used to convert voice input content into text information based on speech recognition;

[0097] In this embodiment, the semantic recognition module is used to perform semantic recognition on text information based on NLP algorithms to obtain semantic information;

[0098] In this embodiment, the corpus construction module is used to construct a keyword command library for structural design, a key parameter command library for structural design, and an important information command library for structural analysis as a corpus for NLP algorithm training;

[0099] In this embodiment, the information extraction module is used to compare semantic information with the corpus and extract keywords for structural design, key parameters for structural design, and important information for structural analysis from the semantic information.

[0100] In this embodiment, the semantic information analysis module is used to determine whether the extracted semantic information is missing or whether the parameters required for structural design analysis and optimization are modified based on preset parameters for structural design analysis and optimization. If it is determined that the parameters are missing or modified, the currently identified information is stored, and the missing information is returned or the identified information is modified. If it is determined that the parameters are not missing or modified, an optimization program is generated.

[0101] In this embodiment, the optimization program generation module is used to generate an optimization program, which includes: determining the optimization method, design domain, design objective, constraints and parameters required for structural design analysis and optimization; constructing a topology optimization model; performing numerical solution based on the finite element method; performing sensitivity analysis and suppressing numerical instability based on a filtering method; iteratively updating the design variables until the convergence condition is met and then outputting the optimal material distribution scheme.

[0102] In this embodiment, the executable program generation module is used to match the corresponding optimization program according to the optimization method, design goal and constraint conditions in the acquired semantic information, and fill the parameter data in the semantic information into the optimization program to generate an executable program.

[0103] In this embodiment, the CAD modeling module is used for automated modeling of the CAD system;

[0104] In this embodiment, the mesh generation module is used to perform mesh generation;

[0105] In this embodiment, the structural performance analysis module is used to perform structural performance analysis using the constant geometric scale boundary finite element method;

[0106] In this embodiment, the topology optimization module is used to call the optimization program corresponding to the design goal to perform topology optimization until the topology optimization convergence condition is met.

[0107] In this embodiment, the result output module is used to smooth the topology and output the optimized topology.

[0108] In this embodiment, the mesh generation module is used to perform mesh generation, and the structural performance analysis module is used to perform structural performance analysis using the constant geometric scale boundary finite element method, specifically including:

[0109] The structural model is built using NURBS, and the mesh is generated based on quadtree decomposition technology. The stiffness matrix of the elements inside the design domain is pre-stored. For the elements on the NURBS boundary, the node vector corresponding to the intersection point is obtained by inverse calculation of the points based on the coordinates of the intersection point of the NURBS curve and the quadtree mesh through the physical point information. The new NURBS curve is obtained by inserting the nodes. The control points of the reconstructed NURBS curve are distributed at the intersection point of the NURBS curve and the quadtree mesh.

[0110] A scaled boundary coordinate system is established based on the scaled boundary finite element method. The boundary elements are reduced in dimension and discretized into NURBS line elements and ordinary line elements. The displacement fields of the two are constructed by NURBS shape functions and traditional Lagrange shape functions, respectively. Numerical solutions are performed in the circumferential direction and analytical solutions are performed in the radial direction.

[0111] For three-dimensional problems, meshing is performed based on octree decomposition. The internal hexahedral elements are pre-stored using stiffness matrices. For elements on the boundary, the NURBS surface information is reconstructed by inverse point calculation and node insertion based on the intersection of the NURBS surface and the octree mesh. The boundary elements are then discretized into NURBS surfaces and ordinary planes using proportional boundary finite element method, and analyzed and solved separately to complete the topology optimization of the equal geometric scale boundary finite element method.

[0112] The above embodiments are preferred embodiments of the present invention, but the embodiments of the present invention are not limited to the above embodiments. Any changes, modifications, substitutions, combinations, or simplifications made without departing from the spirit and principle of the present invention shall be considered equivalent substitutions and shall be included within the protection scope of the present invention.

Claims

1. A method for integrated structural design, analysis, and optimization based on intelligent interaction, characterized in that, Includes the following steps: Obtain the designer's voice or text input; Based on speech recognition, voice input is converted into text information; Semantic information is obtained by performing semantic recognition on text information based on NLP algorithms; We construct a keyword command library for structural design, a key parameter command library for structural design, and an important information command library for structural analysis as a corpus for training NLP algorithms. By comparing semantic information with a corpus, keywords for structural design, key parameters for structural design, and important information for structural analysis can be extracted from the semantic information. Based on the preset parameters required for structural design analysis and optimization, it is determined whether the extracted semantic information is missing or whether the parameters required for structural design analysis and optimization are modified. If it is determined that the parameters are missing or modified, the currently identified information is stored, and the missing information is returned or the identified information is modified. If it is determined that the parameters are not missing or modified, the optimization program is generated. The optimization procedure includes: determining the optimization method, design domain, design objective, constraints, and parameters required for structural design analysis and optimization; constructing a topology optimization model; performing numerical solutions based on the finite element method; conducting sensitivity analysis and suppressing numerical instability based on a filtering method; iteratively updating the design variables until the convergence conditions are met, and then outputting the optimal material distribution scheme. Based on the optimization method, design goal, and constraints obtained from the semantic information, the corresponding optimization program is matched, and the parameter data from the semantic information is filled into the optimization program to generate an executable program. The process involves automated modeling using a CAD system, mesh generation, and structural performance analysis using constant geometric scale boundary finite element method. Topology optimization is then performed by calling the optimization program corresponding to the design objectives. Specifically, this includes: The structural model is built using NURBS, and the mesh is generated based on quadtree decomposition technology. The stiffness matrix of the elements inside the design domain is pre-stored. For the elements on the NURBS boundary, the node vector corresponding to the intersection point is obtained by inverse calculation of the points based on the coordinates of the intersection point of the NURBS curve and the quadtree mesh through the physical point information. The new NURBS curve is obtained by inserting the nodes. The control points of the reconstructed NURBS curve are distributed at the intersection point of the NURBS curve and the quadtree mesh. A scaled boundary coordinate system is established based on the scaled boundary finite element method. The boundary elements are reduced in dimension and discretized into NURBS line elements and ordinary line elements. The displacement fields of the two are constructed by NURBS shape functions and traditional Lagrange shape functions, respectively. Numerical solutions are performed in the circumferential direction and analytical solutions are performed in the radial direction. For three-dimensional problems, meshing is performed based on octree decomposition. The internal hexahedral elements are pre-stored using stiffness matrices. For the elements on the boundary, the NURBS surface information is reconstructed by inverse calculation of points and node insertion based on the intersection of the NURBS surface and the octree mesh. The boundary elements are reduced in dimension and discretized into NURBS surfaces and ordinary planes using proportional boundary finite element method, and analyzed and solved separately to complete the topology optimization of the equal geometric scale boundary finite element method. When the convergence condition for topology optimization is met, the topology structure is smoothed and the optimized topology structure is output.

2. The integrated method for structural design, analysis, and optimization based on intelligent interaction as described in claim 1, characterized in that, The keyword command library for structural design includes: design objectives, load conditions, and constraint conditions; The key parameter command library for structural design includes: material elastic modulus or Young's modulus, Poisson's ratio, density, heat transfer coefficient, thermoelastic matrix, fluid viscosity, and stress; The essential command library for structural analysis includes static analysis, dynamic analysis, thermodynamic analysis, heat dissipation analysis, heat conduction analysis, and fluid analysis.

3. The integrated method for structural design, analysis, and optimization based on intelligent interaction as described in claim 2, characterized in that, The design objectives include: minimum flexibility, maximum stiffness, structural flexibility, and minimum heat dissipation weakness; the boundary conditions include: the upper / lower / left / right boundaries are completely fixed, the upper / lower / left / right boundaries are fixed only in the x-direction, y-direction, or z-direction, constraining an endpoint, constraining a feature point, and insulating boundaries; The load conditions include: force, pressure, load, thermal load, heat source, and flow velocity. The constraints include: volume fraction, stress constraint, vibration frequency constraint, and displacement constraint.

4. The integrated method for structural design, analysis, and optimization based on intelligent interaction as described in claim 1, characterized in that, The process of returning missing information or modifying identified information involves supplementing missing information by calling default values ​​and NLP semantic recognition. In CAD modeling, determine whether the model is fully defined; In isogeometric statics analysis, it is determined whether all material properties, boundary conditions, and load conditions are given; in isogeometric topology optimization, it is determined whether the design domain, objective function, and constraints are given.

5. The integrated method for structural design, analysis, and optimization based on intelligent interaction according to claim 1, characterized in that, The optimization methods include: topology optimization, shape optimization, and size optimization. The design objectives include: minimum structural flexibility, lightest weight, smallest volume, minimum heat dissipation, minimum energy consumption, and multi-objective coupling; The constraints include: volume constraints, displacement constraints, stress constraints, temperature constraints, and multiple constraints.

6. The integrated method for structural design, analysis, and optimization based on intelligent interaction as described in claim 1, characterized in that, The automated modeling of the CAD system uses isogeometric splines to construct spline boundary representation models, including NURBS, T-splines, and B-splines. The meshing is performed using spline basis functions or quadtree / octree meshes.

7. The integrated method for structural design, analysis, and optimization based on intelligent interaction according to claim 1, characterized in that, The executable program includes: Establish a CAD model, generate NURBS surfaces, decompose and divide the design domain using a quadtree and construct NURBS elements, and obtain a CAE model based on constraints and load conditions. The CAE model is used to call the finite element topology optimization program of equal geometric scale boundary based on control point density to obtain the optimization results, which are then smoothed.

8. A structural design, analysis, and optimization integrated system based on intelligent interaction, characterized in that, The method for implementing the integrated structural design, analysis and optimization method based on intelligent interaction as described in any one of claims 1-7 includes: an input content acquisition module, an input content conversion module, a semantic recognition module, a corpus construction module, an information extraction module, a semantic information analysis module, an optimization program generation module, an executable program generation module, a CAD modeling module, a mesh generation module, a structural performance analysis module, a topology optimization module, and a result output module. The input content acquisition module is used to acquire the voice input or text input content of the designer; The input content conversion module is used to convert voice input content into text information based on speech recognition; The semantic recognition module is used to perform semantic recognition on text information based on NLP algorithms to obtain semantic information; The corpus construction module is used to construct a keyword command library for structural design, a key parameter command library for structural design, and an important information command library for structural analysis as a corpus for NLP algorithm training. The information extraction module is used to compare semantic information with the corpus and extract keywords for structural design, key parameters for structural design, and important information for structural analysis from the semantic information. The semantic information analysis module is used to determine whether the extracted semantic information is missing or whether the parameters required for structural design analysis and optimization are modified based on preset parameters for structural design analysis and optimization. If it is determined that the parameters are missing or modified, the currently identified information is stored, and the missing information is returned or the identified information is modified. If it is determined that the parameters are not missing or modified, the optimization program is generated. The optimization program generation module is used to generate an optimization program, which includes: determining the optimization method, design domain, design objective, constraints and parameters required for structural design analysis and optimization; constructing a topology optimization model; performing numerical solution based on the finite element method; performing sensitivity analysis and suppressing numerical instability based on a filtering method; iteratively updating the design variables until the convergence condition is met, and then outputting the optimal material distribution scheme. The executable program generation module is used to match the corresponding optimization program according to the optimization method, design goal and constraint conditions in the acquired semantic information, and at the same time fill the parameter data in the semantic information into the optimization program to generate an executable program. The CAD modeling module is used for automated modeling of the CAD system. The grid division module is used to perform grid division; The structural performance analysis module is used to perform structural performance analysis using the constant geometric scale boundary finite element method. The topology optimization module is used to call the optimization program corresponding to the design target to perform topology optimization until the topology optimization convergence condition is met. The result output module is used to smooth the topology and output the optimized topology.

Images