AI-based method and device for quickly generating positive section of vehicle body sheet metal

Abstract

The application discloses a kind of based on AI's car body sheet metal positive section quick generation method and device.Therein, the method includes: determining the boundary constraint data of car body section based on the total arrangement data of car body, and determining section position data based on the design requirement of car body section;Based on training data set, boundary constraint data and section position data, build and train car body positive section AI generation model, and utilize car body positive section AI generation model to generate multiple section schemes, obtain the section model data and section parameter data corresponding to multiple section schemes;Based on the parameterization processing of the section model data and section parameter data corresponding to optimal section scheme, obtain parameterized section model;Based on parameterized section model, extract section design parameters, and determine the adaptation relationship between parameterized section model and total arrangement data of car body.The application solves the technical problems that section design efficiency is low and cannot adapt to the rapid iteration demand of vehicle model in the prior art.

Description

Technical Field

[0001] This invention relates to the field of automobile manufacturing, and more specifically, to an AI-based method and apparatus for rapidly generating forward cross-sections of car body sheet metal. Background Technology

[0002] The front section of the body sheet metal is a core foundation in the front design of the vehicle body, widely used in the overall body structure design to represent the structural layout of key areas of the body (such as the front door area, encompassing the front door body, door frame, door hinges, glass rails, anti-collision beams, and other related components). It is defined as: a 3D sectional model with complete, reusable, and referable parameters, constructed based on the A-plane level basic data of the vehicle body model, combined with overall layout requirements, vehicle performance targets, and manufacturing standards. Modeling is typically completed using mainstream CAD tools such as CATIA. The core parameters of this model include dimensions, material grades, and performance indicators, directly determining the production feasibility of body sheet metal parts and the overall vehicle performance. It is a key technical link connecting the conceptual design and detailed design of the vehicle body.

[0003] Currently, the industry's process for fabricating frontal sections of body sheet metal still relies entirely on a combination of manual operation and CAD tools. The specific process is as follows: First, based on the section distribution diagram provided by the general layout input, the engineer manually determines the reference plane and reference point for section generation, and manually imports the general layout input data such as the coordinates of the installation points; then, based on their own design experience, they manually complete the section structure drawing and core parameter setting to form an initial section model; after repeated adjustments and optimizations, the final output is the section design data with parameter annotations.

[0004] However, with the accelerating pace of mass production in the automotive industry and the continuously increasing demands for precision in body design, the traditional manual-driven method of front section fabrication has gradually revealed numerous shortcomings, making it difficult to meet the demands of modern design for high efficiency and precision. Meanwhile, the mature application of artificial intelligence (AI) technology in industrial design provides feasible technical support for the intelligent generation of front sections in body sheet metal. Therefore, there is an urgent need for a technical solution that can deeply integrate AI technology with existing CAD / CAA tools to achieve rapid and accurate generation of front sections. Summary of the Invention

[0005] This invention provides an AI-based method and apparatus for rapidly generating forward cross-sections of vehicle body sheet metal, which at least solves the technical problems of low efficiency in cross-section design and inability to adapt to the rapid iteration needs of vehicle models in the prior art.

[0006] According to one aspect of the present invention, an AI-based method for rapidly generating forward cross-sections of vehicle body sheet metal is provided, comprising: acquiring historical vehicle body cross-section data, historical benchmark vehicle cross-section data, and vehicle body layout data; establishing a vehicle body cross-section design standard based on vehicle body cross-section design requirements; and filtering and labeling the historical vehicle body cross-section data and the historical benchmark vehicle cross-section data based on the vehicle body cross-section design standard to obtain a training dataset; determining boundary constraint data of the vehicle body cross-section based on the vehicle body layout data, and determining cross-section position data based on the vehicle body cross-section design requirements; and constructing and training a vehicle body based on the training dataset, the boundary constraint data, and the cross-section position data. An AI-generated model for forward cross-sections is used to generate multiple cross-section schemes, obtaining cross-section model data and cross-section parameter data corresponding to the multiple cross-section schemes. The optimal cross-section scheme selected from the multiple cross-section schemes is obtained, and parameterization processing is performed based on the cross-section model data and cross-section parameter data corresponding to the optimal cross-section scheme to obtain a parameterized cross-section model. Cross-section design parameters are extracted based on the parameterized cross-section model, and the adaptation relationship between the parameterized cross-section model and the overall vehicle layout data is determined. The parameterized cross-section model, the cross-section design parameters, and the adaptation relationship are associated to obtain forward cross-section design data for the vehicle sheet metal.

[0007] According to another aspect of the present invention, an AI-based device for rapid generation of forward cross-sections of vehicle body panels is also provided, comprising: an acquisition module configured to acquire historical vehicle body cross-section data, historical benchmark vehicle cross-section data, and vehicle body layout data; establish vehicle body cross-section design standards based on vehicle body cross-section design requirements; and filter and label the historical vehicle body cross-section data and the historical benchmark vehicle cross-section data based on the vehicle body cross-section design standards to obtain a training dataset; a determination module configured to determine boundary constraint data of the vehicle body cross-section based on the vehicle body layout data, and determine cross-section position data based on the vehicle body cross-section design requirements; and a cross-section generation module configured to construct and train an AI generator for forward cross-sections of the vehicle body based on the training dataset, the boundary constraint data, and the cross-section position data. A model is generated to obtain a trained AI-generated model of the vehicle body's front cross-section. Multiple cross-section schemes are generated using this model, yielding cross-section model data and cross-section parameter data corresponding to each scheme. A model generation module is configured to acquire the optimal cross-section scheme selected from the multiple schemes and perform parameterization processing based on the cross-section model data and parameter data corresponding to the optimal scheme to obtain a parameterized cross-section model. A design data generation module is configured to extract cross-section design parameters based on the parameterized cross-section model, determine the adaptation relationship between the parameterized cross-section model and the vehicle body layout data, and associate the parameterized cross-section model, the cross-section design parameters, and the adaptation relationship to obtain the vehicle body sheet metal front cross-section design data.

[0008] According to another aspect of the present invention, a computer-readable storage medium is also provided, the computer-readable storage medium including a stored program, wherein, when the program is executed, it controls the device where the computer-readable storage medium is located to perform the method described above.

[0009] According to another aspect of the present invention, a computer device is also provided, comprising: a memory and a processor, the memory storing a computer program; the processor being configured to execute the computer program stored in the memory, wherein the computer program, when executed, causes the processor to perform the method described above.

[0010] According to another aspect of the present invention, a computer program product is also provided, including a computer program that, when executed by a processor, implements the steps of the method described above.

[0011] In this embodiment of the invention, the above solution solves the technical problems of low efficiency in surface design and inability to adapt to the rapid iteration requirements of vehicle models in the prior art. Attached Figure Description

[0012] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this application, illustrate exemplary embodiments of the invention and, together with their description, serve to explain the invention and do not constitute an undue limitation thereof. In the drawings:

[0013] Figure 1 This is a flowchart of an optional AI-based method for rapidly generating front cross-sections of car body sheet metal according to an embodiment of the present invention;

[0014] Figure 2 This is a flowchart of another optional AI-based method for rapid generation of front cross-sections of body sheet metal according to an embodiment of the present invention;

[0015] Figure 3 This is a flowchart of an optional calibration method according to an embodiment of the present invention;

[0016] Figure 4 This is a flowchart of an optional method for generating a cross-section using an AI-generated model of a vehicle body front cross-section according to an embodiment of the present invention;

[0017] Figure 5 This is a flowchart of an optional method for CAA parameterization processing according to an embodiment of the present invention;

[0018] Figure 6 This is a flowchart of an optional method for outputting results according to an embodiment of the present invention;

[0019] Figure 7 This is a schematic diagram of an optional AI-based rapid generation device for front cross-section of car body sheet metal according to an embodiment of the present invention.

[0020] Figure 8 A schematic diagram of the structure of a computer device suitable for implementing embodiments of the present disclosure is shown. Detailed Implementation

[0021] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0022] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0023] According to an embodiment of the present invention, a method embodiment for rapidly generating forward cross-sections of body sheet metal based on AI is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0024] Figure 1 This is an AI-based method for rapid generation of front cross-sections of car body sheet metal according to an embodiment of the present invention, such as... Figure 1 As shown, the method includes the following steps:

[0025] Step S102: Obtain historical vehicle body cross-section data, historical benchmark vehicle cross-section data, and overall vehicle layout data. Establish vehicle body cross-section design standards based on vehicle body cross-section design requirements. Filter and label the historical vehicle body cross-section data and the historical benchmark vehicle cross-section data based on the vehicle body cross-section design standards to obtain a training dataset.

[0026] For example, based on the body area and section type in the body section design requirements, performance targets and process requirements corresponding to the body area and section type are determined, and parameter output specifications are determined based on the usage requirements of section parameters, and body layout data import specifications are determined based on the usage requirements of the body layout data; the body section design standard is established based on the performance targets, process requirements, parameter output specifications, and body layout data import specifications; based on the body section design standard, the historical body section data and the historical benchmark vehicle section data are screened to remove abnormal data that does not conform to the body section design standard, and valid section data is obtained; section structure parameters are extracted from the valid section data, and performance results corresponding to the section structure parameters are obtained; the section structure parameters and performance results are labeled accordingly to obtain labeled valid section data, and the training dataset is formed based on the labeled valid section data.

[0027] Step S104: Determine the boundary constraint data of the vehicle body section based on the overall vehicle body layout data, and determine the section position data based on the vehicle body section design requirements.

[0028] Based on the overall vehicle layout data, determine the assembly relationship, tolerance requirements, and positioning datum corresponding to the vehicle body section; acquire the sealing strip data and mounting point data from the overall vehicle layout data, and determine the spatial constraint conditions of the vehicle body section based on the sealing strip data and the mounting point data; form the boundary constraint data based on the assembly relationship, the tolerance requirements, the positioning datum, and the spatial constraint conditions; determine the key areas of the vehicle body section according to the vehicle body section design requirements, determine the section position in the key areas, and determine the corresponding section reference plane and reference point based on the section position to obtain the section position data.

[0029] Step S106: Construct and train a vehicle body front cross-section AI generation model based on the training dataset, the boundary constraint data, and the cross-section position data, and use the vehicle body front cross-section AI generation model to generate multiple cross-section schemes.

[0030] First, an AI generation model for the front cross-section of the vehicle body is constructed and trained. Based on an improved U-Net neural network, the model is built, and training feature data is obtained by extracting cross-section features, correlation features between cross-section parameters and performance, and matching features between cross-section parameters and the overall vehicle layout data from the training dataset. Assembly constraint features and spatial constraint features of the cross-section are determined based on the boundary constraint data, and spatial position features are determined based on the cross-section position data. These features are then associated with the training feature data to obtain constraint association features. The training feature data and constraint association features are input into the AI ​​generation model, and the attention mechanism in the improved U-Net neural network is used to process these features, obtaining similarity features corresponding to each cross-section feature. The similarity score is used to determine the feature weights corresponding to each cross-section feature. These weights are then used to weight the corresponding cross-section features, resulting in attention-enhanced cross-section features. Cross-section parameter prediction and cross-section structure prediction are performed based on these enhanced features, yielding prediction results. The model training loss is determined based on the parameter prediction deviation between the predicted and actual cross-section parameters, and the structural matching degree between the predicted and actual cross-section structures. The model parameters of the vehicle body front cross-section AI generation model are adjusted based on this training loss. The adjusted model parameters are then used to iteratively train the training feature data and the constraint-related features until the vehicle body front cross-section AI generation model meets the preset model accuracy requirements, resulting in the trained vehicle body front cross-section AI generation model. Finally, multiple cross-section schemes are generated using the vehicle body front cross-section AI generation model.

[0031] Step S108: Obtain the optimal cross-section scheme selected from the multiple cross-section schemes, and perform parameterization processing based on the cross-section model data and cross-section parameter data corresponding to the optimal cross-section scheme to obtain a parameterized cross-section model.

[0032] Obtain the cross-section model data and cross-section parameter data corresponding to the optimal cross-section scheme; determine the key structural dimensions in the cross-section sketch based on the cross-section parameter data, and perform parameterized constraint and parameter binding processing on the key structural dimensions to establish a relationship between the cross-section structural dimensions in the cross-section model data and the cross-section parameter data, thereby obtaining a parameterized cross-section structure; generate the parameterized cross-section model based on the parameterized cross-section structure, wherein the parameterized cross-section model retains the relationship between the cross-section structural dimensions and the cross-section parameter data.

[0033] Step S110: Extract section design parameters based on the parametric section model, determine the adaptation relationship between the parametric section model and the vehicle body layout data, and associate the parametric section model, the section design parameters and the adaptation relationship to obtain the vehicle body sheet metal forward section design data.

[0034] The cross-sectional dimension parameters, material parameters, performance parameters, and process parameters are extracted from the parametric cross-sectional model, and the cross-sectional dimension parameters, material parameters, performance parameters, and process parameters are correlated and organized to obtain the cross-sectional design parameters; the sealing strip data and installation point data in the vehicle body layout data are obtained, and the matching relationship between the sealing strip, installation point, and cross-section is determined based on the sealing strip data and installation point data to obtain the adaptation relationship; the parametric cross-sectional model, the cross-sectional design parameters, and the vehicle body layout data adaptation relationship are correlated to obtain the vehicle body sheet metal forward cross-sectional design data.

[0035] This invention focuses on the vehicle body area and constructs a rapid generation system for frontal cross-sections of body sheet metal through three stages: calibration, AI generation, and CAA implementation. By embedding the trained AI model into CAD tools such as CATIA, collaborative processing is achieved from determining the cross-section location and generating cross-section schemes to outputting parametric cross-sections. Simultaneously, by pre-establishing unified design standards and verifying the generated results, the AI ​​generation process has a clear design basis, and the rationality and feasibility of the generated cross-section schemes are controlled, ensuring that the final generated frontal cross-sections meet performance, process, and overall layout requirements.

[0036] The core technologies of this invention mainly include the following aspects: 1) Standardization system for vehicle body cross-section design. Unify existing vehicle body cross-section design requirements and formulate vehicle body cross-section design standards, including performance targets, process requirements, parameter output specifications, and data import specifications. 2) Construction of AI generation model. Construct an AI generation model based on the structural characteristics of the vehicle body sheet metal cross-section. Train the AI ​​generation model based on historical valid cross-section data, historical benchmark vehicle cross-section data, and general layout data, enabling the AI ​​generation model to generate multiple cross-section schemes that meet design requirements according to the input design conditions. Simultaneously, standardize the input and output data of the AI ​​model to facilitate data interaction with subsequent CAD tools. 3) AI and CAD tool integration technology. Interface the trained AI generation model with CAD tools such as CATIA, enabling the transfer of design data such as cross-section position data and general layout data between the AI ​​model and CAD tools. The AI-generated cross-section schemes can be directly imported into CAD tools for visualization and further processing, reducing manual data conversion operations. 4) CAA parametric processing. Based on the CAA interface and corresponding commands, the manually selected optimal cross-section scheme is parametrically processed, converting the cross-section model into an editable and adjustable parametric model. On this basis, a cross-section sketch is regenerated according to the cross-section reference plane, and features such as rounded corners and sharp corners in the cross-section are identified and their parameters are inferred in reverse. Furthermore, through parameter association, the overall linkage adjustment of the cross-section structure is achieved, thus forming editable and reusable parametric forward cross-section data.

[0037] Figure 2 This is a flowchart illustrating an AI-based method for rapid generation of front cross-sections of vehicle body sheet metal according to an embodiment of the present invention. Figure 2 As shown, the method includes the following steps:

[0038] Step S202: Perform calibration.

[0039] This step is fundamental to the rapid generation of front cross-sections of the vehicle body sheet metal. Its main purpose is to establish unified standards for vehicle body cross-section design, create a training dataset for AI model training, determine the design boundaries and spatial locations required for cross-section generation, and provide a design basis for subsequent AI generation and result verification. Specifically, such as... Figure 3 As shown, it includes the following steps:

[0040] Step S2022: Develop body cross-section design standards.

[0041] Based on the design requirements of different areas and cross-sectional types of the vehicle body, corresponding performance targets and process requirements are determined. These cross-sectional types include the body itself, anti-collision beams, door hinges, and glass rails. Further, output specifications for cross-sectional parameters are determined, including dimensional accuracy, material labeling methods, and parameter labeling methods. Simultaneously, import specifications for the overall vehicle layout data are determined based on the usage requirements of the overall vehicle layout data, which includes sealing strip data and installation point data. Based on the determined performance targets, process requirements, parameter output specifications, and overall vehicle layout data import specifications, combined with existing design standards and the design requirements of different OEMs, a vehicle body cross-sectional design standard is formed.

[0042] Step S2024: Filter and label the training data.

[0043] Historical vehicle body cross-section data and historical benchmark vehicle cross-section data are acquired, and the historical vehicle body cross-section data and historical benchmark vehicle cross-section data are filtered according to the vehicle body cross-section design standards to remove abnormal data with dimensional deviations, performance indicators or data formats that do not meet the requirements, so as to obtain valid cross-section data.

[0044] The effective cross-sectional data is labeled with structural parameters and performance results. The structural parameters include cross-sectional thickness, fillet radius, and cross-sectional shape, while the performance results include stiffness and collision energy absorption data corresponding to the structural parameters. Simultaneously, a matching relationship is established between the vehicle body layout data and the cross-sectional parameters to obtain labeled effective cross-sectional data. A training dataset is formed based on this labeled effective cross-sectional data for training the subsequent AI-generated vehicle body front cross-section model.

[0045] Step S2026: Generate boundary constraint data.

[0046] Obtain the overall vehicle body layout data, and determine the assembly relationship, tolerance requirements, and positioning datum corresponding to the vehicle body section based on the overall vehicle body layout data; further obtain the sealing strip data and installation point data, and determine the spatial constraint conditions of the vehicle body section based on the sealing strip data and installation point data.

[0047] Based on the assembly relationship, tolerance requirements, positioning reference, and spatial constraints, corresponding boundary constraint data is formed. This boundary constraint data is then used as the constraint condition for subsequent AI-generated cross-section schemes, so that the generated cross-section schemes meet the vehicle assembly requirements and avoid interference with the overall vehicle layout data.

[0048] Step S2028: Determine the location of the cross-section.

[0049] Obtain the CATIA vehicle body model, and determine the key areas of the vehicle body section in the CATIA vehicle body model according to the vehicle body section design requirements and the connection relationship of parts. The key areas include the middle of the anti-collision beam, the upper and lower ends of the door frame, and the door hinge installation area.

[0050] The corresponding cross-sectional positions are determined in the key areas, and the corresponding reference planes and reference points are determined based on the cross-sectional positions to obtain cross-sectional position data and associated part data information. The cross-sectional position data is associated with the vehicle body cross-sectional design standards, the vehicle body overall layout data, and the boundary constraint data as the spatial location basis for subsequent AI-generated cross-sectional schemes.

[0051] Step S204: Perform AI generation.

[0052] This step is the core execution stage of the rapid generation process of the front cross-section of the vehicle body sheet metal. It is mainly used to build and train an AI generation model for the front cross-section of the vehicle body based on the vehicle body cross-section design standards, training dataset, boundary constraint data, and cross-section position data obtained in step S202. The trained AI generation model is then used to generate cross-section schemes, while simultaneously enabling data interaction and cross-section scheme fusion between the AI ​​generation model and the CATIA tool. Specifically... Figure 4 As shown, it includes the following steps:

[0053] Step S2041: Build an AI-generated model of the vehicle body's front cross-section.

[0054] An AI-generated model of the vehicle body's forward cross-section is built based on an improved U-Net neural network algorithm, and an attention mechanism is integrated into the improved U-Net neural network. The improved U-Net neural network is optimized for the continuity of the vehicle body cross-section structure and the correlation between cross-section parameters, used to extract cross-section features and establish an adaptation relationship between cross-section features and the overall vehicle body layout data.

[0055] When using the attention mechanism for cross-sectional feature processing, the similarity score corresponding to each cross-sectional feature is obtained, and the weight of each cross-sectional feature is determined based on the similarity score. Then, the corresponding cross-sectional features are weighted according to the weight to obtain cross-sectional features enhanced by the attention mechanism, thereby strengthening the identification of key structural features, including the anti-collision beam cross-section, hinge mounting position, etc.

[0056] Specifically, the formula for calculating the attention mechanism weights is as follows: ,in The weight of the feature of the i-th cross section, and These are the similarity scores for the i-th and j-th cross-sectional features, respectively, where n is the total number of cross-sectional features. This is used to enhance the recognition accuracy of key structural features (such as the cross-section of the crash beam and the hinge mounting position). Weights for the attention mechanism.

[0057] Based on the cross-sectional features enhanced by the attention mechanism, cross-sectional parameters are predicted to obtain a cross-sectional parameter vector, which includes cross-sectional structural parameters such as cross-sectional thickness and fillet radius. During the cross-sectional parameter prediction process, the cross-sectional features are parameter-mapped using a model weight matrix, input feature vector, and bias term. The parameter mapping result is then processed using an activation function to obtain the cross-sectional parameter vector.

[0058] Specifically, the formula for predicting cross-sectional parameters is as follows: ,in This is the predicted cross-sectional parameter vector (including thickness, fillet radius, etc.). This is the model weight matrix. For the input feature vector, For bias terms, Use the Sigmoid activation function to ensure that the output parameters are within a reasonable range.

[0059] Furthermore, during the training process of the AI ​​generation model for the front cross section of the vehicle body, the model training loss is determined based on the parameter prediction deviation between the predicted cross section parameters (cross section parameter vector) and the actual cross section parameters, as well as the structural matching degree between the predicted cross section structure and the actual cross section structure. The AI ​​generation model for the front cross section of the vehicle body is then optimized based on the model training loss, so that the AI ​​generation model for the front cross section of the vehicle body can balance the accuracy of cross section parameter prediction and the adaptability of cross section structure.

[0060] Specifically, the model loss function is (used for training optimization): ,in Mean absolute error (a measure of parameter prediction bias). The intersection-to-exchange ratio (a measure of the degree of matching of cross-sectional structures) is used. This is a weighting coefficient (with a value of 0.6-0.8), used to balance parameter accuracy and structural adaptability.

[0061] Step S2042: Construct the input and output data formats for the AI ​​model.

[0062] Obtain the vehicle body cross-section design standard, boundary constraint data, cross-section position data, and vehicle body overall layout data obtained in step S202, and encapsulate the vehicle body cross-section design standard, boundary constraint data, cross-section position data, and vehicle body overall layout data to form the input data for the vehicle body forward cross-section AI generation model.

[0063] In this embodiment, the input data is organized in JSON format to ensure compatibility with CATIA's data export format. The input data includes performance targets, boundary constraints, section location coordinates, vehicle body layout data, and design standards. Specifically, the performance targets characterize the performance requirements of the section to be generated; the boundary constraints characterize the spatial and assembly constraints of the section; the section location coordinates characterize the spatial location of the section; the vehicle body layout data characterizes the fit between the section and other vehicle body components; and the design standards provide the design requirements upon which the section generation is based.

[0064] In this embodiment, the input data in JSON format can be organized as follows:

[0065] {

[0066] "Performance Goals": {

[0067] Collision energy absorption of the crash beam: ≥800J

[0068] Door assembly stiffness: ≥200MPa

[0069] },

[0070] Boundary constraints: {

[0071] "Clearance": ≥5mm

[0072] Positioning reference: X=1250mm, Y=450mm, Z=1800mm

[0073] },

[0074] "Cross-section location coordinates": [

[0075] {

[0076] "Area": ​​"Middle of the crash beam",

[0077] "X": 1250,

[0078] "Y": 450,

[0079] "Z": 1800

[0080] }, ...

[0082] ],

[0083] "Overall vehicle layout data": {

[0084] Format: "CATProduct",

[0085] "Path": ". / data / seal.CATProduct",

[0086] Matching error: ≤ ±0.1mm

[0087] },

[0088] Design Standards: ". / standard / body_front_door_standard.pdf"

[0089] }

[0090] The vehicle body front section AI generation model generates a section scheme based on the input data, and outputs the section model data and section parameter data corresponding to the section scheme according to the preset output data format.

[0091] The cross-sectional model data adopts the CATPart format compatible with CATIA, and establishes a parameter binding relationship between the cross-sectional parameters and the cross-sectional model, so that the cross-sectional model data can be directly imported into CATIA tools for visualization and structural viewing, and the corresponding cross-sectional parameters can be obtained while viewing the cross-sectional structure.

[0092] The cross-sectional parameter data is output in Excel format to form a parameter list. The parameter list includes information such as parameter name, parameter value, precision, unit, corresponding cross-sectional area, and process requirements, which is used to parametrically describe the generated cross-sectional scheme.

[0093] Step S2043: Construct and process the AI ​​model training dataset.

[0094] Obtain the training dataset formed in step S202, and perform data cleaning on the training dataset to remove duplicate and abnormal data, thereby obtaining effective training data.

[0095] Feature extraction is performed on the effective training data to extract vehicle body cross-sectional features, correlation features between cross-sectional parameters and performance, and matching features between vehicle body layout data and cross-sectional parameters, thereby obtaining training feature data.

[0096] The training dataset is divided based on the training feature data to obtain a training dataset for training the AI ​​generation model of the front section of the vehicle body and a verification dataset for testing the accuracy of the AI ​​generation model of the front section of the vehicle body, thereby providing a data foundation for subsequent model training and accuracy debugging.

[0097] Step S2044: Train and debug the AI ​​generation model for the front cross section of the vehicle body.

[0098] The training dataset is input into the vehicle body front section AI generation model, and the vehicle body front section AI generation model is trained based on the vehicle body section features, internal structural relationships, correlation features between section parameters and performance, and matching features between vehicle body layout data and section parameters in the training dataset, to obtain the trained vehicle body front section AI generation model.

[0099] The verification dataset is input into the trained vehicle body front section AI generation model to obtain the section parameter prediction results and section structure prediction results output by the trained vehicle body front section AI generation model. The parameter prediction accuracy and section structure matching accuracy of the model are determined based on the section parameter prediction results and section structure prediction results, respectively.

[0100] The parameter prediction accuracy and the cross-section structure matching accuracy are compared with the preset accuracy requirements. When the parameter prediction accuracy or the cross-section structure matching accuracy fails to meet the accuracy requirements, the internal parameters of the AI-generated model of the vehicle body front cross section are adjusted, and the model training and verification are re-executed based on the adjusted model parameters.

[0101] Repeat the process of model training, accuracy verification, and model parameter adjustment until the model accuracy of the AI-generated front cross-section of the vehicle body meets the accuracy requirements, and obtain the trained AI-generated front cross-section of the vehicle body, which can be used for subsequent cross-section scheme generation.

[0102] Step S2045: Generate a cross-section scheme using the trained AI generation model for the front cross-section of the vehicle body.

[0103] Input the AI ​​model input data into the trained AI generation model of the vehicle body front cross section obtained in step S2044.

[0104] The AI ​​generation model for the front cross-section of the vehicle body predicts the cross-sectional features and parameters of the cross-section to be generated based on the performance targets, boundary constraints, cross-section location data, overall vehicle layout data, and vehicle cross-section design standards in the input data of the AI ​​model, and generates a cross-section scheme based on the predicted cross-sectional features and parameters.

[0105] In this embodiment, the vehicle body front section AI generation model can generate multiple section schemes based on the same set of AI model input data to form section schemes with different structural forms or different parameter combinations, and output the CATPart format section model data and Excel format parameter list corresponding to the multiple section schemes according to the output data format determined in step S2042.

[0106] Step S2046: Integrate the AI-generated model of the vehicle body front cross section with the CATIA tool.

[0107] The trained AI-generated model of the vehicle body front section is embedded into the CATIA plugin. The CATIA plugin is developed based on the CATIA CAA API, and the data interaction relationship between the AI-generated model of the vehicle body front section and the CATIA tool is established through the CATIA plugin.

[0108] The cross-section location data and vehicle body layout data determined in step S202 are obtained through the CATIA data export interface, and the cross-section location data and vehicle body layout data are transmitted to the vehicle body front cross-section AI generation model, so that the vehicle body front cross-section AI generation model generates cross-section schemes based on the design data obtained in real time.

[0109] Obtain the CATPart format cross-sectional model data output by the AI-generated model of the vehicle body's front cross-section, and directly import the CATPart format cross-sectional model data into the CATIA tool, so that the CATIA tool can visualize and view the structure of the cross-sectional model data.

[0110] Furthermore, by establishing the binding relationship between section parameters and section models, the linkage relationship between section positions and vehicle body digital models, and the linkage relationship between vehicle body layout data and section parameters through the CATIA plugin, engineers can view the adaptation relationship of section parameters, section positions, and vehicle body layout data corresponding to the section scheme in the CATIA tool, and make preliminary adjustments to the generated section scheme.

[0111] Step S206: Perform CAA parameterization processing.

[0112] This step primarily obtains the optimal cross-section scheme determined in the preceding steps, and utilizes the CATIA CAA V5 API to perform parametric settings, sketch regeneration, reverse inference of cross-section features, and overall linkage adjustments on the optimal cross-section scheme. This transforms the optimal cross-section scheme into an editable, adjustable, and reusable parametric cross-section model to meet the needs of subsequent design adjustments and data reuse. Specifically, as follows... Figure 5 As shown, it includes the following steps:

[0113] Step S2061: Obtain the optimal cross-section scheme and perform parameter settings.

[0114] Obtain the optimal cross-section scheme determined in the preceding steps, and read the CATPart format cross-section model data corresponding to the optimal cross-section scheme from the CATIA tool.

[0115] The CATPart format cross-sectional model data is read through the model reading interface in the CATIA CAA V5 API. The model reading interface uses the CATIPart interface to obtain the Part object corresponding to the cross-sectional model.

[0116] Furthermore, the cross-sectional model data is parameterized through a parameterization setting interface, which includes the CATISketchEditor interface and the CATIDimension interface. The CATISketchEditor interface is used to enter the cross-sectional sketch editing state, and the CATIDimension interface is used to establish parameterized constraints and parameter binding relationships for the key structural dimensions of the cross-sectional structure, thus associating the structural dimensions in the cross-sectional model with their corresponding parameters.

[0117] In this embodiment, the parameterization setting process can be executed according to the following command flow:

[0118] / / 1. Read the cross-sectional model

[0119] CATIDocument_var spDocument =

[0120] spApplication->OpenDocument(". / output / section.CATPart");

[0121] CATIPart_var spPart = spDocument->GetPart();

[0122] / / 2. Launch the sketch editor and enter parametric mode.

[0123] CATISketchEditor_var spSketchEditor =

[0124] spPart->GetSketchEditor();

[0125] spSketchEditor->StartEditing();

[0126] / / 3. Parametrically bind key dimensions of the cross-section

[0127] CATIDimension_var spDimension =

[0128] spSketchEditor->CreateDimension(

[0129] spLine,

[0130] CATDimension::CAT_DIMENSION_LENGTH,

[0131] "Thickness");

[0132] spDimension->SetValue(2.0);

[0133] Through the above parametric processing, the original cross-sectional model is converted into a parametric cross-sectional model with editable parameters, so that structural dimensions such as cross-sectional thickness and fillet radius can be adjusted independently, and the parameter adjustment can be applied to the corresponding cross-sectional structure.

[0134] Step S2062: Generate a cross-sectional sketch.

[0135] Obtain the cross-sectional position data of the optimal cross-sectional scheme, and obtain the cross-sectional reference plane or reference plane from the cross-sectional position data.

[0136] The cross-sectional reference plane is read through CAA's CATISketchEditor interface, and the cross-sectional sketch is regenerated based on the cross-sectional reference plane and the cross-sectional line data corresponding to the optimal cross-sectional scheme, so that the regenerated cross-sectional sketch establishes a correspondence with the cross-sectional reference plane.

[0137] In this embodiment, the cross-sectional sketch is generated based on the cross-sectional reference plane and the cross-sectional line data using the CreateSketchFromReference command. The command flow is as follows:

[0138] CATISketch_var spSketch =

[0139] spSketchEditor->CreateSketchFromReference(

[0140] spReferencePlane,

[0141] spSectionLines);

[0142] By regenerating the cross-section sketch based on the cross-section reference plane, the cross-section sketch is matched with the reference of the vehicle body model, providing a unified reference for subsequent cross-section parameter adjustment and cross-section feature recognition, and reducing the possibility of reference deviation during cross-section adjustment.

[0143] Step S2063: Infer the cross-sectional characteristic parameters in reverse.

[0144] Obtain a cross-sectional sketch regenerated based on the cross-sectional reference plane, and perform feature recognition on the cross-sectional sketch using the CAA feature recognition interface, wherein the feature recognition interface adopts the CATIFeatureRecognizer interface.

[0145] Based on the curvature of the lines and the angular relationship between the lines in the cross-sectional sketch, the sharp corner features in the cross-sectional sketch are identified and rounded corner features are obtained. Based on the identified sharp corner features and rounded corner features, the corresponding feature parameters are determined. The feature parameters of the rounded corner features include the rounded corner radius and the rounded corner position, and the feature parameters of the sharp corner features include the sharp corner angle and the sharp corner position.

[0146] In this embodiment, the corresponding feature recognition command can be called through the CATIFEatureRecognizer interface to identify rounded corner features and sharp corner features in the cross-sectional sketch. The core commands are as follows:

[0147] CATIFeature_var spFillets =

[0148] spFeatureRec->RecognizeFeatures(

[0149] CATFeatureType::CAT_FILLET);

[0150] CATIFeature_var spChamfers =

[0151] spFeatureRec->RecognizeFeatures(

[0152] CATFeatureType::CAT_CHAMFER);

[0153] Furthermore, the identified rounded and sharp corner features are annotated with parameters through the CATIDimension interface, and the feature parameters corresponding to the rounded and sharp corner features are associated with the cross-sectional sketch to obtain feature parameterized data, thereby providing a parameterized basis for subsequent manual adjustments.

[0154] Step S2064: Perform overall linkage adjustment of the cross-section.

[0155] Acquire parameterized feature data and determine the target feature parameters to be adjusted according to process requirements and performance requirements. The target feature parameters include parameters such as fillet radius and sharp corner angle.

[0156] The system receives adjustment operations for the target feature parameters and determines the cross-sectional structure and dimensional parameters associated with the target feature parameters based on the correlation between the target feature parameters and the cross-sectional structure.

[0157] The linkage relationship between the target feature parameters and the cross-sectional structure and dimensional parameters is established through the CAA linkage adjustment interface. The linkage adjustment interface uses the CATIRElation interface, and the linkage adjustment is executed through the SetLinkAdjust command. The core command is as follows:

[0158] spRelation->SetLinkAdjust(

[0159] spFillet,

[0160] spRelatedLines,

[0161] TRUE);

[0162] Based on the established linkage relationship, when the target feature parameter is adjusted, the cross-sectional structure and size parameters associated with the target feature parameter are adjusted synchronously, so that the associated cross-sectional structure automatically adapts to the target feature parameter, and a parameterized cross-sectional model that completes the overall linkage adjustment is obtained.

[0163] In this embodiment, the CATPart format cross-section model data corresponding to the optimal cross-section scheme is obtained, and the Part object corresponding to the cross-section model is obtained through the model reading interface in the CATIA CAA V5 API; the sketch editing state of the cross-section model is entered through the parameterization setting interface, and parameterized constraints and parameter binding relationships are established for the key structural dimensions in the cross-section model, so that the key structural dimensions are associated with the corresponding structural parameters; the cross-section position data corresponding to the optimal cross-section scheme is obtained, and the cross-section reference plane or reference plane is determined according to the cross-section position data; the cross-section sketch is regenerated based on the cross-section reference plane and the cross-section line data corresponding to the optimal cross-section scheme through the sketch editing interface, so that the cross-section sketch is associated with the cross-section reference plane; the cross-section sketch is feature-recognized, and the rounded corner features and sharp corner features in the cross-section sketch are identified according to the line curvature and the angle relationship between the lines in the cross-section sketch, and the rounded corner radius and rounded corner position corresponding to the rounded corner feature and the sharp corner angle and sharp corner position corresponding to the sharp corner feature are determined. The parameter annotation interface associates the feature parameters corresponding to the rounded corner features and the sharp corner features with the cross-sectional sketch to obtain feature parameterization data. Based on process and performance requirements, it determines the target feature parameter to be adjusted from the feature parameterization data, receives adjustment operations for the target feature parameter, and determines the cross-sectional structure and dimensional parameters associated with the target feature parameter based on the association between the target feature parameter and the cross-sectional structure. Through a linkage adjustment interface, it establishes a linkage relationship between the target feature parameter and the cross-sectional structure and dimensional parameters, so that when the target feature parameter is adjusted, the cross-sectional structure and dimensional parameters associated with the target feature parameter are adjusted synchronously, resulting in a parameterized cross-sectional model that has completed overall linkage adjustment. Through the above processing, when adjusting a single cross-sectional parameter, it is not necessary to modify the associated cross-sectional structure and dimensional parameters segment by segment; the overall adjustment of the cross-sectional structure can be achieved while maintaining the integrity and parameter consistency of the cross-sectional model.

[0164] Step S208: Output the results.

[0165] This step is used to organize the parametric section model obtained in step S206 into body sheet metal forward section design data that can be directly used for design and production. Based on the parametric section model, section structure parameters, material parameters, performance parameters, and process parameters are extracted, and combined with the body layout data adaptation results to form a parametric 3D model, a complete parameter list, and body layout data adaptation relationships, thereby obtaining the final body sheet metal forward section design data. Specifically, as follows... Figure 6 The steps shown are as follows:

[0166] Step S2081: Generate a 3D parametric cross-sectional model.

[0167] Obtain the parametric cross-section model obtained in step S206, and save the parametric cross-section model according to the CATIA general format to obtain a 3D parametric cross-section model in CATPart format.

[0168] The 3D parametric section model retains the correlation between the section structure dimensions and parameters, enabling the 3D parametric section model to be edited and modified in CATIA tools and used for subsequent detailed body design.

[0169] Step S2082: Generate a complete parameter list.

[0170] Obtain the cross-sectional dimension parameters, material parameters, performance parameters, and process parameters from the parameterized cross-sectional model, and organize the cross-sectional dimension parameters, material parameters, performance parameters, and process parameters to generate a complete parameter list.

[0171] The cross-sectional dimension parameters include cross-sectional thickness and fillet radius, the material parameters include national standard grade, the performance parameters include stiffness and impact energy absorption, and the process parameters include welding gap.

[0172] The complete parameter list is output in Excel format to obtain a parameter list in .xlsx format for subsequent design and manufacturing.

[0173] Step S2083: Generate the vehicle body layout data adaptation relationship.

[0174] Obtain the vehicle body layout data and the parametric section model from the aforementioned steps, and obtain the sealing strip data and installation point data from the vehicle body layout data.

[0175] Based on the sealing strip data, the mounting point data, and the parameterized cross-section model, the matching relationship between the sealing strip, mounting point, and cross-section is analyzed to obtain the vehicle body layout data adaptation relationship. The vehicle body layout data adaptation relationship includes the matching parameters between the sealing strip and the cross-section, the matching parameters between the mounting point and the cross-section, and the corresponding adaptation analysis results, which are used to characterize the adaptation relationship between the parameterized cross-section model and the vehicle body layout data.

[0176] Step S2084: Generate the front cross-sectional design data of the body sheet metal.

[0177] The parametric cross-sectional model of the 3D model, the cross-sectional design parameters in the complete parameter list, and the vehicle body layout data adaptation relationship are obtained. The parametric cross-sectional model of the 3D model, the cross-sectional design parameters, and the vehicle body layout data adaptation relationship are associated to form the positive cross-sectional design data of the body sheet metal.

[0178] The body sheet metal front section design data includes a 3D parametric section model in CATPart format, a complete parameter list in Excel format, and body layout data adaptation relationships in Word format.

[0179] The body sheet metal forward section design data can be used for the subsequent design and manufacturing of body sheet metal parts, and can be reused as section design data for subsequent models, thereby reducing repetitive design and parameter labeling work.

[0180] This application has the following beneficial effects:

[0181] 1) Significantly improved design efficiency. The AI ​​model can generate 3 to 5 cross-sectional designs that meet design requirements within 1 to 2 hours, essentially replacing the core work of manual drawing and repeated parameter tuning. Simultaneously, through deep integration with CAD tools, data can flow seamlessly between platforms, avoiding the extra time spent on manual export and re-import. Combined with the cross-sectional overall linkage adjustment mechanism implemented by the CAA tool, parameter modification becomes more efficient, with single-item modification efficiency improved by over 80%. In actual project verification, the cross-sectional design cycle for the front door area was reduced from 3 to 5 working days to within 1 working day, an overall efficiency improvement of over 70%, effectively adapting to the rapid iteration development pace of current vehicle models.

[0182] 2) Design accuracy has been significantly improved.

[0183] The AI ​​model is trained on a training dataset, employing an improved U-Net neural network structure, and strictly controls the accuracy threshold (parameter deviation not exceeding ±0.1mm, position deviation not exceeding ±0.2mm). The resulting cross-sectional design parameters are accurate and closely match the overall layout input data, effectively reducing design deviations caused by differences in human experience and minimizing potential problems in subsequent production and assembly stages.

[0184] 3) Data reusability and standardization have been greatly improved.

[0185] The standardized design standards and parameter output specifications, uniformly defined by humans, ensure that the AI-generated cross-sections contain standardized parameter information. The parametric 3D models output by the CAA tool achieve integrated binding of "cross-section + parameters," with neat, uniform, and compliant parameter annotations. This data not only facilitates the delivery of the current project but can also be directly reused for cross-sections in subsequent vehicle models, eliminating the need for repeated modeling and re-annotation. This reduces secondary development costs by more than 60% overall and effectively solves the problem of data reuse difficulties in existing processes.

[0186] This application also provides an AI-based device for rapid generation of front cross-sections of automotive body sheet metal, such as... Figure 7As shown, it includes: an acquisition module 72, configured to acquire historical vehicle body cross-section data, historical benchmark vehicle cross-section data, and vehicle body layout data; establish a vehicle body cross-section design standard based on vehicle body cross-section design requirements; and filter and label the historical vehicle body cross-section data and the historical benchmark vehicle cross-section data based on the vehicle body cross-section design standard to obtain a training dataset; a determination module 74, configured to determine the boundary constraint data of the vehicle body cross-section based on the vehicle body layout data, and determine the cross-section position data based on the vehicle body cross-section design requirements; and a cross-section generation module 76, configured to construct and train a vehicle body forward cross-section AI generation model based on the training dataset, the boundary constraint data, and the cross-section position data to obtain a trained vehicle body forward cross-section. An AI-generated model is used to generate multiple cross-section schemes, obtaining cross-section model data and cross-section parameter data corresponding to the multiple cross-section schemes. The model generation module 78 is configured to obtain the optimal cross-section scheme selected from the multiple cross-section schemes, and perform parameterization processing based on the cross-section model data and cross-section parameter data corresponding to the optimal cross-section scheme to obtain a parameterized cross-section model. The design data generation module 79 is configured to extract cross-section design parameters based on the parameterized cross-section model, determine the adaptation relationship between the parameterized cross-section model and the overall vehicle layout data, and associate the parameterized cross-section model, the cross-section design parameters, and the adaptation relationship to obtain the body sheet metal front cross-section design data.

[0187] It should be noted that the AI-based rapid generation device for front cross-sections of body sheet metal provided in the above embodiments is only an example of the division of the above functional modules. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above. In addition, the AI-based rapid generation device for front cross-sections of body sheet metal provided in the above embodiments and the AI-based rapid generation method for front cross-sections of body sheet metal belong to the same concept. The specific implementation process is detailed in the method embodiments and will not be repeated here.

[0188] Figure 8 A schematic diagram of a computer device suitable for implementing embodiments of the present disclosure is shown. It should be noted that... Figure 8 The computer device shown is merely an example and should not be construed as limiting the functionality and scope of the embodiments disclosed herein.

[0189] like Figure 8As shown, the computer device includes a central processing unit (CPU) 1001, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 1002 or a program loaded from a storage section 1008 into a random access memory (RAM) 1003. The RAM 1003 also stores various programs and data required for system operation. The CPU 1001, ROM 1002, and RAM 1003 are interconnected via a bus 1004. An input / output (I / O) interface 1005 is also connected to the bus 1004.

[0190] The following components are connected to I / O interface 1005: an input section 1006 including a keyboard, mouse, etc.; an output section 1007 including a cathode ray tube (CRT), liquid crystal display (LCD), etc., and speakers, etc.; a storage section 1008 including a hard disk, etc.; and a communication section 1009 including a network interface card such as a LAN card, modem, etc. The communication section 1009 performs communication processing via a network such as the Internet. A drive 1010 is also connected to I / O interface 1005 as needed. A removable medium 1011, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., is installed on drive 1010 as needed so that computer programs read from it can be installed into storage section 1008 as needed.

[0191] The above description is only a preferred embodiment of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A method for rapid generation of front cross-sections of car body sheet metal based on AI, characterized in that, include: Acquire historical vehicle body cross-section data, historical benchmark vehicle cross-section data, and overall vehicle layout data. Establish vehicle body cross-section design standards based on vehicle body cross-section design requirements. Filter and label the historical vehicle body cross-section data and the historical benchmark vehicle cross-section data based on the aforementioned vehicle body cross-section design standards to obtain a training dataset. The boundary constraint data of the vehicle body section is determined based on the overall vehicle body layout data, and the section position data is determined based on the vehicle body section design requirements. Based on the training dataset, the boundary constraint data, and the cross-section location data, a vehicle body front cross-section AI generation model is constructed and trained, and multiple cross-section schemes are generated using the vehicle body front cross-section AI generation model. The optimal cross-section scheme selected from the multiple cross-section schemes is obtained, and parameterization processing is performed based on the cross-section model data and cross-section parameter data corresponding to the optimal cross-section scheme to obtain a parameterized cross-section model. Based on the parametric section model, section design parameters are extracted, and the adaptation relationship between the parametric section model and the vehicle body layout data is determined. The parametric section model, the section design parameters, and the adaptation relationship are then associated to obtain the body sheet metal forward section design data.

2. The method according to claim 1, characterized in that, A body cross-section design standard is established based on the body cross-section design requirements. Based on this standard, the historical body cross-section data and the historical benchmark vehicle cross-section data are filtered and labeled to obtain the training dataset, which includes: Based on the body area and section type in the body section design requirements, determine the performance targets and process requirements corresponding to the body area and section type, determine the parameter output specifications based on the usage requirements of the section parameters, and determine the body layout data import specifications based on the usage requirements of the body layout data. The body section design standard is established based on the performance objectives, process requirements, parameter output specifications, and body layout data import specifications. Based on the aforementioned vehicle body cross-section design standards, the historical vehicle body cross-section data and the historical benchmark vehicle cross-section data are filtered to remove abnormal data that do not conform to the aforementioned vehicle body cross-section design standards, thereby obtaining valid cross-section data. Cross-sectional structural parameters are extracted from the effective cross-sectional data, and performance results corresponding to the cross-sectional structural parameters are obtained. The cross-sectional structural parameters and the performance results are labeled accordingly to obtain labeled effective cross-sectional data, and the training dataset is formed based on the labeled effective cross-sectional data.

3. The method according to claim 1, characterized in that, Based on the overall vehicle layout data, boundary constraint data for the vehicle body section is determined, and based on the vehicle body section design requirements, section location data is determined, including: Based on the overall vehicle layout data, determine the assembly relationships, tolerance requirements, and positioning datums corresponding to the vehicle cross-sections; Obtain the sealing strip data and mounting point data from the overall vehicle layout data, and determine the spatial constraint conditions of the vehicle body section based on the sealing strip data and the mounting point data; The boundary constraint data is formed based on the assembly relationship, the tolerance requirements, the positioning datum, and the spatial constraints. Based on the vehicle body cross-section design requirements and the component overlap relationship, the key areas of the vehicle body cross-section are determined, the cross-section position is determined in the key areas, and the corresponding cross-section reference plane and reference point are determined based on the cross-section position to obtain the cross-section position data.

4. The method according to claim 1, characterized in that, Based on the training dataset, the boundary constraint data, and the cross-section location data, an AI generation model for the front cross-section of the vehicle body is constructed and trained, including: The AI ​​generation model of the front cross section of the vehicle body is built based on the improved U-Net neural network. The cross section features, internal structural relationships, correlation features between cross section parameters and performance, and matching features between the cross section parameters and the overall layout data of the vehicle body are extracted from the training dataset to obtain training feature data. Based on the boundary constraint data, the assembly constraint features and spatial constraint features of the vehicle body section are determined, and based on the section position data, the spatial position features of the vehicle body section are determined. The assembly constraint features, the spatial constraint features, and the spatial position features are associated with the training feature data to obtain constraint association features. The training feature data and the constraint-related features are input into the vehicle body front section AI generation model, and the attention mechanism in the improved U-Net neural network is used to perform feature processing on the training feature data and the constraint-related features to obtain the similarity score corresponding to each section feature. The feature weight corresponding to each section feature is determined according to the similarity score, and the corresponding section features are weighted using the feature weight to obtain the attention mechanism-enhanced section features. Based on the enhanced cross-sectional features of the attention mechanism, cross-sectional parameters and cross-sectional structures are predicted to obtain cross-sectional parameter prediction results and cross-sectional structure prediction results. The model training loss is determined according to the parameter prediction deviation between the cross-sectional parameter prediction results and the actual cross-sectional parameters, and the structural matching degree between the cross-sectional structure prediction results and the actual cross-sectional structures. The model parameters of the AI ​​generation model for the front cross section of the vehicle body are adjusted based on the model training loss, and the training feature data and the constraint association features are iteratively trained based on the adjusted model parameters until the AI ​​generation model for the front cross section of the vehicle body meets the preset model accuracy requirements, thus obtaining the trained AI generation model for the front cross section of the vehicle body.

5. The method according to claim 1, characterized in that, Based on the cross-section model data and cross-section parameter data corresponding to the optimal cross-section scheme, parameterization processing is performed to obtain the parameterized cross-section model, including: Obtain the cross-section model data and cross-section parameter data corresponding to the optimal cross-section scheme; Based on the cross-sectional parameter data, the key structural dimensions of the cross-sectional structure in the cross-sectional sketch are determined, and the key structural dimensions are subjected to parameterized constraints and parameter binding to obtain the parameterized cross-sectional model.

6. The method according to claim 1, characterized in that, Based on the parametric section model, the section design parameters are extracted, and the adaptation relationship between the parametric section model and the vehicle body layout data is determined. The parametric section model, the section design parameters, and the adaptation relationship are then associated to obtain the vehicle body sheet metal forward section design data, including: The cross-sectional dimension parameters, material parameters, performance parameters, and process parameters are extracted from the parametric cross-sectional model, and the cross-sectional dimension parameters, material parameters, performance parameters, and process parameters are correlated and organized to obtain the cross-sectional design parameters. Obtain the sealing strip data and mounting point data from the vehicle body layout data, and determine the matching relationship between the sealing strip, mounting point and cross-sectional structure in the parameterized cross-sectional model based on the sealing strip data and mounting point data to obtain the adaptation relationship; By associating the parametric section model, the section design parameters, and the adaptation relationship, the positive section design data of the vehicle body sheet metal is obtained.

7. A rapid generation device for front cross-section of car body sheet metal based on AI, characterized in that, include: The acquisition module is configured to acquire historical vehicle body cross-section data, historical benchmark vehicle cross-section data, and vehicle body layout data; establish vehicle body cross-section design standards based on vehicle body cross-section design requirements; and filter and label the historical vehicle body cross-section data and the historical benchmark vehicle cross-section data based on the vehicle body cross-section design standards to obtain a training dataset. The determination module is configured to determine the boundary constraint data of the vehicle body section based on the overall vehicle body layout data, and to determine the section position data based on the vehicle body section design requirements; The cross-section generation module is configured to construct and train a vehicle body front cross-section AI generation model based on the training dataset, the boundary constraint data and the cross-section position data, to obtain the trained vehicle body front cross-section AI generation model, and to generate multiple cross-section schemes using the vehicle body front cross-section AI generation model. The model generation module is configured to obtain the optimal cross-section scheme selected from the plurality of cross-section schemes, and perform parameterization processing based on the cross-section model data and cross-section parameter data corresponding to the optimal cross-section scheme to obtain a parameterized cross-section model. The design data generation module is configured to extract section design parameters based on the parametric section model, determine the adaptation relationship between the parametric section model and the vehicle body layout data, and associate the parametric section model, the section design parameters and the adaptation relationship to obtain the body sheet metal forward section design data.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored program, wherein, when the program is executed, it controls the device on which the computer-readable storage medium is located to perform the method according to any one of claims 1 to 6.

9. A computer device, characterized in that, include: Memory and processor The memory stores computer programs; The processor is configured to execute a computer program stored in the memory, wherein when the computer program is executed, the processor performs the method according to any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

Images