Method, device and equipment for generating structure reinforcement data of a body die casting

Abstract

The application discloses a kind of structure reinforcement data generation methods, device and equipment of vehicle body die casting, comprising: constructing the basic finite element model of vehicle body die casting, and defining data enhancement design area;Control node set is set in data enhancement design area and establishes reinforcement rib set, and parameterized finite element model set is obtained;Parameterized finite element model set is carried out finite element simulation, and simulation data information is obtained;According to simulation data information, construct reinforcement rib real structure data pair;Reinforcement rib real structure data pair is input into pre-trained reinforcement rib structure graph perception generator module, and reinforcement rib structure generation data pair is obtained;Feasibility check is carried out to reinforcement rib structure generation data pair, and structure enhancement data are obtained.The problem that traditional test design and finite element simulation combined mode data enhancement mode is difficult to improve model generalization ability under very small amount of real sample is solved, and topology, performance is preserved Structure data enhancement is realized.

Description

Technical Field

[0001] This invention relates to the field of vehicle engineering technology, and in particular to a method, apparatus and equipment for generating structural reinforcement data for die-cast body parts. Background Technology

[0002] Currently, optimizing vehicle body structures using machine learning methods faces bottlenecks such as a scarcity of high-quality simulation data, complex topological relationships between design parameters, and poor engineering rationality in generating new solutions. Structural representation methods are still based on traditional parametric modeling, relying on manually defined design variables (such as plate thickness and beam dimensions), failing to fully utilize the natural topological relationships of the vehicle body structure (such as nodes, connections, and dependencies between components). For complex and irregular structural forms, parametric modeling has limited flexibility and is difficult to adapt to the design requirements of new vehicle bodies.

[0003] Furthermore, regarding data augmentation methods, relying solely on experimental design and finite element simulation to obtain initial data lacks generative data augmentation mechanisms, making it difficult to improve the model's generalization ability with a very small number of real samples. Existing data augmentation methods cannot effectively understand and maintain the intrinsic connections and spatial constraints of the reinforcing rib structure network in die-cast body parts, resulting in poor effectiveness of the generated solutions. Summary of the Invention

[0004] In view of the above, the present invention aims to provide a method, apparatus and equipment for generating structural reinforcement data of die-cast body parts, so as to solve the aforementioned technical problems.

[0005] The technical solution adopted in this invention is as follows:

[0006] This invention provides a method for generating structural reinforcement data for die-cast automotive body parts, including:

[0007] Construct the basic finite element model of the vehicle body die casting and define the data-enhanced design region;

[0008] Within the data-enhanced design area, a set of control nodes and a set of stiffeners are set to obtain a parametric finite element model set.

[0009] Finite element simulations were performed on the parameterized finite element model set to obtain simulation data information;

[0010] Based on the simulation data, construct a pair of real structural data for the reinforcing ribs;

[0011] Input the real structural data of the stiffener into the pre-trained stiffener structure graph perception generator module to obtain stiffener structure generated data pairs.

[0012] The feasibility of the data generated for the stiffening rib structure is verified to obtain structural reinforcement data.

[0013] Optionally, control nodes and reinforcing ribs are set within the data augmentation design area, including:

[0014] Extract data to enhance the upper boundary line, center line, and lower boundary line of the design area;

[0015] A first set of control nodes is set at equal intervals along the upper boundary line;

[0016] A second set of control nodes is set at equal intervals along the centerline;

[0017] A third set of control nodes is set at equal intervals along the lower boundary line;

[0018] A set of reinforcing ribs is established by taking the second set of control nodes as the starting point of the reinforcing ribs and the first and third sets of control nodes as the ending points of the reinforcing ribs.

[0019] Optionally, control nodes and stiffeners are set within the data-enhanced design area to obtain a parametric finite element model set, including:

[0020] Based on the set of control nodes, spatial position perturbation is performed to obtain the finite element model set of the integrated die-cast floor of the vehicle body after node perturbation deformation;

[0021] The thickness of the stiffener set of the finite element model set of the integrated die-cast floor of the vehicle body after nodal disturbance deformation is parameterized to obtain the parameterized finite element model set.

[0022] Optionally, finite element simulation is performed on the parametric finite element model set to obtain simulation data information, including:

[0023] Finite element simulation is performed on the parametric finite element model set to obtain the maximum stress value of each stiffener, and the maximum stress value is used as the stress result of the stiffener.

[0024] The simulation data information is obtained by mapping the actual structural data of the reinforcing ribs of the die-cast body parts with the maximum stress value;

[0025] The simulation data includes the node number and spatial coordinates of the enhanced design area, the stiffener number, stiffener thickness, stiffener stress results, and die-cast part quality.

[0026] Optionally, based on the simulation data information, a pair of real structural data for the stiffener is constructed, including:

[0027] Extract the stiffener diagram data and stress results from the simulation data to construct a pair of real structural data for the stiffener.

[0028] Optionally, input the actual structural data of the stiffener into the stiffener structural diagram perception generator module to obtain stiffener structural generation data pairs, including:

[0029] The node features in the real structural data of the stiffener are input into the node feature encoding branch of the stiffener structural diagram perception generator module, and then encoded into node features.

[0030] The edge features in the stiffener structure diagram data pair are input into the edge feature encoding branch of the stiffener structure diagram perception generator module to obtain the edge feature encoding;

[0031] The node encoding features and the edge encoding features are fused and input into the convolutional encoding layer of the reinforcing rib structure graph perception generator module to obtain graph convolutional features;

[0032] The graph convolutional features are input into the fully connected layer of the reinforcing rib structure graph perception generator module to obtain the latent vector;

[0033] The potential vector is input into the node generation branch of the stiffener structure graph perception generator module to obtain node features;

[0034] The node features are traversed and combined with the potential vectors, then input into the edge generation branch of the stiffener structure graph perception generator module to obtain the probability value of the existence of edges for node pairs and the thickness features.

[0035] The potential vector is input into the performance prediction branch of the stiffener structure diagram perception generator module to obtain the predicted stress value;

[0036] By combining node features, existing edges that meet preset probability values, thickness features, and predicted stress values, a data pair for generating stiffener structures is obtained.

[0037] Optionally, the feasibility of generating data pairs for the stiffener structure is verified to obtain structural reinforcement data, including:

[0038] Data pairs of stiffener structures with node spacing less than or equal to a preset minimum value and thickness characteristics within a preset thickness range are generated and identified as structural reinforcement data.

[0039] Optionally, the node features are traversed and combined with the latent vectors, then input into the edge generation branch of the stiffener structure graph perception generator module to obtain the probability values ​​of the existence of edges for node pairs and thickness features, including:

[0040] In the edge generation branch, a spatial adjacency-inspired edge probability constraint mechanism is introduced;

[0041] When the edge probability constraint mechanism is executed, it only calculates the probability of each starting node forming an edge connection with other nodes in its local neighborhood to obtain the local probability; the local neighborhood is defined as all other nodes whose spatial distance from the starting node is less than a preset neighborhood radius.

[0042] Based on the local probabilities, the output node pairs are assigned final probability values ​​for the existence of edges and thickness features.

[0043] The present invention also provides a computing device, including: a processor and a memory storing a computer program, wherein the computer program, when run by the processor, executes the method described above.

[0044] The present invention also provides a computer-readable storage medium storing instructions that, when executed on a computer, cause the computer to perform the method described above.

[0045] The above-described solution of the present invention has at least the following beneficial effects:

[0046] The above-described solution of the present invention constructs a basic finite element model of a die-cast vehicle body and defines a data augmentation design region; within the data augmentation design region, a set of control nodes and a set of stiffeners are established to obtain a parametric finite element model set; finite element simulation is performed on the parametric finite element model set to obtain simulation data information; based on the simulation data information, a pair of real structural data for stiffeners is constructed; the pair of real structural data for stiffeners is input into a pre-trained stiffener structure diagram perception generator module to obtain a pair of generated stiffener structure data; the feasibility of the generated stiffener structure data pair is verified to obtain structural augmentation data. This solves the problem that traditional data augmentation methods combining experimental design and finite element simulation are difficult to improve the generalization ability of the model with a very small number of real samples, and achieves structural data augmentation that preserves topology and performance. Attached Figure Description

[0047] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described below with reference to the accompanying drawings, wherein:

[0048] Figure 1 A flowchart illustrating a method for generating structural reinforcement data for vehicle body die-cast parts provided in an embodiment of the present invention.

[0049] Figure 2 A schematic diagram illustrating the construction process of a vehicle body die-casting structural data enhancement system provided in an embodiment of the present invention.

[0050] Figure 3 A schematic diagram of the system architecture for enhancing the structural data of die-cast parts for vehicle bodies provided in an embodiment of the present invention.

[0051] Figure 4This is a schematic diagram of the data flow of the stiffener structure diagram perception generator module provided in an embodiment of the present invention.

[0052] Figure 5 This is a schematic diagram illustrating the process of data augmentation design region node perturbation and thickness parameterization provided in an embodiment of the present invention.

[0053] Figure 6 A schematic diagram of a module for generating structural reinforcement data of a vehicle body die-casting part provided in an embodiment of the present invention. Detailed Implementation

[0054] Embodiments of the present invention are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0055] This invention proposes an embodiment of a method for generating structural reinforcement data for die-cast automotive body parts, specifically, as follows: Figure 1 and Figure 2 As shown, it includes:

[0056] Step 11: Construct the basic finite element model of the vehicle body die casting and define the data augmentation design region;

[0057] Step 12: Set a set of control nodes and establish a set of stiffeners within the data augmentation design area to obtain a set of parametric finite element models;

[0058] Step 13: Perform finite element simulation on the parameterized finite element model set to obtain simulation data information;

[0059] Step 14: Based on the simulation data information, construct a pair of real structural data for the reinforcing ribs;

[0060] Step 15: Input the real structural data pairs of the reinforcing ribs into the pre-trained reinforcing rib structure graph perception generator module to obtain the generated data pairs of the reinforcing rib structures.

[0061] Step 16: Perform a feasibility verification on the data generated by the stiffener structure to obtain structural reinforcement data.

[0062] In this embodiment, in step 11, as follows: Figure 3 As shown, the vehicle body die-casting structural data enhancement system of this embodiment includes a stiffener structure diagram perception generator module, a vehicle body die-casting model construction module, a high-performance parallel simulation calculation module, a stiffener diagram perception generator module, and a stiffener diagram structure discriminator module.

[0063] In the CAE environment, a finite element model of the basic die-cast rear floor of the vehicle body is established. The distribution area of ​​the side stiffeners of the finite element model of the floor after the unibody die casting is defined as the data augmentation design region. This region has a variable cross-section, rectangular-like structure. The upper boundary line, center line, and lower boundary line of the data augmentation design region are extracted, and a row of control nodes is set at equal intervals on the upper boundary line. , The spacing between them is 50mm; similarly, a row of control nodes is set at the same interval on the center line and the lower boundary line. , :

[0064] ;

[0065] .

[0066] In step 12, the centerline control node is... As the starting point for the reinforcing ribs, the control nodes of the upper and lower boundary lines are used. and As the endpoint of the reinforcing rib, each centerline control node Control nodes with the same number and adjacent upper and lower boundary lines and Perform pairing, for example , , , , At the same time, control nodes of the center line Pair them in sequence, for example , Then, stiffeners are created; the created stiffeners are assigned numbers to obtain a stiffener set. , .

[0067] like Figure 5 As shown, with , , To control the nodes, a mesh deformation method is used to perturb them in the XY plane, with the perturbation range for the control nodes extending in the X and Y directions. For the upper and lower boundary line nodes , During disturbance deformation, the smoothness of the local structure must be ensured, resulting in a finite element model set of the integral die-cast floor of the vehicle body after nodal disturbance deformation. , .

[0068] Furthermore, a finite element model set for the die-cast rear floor of the vehicle body was developed. For each model in the model, the stiffener set is parameterized for thickness, and the range of thickness parameterization is defined as follows: Where 0.1mm indicates that the reinforcing rib does not exist, thus obtaining the finite element model set of the one-piece die-cast floor of the car body after thickness parameterization. , .

[0069] In step 13, the established parameterized finite element model set of the die-cast integrated floor of the vehicle body is used for simulation calculation to obtain the maximum stress value corresponding to each model, and to form simulation data information with a one-to-one correspondence between structural data and performance response, and to build a structural database. The simulation data information includes the node number, node spatial coordinates, stiffener number, thickness, and corresponding maximum stress value of the control nodes in the data augmentation design area, as well as the mass of the die-cast part.

[0070] In step 14, the stiffener diagram data is extracted from the simulation data. and stress results Construct a pair of real structural data for stiffeners that includes stiffener diagram data and stress result data.

[0071] Among them, the reinforcing rib diagram data The construction rules are as follows:

[0072] Node features : Put all control nodes , , As nodes in a graph, each node The feature vector is defined as a two-dimensional vector ( )

[0073] Edge features Based on the established set of stiffeners, an undirected edge is created for the two connecting nodes corresponding to each stiffener. The initial characteristic of each edge is the thickness reference value of the stiffener it represents.

[0074] In step 15, the real structural data of the reinforcing ribs is input into the pre-trained reinforcing rib structure graph perception generator module, which outputs new graph data. and new stress results And form a reinforcing rib structure to generate data pairs .

[0075] Specifically, such as Figure 4 As shown, the stiffener structure image perception generator module includes an encoder unit and a decoder unit. It generates node features from the real stiffener structure data pairs. The input is fed into the node feature encoding branch to obtain the node encoded features. The specific structure of the node feature encoding branch is as follows: Input layer ( ,3), linear connection layer (3,64), batch normalized layer (64), linear connection layer (64,128), batch normalized layer (128), all with ReLU activation function, where To increase the number of nodes in the ribbed structure;

[0076] The edge features of the real structural data of the stiffener The input is fed into the edge feature encoding branch to obtain the edge feature encoding. The edge feature encoding branch is used to encode the edge features in the stiffener diagram structure. The encoding process is structured as follows: input layer ( ,1), linear connection layer (1, 32), batch normalized layer (32), linear connection layer (32, 64), batch normalized layer (64), all with ReLU activation function. The number of edges in the stiffener diagram structure;

[0077] Encode node features Edge feature encoding The input is fused into a convolutional coding layer to obtain graph convolutional features. The convolutional coding layer here is used to perform graph convolution operations on node feature encoding and edge feature encoding information. Its specific structure is as follows: first graph convolutional layer (128, 256), second graph convolutional layer (256, 256), and third graph convolutional layer (256, 128).

[0078] Convolution features of the graph The input is fed into a fully connected layer after global pooling to obtain a latent vector. The specific structure of the fully connected layer is as follows: a linear connected layer (128, 128), a linear connected layer (128, 64), and the activation function for both layers is the ReLU function.

[0079] latent vectors Input to a node generates a branch, resulting in new node features. Node generation branches are used to process latent vectors. Decode to obtain node information Its specific structure is as follows: input layer ( ,128), batch normalization layer (128), linear connection layer (128, 256), batch normalization layer (256), linear connection layer (256, *3), the activation function is ReLU;

[0080] Iterate through all possible node pairs and match them with the latent vector. By combining the inputs and generating branches from the edges, we can obtain the probability that each node pair has an edge. and thickness characteristics ;

[0081] latent vectors Decode the input and feed it into the performance prediction branch to predict the corresponding stress value. ; Node features The probability of existence Edges with a value of 1 and their thickness features stress value Combined to generate data pairs for stiffener structures .

[0082] Edge generation branches are used to process latent vectors Decoding yields edge feature information. Its specific structure is as follows: input layer ( ,256), batch normalization layer (256), linear connection layer (256, 128), batch normalization layer (128), linear connection layer (128, 2), containing edge existence probability and thickness features, where the activation function for edge existence probability is Sigmoid and the activation function for thickness features is ReLU;

[0083] The performance prediction branch is used based on the latent vector Predict the corresponding stress results The specific structure is as follows: Input layer ( (64), linear connection layer (64, 32), linear connection layer (32, 1), all with ReLU activation function;

[0084] In this embodiment, the training process of the stiffener structure diagram perception generator module is as follows:

[0085] The potential vector obtained by comparing the real structural data of the stiffeners The data is input into the stiffener diagram structure discriminator module to calculate the probability of data authenticity. If the probability value of a node pair is greater than or equal to 0.5, the edge is determined to exist, and its generated thickness feature is retained. If the probability value is less than 0.5, the thickness feature of the edge is forcibly set to 0.0mm, so that the reinforcing rib it represents does not physically exist. The reinforcing rib diagram structure discriminator is used to determine whether the input reinforcing rib structure diagram is a real and reasonable scheme or a fictitious scheme. Its specific structure is as follows: Input layer ( (128), linear connection layer (128, 64), linear connection layer (64, 1), with the activation function being the Sigmoid function.

[0086] Probability of data authenticity The system makes a judgment: if the value is 1, the data is considered real; if the value is 0, the data is considered fake.

[0087] Furthermore, the loss of the stiffener-based image perception generator is calculated. Calculate according to the following formula:

[0088] ;

[0089] Calculate the loss of the stiffener diagram structure discriminator module Calculate according to the following formula:

[0090] ;

[0091] in, The data authenticity loss, used to make the score of the true structured data close to 1, is calculated according to the following formula:

[0092] ;

[0093] The data falsity loss, used to make the generated structured data score close to 0, is calculated according to the following formula:

[0094] ;

[0095] in, To use random vectors The false data obtained by inputting into the stiffener diagram perception generator module.

[0096] The structural data in the aforementioned structural database is input into the stiffener diagram perception generator and the stiffener diagram structure discriminator in batches. The generator and discriminator are then trained iteratively, and their parameters are updated.

[0097] Determine the loss function of the rib map perception generator Loss function of stiffener diagram structure discriminator If convergence is achieved, output the trained ribbed graph perception generator and the trained ribbed graph structure discriminator.

[0098] In step 16, the feasibility of generating data pairs for the reinforcing rib structure is verified, including:

[0099] Perform geometric validity verification: check the structure Check whether the spatial distance between the two nodes connected by each edge is greater than the preset minimum spacing (5mm) to prevent the formation of excessively short and meaningless reinforcing ribs; at the same time, check whether the thickness characteristic value of each edge falls within the preset reasonable engineering range (2mm, 5mm).

[0100] Perform topology continuity verification: check the structure The graph formed by nodes and edges is checked for connectivity, and invalid generation results that produce isolated nodes or isolated reinforcing subgraphs are filtered out.

[0101] If the geometric check passes, the newly generated stiffener structure will be... and the corresponding stress results , as structural enhancement data.

[0102] The method for generating structural reinforcement data for die-cast body parts in this embodiment combines the structural diagram data of the reinforcing ribs of the die-cast body parts with performance parameters into a data pair, which is then input into the graph structure discriminator module for discrimination. This method can learn physical laws from the original small amount of real data and effectively filter out data combinations that are structurally reasonable but have mismatched performance or have reasonable performance but invalid structure. This representation method completely preserves the topological connection relationship between the reinforcing ribs, ensuring that any new structure generated automatically meets the geometric continuity constraint, fundamentally solving the problems of structural discontinuity and missing connection relationship caused by the flattened vector representation in traditional generation methods.

[0103] When expanding the data, the constructed graph perception generator module can sample and interpolate based on the basic model and generate a reasonable structure that matches it. The entire process does not require calling any external simulation programs, thereby reducing the computational cost of a single data generation from several hours to milliseconds, and realizing that the cost of expanding the data of the reinforcing rib structure of the die-cast body parts is close to zero.

[0104] By fusing the reconstruction capability of the encoder with the generation quality of the generative adversarial network, a multi-objective training framework that integrates reconstruction loss and adversarial loss is formed. This framework can constrain the generated reinforcement schemes to approach the original topology, making the augmented data more reasonable.

[0105] A topology-preserving generation framework for vehicle body die-casting stiffener structure based on graph variational autoencoder: The geometric structure of the vehicle body die-casting stiffener is innovatively represented as a graph data structure. The topological features are extracted by a graph neural network encoder and then reconstructed by a graph structure perception decoder, realizing the topology-preserving generation from random latent vectors to a reasonable engineering structure.

[0106] Addressing the engineering reality of scarce CAE simulation data for die-cast automotive parts, this paper constructs a latent space with clear physical meaning through graph variational autoencoders, achieving dimensional decoupling of feature control. This supports stable learning and precise, controllable generation in small sample scenarios, breaking the dependence of traditional generation methods on large amounts of data.

[0107] like Figure 6 As shown, this embodiment also provides a device 60 for generating structural reinforcement data for vehicle body die-cast parts, including:

[0108] Module 61 is used to construct the basic finite element model of the body die casting and define the data augmentation design region; within the data augmentation design region, a set of control nodes and a set of stiffeners are set to obtain a parametric finite element model set.

[0109] The processing module 62 is used to perform finite element simulation on the parameterized finite element model set to obtain simulation data information; construct a pair of real structural data for stiffeners based on the simulation data information; input the pair of real structural data for stiffeners into the stiffener structure diagram perception generator module to obtain a pair of generated stiffener structure data; and perform feasibility verification on the pair of generated stiffener structure data to obtain structural reinforcement data.

[0110] Optionally, control nodes and reinforcing ribs are set within the data augmentation design area, including:

[0111] Extract data to enhance the upper boundary line, center line, and lower boundary line of the design area;

[0112] A first set of control nodes is set at equal intervals along the upper boundary line;

[0113] A second set of control nodes is set at equal intervals along the centerline;

[0114] A third set of control nodes is set at equal intervals along the lower boundary line;

[0115] A set of reinforcing ribs is established by taking the second set of control nodes as the starting point of the reinforcing ribs and the first and third sets of control nodes as the ending points of the reinforcing ribs.

[0116] Optionally, control nodes and stiffeners are set within the data-enhanced design area to obtain a parametric finite element model set, including:

[0117] Based on the set of control nodes, spatial position perturbation is performed to obtain the finite element model set of the integrated die-cast floor of the vehicle body after node perturbation deformation;

[0118] The thickness of the stiffener set of the finite element model set of the integrated die-cast floor of the vehicle body after nodal disturbance deformation is parameterized to obtain the parameterized finite element model set.

[0119] Optionally, finite element simulation is performed on the parametric finite element model set to obtain simulation data information, including:

[0120] Finite element simulation is performed on the parametric finite element model set to obtain the maximum stress value of each stiffener, and the maximum stress value is used as the stress result of the stiffener.

[0121] The simulation data information is obtained by mapping the actual structural data of the reinforcing ribs of the die-cast body parts with the maximum stress value;

[0122] The simulation data includes the node number and spatial coordinates of the enhanced design area, the stiffener number, stiffener thickness, stiffener stress results, and die-cast part quality.

[0123] Optionally, based on the simulation data information, a pair of real structural data for the stiffener is constructed, including:

[0124] Extract the stiffener diagram data and stress results from the simulation data to construct a pair of real structural data for the stiffener.

[0125] Optionally, input the actual structural data of the stiffener into the stiffener structural diagram perception generator module to obtain stiffener structural generation data pairs, including:

[0126] The node features in the real structural data of the stiffener are input into the node feature encoding branch of the stiffener structural diagram perception generator module, and then encoded into node features.

[0127] The edge features in the stiffener structure diagram data pair are input into the edge feature encoding branch of the stiffener structure diagram perception generator module to obtain the edge feature encoding;

[0128] The node encoding features and the edge encoding features are fused and input into the convolutional encoding layer of the reinforcing rib structure graph perception generator module to obtain graph convolutional features;

[0129] The graph convolutional features are input into the fully connected layer of the reinforcing rib structure graph perception generator module to obtain the latent vector;

[0130] The potential vector is input into the node generation branch of the stiffener structure graph perception generator module to obtain node features;

[0131] The node features are traversed and combined with the potential vectors, then input into the edge generation branch of the stiffener structure graph perception generator module to obtain the probability value of the existence of edges for node pairs and the thickness features.

[0132] The potential vector is input into the performance prediction branch of the stiffener structure diagram perception generator module to obtain the predicted stress value;

[0133] By combining node features, existing edges that meet preset probability values, thickness features, and predicted stress values, a data pair for generating stiffener structures is obtained.

[0134] Optionally, the feasibility of generating data pairs for the stiffener structure is verified to obtain structural reinforcement data, including:

[0135] Data pairs of stiffener structures with node spacing less than or equal to a preset minimum value and thickness characteristics within a preset thickness range are generated and identified as structural reinforcement data.

[0136] It should be noted that this device is the same as the method described above. All implementations in the above method embodiments are applicable to the embodiments of this device and can achieve the same technical effect.

[0137] An embodiment of the present invention also provides a computing device, including: a processor and a memory storing a computer program, wherein the computer program, when executed by the processor, performs the method described in the above embodiments. All implementations in the above method embodiments are applicable to this embodiment and can achieve the same technical effect.

[0138] In this embodiment of the invention, a computer-readable storage medium is also provided, storing instructions that, when executed on a computer, cause the computer to perform the method described in the above embodiments. All implementations of the methods described in the above embodiments are applicable to this embodiment and can achieve the same technical effect.

[0139] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0140] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0141] In the embodiments provided by this invention, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative. For instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0142] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0143] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0144] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, essentially, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, ROM, RAM, magnetic disks, or optical disks.

[0145] Furthermore, it should be noted that in the apparatus and method of the present invention, it is obvious that the components or steps can be decomposed and / or recombined. These decompositions and / or recombinations should be considered equivalent solutions of the present invention. Moreover, the steps performing the above series of processes can naturally be executed in the order described, but are not necessarily required to be executed in chronological order; some steps can be executed in parallel or independently of each other. Those skilled in the art will understand that all or any step or component of the method and apparatus of the present invention can be implemented in any computing device (including processors, storage media, etc.) or network of computing devices, in hardware, firmware, software, or a combination thereof. This is something that those skilled in the art can achieve by using their basic programming skills after reading the description of the present invention.

[0146] Therefore, the object of the present invention can also be achieved by running a program or a set of programs on any computing device. The computing device can be a known general-purpose device. Therefore, the object of the present invention can also be achieved simply by providing a program product containing program code implementing the method or apparatus. That is, such a program product also constitutes the present invention, and the storage medium storing such a program product also constitutes the present invention. Obviously, the storage medium can be any known storage medium or any storage medium developed in the future. It should also be noted that in the apparatus and method of the present invention, it is obvious that the components or steps can be decomposed and / or recombined. These decompositions and / or recombinations should be considered equivalent to the present invention. Furthermore, the steps performing the above series of processes can naturally be performed in the order described, but are not necessarily required to be performed in chronological order. Some steps can be performed in parallel or independently of each other.

[0147] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for generating structural reinforcement data for a vehicle body die-casting part, characterized in that, include: Construct the basic finite element model of the vehicle body die casting and define the data-enhanced design region; Within the data-enhanced design area, a set of control nodes and a set of stiffeners are set to obtain a parametric finite element model set. Finite element simulations were performed on the parameterized finite element model set to obtain simulation data information; Based on the simulation data, construct a pair of real structural data for the reinforcing ribs; The real structural data of the reinforcing ribs are input into a pre-trained reinforcing rib structure graph perception generator module to obtain reinforcing rib structure generated data pairs. The reinforcing rib structure graph perception generator module includes an encoding unit and a decoding unit. Node features from the real structural data pairs are input to the node feature encoding branch of the reinforcing rib structure graph perception generator module to obtain node encoded features. Edge features from the real structural data pairs are input to the edge feature encoding branch of the reinforcing rib structure graph perception generator module to obtain edge feature encodings. The node encoded features and the edge feature encodings are fused and input into the convolutional encoding layer of the reinforcing rib structure graph perception generator module to obtain a graph convolution. The graph convolution features are input into the fully connected layer of the stiffener structure graph perception generator module to obtain a latent vector. The latent vector is then input into the node generation branch of the stiffener structure graph perception generator module to obtain node features. The node features are traversed and combined with the latent vector, then input into the edge generation branch of the stiffener structure graph perception generator module to obtain the probability value of the existence of edges for node pairs and the thickness features. The latent vector is then input into the performance prediction branch of the stiffener structure graph perception generator module to obtain the predicted stress value. The node features, the existence of edges that meet the preset probability value, the thickness features, and the predicted stress value are combined to obtain stiffener structure generation data pairs. The feasibility of the data generated for the stiffening rib structure is verified to obtain structural reinforcement data.

2. The method for generating structural reinforcement data for vehicle body die-cast parts according to claim 1, characterized in that, Setting control nodes and establishing reinforcing ribs within the data augmentation design area includes: Extract data to enhance the upper boundary line, center line, and lower boundary line of the design area; A first set of control nodes is set at equal intervals along the upper boundary line; A second set of control nodes is set at equal intervals along the centerline; A third set of control nodes is set at equal intervals along the lower boundary line; A set of reinforcing ribs is established by taking the second set of control nodes as the starting point of the reinforcing ribs and the first and third sets of control nodes as the ending points of the reinforcing ribs.

3. The method for generating structural reinforcement data for vehicle body die-cast parts according to claim 2, characterized in that, Control nodes and stiffeners are set within the data augmentation design region to obtain a parametric finite element model set, including: Based on the set of control nodes, spatial position perturbation is performed to obtain the finite element model set of the integrated die-cast floor of the vehicle body after node perturbation deformation; The thickness of the stiffener set of the finite element model set of the integrated die-cast floor of the vehicle body after nodal disturbance deformation is parameterized to obtain the parameterized finite element model set.

4. The method for generating structural reinforcement data for vehicle body die-cast parts according to claim 1, characterized in that, Finite element simulations are performed on the parametric finite element model set to obtain simulation data information, including: Finite element simulation is performed on the parametric finite element model set to obtain the maximum stress value of each stiffener, and the maximum stress value is used as the stress result of the stiffener. The simulation data information is obtained by mapping the actual structural data of the reinforcing ribs of the die-cast body parts with the maximum stress value; The simulation data includes the node number and spatial coordinates of the enhanced design area, the stiffener number, stiffener thickness, stiffener stress results, and die-cast part quality.

5. The method for generating structural reinforcement data for vehicle body die-cast parts according to claim 1, characterized in that, Based on the simulation data, construct a pair of real structural data for the stiffeners, including: Extract the stiffener diagram data and stress results from the simulation data to construct a pair of real structural data for the stiffener.

6. The method for generating structural reinforcement data for vehicle body die-cast parts according to claim 1, characterized in that, The feasibility of generating data pairs for the stiffener structure is verified to obtain structural reinforcement data, including: Data pairs of stiffener structures with node spacing less than or equal to a preset minimum value and thickness characteristics within a preset thickness range are generated and identified as structural reinforcement data.

7. The method for generating structural reinforcement data for vehicle body die-cast parts according to claim 1, characterized in that, The node features are traversed and combined with the latent vectors, then input into the edge generation branch of the stiffener structure graph perception generator module to obtain the probability values ​​of the existence of edges for node pairs and thickness features, including: In the edge generation branch, a spatial adjacency-inspired edge probability constraint mechanism is introduced; When the edge probability constraint mechanism is executed, it only calculates the probability of each starting node forming an edge connection with other nodes in its local neighborhood to obtain the local probability. The local neighborhood is defined as all other nodes whose spatial distance from the starting node is less than the preset neighborhood radius; Based on the local probabilities, the output node pairs are assigned final probability values ​​for the existence of edges and thickness features.

8. A computing device, characterized in that, include: A processor, a memory storing a computer program, wherein the computer program, when executed by the processor, performs the method as described in any one of claims 1 to 7.

9. A computer-readable storage medium, characterized in that, The system stores instructions that, when executed on a computer, cause the computer to perform the method as described in any one of claims 1 to 7.

Images