Hub style generation method and device based on topological optimization and server

Abstract

The invention provides a hub style generation method and device based on topological optimization and a server, and relates to the technical field of artificial intelligence computer aided design, and the method comprises the steps: obtaining an original data set; based on the hub reference pattern, performing topological optimization processing on the initial hub pattern to reduce the flexibility of the initial hub pattern and the difference between the initial hub pattern and the hub reference pattern, and determining a first hub optimization pattern; through a preset L1 loss function, the hub optimization styles are screened, a second hub optimization style is determined, and a new style proportion corresponding to the second hub optimization style is obtained; when the new style proportion is smaller than a preset proportion threshold value, the second hub optimization style is determined as a target hub style, style evaluation processing is conducted on the target hub style, and a visual evaluation result is generated. According to the method, the diversity and engineering performance of generated hub styles can be remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of artificial intelligence computer-aided design, and in particular to a wheel hub style generation method, device and server based on topology optimization. Background Art

[0002] The generation of automobile wheel styles is an important part of automobile styling design. For the design of wheel styles, it is necessary to first ensure the engineering performance of the wheel. In addition, for users, the aesthetic design of the wheel is also an indispensable factor. At present, the existing technology proposes that the wheel styles can be manually designed by CAD designers or new wheel styles can be automatically generated by automatic encoders. Among them, the manually designed wheel styles have low diversity, and the automatic design of the encoder can only focus on the engineering performance of the wheel, and cannot create aesthetic design, and thus cannot achieve a balance between the aesthetics and engineering performance of the wheel design. Summary of the invention

[0003] In view of this, an object of the present invention is to provide a wheel hub style generation method, device and server based on topology optimization, which can significantly improve the diversity and engineering performance of generated wheel hub styles.

[0004] In a first aspect, an embodiment of the present invention provides a method for generating a wheel hub style based on deep learning and topology optimization, the method comprising: obtaining an original data set, wherein the original data set comprises: a preset number of wheel hub reference styles based on existing vehicle wheel hub styles; based on the wheel hub reference style, performing topology optimization processing on the initial wheel hub style to reduce the flexibility of the initial wheel hub style and the difference between the initial wheel hub style and the wheel hub reference style, and determining a first wheel hub optimized style, wherein the flexibility is used to represent engineering performance, and the smaller the flexibility, the better the engineering performance of the wheel hub style; screening the wheel hub optimized style through a preset L1 loss function, determining a second wheel hub optimized style, and obtaining a new style ratio corresponding to the second wheel hub optimized style, wherein the new style ratio is the ratio of the number of new wheel hub styles in the current iteration to the total number of wheel hub styles in the previous iteration; when the new style ratio is less than a preset ratio threshold, determining the second wheel hub optimized style as the target wheel hub style, and performing style evaluation processing on the target wheel hub style to generate a visual evaluation result.

[0005] In one embodiment, based on the hub reference style, the initial hub style is topologically optimized to reduce the flexibility of the initial hub style and the difference between the initial hub style and the hub reference style, and the step of determining the first hub optimized style includes: performing data analysis on the hub reference style to determine the displacement vector, the overall stiffness matrix and the density vector contained in the hub reference style, and determining the optimization model based on the displacement vector, the overall stiffness matrix and the density vector; topologically optimizing the initial hub style using the optimization model to determine the first hub optimized style, wherein the optimization model includes: flexibility terms and difference terms, and the optimization model is:

[0006] in, is the design variable, representing the unit The density of is the density vector, is the displacement vector, T is the transpose of the matrix, is the overall stiffness matrix, is the flexibility term, is the difference term, For the wheel hub reference style, is the L1 norm, is the similarity parameter, The larger the value, the closer the optimal design is to the hub reference style. , the hub reference style is ignored and optimized to minimize compliance. is the volume fraction, is the material volume, is the design domain volume.

[0007] In one embodiment, the flexibility term in the optimization model is:

[0008] in, is the displacement vector, K is the overall stiffness matrix, For softness, is the unit displacement vector, is the element stiffness matrix, is the volume fraction, is the number of units, for The density of the unit, is the material volume, is the design domain volume.

[0009] In one embodiment, after the step of obtaining the new style ratio corresponding to the second wheel hub optimization style, it includes: when the new style ratio is not less than a preset ratio threshold, using a preset generation model to perform auxiliary design processing on the second wheel hub optimization style to determine a third wheel hub optimization style.

[0010] In one embodiment, the generative model includes a preset generative adversarial network, which includes: a generator and a discriminator. Before using the preset generative model to perform auxiliary design processing on the second wheel hub optimization style and determining the third wheel hub optimization style, it includes: obtaining a random noise vector and sending the random noise vector to the generator to provide randomness for wheel hub style generation and determine random wheel hub styles of different styles; sending the random wheel hub style to the discriminator to perform adversarial training on the generator and the discriminator to determine a target generative adversarial network, so as to use the target generative adversarial network to perform auxiliary design processing on the second wheel hub optimization style and determine the third wheel hub optimization style.

[0011] In one embodiment, after using a preset generation model to perform auxiliary design processing on the second hub optimization style and determining the third hub optimization style, it includes: performing secondary screening processing on the third hub optimization style through a preset L1 loss function, determining the fourth hub optimization style, and ending this round of topology optimization, and re-performing topology optimization processing on the fourth hub optimization style.

[0012] In one embodiment, the steps of performing style evaluation processing on a target wheel hub style and generating a visual evaluation result include: quantitatively evaluating the innovation of the target wheel hub style by using a preset autoencoder, determining a structural innovation evaluation result of the target wheel hub style based on the difference between the target wheel hub style and a wheel hub reference style, and performing a physical property test on the target wheel hub style to determine a feasibility evaluation result of the target wheel hub style; integrating the structural innovation evaluation result and the feasibility evaluation result to determine a visual evaluation result.

[0013] In a second aspect, an embodiment of the present invention further provides a wheel hub style generation device based on topology optimization, the device comprising: a data acquisition module, acquiring an original data set, wherein the original data set comprises: a preset number of wheel hub reference styles based on existing vehicle wheel hub styles; a topology optimization module, based on the wheel hub reference style, performing topology optimization processing on the initial wheel hub style to reduce the flexibility of the initial wheel hub style and the difference between the initial wheel hub style and the wheel hub reference style, and determining a first wheel hub optimization style, wherein the flexibility is used to represent engineering performance, and the smaller the flexibility, the better the engineering performance of the wheel hub style; a screening and filtering module, screening the wheel hub optimization style through a preset L1 loss function, determining a second wheel hub optimization style, and obtaining a new style ratio corresponding to the second wheel hub optimization style, wherein the new style ratio is the ratio of the number of new wheel hub styles in the current iteration to the total number of wheel hub styles in the previous iteration; a wheel hub style evaluation module, determining the second wheel hub optimization style as the target wheel hub style when the new style ratio is less than a preset ratio threshold, and performing style evaluation processing on the target wheel hub style to generate a visual evaluation result.

[0014] In a third aspect, an embodiment of the present invention further provides a server, comprising a processor and a memory, wherein the memory stores computer executable instructions that can be executed by the processor, and the processor executes the computer executable instructions to implement any one of the methods provided in the first aspect.

[0015] In a fourth aspect, an embodiment of the present invention further provides a computer-readable storage medium, which stores computer-executable instructions. When the computer-executable instructions are called and executed by a processor, the computer-executable instructions prompt the processor to implement any one of the methods provided in the first aspect.

[0016] The embodiments of the present invention bring the following beneficial effects: A method, device and server for generating a hub style based on topology optimization are provided in an embodiment of the present invention. After acquiring an original data set, the method performs topology optimization processing on an initial hub style based on a hub reference style to reduce the flexibility of the initial hub style and the difference between the initial hub style and the hub reference style, and determines a first hub optimization style. Thereafter, the hub optimization style is screened through a preset L1 loss function to filter out similar styles, determine a second hub optimization style, and obtain a new style ratio corresponding to the second hub optimization style. When the new style ratio is less than a preset ratio threshold, the second hub optimization style is determined to be a target hub style, and a style evaluation processing is performed on the target hub style to generate a visualized evaluation result. The embodiment of the present invention ensures the engineering performance of the generated hub through topology optimization processing, and at the same time combines deep learning to generate a variety of hub styles, which can significantly improve the diversity and engineering performance of the generated hub styles.

[0017] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention are realized and obtained by the structures particularly pointed out in the description, claims and drawings.

[0018] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, preferred embodiments are given below and described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0020] Figure 1 A schematic diagram of the structure of a wheel design area provided by an embodiment of the present invention; Figure 2 A schematic flow chart of a method for generating a hub style based on topology optimization provided by an embodiment of the present invention; Figure 3 A schematic diagram of a specific process of a hub style generation method based on topology optimization provided by an embodiment of the present invention; Figure 4 A schematic structural diagram of a hub style generation device based on topology optimization provided by an embodiment of the present invention; Figure 5 A schematic diagram of the structure of a server provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0021] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution of the present invention will be clearly and completely described in combination with the embodiments below. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0022] At present, the generation of automobile wheel styles is an important part of automobile styling design. For the design of wheel styles, it is necessary to ensure the engineering performance of the wheel first. In addition, for users, the aesthetic design of the wheel is also an indispensable factor. The existing technology proposes that the wheel style can be manually designed by a CAD designer or a new wheel style can be automatically generated by an automatic encoder. Among them, the manually designed wheel style has a low diversity, and the automatic design of the encoder can only focus on the engineering performance of the wheel, and cannot create an aesthetic design, and thus cannot achieve a balance between the aesthetics and engineering performance of the wheel design. In addition, the existing generative models are limited when directly generating engineering designs. First, the generative model requires a large amount of data, but the training data accumulated by the industry for various designs are confidential. Secondly, the generative model cannot guarantee the feasibility of the project. Finally, mode collapse is one of the main problems of the generative model. Low-quality designs can be improved through topology optimization processing.

[0023] Based on this, the present invention implements the method, device and server for generating wheel hub styles based on topology optimization, which first ensure the engineering performance of the generated wheel hub through topology optimization processing, and then generate various wheel hub styles from limited previous data designs through the generative model combined with deep learning. Compared with previous generative design methods, this framework has better generative design aesthetics, diversity and robustness, and can significantly improve the diversity and engineering performance of generated wheel hub styles.

[0024] To facilitate understanding of this embodiment, a method for generating a hub style based on topology optimization disclosed in an embodiment of the present invention is first introduced in detail. Before topology training, it is necessary to first collect hub reference styles, collect previous designs in the market and industry, and use them as the benchmark design for subsequent topology optimization. Then, based on the hub reference style, a new design is generated through topology optimization. The optimization goal is to minimize flexibility (representing engineering performance) and the difference from the hub reference style (aesthetics and diversity), where topology optimization is typically used in structural design, where the design area is divided into elements, and the optimal material density of the elements is determined under given loads and boundary conditions to minimize compliance. An embodiment of the present invention provides a structural schematic diagram of a wheel design area, such as Figure 1 As shown in the figure, the design domain and boundary conditions of the two-dimensional wheel design, the original design domain is 128 × 128 elements, and the hub reference style domain is also 128 × 128 pixels (that is, the format of the input hub reference style is unified to 128 × 128 pixels, which refers to the sum of D1, D2 and D3). The outer ring D1 of the wheel is set as the non-design area to maintain the shape of the rim, and the inner area D3 is set as a fixed boundary condition to connect the parts. The spoke D2 is the main component of the design domain.

[0025] based on Figure 1The structural schematic diagram of the wheel design area shown in FIG. 1 is a schematic diagram of the wheel design area shown in FIG. 1 . The embodiment of the present invention introduces in detail the wheel hub style generation method based on topology optimization. Figure 2 A schematic flow chart of a method for generating a hub style based on topology optimization is shown, and the method mainly includes the following steps S202 to S208: Step S202, obtaining an original data set, wherein the original data set includes: a preset number of wheel hub reference patterns based on existing vehicle wheel hub patterns.

[0026] Step S204: Based on the hub reference style, the initial hub style is topologically optimized to reduce the flexibility of the initial hub style and the difference between the initial hub style and the hub reference style, and the first hub optimized style is determined, wherein flexibility is used to represent engineering performance, and the smaller the flexibility, the better the engineering performance of the hub style. In one embodiment, data analysis can be performed on the hub reference style to determine the displacement vector, the overall stiffness matrix and the density vector contained in the hub reference style, and an optimization model is determined based on the displacement vector, the overall stiffness matrix and the density vector, and the initial hub style is topologically optimized using the optimization model to determine the first hub optimized style, wherein the optimization model includes: flexibility terms and difference terms, and the optimization model is:

[0027] in, is the design variable, representing the unit The density of is the density vector, is the displacement vector, T is the transpose of the matrix, is the overall stiffness matrix, is the flexibility term, is the difference term, For the wheel hub reference style, is the L1 norm, is the similarity parameter, The larger the value, the closer the optimal design is to the hub reference style. , the hub reference style is ignored and optimized to minimize compliance. is the volume fraction, is the material volume, is the design domain volume.

[0028] Furthermore, the flexibility term in the optimization model is:

[0029] in, is the displacement vector, K is the overall stiffness matrix, For softness, is the unit displacement vector, is the element stiffness matrix, is the volume fraction, is the number of units, for The density of the unit, is the material volume, is the design domain volume.

[0030] In addition, in the above expression, the density directly related to Young's modulus can be expressed as:

[0031] in, As a penalty factor, to ensure a black-and-white design, This is to avoid numerical instabilities when the element density is zero.

[0032] Step S206, through the preset L1 loss function, the wheel hub optimization style is screened to filter out similar styles, determine the second wheel hub optimization style, and obtain the new style ratio corresponding to the second wheel hub optimization style, wherein the new style ratio is the ratio of the number of new wheel hub styles in the current iteration to the total number of wheel hub styles in the previous iteration. In one embodiment, after calculating the ratio of the number of new designs in the current iteration to the total number of designs in the previous iteration, if it is less than a user-specified threshold, the iterative design exploration is exited and the design scheme evaluation is performed; otherwise, the preset generation model is used to perform auxiliary design processing on the second wheel hub optimization style to further expand the wheel hub design style. In actual applications, the threshold can be set to 0.3, and the threshold can be adjusted according to the number of wheel hub styles to be generated.

[0033] Step S208, when the new style ratio is less than a preset ratio threshold, the second wheel hub optimization style is determined to be the target wheel hub style, and a style evaluation process is performed on the target wheel hub style to generate a visual evaluation result. In one embodiment, the innovation of the target wheel hub style is quantitatively evaluated by a preset autoencoder, and the structural innovation evaluation result of the target wheel hub style is determined based on the difference between the target wheel hub style and the hub reference style. A physical performance test is performed on the target wheel hub style to determine the feasibility evaluation result of the target wheel hub style. Finally, the structural innovation evaluation result and the feasibility evaluation result are integrated to determine the visual evaluation result.

[0034] Specifically, we can learn the designs in the dataset and train the autoencoder to construct a loss function, which can help quantify the degree of innovation of the new design so that it can be compared with the previous design. In the design scheme evaluation stage, we can comprehensively evaluate the design schemes obtained through iterative design exploration, which includes not only the novelty evaluation of the design scheme, but also the actual physical properties such as the volume and compliance of the design. Finally, we can make trade-offs between different design attributes, draw the design schemes on each design attribute axis, and select the optimal design based on the relative importance of these attributes.

[0035] The above-mentioned wheel hub style generation method based on topology optimization provided in an embodiment of the present invention ensures the engineering performance of the generated wheel hub through topology optimization processing, and at the same time combines deep learning to generate various wheel hub styles, which can significantly improve the diversity and engineering performance of the generated wheel hub styles.

[0036] In one implementation, the present invention also provides a wheel hub style optimization solution when the new style ratio is not less than a preset ratio threshold: see (A) to (B) below: (A) When the new style ratio is not less than a preset ratio threshold, the preset generation model is used to assist in the design of the second wheel hub optimization style to determine the third wheel hub optimization style, wherein the generation model includes a preset generation adversarial network, and the preset generation adversarial network includes: a generator and a discriminator. In one embodiment, a random noise vector can be obtained and sent to the generator to provide randomness for wheel hub style generation, determine random wheel hub styles of different styles, and then send the random wheel hub styles to the discriminator, so that the generator and the discriminator are subjected to adversarial training to determine the target generation adversarial network, so as to utilize the target generation adversarial network to assist in the design of the second wheel hub optimization style to determine the third wheel hub optimization style.

[0037] In one implementation, a generative model can be used to generate a new design based on the current iteration of the design and serve as a new wheel hub reference style. The generative model is mainly implemented by a generative adversarial network (GAN):

[0038] in, is random noise, is random noise The distribution of is a generator, with input z and a weight parameter , is the discriminator, the input is the image , whose weight parameter is For the Generative Adversarial Network (GAN), the training process involves two models that compete with each other, so its objective function includes the process of finding the maximum value (max) and the minimum value (min).

[0039] Specifically, first sample a random noise vector from a normal distribution or a uniform distribution , as the input of the generator, the noise vector provides randomness for generating wheel images of different styles, where the generator It is a deep neural network model with input noise Then, a simulated wheel image is generated through the learned weight parameters. The goal of the generator is to generate an image that is as close to the real data distribution as possible so that the discriminator cannot easily distinguish it.

[0040] Discriminator It is responsible for receiving input images, including real wheel images and images generated by the generator, and outputting a probability value to determine whether the image comes from the real data distribution. By maximizing the output probability of the real image and minimizing the output probability of the generated image, the discriminator gradually improves its ability to distinguish fake images.

[0041] The generator and discriminator are then optimized through adversarial training. The generator tries to minimize the discriminator's judgment error on the generated images, while the discriminator tries to maximize the accuracy of distinguishing between real and generated images. As the training progresses, the wheel images generated by the generator gradually approach the distribution of real images, and eventually it will be able to generate a variety of different styles of highly realistic wheel images.

[0042] (B) By presetting the L1 loss function, the third hub optimization pattern is subjected to secondary screening to filter out similar patterns, determine the fourth hub optimization pattern, and end the current round of topology optimization. The fourth hub optimization pattern is subjected to topology optimization again. In one embodiment, the pixel-by-pixel L1 distance can be used as a standard, and the boundary value can be set to , filter out similar designs and reduce computational costs.

[0043] See also Figure 3 The specific process diagram of a hub style generation method based on topology optimization is shown. The embodiment of the present invention also provides a specific implementation method for generating a hub style. For details, see (1) to (7) below: (1) Collection of wheel hub reference styles (original data sets): First, 1,000 existing wheel hub designs were collected from the market and industry as benchmarks. These wheel hub reference styles were input into the system in the form of images and used to set initial conditions in subsequent topology optimization. This step provided a variety of design references for subsequent optimization and helped generate solutions that met engineering requirements and were aesthetically pleasing.

[0044] (2) Topology optimization: Based on the collected wheel reference styles, topology optimization is performed. The goal of topology optimization is not limited to improving the mechanical properties of the wheel hub, but also to ensure the diversity of design styles. Optimization is achieved by minimizing design flexibility (i.e., improving structural stiffness) and controlling the difference between the design and the reference image.

[0045] Specifically, the design area is divided into a 128 × 128 element grid, and different design areas and boundary conditions are set. The outer ring of the wheel is defined as a non-design area to ensure the stability of the rim shape, and the inner ring area is set as a fixed boundary for connecting components. During the optimization process, the spoke area of ​​the wheel is the core of the design, and the adjustment of the element density determines the material distribution in this area.

[0046] During this stage, specific mathematical models (such as flexibility minimization and hub reference style difference control) are used to constrain the design and the material distribution of each unit is calculated through finite element analysis (FEA) to obtain a preliminary design that meets the structural performance requirements.

[0047] (3) Design screening, filtering similar designs: In order to avoid generating redundant or similar designs, the system will automatically screen after each optimization. In one embodiment, the pixel-by-pixel L1 distance can be used as the screening criterion, and the threshold is set to 10. -3 , filtering out solutions that are too similar to the previous design and retaining designs with large differences. This process significantly reduces the redundancy of subsequent calculations.

[0048] (4) Calculation of the proportion of new designs to determine whether the threshold is reached: The system evaluates the proportion of new designs generated in the current iteration and calculates the ratio of the new design schemes to the total number of schemes in the previous iteration. Since the design process usually involves a large number of parameter adjustments and multi-dimensional optimizations, especially when generating new designs, the system needs to balance innovation and exploration costs. Therefore, if the proportion of new designs generated during the iteration is too high, it may mean that the design space has not been fully covered and there are still many unexplored areas, which need to be further assisted by the generative model; however, if the proportion is too low, it may mean that the design has become stable and further iterations are not very effective. Continuing exploration will lead to a waste of resources. After a large number of experiments and empirical analysis, the threshold for the proportion of new designs is 0.3. As a reasonable critical point, this threshold can balance the requirements of system efficiency and design diversity. If the proportion of new designs is lower than the set threshold, it is considered that the design space has been fully explored, and further iterations are stopped and the final evaluation stage is entered; if the proportion is higher than the threshold, iterations are continued to generate more schemes.

[0049] (5) Generate model-assisted design to generate new wheel styles: To further improve the diversity and innovation of design solutions, this step generates new designs through a generative adversarial network (GAN). The GAN model generates new wheel design images through random noise. The generator G generates new design images based on the input random noise, while the discriminator D is responsible for distinguishing between the generated images and the real images. After multiple rounds of adversarial training, the generator gradually learns to generate designs that are close to the real data distribution. Ultimately, GAN can generate a variety of wheel design images with different styles and high realism.

[0050] (6) Secondary design screening to filter similar designs: The new designs generated by the generative adversarial network are screened again to ensure the diversity of the designs. The pixel-by-pixel L1 distance is still used for difference judgment, which effectively ensures that the generated design schemes are diverse while maintaining structural rationality.

[0051] (7) Design evaluation: In this stage, the autoencoder is used to quantitatively evaluate the design innovation. By calculating the loss function, the difference between the newly generated design and the previous design is compared to evaluate its contribution to structural innovation. In addition, the design scheme must also undergo rigorous physical performance testing, including material volume, flexibility, stress resistance, etc., to ensure its feasibility in practical applications.

[0052] In summary, the above scheme combines topology optimization with generative adversarial networks. Through multiple rounds of iterative design and intelligent screening, it effectively improves design efficiency and reduces manual intervention. During the design evaluation process, the system conducts a multi-dimensional analysis of each design attribute, comprehensively considers the physical performance and aesthetics of the design, and selects the optimal design through attribute weight setting. Specifically: During the design evaluation process, the system first assigns initial weights to each design attribute. These weights are set according to the historical data and design goals in the wheel reference style collection phase. To improve flexibility, the system also introduces adjustable parameters, which reflect the weight relationship between different attributes and are dynamically adjusted during topology optimization and generative adversarial network iterations. For example, the system can adjust the weight according to the wheel hub design effect after topology optimization, and test the influence of different attribute combinations on physical performance and design style. After the design generation and evaluation are completed, the system collects the performance of each design scheme through a feedback mechanism, especially the balance between physical performance and aesthetics, and evaluates the rationality of the current weight setting.

[0053] For the hub style generation method based on topology optimization provided in the above embodiment, the embodiment of the present invention provides a hub style generation device based on topology optimization, see Figure 4 A schematic diagram of the structure of a hub style generation device based on topology optimization is shown, and the device includes the following parts: The data acquisition module 402 acquires an original data set, wherein the original data set includes: a preset number of wheel hub reference patterns based on existing vehicle wheel hub patterns; A topology optimization module 404 performs topology optimization processing on the initial hub style based on the hub reference style to reduce the flexibility of the initial hub style and the difference between the initial hub style and the hub reference style, and determines a first hub optimization style, wherein flexibility is used to represent engineering performance, and the smaller the flexibility, the better the engineering performance of the hub style; The screening and filtering module 406 screens the wheel hub optimization patterns by using a preset L1 loss function to filter out similar patterns, determine a second wheel hub optimization pattern, and obtain a new pattern ratio corresponding to the second wheel hub optimization pattern, wherein the new pattern ratio is a ratio of the number of new wheel hub patterns in the current iteration to the total number of wheel hub patterns in the previous iteration; The wheel hub style evaluation module 408 determines that the second wheel hub optimization style is the target wheel hub style when the new style ratio is less than a preset ratio threshold, performs style evaluation processing on the target wheel hub style, and generates a visual evaluation result.

[0054] The above-mentioned wheel hub style generation device based on topology optimization provided in the embodiment of the present application ensures the engineering performance of the generated wheel hub through topology optimization processing, and at the same time combines deep learning to generate various wheel hub styles, which can significantly improve the diversity and engineering performance of the generated wheel hub styles.

[0055] In one embodiment, when performing a topological optimization process on the initial hub style based on the hub reference style to reduce the flexibility of the initial hub style and the difference between the initial hub style and the hub reference style to determine the first hub optimized style, the topological optimization module 404 is further used to: perform data analysis on the hub reference style to determine the displacement vector, the overall stiffness matrix and the density vector contained in the hub reference style, and determine the optimization model based on the displacement vector, the overall stiffness matrix and the density vector; perform topological optimization process on the initial hub style using the optimization model to determine the first hub optimized style, wherein the optimization model includes: a flexibility term and a difference term, and the optimization model is:

[0056] in, is the design variable, representing the unit The density of is the density vector, is the displacement vector, T is the transpose of the matrix, is the overall stiffness matrix, is the flexibility term, is the difference term, For the wheel hub reference style, is the L1 norm, is the similarity parameter, The larger the value, the closer the optimal design is to the hub reference style. , the hub reference style is ignored and optimized to minimize compliance. is the volume fraction, is the material volume, is the design domain volume.

[0057] In one implementation, the topology optimization module 404 is further used to optimize the flexibility term in the model as follows:

[0058] in, is the displacement vector, K is the overall stiffness matrix, For softness, is the unit displacement vector, is the element stiffness matrix, is the volume fraction, is the number of units, for The density of the unit, is the material volume, is the design domain volume.

[0059] In one embodiment, after the step of obtaining the new style ratio corresponding to the second wheel hub optimization style, the screening and filtering module 406 is also used to: when the new style ratio is not less than a preset ratio threshold, use a preset generation model to perform auxiliary design processing on the second wheel hub optimization style to determine a third wheel hub optimization style.

[0060] In one embodiment, the generative model includes a preset generative adversarial network, which includes: a generator and a discriminator. Before using the preset generative model to perform auxiliary design processing on the second wheel hub optimization style to determine the third wheel hub optimization style, the above-mentioned screening and filtering module 406 is also used to: obtain a random noise vector, and send the random noise vector to the generator to provide randomness for wheel hub style generation and determine random wheel hub styles of different styles; send the random wheel hub style to the discriminator, so that the generator and the discriminator are trained in adversarial training, and a target generative adversarial network is determined, so as to use the target generative adversarial network to perform auxiliary design processing on the second wheel hub optimization style and determine the third wheel hub optimization style.

[0061] In one embodiment, after performing auxiliary design processing on the second hub optimization style using a preset generation model to determine the third hub optimization style, the screening and filtering module 406 is also used to: perform secondary screening processing on the third hub optimization style through a preset L1 loss function to filter out similar styles, determine the fourth hub optimization style, end this round of topology optimization, and re-perform topology optimization processing on the fourth hub optimization style.

[0062] In one embodiment, when performing style evaluation processing on the target wheel hub style and generating a visual evaluation result, the wheel hub style evaluation module 408 is also used to: quantitatively evaluate the innovation of the target wheel hub style by using a preset autoencoder, determine the structural innovation evaluation result of the target wheel hub style based on the difference between the target wheel hub style and the wheel hub reference style, and perform physical performance testing on the target wheel hub style to determine the feasibility evaluation result of the target wheel hub style; integrate the structural innovation evaluation result and the feasibility evaluation result to determine the visual evaluation result.

[0063] The device provided in the embodiment of the present invention has the same implementation principle and technical effects as those of the aforementioned method embodiment. For the sake of brief description, for matters not mentioned in the device embodiment, reference may be made to the corresponding contents in the aforementioned method embodiment.

[0064] An embodiment of the present invention provides a server. Specifically, the server includes a processor and a storage device. The storage device stores a computer program, and when the computer program is executed by the processor, it executes the method described in any one of the above-mentioned embodiments.

[0065] Figure 5 A structural diagram of a server provided in an embodiment of the present invention, the server 100 includes: a processor 50, a memory 51, a bus 52 and a communication interface 53, wherein the processor 50, the communication interface 53 and the memory 51 are connected via the bus 52; the processor 50 is used to execute an executable module stored in the memory 51, such as a computer program.

[0066] The memory 51 may include a high-speed random access memory (RAM), and may also include a non-volatile memory, such as at least one disk memory. The communication connection between the system network element and at least one other network element is realized through at least one communication interface 53 (which may be wired or wireless), and the Internet, wide area network, local area network, metropolitan area network, etc. may be used.

[0067] The bus 52 may be an ISA bus, a PCI bus, or an EISA bus, etc. The bus may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 5 Only one bidirectional arrow is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0068] Among them, the memory 51 is used to store programs, and the processor 50 executes the program after receiving the execution instruction. The method executed by the device for flow process definition disclosed in any embodiment of the above-mentioned embodiment of the present invention can be applied to the processor 50 or implemented by the processor 50.

[0069] The processor 50 may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by the hardware integrated logic circuit or software instructions in the processor 50. The above processor 50 can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The methods, steps and logic block diagrams disclosed in the embodiments of the present invention can be implemented or executed. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor. The steps of the method disclosed in the embodiments of the present invention can be directly embodied as a hardware decoding processor to be executed, or can be executed by a combination of hardware and software modules in the decoding processor. The software module can be located in a mature storage medium in the field such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory, or an electrically erasable programmable memory, a register, etc. The storage medium is located in the memory 51, and the processor 50 reads the information in the memory 51 and completes the steps of the above method in combination with its hardware.

[0070] The computer program product of the readable storage medium provided in the embodiment of the present invention includes a computer-readable storage medium storing program code, and the instructions included in the program code can be used to execute the methods described in the previous method embodiments. The specific implementation can be referred to the previous method embodiments, which will not be repeated here.

[0071] If the functions are implemented in the form of 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 the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc., which can store program codes.

[0072] Finally, it should be noted that the above-described embodiments are only specific implementations of the present invention, which are used to illustrate the technical solutions of the present invention, rather than to limit them. The protection scope of the present invention is not limited thereto. Although the present invention is described in detail with reference to the above-described embodiments, ordinary technicians in the field should understand that any technician familiar with the technical field can still modify the technical solutions recorded in the above-described embodiments within the technical scope disclosed by the present invention, or can easily think of changes, or make equivalent replacements for some of the technical features therein; and these modifications, changes or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.

Claims

1. A wheel hub style generation method based on deep learning and topology optimization, characterized in that: The method comprises: Acquire an original data set, wherein the original data set includes: a preset number of wheel hub reference patterns based on existing vehicle wheel hub patterns; Based on the hub reference pattern, topologically optimize the initial hub pattern to reduce the flexibility of the initial hub pattern and the difference between the initial hub pattern and the hub reference pattern, and determine a first hub optimized pattern, wherein flexibility is used to represent engineering performance, and the smaller the flexibility, the better the engineering performance of the hub pattern; The wheel hub optimization patterns are screened by a preset L1 loss function to determine a second wheel hub optimization pattern, and a new pattern ratio corresponding to the second wheel hub optimization pattern is obtained, wherein the new pattern ratio is a ratio of the number of new wheel hub patterns in a current iteration to the total number of wheel hub patterns in a previous iteration; When the new style ratio is less than a preset ratio threshold, the second wheel hub optimization style is determined to be a target wheel hub style, and a style evaluation process is performed on the target wheel hub style to generate a visual evaluation result.

2. The method for generating a hub style based on deep learning and topology optimization according to claim 1, characterized in that: The step of performing topological optimization processing on the initial hub pattern based on the hub reference pattern to reduce the flexibility of the initial hub pattern and the difference between the initial hub pattern and the hub reference pattern, and determining a first hub optimized pattern comprises: Performing data analysis on the hub reference pattern to determine a displacement vector, an overall stiffness matrix, and a density vector contained in the hub reference pattern, and determining an optimization model based on the displacement vector, the overall stiffness matrix, and the density vector; The initial hub style is topologically optimized using the optimization model to determine the first hub optimization style, wherein the optimization model includes: a flexibility term and a difference term, and the optimization model is: in, is the design variable, representing the unit The density of is the density vector, is the displacement vector, T is the transpose of the matrix, is the overall stiffness matrix, is the flexibility term, is the difference term, For the wheel hub reference style, is the L1 norm, is the similarity parameter, The larger the value, the closer the optimal design is to the hub reference style. , the hub reference style is ignored and optimized to minimize compliance. is the volume fraction, is the material volume, is the design domain volume.

3. The method for generating a hub style based on deep learning and topology optimization according to claim 2, characterized in that: The flexibility term in the optimization model is: in, is the displacement vector, K is the overall stiffness matrix, For softness, is the unit displacement vector, is the element stiffness matrix, is the volume fraction, is the number of units, for The density of the unit, is the material volume, is the design domain volume.

4. The method for generating a hub style based on deep learning and topology optimization according to claim 1, characterized in that: After the step of obtaining the new style ratio corresponding to the second wheel optimization style, the method includes: When the new style ratio is not less than a preset ratio threshold, the second hub optimized style is subjected to auxiliary design processing using a preset generation model to determine a third hub optimized style.

5. The method for generating a hub style based on deep learning and topology optimization according to claim 4, characterized in that: The generation model includes a preset generation adversarial network, and the preset generation adversarial network includes: a generator and a discriminator. Before the step of using the preset generation model to perform auxiliary design processing on the second wheel hub optimization style to determine the third wheel hub optimization style, the method includes: Obtaining a random noise vector, and sending the random noise vector to the generator to provide randomness for generating a wheel style, and determining random wheel styles of different styles; The random wheel hub pattern is sent to the discriminator, so that the generator and the discriminator are subjected to adversarial training, and a target generative adversarial network is determined, so as to utilize the target generative adversarial network to perform auxiliary design processing on the second wheel hub optimization pattern and determine a third wheel hub optimization pattern.

6. The method for generating a hub style based on deep learning and topology optimization according to claim 4, characterized in that: After the step of performing auxiliary design processing on the second hub optimization pattern by using the preset generation model to determine the third hub optimization pattern, the method includes: By presetting the L1 loss function, the third wheel hub optimization style is subjected to secondary screening processing, the fourth wheel hub optimization style is determined, and the current round of topology optimization is terminated, and the topology optimization processing is performed again on the fourth wheel hub optimization style.

7. The method for generating wheel hub styles based on deep learning and topology optimization according to claim 1, characterized in that: The step of performing style evaluation processing on the target wheel hub style and generating a visual evaluation result comprises: The innovation of the target wheel hub style is quantitatively evaluated by a preset autoencoder, and the structural innovation evaluation result of the target wheel hub style is determined according to the difference between the target wheel hub style and the wheel hub reference style, and the physical performance test is performed on the target wheel hub style to determine the feasibility evaluation result of the target wheel hub style; The structural innovation evaluation result and the feasibility evaluation result are integrated to determine the visualization evaluation result.

8. A wheel hub style generation device based on topology optimization, characterized in that: The device comprises: A data acquisition module, acquiring an original data set, wherein the original data set includes: a preset number of wheel hub reference patterns based on existing vehicle wheel hub patterns; a topology optimization module, which performs topology optimization processing on an initial hub pattern based on the hub reference pattern to reduce flexibility of the initial hub pattern and a difference between the initial hub pattern and the hub reference pattern, and determines a first hub optimization pattern, wherein flexibility is used to represent engineering performance, and the smaller the flexibility, the better the engineering performance of the hub pattern; a screening and filtering module, which screens the wheel hub optimization patterns by using a preset L1 loss function, determines a second wheel hub optimization pattern, and obtains a new pattern ratio corresponding to the second wheel hub optimization pattern, wherein the new pattern ratio is a ratio of the number of new wheel hub patterns in a current iteration to the total number of wheel hub patterns in a previous iteration; The wheel hub style evaluation module determines that the second wheel hub optimization style is a target wheel hub style when the new style ratio is less than a preset ratio threshold, performs style evaluation processing on the target wheel hub style, and generates a visual evaluation result.

9. A server, characterized in that: The method comprises a processor and a memory, wherein the memory stores computer executable instructions that can be executed by the processor, and the processor executes the computer executable instructions to implement the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-executable instructions, and when the computer-executable instructions are called and executed by a processor, the computer-executable instructions prompt the processor to implement the method according to any one of claims 1 to 7.

Images