Vehicle body grid parameter dynamic optimization method and system based on model simulation

Abstract

The invention discloses a vehicle body grid parameter dynamic optimization method and system based on model simulation, and belongs to the crossing field of computer-aided engineering and artificial intelligence technology. According to an existing vehicle body grid optimization scheme, deep mining and analysis of grid data cannot be achieved, and consequently the accuracy and reliability of a grid optimization result are not high. According to the vehicle body grid parameter dynamic optimization method based on model simulation, a vehicle body historical database module, a vehicle body grid optimization module and a domain knowledge base module are utilized to match vehicle body grid data to be optimized with vehicle body historical data to obtain the most similar vehicle body historical data; the information is used as experience information and verification basis of vehicle body grid parameter optimization; and then experience information and a verification basis are reused, the to-be-optimized grid data of the vehicle body are analyzed and processed, and optimization suggestion information is generated, so that historical data can be deeply mined and analyzed, and the accuracy and reliability of a grid optimization result can be effectively improved.

Description

Technical Field

[0001] This invention relates to a method and system for dynamic optimization of vehicle body mesh parameters based on model simulation, belonging to the interdisciplinary field of computer-aided engineering and artificial intelligence technology. Background Technology

[0002] In the field of automotive body design and optimization, adjusting the body mesh parameters is crucial. Traditional body mesh parameter tuning primarily relies on manual experience and trial-and-error methods. Engineers need to manually adjust the body mesh parameters in design software based on their professional knowledge and experience, and then conduct multiple simulation experiments to observe whether the mesh optimization results meet design requirements. This method requires a large amount of manual operation and numerous experiments, resulting in very low optimization efficiency. Furthermore, manual experience has inherent subjectivity and limitations, affecting the accuracy of body mesh optimization.

[0003] Chinese Patent (Announcement No.: CN109977460B) discloses a multi-objective optimization design method based on vehicle body cross-section parameterization, including vehicle body cross-section parameterization modeling and multi-objective optimization of vehicle body cross-sections. Based on the vehicle body cross-section input, the overlapping joints of the main cross-sections of the vehicle body parameterization model are established. Based on the styling CAS data, an external feature model of the parameterized vehicle body is established. Using the previous generation vehicle body finite element model or competitor vehicle body data, a detailed parameterized vehicle body model is established. Variables are recorded for the main cross-sections of the vehicle body to complete the vehicle body parameterization modeling. Based on the parameterized model, the finite element mesh of the DOE test matrix is ​​completed. The finite element model of the test matrix is ​​solved one by one to obtain the vehicle body mass, modal, torsional stiffness, and bending stiffness performance. An approximate response surface model is established. Based on the approximate model, the dimensions of the main cross-sections of the vehicle body are optimized without increasing the weight to obtain the optimal vehicle body stiffness performance.

[0004] The above solution is based on the previous generation of vehicle body finite element model or competitor vehicle body data to complete the parametric modeling of the vehicle body. It has a narrow range of applicable scenarios and cannot flexibly optimize various vehicle body mesh parameters.

[0005] Furthermore, the above scheme relies on an approximate model to optimize the main cross-sectional dimensions of the vehicle body by fitting functional relationships. This approach fails to achieve in-depth mining and analysis of grid data and makes it difficult to fully consider various complex factors and conditions, resulting in low accuracy and reliability of the grid optimization results.

[0006] The information disclosed in this background section is only for understanding the background of the inventive concept, and therefore may include information that does not constitute prior art. Summary of the Invention

[0007] To address the aforementioned problems, or one of them, the present invention aims to provide a method and system for dynamic optimization of vehicle body mesh parameters based on model simulation. This method utilizes a vehicle body historical database module, a vehicle body mesh optimization module, and a domain knowledge base module to match the vehicle body mesh data to be optimized with historical vehicle body data, obtaining the most similar historical data as empirical information and verification basis for vehicle body mesh parameter optimization. Then, the empirical information and verification basis are reused to analyze and process the vehicle body mesh data to be optimized, generating optimization suggestion information. This allows for in-depth mining and analysis of historical data, comprehensively considering various complex factors and conditions. Furthermore, based on the optimization suggestion information, the vehicle body mesh data to be optimized is precisely adjusted and optimized, thus effectively improving the accuracy and reliability of mesh optimization results. This method is applicable to various vehicle body design scenarios and can flexibly optimize various vehicle body mesh parameters.

[0008] To address the aforementioned problems, or one of them, the second objective of this invention is to provide a method and system for dynamic optimization of vehicle body mesh parameters based on model simulation. This method effectively solves the problem of insufficient utilization of real-time data and historical vehicle body data in traditional methods, accurately uncovering potential patterns and trends in the data. By combining historical vehicle body data with domain knowledge, it generates specific and accurate optimization suggestions for minor adjustments to vehicle body mesh parameters, forming a closed loop of "simulation-analysis-optimization." This significantly improves the efficiency and accuracy of vehicle body design and optimization in detail processing, and shortens the time for tuning vehicle body mesh parameters during the automotive R&D cycle.

[0009] To achieve one of the above objectives, the first technical solution of the present invention is as follows: A method for dynamic optimization of vehicle body mesh parameters based on model simulation includes the following: Collect the grid data of the vehicle body to be optimized, and obtain several historical data of the vehicle body from the pre-built historical database module of the vehicle body; Based on the pre-built vehicle mesh optimization module, the vehicle mesh data to be optimized is matched with the vehicle historical data to obtain the most similar vehicle historical data, which serves as empirical information and verification basis for vehicle mesh parameter optimization. Based on the domain knowledge in the pre-built domain knowledge base module, and by reusing experience information and verification basis, the vehicle body mesh data to be optimized is analyzed and processed to generate optimization suggestion information; Based on the optimization suggestions, the mesh data of the vehicle body to be optimized is dynamically adjusted and optimized to obtain the optimized mesh data of the vehicle body.

[0010] This invention utilizes a vehicle body history database module, a vehicle body mesh optimization module, and a domain knowledge base module to match the vehicle body mesh data to be optimized with historical vehicle body data, obtaining the most similar historical vehicle body data as empirical information and verification basis for vehicle body mesh parameter optimization. Then, the empirical information and verification basis are reused to analyze and process the vehicle body mesh data to be optimized, generating optimization suggestion information. This allows for in-depth mining and analysis of historical data, comprehensively considering various complex factors and conditions. Furthermore, based on the optimization suggestion information, the vehicle body mesh data to be optimized is precisely adjusted and optimized, thus effectively improving the accuracy and reliability of mesh optimization results. It is applicable to various vehicle body design scenarios and can flexibly optimize various vehicle body mesh parameters.

[0011] Furthermore, this invention combines historical vehicle data with domain knowledge to generate optimization suggestions and dynamically adjusts the vehicle mesh parameters, forming a closed loop of "simulation-analysis-optimization," thereby significantly improving the efficiency and accuracy of vehicle design and optimization and shortening the automotive R&D cycle.

[0012] As a preferred technical measure: The method for collecting the vehicle body mesh data to be optimized is as follows: Step 11: Obtain the 3D model of the car to be optimized; Step 12: Perform preliminary geometric analysis on the 3D model of the car body using the vehicle body mesh generation module to identify different regions and features of the vehicle body and obtain information on key parts of the vehicle body; Step 13: Based on the overall mesh size, mesh the 3D model of the car to generate the initial body mesh model; Step 14: Calculate the area of ​​each key part based on the information of the key parts of the vehicle body; Step 15: Calculate the number of grids for the critical part based on the area and grid density; Step 16: Based on the number of grids, perform grid refinement on the initial body grid model related to the key parts to generate body grid data to be optimized, which includes the overall grid size, the grid density of each key part, and the specific location and shape information of each grid.

[0013] As a preferred technical measure: The method for obtaining several vehicle history data points from a pre-built vehicle history database module is as follows: Obtain historical vehicle body design information accumulated during previous vehicle body design and simulation processes that can be used for mesh parameter optimization; The vehicle body historical design information is text-based data, which includes historical mesh adjustment parameters and their corresponding performance indicators, as well as simulation context information; Historical mesh adjustment parameters include mesh size, density, and shape; Performance metrics include vehicle body stress, strain, and stiffness; Simulation context information includes vehicle model, material type, and simulation boundary conditions; Based on the simulation context information, the historical grid adjustment parameters and their corresponding performance indicators are broken down to obtain several text-type historical vehicle data.

[0014] As a preferred technical measure: The method for matching the vehicle body mesh data to be optimized with historical vehicle body data based on the pre-built vehicle body mesh optimization module to obtain the most similar historical vehicle body data is as follows: Step 21: Convert several text-type vehicle history data into multiple word vectors using a word vector algorithm; Step 22: Average pooling is performed on multiple word vectors to obtain a fixed-dimensional vehicle body history vector; Step 23 involves calculating and processing the vehicle body historical vector and the vehicle body mesh data to be optimized to construct a unified vehicle body feature vector, which includes the following steps: Step 231: Calculate the mean and standard deviation of the vehicle body history vector and the vehicle body grid data to be optimized, respectively; Step 232: Based on the mean and standard deviation, correct or remove outliers in the vehicle history vector and the vehicle grid data to be optimized; then calculate the minimum and maximum values ​​in the vehicle history vector and the vehicle grid data to be optimized, respectively. Step 233: Based on the minimum and maximum values, normalize the vehicle body history vector and the vehicle body grid data to be optimized respectively, and map the data to the [0,1] interval to obtain the vehicle body history standard vector and the vehicle body standard vector to be optimized. Step 234: Combine the historical standard vector of the vehicle body with the standard vector of the vehicle body to be optimized to obtain a unified vehicle body feature vector; Step 24: Calculate the similarity of the vehicle body feature vectors to obtain the cosine similarity between the grid data of the vehicle body to be optimized and each historical data of the vehicle body, and select the historical data of the vehicle body with the highest cosine similarity as the optimization reference.

[0015] As a preferred technical measure: Based on domain knowledge from a pre-built domain knowledge base module, and reusing experiential information and verification criteria, the following method is used to analyze and process the vehicle body mesh data to be optimized, and generate optimization suggestions: Based on the historical grid adjustment parameters and their corresponding performance indicators in the vehicle's historical data, empirical information and verification criteria are established. Several types of textual experience information and verification criteria are transformed into multiple semantic vectors using word vector algorithms; Multiple semantic vectors are averaged and pooled to obtain a fixed-dimensional vehicle body pattern vector; Acquire domain knowledge from the domain knowledge base module, including knowledge of vehicle body structure mechanics, mesh generation principles, and material properties; knowledge of vehicle body structure mechanics is used to understand the stress characteristics and deformation laws of different parts of the vehicle body under various working conditions; mesh generation principles are used to generate meshes according to the geometry and performance requirements of the vehicle body; and knowledge of material properties is used to obtain the mechanical properties of different vehicle body materials. Based on knowledge of vehicle body structural mechanics, mesh generation principles, and material properties, and combined with vehicle body regularity vectors, the mesh data of the vehicle body to be optimized is analyzed and processed to generate targeted optimization suggestions, thereby achieving closed-loop iterative optimization based on experience and real-time simulation.

[0016] As a preferred technical measure: Based on knowledge of vehicle body structural mechanics, mesh generation principles, and material properties, and combined with vehicle body regularity vectors, the following method is used to analyze and process the vehicle body mesh data to be optimized, and generate targeted optimization suggestions: The grid data of the vehicle body to be optimized is decomposed to obtain several index values; Based on the numerical values ​​of the indicators and combined with the body shape vectors, multiple vehicle performance indicators are constructed, which include at least the aerodynamic drag coefficient and body stiffness. Calculate the numerical values ​​of the indicators and the average values ​​of the vehicle performance indicators, respectively. Based on the mean values ​​of the indicators, correlation analysis is used to calculate the correlation coefficient between the indicator values ​​and the vehicle performance indicators. Based on the correlation coefficient, the correspondence between a certain indicator value of the vehicle body and a certain vehicle performance indicator is determined, and information on the direction of indicator optimization is obtained. Based on the information on the direction of indicator optimization, determine the indicator values ​​and the corresponding performance indicators; Based on knowledge of vehicle body structural mechanics, mesh generation principles, and material properties, the numerical values ​​of the indicators and their corresponding performance indicators are evaluated to determine the quality of the mesh data to be optimized for the vehicle body. Then, a causal inference method is used to generate targeted optimization suggestions.

[0017] As a preferred technical measure: The method for evaluating the index values ​​and corresponding performance indicators to determine the quality of the vehicle body mesh data to be optimized, and generating targeted optimization suggestions using causal inference, is as follows: Based on knowledge of vehicle body structure mechanics, mesh generation principles, and material properties, each performance indicator is compared with the preset standard value to obtain the performance status information of the current vehicle body design. Based on the importance of different performance indicators, each performance indicator is assigned a corresponding weight, and then a weighted average is calculated as the comprehensive performance score to judge the merits of the current vehicle body design. A density-based clustering algorithm is used to cluster the vehicle body mesh data to be optimized, dividing the mesh data into different clusters. If a data point in the vehicle body mesh data does not belong to any cluster, it is considered an outlier. By analyzing the mesh parameters and performance indicators corresponding to these outliers, the problematic mesh regions are located. For the identified abnormal grid regions, a causal inference method is used to determine whether the change of a certain grid parameter is the cause of the abnormal performance index, and to determine the causal relationship between the change of grid parameter and the performance index. Based on performance status information, comprehensive performance score, and causal relationship information, targeted optimization suggestions are generated.

[0018] As a preferred technical measure: Based on the optimization suggestions, the method for dynamically adjusting and optimizing the vehicle body mesh data to obtain the optimized vehicle body mesh data is as follows: Based on the optimization suggestions, the grid parameters of the corresponding area of ​​the vehicle body grid data to be optimized are adjusted, including at least the adjustment of grid size, grid density and grid shape; After adjustment, the vehicle body mesh data is obtained; according to the working condition requirements, the boundary conditions for simulation are set. Based on the vehicle body mesh data, material properties, and boundary conditions, new performance indicators of the vehicle body are obtained. By associating vehicle body mesh data with performance indicators, optimized vehicle body mesh data is generated.

[0019] To achieve one of the above objectives, the second technical solution of the present invention is as follows: A method for dynamic optimization of vehicle body mesh parameters based on model simulation includes the following steps: Step 1: Collect the grid data of the vehicle body to be optimized in real time using a pre-built data acquisition model; Step 2: Using a pre-built data analysis model, relevant vehicle history data is obtained from the vehicle history database module. Combined with domain knowledge from the domain knowledge base module, the vehicle grid data to be optimized is analyzed and processed to obtain empirical information and verification basis. Step 3: Using a pre-built optimization suggestion generation model, based on empirical information and verification criteria, optimization suggestion information is generated for the current vehicle body mesh data to be optimized; Step 4: Using the pre-built parameter adjustment model, dynamically adjust the mesh data of the vehicle body to be optimized according to the optimization suggestion information, and then conduct a new round of simulation experiments to obtain simulation results; Step 5: Based on the pre-built closed-loop iterative optimization model and the simulation results, iteratively optimize the vehicle body mesh parameters until the simulation results meet the design requirements or reach the preset optimization target, thereby realizing dynamic optimization of vehicle body mesh parameters based on model simulation.

[0020] This invention effectively addresses the problem of insufficient utilization of real-time and historical vehicle data in traditional methods, accurately uncovering potential patterns and trends within the data. By combining historical vehicle data with domain knowledge, it generates specific and precise optimization suggestions for minor adjustments to vehicle mesh parameters, such as fine-tuning the mesh size in specific areas, precisely modifying the mesh shape, and subtly changing the mesh density. This dynamically adjusts the vehicle mesh parameters, forming a closed loop of "simulation-analysis-optimization." This significantly improves the efficiency and accuracy of vehicle design and optimization in handling details, shortening the time required for mesh parameter tuning during the automotive R&D cycle.

[0021] To achieve one of the above objectives, the third technical solution of the present invention is as follows: A model simulation-based dynamic optimization system for vehicle body mesh parameters, comprising: One or more processing units; Storage device for storing one or more programs; When the one or more programs are executed by the one or more processing units, the one or more processing units implement the above-described method for dynamic optimization of vehicle body mesh parameters based on model simulation.

[0022] Compared with existing technical solutions, the present invention has the following beneficial effects: This invention utilizes a vehicle body history database module, a vehicle body mesh optimization module, and a domain knowledge base module to match the vehicle body mesh data to be optimized with historical vehicle body data, obtaining the most similar historical vehicle body data as empirical information and verification basis for vehicle body mesh parameter optimization. Then, the empirical information and verification basis are reused to analyze and process the vehicle body mesh data to be optimized, generating optimization suggestion information. This allows for in-depth mining and analysis of historical data, comprehensively considering various complex factors and conditions. Furthermore, based on the optimization suggestion information, the vehicle body mesh data to be optimized is precisely adjusted and optimized, thus effectively improving the accuracy and reliability of mesh optimization results. It is applicable to various vehicle body design scenarios and can flexibly optimize various vehicle body mesh parameters.

[0023] This invention effectively addresses the problem of insufficient utilization of real-time and historical vehicle data in traditional methods, accurately uncovering potential patterns and trends within the data. By combining historical vehicle data with domain knowledge, it generates specific and precise optimization suggestions for minor adjustments to vehicle mesh parameters, such as fine-tuning the mesh size in specific areas, precisely modifying the mesh shape, and subtly changing the mesh density. This dynamically adjusts the vehicle mesh parameters, forming a closed loop of "simulation-analysis-optimization." This significantly improves the efficiency and accuracy of vehicle design and optimization in handling details, shortening the time required for mesh parameter tuning during the automotive R&D cycle. Attached Figure Description

[0024] Figure 1 This is a flowchart illustrating the dynamic optimization method for vehicle body mesh parameters according to the present invention. Figure 2 This is a structural block diagram of the vehicle body mesh parameter dynamic optimization system of the present invention; Figure 3 This is a schematic diagram illustrating a process for dynamically optimizing vehicle body mesh parameters using this invention. Detailed Implementation

[0025] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present application, and not all of them. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this application. This invention covers any substitutions, modifications, equivalent methods, and solutions made within the spirit and scope of the invention as defined by the claims.

[0026] like Figure 1 As shown, this is the first specific embodiment of the dynamic optimization method for vehicle body mesh parameters based on model simulation of the present invention: A method for dynamic optimization of vehicle body mesh parameters based on model simulation includes the following: Collect the grid data of the vehicle body to be optimized, and obtain several historical data of the vehicle body from the pre-built historical database module of the vehicle body; Based on the pre-built vehicle mesh optimization module, the vehicle mesh data to be optimized is matched with the vehicle historical data to obtain the most similar vehicle historical data, which serves as empirical information and verification basis for vehicle mesh parameter optimization. Based on the domain knowledge in the pre-built domain knowledge base module, and by reusing experience information and verification basis, the vehicle body mesh data to be optimized is analyzed and processed to generate optimization suggestion information; Based on the optimization suggestions, the mesh data of the vehicle body to be optimized is dynamically adjusted and optimized to obtain the optimized mesh data of the vehicle body.

[0027] A second specific embodiment of the dynamic optimization method for vehicle body mesh parameters based on model simulation of the present invention: A method for dynamic optimization of vehicle body mesh parameters based on model simulation includes the following steps: Step 1: Collect the grid data of the vehicle body to be optimized in real time using a pre-built data acquisition model; Step 2: Using a pre-built data analysis model, relevant vehicle history data is obtained from the vehicle history database module, and combined with domain knowledge from the domain knowledge base module, the vehicle grid data to be optimized is analyzed and processed to obtain the analysis results. Step 3: Using a pre-built optimization suggestion generation model, optimization suggestion information is generated for the current vehicle body mesh parameters based on the analysis results. This includes adjusting the mesh size, modifying the mesh shape, and changing the mesh density. Step 4: Using the pre-built parameter adjustment model, dynamically adjust the vehicle body mesh parameters according to the optimization suggestions, and then conduct a new round of simulation experiments to obtain simulation results; Step 5: Based on the pre-built closed-loop iterative optimization model and the simulation results, iteratively optimize the vehicle body mesh parameters until the simulation results meet the design requirements or reach the preset optimization target, thus completing the dynamic optimization of the vehicle body mesh parameters based on model simulation.

[0028] like Figure 2 As shown, a specific embodiment of the vehicle body mesh parameter dynamic optimization system based on model simulation of the present invention is as follows: A model simulation-based dynamic optimization system for vehicle body mesh parameters includes a vehicle body mesh generation module, a data transmission and feedback module, a vehicle body mesh optimization module, a vehicle body historical database module, and a domain knowledge base module.

[0029] Vehicle Body Mesh Generation Module: Responsible for generating vehicle body meshes and conducting automotive performance simulation experiments, producing real-time vehicle body mesh data for optimization. This module can be integrated with professional automotive engineering simulation software ANSA for precise processing of vehicle body meshes.

[0030] Data transmission and feedback module: This module establishes a real-time communication channel between the vehicle body mesh generation module and the vehicle body mesh optimization module, enabling the transmission of vehicle body mesh data to be optimized and the feedback of optimization suggestions. This module includes a client and a server. The client is deployed within the vehicle body mesh generation module and is responsible for sending the vehicle body mesh data to be optimized to the server. The server, deployed in the cloud or locally, receives the data sent by the client and forwards it to the vehicle body mesh optimization module, while simultaneously feeding back the optimization suggestions generated by the vehicle body mesh optimization module to the client.

[0031] The vehicle body mesh optimization module receives the vehicle body mesh data to be optimized from the data transmission and feedback module, analyzes it in conjunction with historical vehicle body data and domain knowledge, and generates optimization suggestions. It can compare and correlate the received vehicle body mesh data to be optimized with historical vehicle body data to uncover potential patterns and trends. Simultaneously, the vehicle body mesh optimization module can be integrated with the Qianwen Big Model module for efficient understanding and analysis of complex automotive engineering data.

[0032] Vehicle Body History Database Module: Stores historical vehicle body data from the design and simulation process, including body mesh parameters, optimization results, and optimization records. The vehicle body mesh optimization module can retrieve relevant data from the vehicle body history database for analysis and to generate optimization suggestions.

[0033] Domain Knowledge Base Module: Stores professional knowledge and experience in the automotive engineering field, such as vehicle body mesh optimization specifications and optimization methods for different materials. The vehicle body mesh optimization module can utilize the knowledge in the domain knowledge base to perform more accurate analysis and judgment on the vehicle body mesh data to be optimized.

[0034] The third specific embodiment of the dynamic optimization method for vehicle body mesh parameters based on model simulation of the present invention: A model simulation-based dynamic optimization method for vehicle body mesh parameters utilizes a vehicle body mesh generation module to collect real-time vehicle body mesh data to be optimized. This data is then transmitted to a vehicle body mesh optimization module via a data transmission and feedback module. The optimization module matches the mesh data with historical vehicle body data, vectorizing mesh parameters and performance indicators, calculating the cosine similarity between real-time and historical data, selecting the most similar historical data as an optimization reference, and combining this with engineering constraints from a domain knowledge base to generate targeted optimization suggestions. This achieves closed-loop iterative optimization based on experience and real-time simulation. The method includes the following steps: Step 1. Collect and initially organize the vehicle body simulation data, which includes the following: The vehicle body mesh generation module collects vehicle body mesh parameters (such as size, shape, and density) and vehicle performance simulation indicators (such as stress and strain) in real time, as well as other vehicle body mesh data to be optimized. The 3σ criterion is used to remove outliers from the data, ensuring data quality and yielding the vehicle body specification data.

[0035] Step 2. Use a linear normalization algorithm to encode and structure the vehicle body specification data, mapping it to the [0,1] interval to obtain the vehicle body coded data. The calculation formula is as follows: in It is the raw data. and These are the minimum and maximum values ​​of the vehicle's historical data for this parameter.

[0036] Vehicle history data is text-based and needs to be converted to numerical values, which includes the following processes: First, the text is converted into word vectors using the Word2Vec algorithm. Then, average pooling is used to obtain a fixed-dimensional overall vector representation. This overall vector is used to encode text information into quantifiable features, which are combined with numerical grid parameter vectors to form a unified feature vector. The vehicle body grid optimization module uses this feature vector to calculate the similarity between the vector and the vector of the vehicle body grid data to be optimized, such as cosine similarity, to determine the degree of similarity between real-time data and historical records.

[0037] Step 3. Merge the processed real-time vehicle body coding data with the vehicle body historical data to form a new vehicle body data matrix. Then, encapsulate the data in JSON object representation for easy transmission.

[0038] In this embodiment, the vehicle body historical data comes from data accumulated in previous automotive body design and simulation processes that can be used for mesh parameter optimization. This includes historical mesh adjustment parameters and their corresponding performance indicators, such as mesh size, density, shape, and performance results such as vehicle body stress, strain, and stiffness. It also incorporates contextual information such as vehicle model, material type, and simulation boundary conditions.

[0039] Data fusion combines real-time vehicle body coding data with historical vehicle body data according to a unified feature dimension to form a new vehicle body data matrix, rather than simply splicing or directly superimposing them. During the fusion process, numerical data is first normalized, and textual data is vectorized and then combined with the numerical data to form a unified feature vector matrix.

[0040] For data from multiple sources that may exist within the same dimension, weighted processing or standardization methods can be used to unify them, ensuring comparability and consistency of data from different sources within the matrix. Forming a new data matrix offers several technical advantages: First, it enables comparison of real-time data with historical vehicle data in a unified feature space, providing a foundation for subsequent similarity calculations and pattern recognition. Second, through matrix representation, the vehicle mesh optimization module can simultaneously process multiple historical and real-time data, achieving efficient parallel analysis and logical reasoning. Third, this matrix structure facilitates secure transmission by encapsulating it into the object representation JSON format, enabling real-time communication between the simulation end and the large model module.

[0041] Step 4. The server of the data transmission and feedback module sends the encapsulated object representation JSON data to the vehicle body mesh optimization module via the HTTP protocol. The AES encryption algorithm is used to ensure data security during transmission, and the encryption formula is C=EK(P).

[0042] Step 5. After receiving the encrypted data, the vehicle body mesh optimization module decrypts it and provides optimization suggestions based on a discrimination mechanism. It then judges the rationality of the suggestions according to the rules and standards in the domain knowledge base. For example, if the vehicle body mesh size has minimum and maximum limits (dmin and dmax), and the suggested size (dsuggest) exceeds these limits, it is deemed unreasonable.

[0043] The similarity between the optimization suggestion information and the vehicle's historical data is calculated using the cosine similarity formula.

[0044] The suggested recommendations are compared with historical vehicle data. If the similarity is greater than a threshold θ, the recommendations are considered reliable, and optimization suggestions are obtained. The threshold θ ranges from 0.7 to 0.95.

[0045] The optimization suggestions are applied to the vehicle body mesh generation module for simulation. Finite element analysis algorithms (such as [K]{δ}={F} to solve the vehicle body stress) are used to compare the simulation results with the design requirements to determine the effectiveness of the suggestions.

[0046] The fourth specific embodiment of the dynamic optimization method for vehicle body mesh parameters based on model simulation of the present invention: A method for dynamic optimization of vehicle body mesh parameters based on model simulation includes the following steps: Step 1: During the process of vehicle body mesh generation and simulation experiment, the vehicle body mesh generation module collects the mesh data to be optimized in real time and sends the data to the server through the client of the data transmission and feedback module.

[0047] Step two: The server forwards the received vehicle body mesh data to be optimized to the vehicle body mesh optimization module. The vehicle body mesh optimization module retrieves relevant historical vehicle body data from the vehicle body history database module and, in conjunction with domain knowledge from the domain knowledge base module, analyzes and processes the vehicle body mesh data to be optimized.

[0048] Step three: Based on the analysis results, the vehicle body mesh optimization module generates optimization suggestions for the current vehicle body mesh parameters. Optimization suggestions may include adjusting mesh size, modifying mesh shape, and changing mesh density.

[0049] Step three: The server feeds back the optimization suggestions generated by the vehicle body mesh optimization module to the client of the data transmission and feedback module. The client then passes the optimization suggestions to the vehicle body mesh generation module. The vehicle body mesh generation module dynamically adjusts the vehicle body mesh parameters based on the optimization suggestions, and then conducts a new round of simulation experiments.

[0050] Step 4: Repeat the above steps to continuously perform the "simulation-analysis-optimization" cycle until the simulation results meet the design requirements or achieve the preset optimization goal.

[0051] In this embodiment, the method for analyzing and processing the vehicle body grid data to be optimized is as follows: The analysis phase plays a crucial role in the entire "simulation-analysis-optimization" closed loop, connecting the preceding and following steps. Based on the real-time data acquired in the simulation phase, it deeply mines the value of the data, accurately identifies problems, and provides a scientific basis for the subsequent optimization phase. It includes the following steps: Step 11. After the server forwards the received vehicle body mesh data to be optimized to the vehicle body mesh optimization module, the module first cleans the data. Due to various factors that may interfere with the simulation process, the data may contain outliers, missing values, etc. Outliers are identified and processed using the statistical 3σ criterion. Let a certain data sequence be... Calculate its mean The formula is as follows: Calculate the standard deviation The formula is as follows: If a certain data point satisfy If a value is missing, it is considered an outlier and corrected or removed. For missing values, methods such as mean imputation, median imputation, or regression prediction are used to fill in the missing values, depending on the characteristics and distribution of the data.

[0052] To eliminate dimensional differences between different data indicators and improve the accuracy of the analysis, the cleaned data is normalized. For numerical data, a linear normalization algorithm is used to map the data to the [0,1] interval, obtaining the normalized corresponding numerical values. The calculation formula is as follows: in The original data, and These are the minimum and maximum values ​​of this data indicator in the vehicle's historical data, respectively.

[0053] Step 12. The vehicle body mesh optimization module compares the preprocessed real-time data with the data in the vehicle body historical database. It analyzes the similarity between the current simulation results and historical data by calculating similarity indices such as cosine similarity and Euclidean distance. Taking cosine similarity as an example, let the real-time data vector be... The vehicle body historical data vector is Then their cosine similarity The calculation formula is as follows: The higher the similarity, the more similar the current situation is to the historical situation, and historical optimization experience can be referenced for processing; if the similarity is low, further in-depth analysis is required.

[0054] While similarity can theoretically be calculated directly based on standard values, this method often only reflects differences under ideal conditions and cannot cover the multi-source disturbances and nonlinear effects present in complex simulations and real engineering environments. Therefore, it is insufficient to support the reliability of optimization recommendations. In contrast, introducing historical optimization experience has significant advantages: First, historical vehicle body data contains a large number of implicit patterns accumulated during real simulations and experiments, which can compensate for the singularity and limitations of standard values; second, by comparing real-time data with historical optimization results, the effective parameter range can be identified more quickly, shortening the optimization iteration time and improving the engineering applicability and convergence efficiency of vehicle body mesh parameter adjustments.

[0055] In this embodiment, the similarity threshold ranges from 0.7 to 0.95.

[0056] The real-time data vector and vehicle history data vector involved in the cosine similarity calculation are both obtained by normalizing and vectorizing the original data.

[0057] The similarity calculation between real-time data and the vehicle's historical database measures the overall matching degree between the current state and historical experience, which can be used to guide subsequent optimization directions, problem localization, and causal analysis. The similarity calculation between optimization suggestions and vehicle historical data measures the feasibility and credibility of the suggested solutions and historical experience, which is used to filter or confirm optimization measures.

[0058] Multi-dimensional data correlation analysis: This involves performing correlation analysis on multiple data indicators, such as studying the relationship between vehicle body mesh parameters and vehicle performance indicators. Correlation analysis methods (such as the Pearson correlation coefficient) are used. Let two data sequences be X and Y, and their Pearson correlation coefficient... The calculation formula is as follows: Where: n represents the number of samples, that is, the total number of data points participating in the correlation analysis; This represents the value of the independent variable X in the i-th sample, such as the value of a certain index in the vehicle body mesh division parameters; This represents the value of the dependent variable Y in the i-th sample, such as the corresponding automotive performance index (e.g., aerodynamic drag coefficient, body stiffness, etc.). This represents the mean of variable X; represents the mean of variable Y; r represents the Pearson correlation coefficient between variables X and Y, and its value range is []. [1,1], where r>0 indicates a positive correlation, r<0 indicates a negative correlation, and |r| is closer to 1, indicating a stronger correlation. The degree and direction of the correlation are determined by analyzing the correlation coefficient.

[0059] Step 13. Combining the professional knowledge and design requirements in the domain knowledge base, compare the various performance indicators obtained from the simulation with the preset standard values. For example, for the strength index of the car body, if the stress value obtained from the simulation exceeds the allowable stress of the material, then the index is considered to fail to meet the requirements. By evaluating multiple performance indicators one by one, a comprehensive understanding of the current performance status of the car body design can be obtained.

[0060] To more intuitively evaluate the overall performance of the vehicle body, a comprehensive performance scoring method is adopted. Based on the importance of different performance indicators, each indicator is assigned a corresponding weight, and then a weighted average is calculated as the comprehensive performance score. Let the performance indicators be... The corresponding weight is The formula for calculating the overall performance score is as follows: By using comprehensive performance scores, the merits and demerits of the current vehicle body design can be quickly determined, providing a reference for subsequent optimization decisions.

[0061] Step 14. Based on data comparison and performance evaluation, the vehicle body mesh optimization module uses machine learning algorithms (such as cluster analysis and anomaly detection algorithms) to detect abnormal features in the data. For example, density-based clustering (DBSCAN) is used to cluster the data, dividing the data points into different clusters. If a data point does not belong to any cluster, it is considered an anomaly. By analyzing the mesh parameters and performance indicators corresponding to these anomalies, potentially problematic areas are located.

[0062] For identified anomalies, further causal relationship analysis is conducted. Causal inference methods (such as structural equation modeling and Granger causality tests) are used to analyze the causal relationships between different factors. For example, Granger causality tests are used to determine whether a change in a certain grid parameter is the cause of the performance anomaly. Through causal relationship analysis, the root cause of the problem is clarified, providing targeted optimization suggestions for subsequent operations.

[0063] This invention combines historical vehicle body data with domain knowledge to generate optimization suggestions and dynamically adjusts vehicle body mesh parameters, significantly shortening the traditional tuning cycle of several weeks to hours, greatly improving the efficiency of vehicle body design and optimization. Simultaneously, it comprehensively and accurately analyzes the vehicle body mesh data to be optimized, generating more reasonable optimization suggestions and significantly improving the accuracy and reliability of simulation results. Furthermore, it fully mines and utilizes real-time and historical vehicle body data during the design and simulation process, achieving in-depth data analysis and value mining, providing strong support for automotive engineering decisions. Moreover, by using unified communication standards and interfaces, it achieves seamless integration of the vehicle body mesh optimization module and the vehicle body mesh generation module, reducing the complexity and cost of system integration.

[0064] like Figure 3 As shown, a specific embodiment of applying the method of the present invention to perform mesh optimization on the sheet metal parts of a certain car is as follows: In a real-world automotive manufacturing scenario, an automaker plans to optimize the mesh of sheet metal parts for a new sedan, such as adjusting mesh size and mesh density in key areas, to improve overall vehicle performance and design efficiency. This invention is then applied to optimize the mesh of the sedan's sheet metal parts, including the following: Step 1, firstly, set a series of initial vehicle body mesh parameters, including the following: Input a 3D CAD model of the car, which includes geometric information of various body parts, such as length, width, height, and curvature; as well as some basic parameters determined based on experience and preliminary design requirements, such as the initial setting of the overall mesh size. And the expected mesh density of key parts (such as car doors, hoods, etc.) .

[0065] Then, based on the input 3D CAD model, a preliminary geometric analysis is performed on the vehicle body to identify different regions and features. This is done according to the set overall mesh size. The entire vehicle body is divided into uniform grids. For critical areas, the grid density is adjusted according to the desired level. A localized mesh refinement method is used for mesh generation. Specifically, for each small region in the critical area, its area S is calculated, and then calculated using the formula n = ... ×S determines the number of grids n that should be divided in the region, and then further refines the grid division of the region based on this number.

[0066] This process generates an initial vehicle body mesh model and records the current mesh parameters, including the overall mesh size. The information includes the mesh density of each key part, as well as the specific location and shape of each mesh.

[0067] Step 2. The client of the data transmission and feedback module (communication module) sends the current vehicle body mesh parameters to the server, and the server forwards the data to the vehicle body mesh optimization module. The vehicle body mesh optimization module needs to obtain relevant data from the vehicle body history database of past vehicle body designs and simulations, and analyze the current mesh data in conjunction with corresponding domain knowledge.

[0068] In this embodiment, the relevant data from previous vehicle body design and simulation include vehicle body mesh parameter data, simulation result data, and optimization record data.

[0069] The body mesh parameter data includes the overall mesh size of different models in previous designs. Grid density in key areas The proportions of grid shapes (such as triangles, quadrilaterals, etc.) .

[0070] Simulation results data include the stress distribution of the vehicle body. Strain distribution Deformation amount Performance indicators, and the specific values ​​of these indicators under different operating conditions.

[0071] The optimization log records measures taken in the past during mesh optimization, such as the magnitude of mesh size adjustments. The methods of changing the mesh shape, and the impact of these measures on the simulation results.

[0072] Domain knowledge includes knowledge of vehicle body structure mechanics, mesh generation principles, and material properties.

[0073] The knowledge of vehicle body structure mechanics is used to understand the stress characteristics and deformation patterns of different parts of the vehicle body under various working conditions. For example, the front of the vehicle body mainly bears the impact force during a collision, requiring a denser and stronger mesh to accurately simulate the stress distribution; while the sides of the vehicle body mainly bear the lateral force during driving, requiring a certain degree of flexibility and stability of the mesh.

[0074] Mesh generation principles are used to determine how to generate reasonable meshes based on the geometry and performance requirements of the vehicle body. For example, smaller mesh sizes should be used for areas with greater curvature to improve simulation accuracy, while larger mesh sizes can be used for areas with uniform stress to reduce computational load.

[0075] Material properties knowledge is used to familiarize oneself with the mechanical properties of different car body materials, such as the elastic modulus E and Poisson's ratio ν of steel and aluminum alloys. These parameters affect the accuracy of mesh generation and simulation results.

[0076] In this embodiment, the process of analyzing the current grid data is as follows: Step 21: Clean the current vehicle body mesh parameter data and relevant data obtained from the vehicle body historical database to remove outliers. Outliers are judged using the 3σ criterion, the calculation formula of which is as follows: in, Given a certain data sequence, Let σ be the mean and σ be the standard deviation.

[0077] If a certain data point satisfy If the value is greater than 3σ, it is considered an outlier and corrected or removed.

[0078] Numerical data is normalized using a linear normalization algorithm, mapping the data to the [0,1] interval. The formula is: = in The original data, and These are the minimum and maximum values ​​of this data indicator in the vehicle's historical data, respectively.

[0079] Step 22: Using the cosine similarity formula, perform data comparison analysis to calculate the similarity between the current grid parameters and the vehicle's historical data. Let the current grid parameter vector be... The vehicle body historical data vector is Then, the formula for calculating their cosine similarity is as follows: Correlation analysis is performed on multiple data indicators, such as studying the relationship between vehicle body mesh parameters and vehicle performance indicators. The Pearson correlation coefficient is used; given two data sequences X and Y, their Pearson correlation coefficient is calculated. The calculation formula is as follows: The correlation coefficient is analyzed to determine the degree and direction of the correlation between indicators.

[0080] Step 23: Combining the professional knowledge and design requirements in the domain knowledge base, compare the various performance indicators obtained from the current simulation with the preset standard values ​​to complete the performance evaluation of the vehicle body.

[0081] For example, regarding the strength indicators of a car body, if the stress value obtained from simulation... If the allowable stress [σ] of the material is exceeded, then the index is considered to be unacceptable.

[0082] A comprehensive performance scoring method is used to evaluate the overall performance of the vehicle body. Let the performance index be... The corresponding weight is The formula for calculating the overall performance score is as follows: Step 3. The vehicle body mesh optimization module, based on the analysis results, found that the mesh division in some areas was unreasonable, resulting in low simulation accuracy. It is necessary to generate optimization suggestions, such as adjusting the mesh size of the part, changing the mesh shape, increasing or decreasing the mesh density, etc.

[0083] The criteria for identifying unreasonable grid division include grid size, grid density, and grid shape.

[0084] Regarding mesh size: If the mesh size d of a certain region is too large, it will lead to inaccurate calculation of the stress and strain distribution in that region, i.e., the stress variation gradient in that region. Exceeding the preset threshold If the grid size of the area is deemed unreasonable, then the grid size of that area is considered unreasonable.

[0085] Regarding mesh density: If the mesh density ρ of a critical part is too low, the deformation and stress concentration of that part cannot be accurately simulated. For example, if the maximum deformation of that part deviates from the theoretical value by more than 10%, then the mesh density of that area is considered unreasonable.

[0086] Regarding grid shape: If the grid shape in a certain area is too irregular, such as a triangular grid with interior angles that are too small (less than 30 degrees), then... If the aspect ratio of a quadrilateral mesh is too large (greater than 3), which affects the stability and accuracy of the calculation, then the mesh shape of that region is considered unreasonable.

[0087] Step 4. Optimize and validate the initial suggestions using historical vehicle data and machine learning algorithms to generate optimized recommendations. Specifically, a regression model is built using historical vehicle data. The inputs are historical grid adjustment parameters and corresponding performance improvement indicators, and the output is the predicted performance improvement. The adjustment parameters from the initial suggestions are input into this regression model to predict the performance improvement. If the predicted performance improvement does not meet the requirements, the suggestions are further adjusted until a satisfactory result is obtained.

[0088] Specific optimization suggestions are generated, including the areas that need adjustment, the adjustment methods (such as adjusting mesh size, changing mesh shape, increasing or decreasing mesh density, etc.), and specific adjustment parameter values. For example, it is suggested to adjust the mesh size of a certain area from... Adjusted to Change the grid density of another region from Increase to And some irregular grid shapes were modified into more regular shapes.

[0089] Step 5. The server feeds back optimization suggestions to the client, which then passes these suggestions to the vehicle body mesh generation module. Based on the optimization suggestions, the vehicle body mesh parameters are dynamically adjusted, and a new round of simulation experiments is conducted.

[0090] In this embodiment, the refinement process of the simulation experiment includes the following: The vehicle body mesh generation module precisely adjusts the mesh parameters of the corresponding areas based on optimization suggestions. For mesh size adjustment, the side length of each mesh in the area is directly modified; for mesh density adjustment, the number of meshes is increased or decreased; for mesh shape adjustment, a mesh re-division algorithm is used to transform irregular meshes into more regular shapes.

[0091] Set the boundary conditions for the simulation based on the actual working conditions. For example, in collision simulation, set parameters such as collision speed, angle, and mass of the colliding object; in driving vibration simulation, set parameters such as road surface unevenness and vehicle speed.

[0092] A finite element method (FEM) solver is selected to calculate the adjusted mesh model. During the calculation, the solver calculates the stress, strain, deformation, and other performance indicators of the vehicle body based on the input mesh parameters, material properties, and boundary conditions. Specifically, the vehicle body is discretized into multiple elements using the finite element method. By establishing the stiffness matrix of each element and the equilibrium equations of the nodes, the response of the entire system is solved. For example, for a linear elastic problem, the equation [K]{δ}={F} is used to solve it, where [K] is the stiffness matrix, {δ} is the nodal displacement vector, and {F} is the nodal force vector.

[0093] After the simulation is completed, new simulation results are output, including performance indicators such as stress distribution, strain distribution, and deformation of the vehicle body. Simultaneously, information such as the mesh parameters, boundary conditions, and solver settings used in this simulation are recorded for subsequent analysis and comparison.

[0094] Step 5. Repeat the above process, continuously performing the "simulation-analysis-optimization" cycle until the performance indicators of the vehicle body mesh under various working conditions meet the design requirements, thus completing the closed-loop iterative optimization.

[0095] In this way, the process of optimizing vehicle body mesh parameters, which originally required several weeks, can be shortened to hours under the method of this invention, greatly improving the efficiency and quality of vehicle body mesh optimization. Throughout the optimization process, simulation results are continuously evaluated and analyzed to ensure that the final vehicle body design meets all design requirements, providing reliable technical support for automobile production and manufacturing. In this way, the process of optimizing vehicle body mesh parameters, which originally required several weeks, can be shortened to hours under the method of this invention, greatly improving the efficiency and quality of vehicle body mesh optimization.

[0096] A server embodiment applying the method of the present invention: A server comprising: One or more processing units; Storage device for storing one or more programs; When the one or more programs are executed by the one or more processing units, the one or more processing units implement the above-described method for dynamic optimization of vehicle body mesh parameters based on model simulation.

[0097] The storage device can be internal memory, external memory, cache memory, or other special-purpose memory. The processing unit may be an integrated circuit chip with signal processing capabilities; it can be a general-purpose processor, a digital signal processor, an application-specific integrated circuit, an off-the-shelf programmable gate array, or other programmable logic device.

[0098] An embodiment of a device applying the method of the present invention: An electronic device is provided with a computer-readable storage medium on which a computer program is stored. When the program is executed by a processing unit, it implements the above-described method for dynamic optimization of vehicle body mesh parameters based on model simulation.

[0099] Computer-readable storage media refers to physical carriers capable of storing computer-recognizable data, instructions, or programs. These media must meet the core characteristic of being "readable by a computer" (i.e., the data exists in the form of electrical, magnetic, or optical signals and can be converted into binary information that a computer can process through appropriate devices). The physical carrier can be a magnetic storage medium, optical storage medium, semiconductor storage medium, or other storage media.

[0100] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, optical storage, etc.) containing computer-usable program code.

[0101] The module in this application is an object that uses physical or virtual representation to form an objective description of form and structure. The object is not the same as a physical object, and is not limited to physical or virtual. It can be a data processing function, software program, processing mode, usage method, operation mode, workflow, application process, electronic hardware, circuit module, processing system, system imitation or simulation object.

[0102] Finally, it should be noted that the above-described embodiments are merely specific implementations of the present invention, used to illustrate the technical solutions of the present invention, and not to limit it. The scope of protection of the present invention is not limited thereto. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that any person skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features within the scope of the technology disclosed in the present invention; and these modifications or substitutions will not cause the substance of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention. Any modifications or equivalent substitutions that do not deviate from the spirit and scope of the present invention should be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for dynamic optimization of body grid parameters based on model simulation, comprising the following steps: collecting body grid data to be optimized and obtaining a plurality of body historical data from a pre-constructed body historical database module; matching the body grid data to be optimized with the body historical data according to a pre-constructed body grid optimization module to obtain the most similar body historical data as experience information and verification basis for optimization of body grid parameters; analyzing and processing the body grid data to be optimized based on domain knowledge in a pre-constructed domain knowledge base module and reusing the experience information and verification basis to generate optimization suggestion information; and dynamically adjusting and optimizing the body grid data to be optimized according to the optimization suggestion information to obtain body optimized grid data. 2.The method for dynamic optimization of body grid parameters based on model simulation according to claim 1, wherein the method for collecting body grid data to be optimized comprises the following steps: Step 11, obtaining a three-dimensional model of a vehicle to be optimized; Step 12, performing preliminary geometric analysis on the three-dimensional model of the vehicle through a body grid generation module to identify different regions and features of the body and obtain body key part information; Step 13, performing grid division on the three-dimensional model of the vehicle based on overall grid size to generate an initial body grid model; Step 14, calculating the area of each key part according to the body key part information; Step 15, calculating the number of grids of the key part based on the area and grid density; Step 16, performing grid encryption division on the initial body grid model related to the key part according to the number of grids to generate body grid data to be optimized, which includes overall grid size, grid density of each key part, and specific position and shape information of each grid. 3.The method for dynamic optimization of body grid parameters based on model simulation according to claim 1, wherein the method for obtaining a plurality of body historical data from a pre-constructed body historical database module comprises the following steps: obtaining body historical design information accumulated in the process of past vehicle body design and simulation which can be used for grid parameter optimization; the body historical design information is text type data, which includes historical grid adjustment parameters and corresponding performance indicators, and simulation context information; the historical grid adjustment parameters include grid size, density and shape; the performance indicators include body stress, strain and stiffness; the simulation context information includes vehicle model, material type and simulation boundary conditions; and the historical grid adjustment parameters and corresponding performance indicators are split according to the simulation context information to obtain a plurality of body historical data in text type. 4.The method for dynamic optimization of body grid parameters based on model simulation according to claim 3, wherein the method for matching the body grid data to be optimized with the body historical data according to a pre-constructed body grid optimization module to obtain the most similar body historical data comprises the following steps: Step 21, converting a plurality of body historical data in text type into a plurality of word vectors through a word vector algorithm; and Step 22, performing average pooling on the plurality of word vectors to obtain a fixed-dimension body historical vector. ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ Step 23, calculate the mean and standard deviation of the body history vector and the body to be optimized grid data respectively; Step 231, calculate the mean and standard deviation of the body history vector and the body to be optimized grid data respectively; Step 232, based on the mean and standard deviation, correct or remove outliers in the body history vector and the body to be optimized grid data; then calculate the minimum and maximum values in the body history vector and the body to be optimized grid data respectively; Step 233, based on the minimum and maximum values, normalize the body history vector and the body to be optimized grid data respectively, and map the data to the [0, 1] interval to obtain the body history standard vector and the body to be optimized standard vector; Step 234, combine the body history standard vector and the body to be optimized standard vector to obtain the unified body feature vector; Step 24, calculate the similarity of the body feature vector to obtain the cosine similarity of the body to be optimized grid data and each body history data, and select the body history data with the highest cosine similarity as the optimization reference.

5. The method according to claim 4, wherein: Based on the domain knowledge in the pre-constructed domain knowledge base module, and reuse of experience information and verification basis, the method for analyzing and processing the body to be optimized grid data to generate optimization suggestion information is as follows: Based on the historical grid adjustment parameters in the body history data and their corresponding performance indicators, an experience information and verification basis is established; A plurality of semantic vectors are obtained by converting the experience information and verification basis of the text type into semantic vectors through a word vector algorithm; A fixed-dimensional body rule vector is obtained by averaging the plurality of semantic vectors; The domain knowledge in the domain knowledge base module includes body structure mechanics knowledge, grid division principles and material performance knowledge; the body structure mechanics knowledge is used to understand the stress characteristics and deformation rules of different parts of the body under various working conditions; The grid division principle is used to divide the grid according to the geometric shape and performance requirements of the body; The material performance knowledge is used to obtain the mechanical properties of different body materials; Based on the body structure mechanics knowledge, grid division principles and material performance knowledge, and combined with the body rule vector, the body to be optimized grid data is analyzed and processed to generate targeted optimization suggestion information, realizing closed-loop iterative optimization based on experience and real-time simulation.

6. The method according to claim 5, wherein: Based on the body structure mechanics knowledge, grid division principles and material performance knowledge, and combined with the body rule vector, the method for analyzing and processing the body to be optimized grid data to generate targeted optimization suggestion information is as follows: Decompose the body to be optimized grid data to obtain a plurality of index values; According to the index values, combined with the body rule vector, a plurality of automobile performance indicators are constructed, including at least the aerodynamic drag coefficient and the body stiffness; Calculate the index mean of the index values and the automobile performance indicators respectively; According to the mean value of the index, the correlation analysis method is used to calculate the correlation coefficient of the index value and the automobile performance index; According to the correlation coefficient, the corresponding relationship between the index value of the vehicle body and the automobile performance index is determined, and the index optimization direction information is obtained; According to the index optimization direction information, the index value and the corresponding performance index are determined; Based on the knowledge of vehicle body structure mechanics, grid division principle and material performance, the index value and the corresponding performance index are evaluated, the advantages and disadvantages of the vehicle body grid data to be optimized are judged, and the causal inference method is used to generate targeted optimization suggestion information.

7. The vehicle body grid parameter dynamic optimization method based on model simulation according to claim 6, characterized in that: The method for evaluating the index value and the corresponding performance index, judging the advantages and disadvantages of the vehicle body grid data to be optimized, and generating targeted optimization suggestion information by using the causal inference method is as follows: Based on the knowledge of vehicle body structure mechanics, grid division principle and material performance, each performance index is compared with the preset standard value one by one to obtain the performance status information of the current vehicle body design; According to the importance of different performance indexes, each performance index is assigned a corresponding weight, and then the weighted average value is calculated as the comprehensive performance score for judging the advantages and disadvantages of the current vehicle body design; The vehicle body grid data to be optimized is clustered by using the density-based clustering algorithm, and the vehicle body grid data to be optimized is divided into different clusters; if a data point in the vehicle body grid data to be optimized does not belong to any cluster, it is regarded as an abnormal point; by analyzing the grid parameters and performance indexes corresponding to these abnormal points, the grid area with problems is located; For the identified abnormal grid area, the causal inference method is used to judge whether the change of a grid parameter is the cause of the abnormal performance index, and the causal relationship information between the change of the grid parameter and the performance index is determined; Based on the performance status information, the comprehensive performance score and the causal relationship information, targeted optimization suggestion information is generated.

8. The vehicle body grid parameter dynamic optimization method based on model simulation according to claim 1, characterized in that: According to the optimization suggestion information, the vehicle body grid data to be optimized is dynamically adjusted and optimized to obtain the vehicle body optimized grid data as follows: According to the optimization suggestion information, the grid parameters of the corresponding area of the vehicle body grid data to be optimized are adjusted, which at least includes the adjustment of grid size, the adjustment of grid density and the adjustment of grid shape; After the adjustment is completed, the vehicle body grid data is obtained; according to the working condition requirement, the boundary conditions of the simulation are set; According to the vehicle body grid data, the material properties and the boundary conditions, the new performance index of the vehicle body is solved; The vehicle body grid data and the performance index are associated to generate the vehicle body optimized grid data.

9. A vehicle body grid parameter dynamic optimization method based on model simulation, characterized in that: It comprises the following steps: Step one, through the pre-constructed data acquisition model, the vehicle body grid data to be optimized is collected in real time; Step two, using the pre-constructed data analysis model, obtain relevant vehicle body history data from the vehicle body history database module, and combine the domain knowledge in the domain knowledge base module to analyze and process the vehicle body grid data to be optimized, to obtain experience information and verification basis; Step three, using the pre-constructed optimization suggestion generation model, generate optimization suggestion information for the current vehicle body grid data to be optimized based on experience information and verification basis; Step four, using the pre-constructed parameter adjustment model, dynamically adjust the vehicle body grid data to be optimized based on the optimization suggestion information, and then perform a new round of simulation experiment to obtain simulation results; Step five, based on the pre-constructed closed-loop iterative optimization model, iteratively optimize the vehicle body grid parameters based on the simulation results until the simulation results meet the design requirements or reach the pre-set optimization target, to realize dynamic optimization of the vehicle body grid parameters based on model simulation.

10. A dynamic optimization system for vehicle body grid parameters based on model simulation, characterized in that: It includes: One or more processing units; Storage device for storing one or more programs; When the one or more programs are executed by the one or more processing units, the one or more processing units implement a dynamic optimization method for vehicle body grid parameters based on model simulation as claimed in any one of claims 1-9.

Images