Wheel design evaluation method and wheel design evaluation system based on model simulation

Abstract

The invention discloses a wheel design evaluation method and a wheel design evaluation system based on model simulation, and belongs to the technical field of wheel design evaluation. According to an existing wheel evaluation scheme, an initial wheel structure needs to be obtained firstly, then wheel optimization can be carried out, and the wheel design evaluation efficiency is affected. According to the wheel design evaluation method based on model simulation, the wheel design acquisition model, the simulation wheel generation model, the numerical simulation model and the wheel evaluation model are established, and the wheel design can be directly evaluated according to the two-dimensional wheel design image, so that the wheel design evaluation efficiency is effectively improved; and the wheel research and development design period is shortened. Furthermore, by applying the method provided by the invention, an industrial designer can evaluate the engineering performance of the two-dimensional wheel design in time, so that an optimal concept design scheme can be directly screened out, and the wheel design scheme can be improved according to an evaluation result, thereby effectively improving the wheel design quality and reducing the wheel research and development time.

Description

Technical Field

[0001] The invention relates to a wheel design evaluation method and a wheel design evaluation system based on model simulation, and belongs to the technical field of wheel design evaluation. Background Art

[0002] A Chinese patent application (publication number: CN117574552A) discloses an optimization integrated design method for wheels, which includes: obtaining an initial wheel structure; based on the initial wheel structure, selecting a first integrated optimization strategy consisting of at least one method selected from a topology optimization method, a parametric shape optimization method, a material structure integration optimization method, etc., to perform a wheel optimization integrated conceptual design, and obtain a wheel model that meets the requirements of a comprehensive weight and reliability evaluation; based on the wheel model that meets the requirements of a comprehensive weight and reliability evaluation, selecting a second integrated optimization strategy consisting of at least one method selected from a free size optimization method, a free shape optimization method, etc., to perform a wheel optimization integrated detailed design; based on the detailed design, performing grid-based rapid load and boundary condition loading to perform wheel shape verification; and obtaining a final lightweight wheel structure.

[0003] The above scheme requires obtaining the initial wheel structure before optimizing the wheel. However, in the existing industrial design model, industrial designers often provide two-dimensional creative wheel design images, which makes it impossible to directly apply the above scheme to perform wheel design optimization evaluation, affecting the efficiency of wheel design evaluation, lengthening the wheel R&D design cycle, and hindering the promotion and use of wheel design evaluation schemes.

[0004] The information disclosed in this Background Art is only for understanding the background of the inventive concept and therefore it may include information that does not constitute the prior art. Summary of the invention

[0005] In response to the above problem or one of the above problems, an object of the present invention is to provide a wheel design evaluation method and a wheel design evaluation system based on model simulation. By creating a wheel design acquisition model, a simulated wheel generation model, a numerical simulation model, and a wheel evaluation model, the wheel design can be evaluated directly based on the two-dimensional wheel design image, thereby effectively improving the efficiency of wheel design evaluation, shortening the wheel research and development design cycle, and facilitating the promotion and use of wheel design evaluation solutions.

[0006] In response to the above problem or one of the above problems, the second object of the present invention is to provide a wheel design evaluation method and a wheel design evaluation system based on model simulation, which can timely evaluate the engineering performance of a two-dimensional wheel design, so as to directly screen out the optimal conceptual design scheme, and can quickly improve the wheel design scheme based on the evaluation results, thereby effectively improving the wheel design quality and reducing the wheel design research and development time.

[0007] To achieve one of the above purposes, the first technical solution of the present invention is: A wheel design evaluation method based on model simulation comprises the following steps: Step 1: Obtain a two-dimensional wheel design image and collect wheel design distribution features through a wheel design acquisition model created in advance; Step 2, using the previously created simulation wheel generation model, based on the edge detection algorithm, the two-dimensional wheel design image and the wheel design distribution characteristics are processed to generate a wheel simulation object; Step 3, using the numerical simulation model created in advance, numerically calculate the wheel simulation object to obtain mass data and natural frequency data of the wheel simulation object; Step 4: Based on the wheel evaluation model created in advance, the wheel stiffness information is calculated according to the natural frequency data and mass data, and the wheel design in the two-dimensional wheel design image is judged based on the stiffness information to complete the wheel design evaluation based on model simulation.

[0008] The present invention creates a wheel design acquisition model, a simulation wheel generation model, a numerical simulation model, and a wheel evaluation model, and can directly evaluate the wheel design based on a two-dimensional wheel design image, thereby effectively improving the efficiency of wheel design evaluation, shortening the wheel research and development design cycle, and facilitating the promotion and use of wheel design evaluation solutions.

[0009] Furthermore, by applying the method of the present invention, compared with the prior art, industrial designers can timely evaluate the engineering performance of the two-dimensional wheel design, so that they can directly screen out the optimal conceptual design scheme, and can quickly improve the wheel design scheme based on the evaluation results, thereby effectively improving the wheel design quality and reducing the wheel design research and development time.

[0010] As the preferred technical measures: Step 1: Using the wheel design acquisition model created in advance, one or more two-dimensional wheel design images are acquired, and the method for acquiring wheel design distribution features is as follows: Use deep learning algorithms to build a generative adversarial network for wheel design; Using a generative adversarial network, the wheel characteristic parameters are learned to generate one or more two-dimensional wheel design images; Based on the distribution characteristics of wheel materials, topology optimization is performed on the two-dimensional wheel design image to obtain a new two-dimensional wheel design image; Feature collection is performed on the new two-dimensional wheel design image to obtain the wheel design distribution characteristics.

[0011] As the preferred technical measures: Step 2: Using the previously created simulated wheel generation model, based on the edge detection algorithm, the two-dimensional wheel design image and the wheel design distribution characteristics are processed to generate a wheel simulation object as follows: Use edge detection algorithm to capture smooth and clear boundaries in the two-dimensional wheel design image and obtain wheel edge coordinate point data; Extract spoke and rim sections based on wheel design distribution features; According to the range distribution of the spoke cross section and the rim cross section, the wheel edge coordinate point data are sorted and grouped to obtain an edge coordinate array; The edge coordinate array includes the spoke cross-section coordinates and the rim cross-section coordinates; According to the spoke cross-section coordinates and using the spline curve algorithm, draw the cross-section of the spoke, and then rotate the cross-section of the spoke to create a spoke body; According to the rim cross-section coordinates and using a spline curve algorithm, the cross-section of the rim is drawn, and then the cross-section of the rim is rotated to create a rim body; The rim body and the spoke body are combined together to generate a wheel simulation object.

[0012] As the preferred technical measures: The method of using edge detection algorithm to capture smooth and clear boundaries in the two-dimensional wheel design image and obtain wheel edge coordinate point data is as follows: Firstly, the image processing algorithm is used to process the two-dimensional wheel design image, and the maximum and minimum values ​​of the horizontal and vertical coordinates of the outer edge of the wheel are obtained to obtain four extreme value coordinates; The four extreme coordinates include a first minimum value and a first maximum value of the abscissa, and a second minimum value and a second maximum value of the ordinate; Based on the four extreme value coordinates, a cropping area is determined in the two-dimensional wheel design image; the cropping area is a rectangle, the coordinates of the upper left corner are the first minimum value and the second minimum value, and the coordinates of the lower right corner are the first maximum value and the second maximum value; The two-dimensional wheel design image is cropped according to the cropping area, and only the image content within the cropping area is retained to obtain a wheel image without blank margins; Using edge detection algorithm, the boundary of wheel image without blank margin is identified according to the brightness change rate, and the gradient information of all pixels is obtained; According to the preset gradient threshold, the gradient information is judged and the pixel points constituting the boundary are screened out to capture the smooth and clear boundary in the two-dimensional wheel design image; The coordinates of the pixel points constituting the boundary are extracted as wheel edge coordinate point data.

[0013] As the preferred technical measures: Using the edge detection algorithm, the method for identifying the boundary of the wheel image without blank margins according to the brightness change rate is as follows: According to the edge detection algorithm, set up a binary image and marking mechanism; The marking mechanism is that edge pixels are marked as white or 1, and non-edge pixels are marked as black or 0; Based on the brightness change rate, two circular masks are set, which include a rim mask and a hub mask; The radius of the rim mask is smaller than the outer diameter of the rim, and the radius of the hub mask is larger than the inner diameter of the hub; According to the rim mask and the hub mask, the binary image is assigned a value, which specifically includes the following contents: Set the pixel values ​​outside the rim mask area to 0, i.e. black, to remove the edge of the rim; The pixel values ​​in the hub mask area are set to 0, i.e. white, to remove the edge of the hub, so that only the edge of the spoke area remains in the wheel image, thereby completing the boundary recognition of the wheel image.

[0014] As the preferred technical measures: The method of combining the rim body and the spoke body to generate a wheel simulation object is as follows: Create a reference section passing through the wheel center and parallel to the wheel axis; Using the reference section, cut the rim body and the spoke body to obtain the wheel body section; Extracting the cross-sectional curve of the wheel body cross section, obtaining the contour lines of the spoke and the rim, and converting the cross-sectional curve into coordinate data to obtain the cross-sectional coordinate data of the spoke and the cross-sectional coordinate data of the rim; Selecting the first and last coordinate points from the cross-sectional coordinate data of the spoke and the cross-sectional coordinate data of the rim, and connecting the first and last coordinate points end to end to form a closed contour line; Rotate the outline to form an original car body; Draw the spoke edge shape based on the cross-sectional coordinate data of the spoke; Create spoke-shaped objects based on the spoke edge shape; According to the spoke-shaped object, the original vehicle body is cut to obtain a vehicle body with a spoke shape, that is, a wheel simulation object.

[0015] As the preferred technical measures: Step 3: Use the numerical simulation model created in advance to perform numerical calculations on the wheel simulation object to obtain the mass data and natural frequency data of the wheel simulation object as follows: According to the wheel simulation object, set the material property data and mesh density; Material property data include wheel material characteristics, Young’s modulus, Poisson’s ratio, density, and shear modulus; According to the mesh density, the wheel simulation object is meshed to obtain a second-order tetrahedral mesh; Based on the material property data, the second-order tetrahedral mesh is processed to obtain quality data; Combined with the mass data, modal analysis is performed on the wheel simulation object to obtain the natural frequency of the lateral mode.

[0016] As the preferred technical measures: Step 4: Based on the wheel evaluation model created in advance, the wheel stiffness information is calculated according to the natural frequency data and mass data, and the wheel design in the two-dimensional wheel design image is judged according to the stiffness information as follows: According to the correlation between wheel stiffness and road noise and the preset stiffness requirement standard, stiffness constraint conditions are set; According to the natural frequency data and mass data, the stiffness information of several wheel simulation objects is calculated; According to the stiffness constraint conditions, each stiffness information is judged to obtain stiffness information that meets the stiffness requirements; The stiffness information that meets the stiffness requirements is matched with the two-dimensional wheel design image to obtain the two-dimensional wheel design image that meets the design requirements, thereby screening out one or more wheel designs that meet the requirements.

[0017] To achieve one of the above purposes, the second technical solution of the present invention is: A wheel design evaluation method based on model simulation includes the following contents: Obtain wheel design distribution characteristics; Based on the wheel design distribution characteristics, several two-dimensional wheel design images are generated to form diversified wheel designs; Using edge detection algorithms, several two-dimensional wheel design images are processed to generate several wheel simulation objects; Perform numerical simulation calculations on each wheel simulation object to obtain mass data and natural frequency data; According to the natural frequency data and mass data, several wheel stiffness information is calculated; One or more wheel designs meeting the stiffness requirement are screened out based on the stiffness information.

[0018] To achieve one of the above purposes, the third technical solution of the present invention is: A wheel design evaluation system based on model simulation, comprising: one or more processors; A storage device for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the above-mentioned wheel design evaluation method based on model simulation.

[0019] Compared with the prior art solutions, the present invention has the following beneficial effects: The present invention creates a wheel design acquisition model, a simulation wheel generation model, a numerical simulation model, and a wheel evaluation model, and can directly evaluate the wheel design based on a two-dimensional wheel design image, thereby effectively improving the efficiency of wheel design evaluation, shortening the wheel research and development design cycle, and facilitating the promotion and use of wheel design evaluation solutions.

[0020] Furthermore, by applying the method of the present invention, compared with the prior art, industrial designers can timely evaluate the engineering performance of the two-dimensional wheel design, so that they can directly screen out the optimal conceptual design scheme, and can quickly improve the wheel design scheme based on the evaluation results, thereby effectively improving the wheel design quality and reducing the wheel design research and development time. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 A schematic diagram of a process flow of the wheel design evaluation method of the present invention; Figure 2 A two-dimensional wheel design image in an embodiment of the present invention; Figure 3 Images are designed for seven two-dimensional wheels in an embodiment of the present invention. DETAILED DESCRIPTION

[0022] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0023] On the contrary, the present invention covers any substitution, modification, equivalent method and scheme made on the essence and scope of the present invention as defined by the claims. Further, in order to make the public have a better understanding of the present invention, some specific details are described in detail in the following detailed description of the present invention. Those skilled in the art can fully understand the present invention without the description of these details.

[0024] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which the present invention belongs. The terms used herein are only for the purpose of describing specific embodiments and are not intended to limit the present invention.

[0025] like Figure 1 As shown, the first specific embodiment of the wheel design evaluation method based on model simulation of the present invention: A wheel design evaluation method based on model simulation comprises the following steps: Step 1: Obtain a two-dimensional wheel design image and collect wheel design distribution features through a wheel design acquisition model created in advance; Step 2, using the previously created simulation wheel generation model, based on the edge detection algorithm, the two-dimensional wheel design image and the wheel design distribution characteristics are processed to generate a wheel simulation object; Step 3, using the numerical simulation model created in advance, numerically calculate the wheel simulation object to obtain mass data and natural frequency data of the wheel simulation object; Step 4: Based on the wheel evaluation model created in advance, the wheel stiffness information is calculated according to the natural frequency data and mass data, and the wheel design in the two-dimensional wheel design image is judged based on the stiffness information to complete the wheel design evaluation based on model simulation.

[0026] A second specific embodiment of the wheel design evaluation method based on model simulation of the present invention: A wheel design evaluation method based on model simulation includes the following contents: Obtain wheel design distribution characteristics; Based on the wheel design distribution characteristics, several two-dimensional wheel design images are generated to form diversified wheel designs; Using edge detection algorithms, several two-dimensional wheel design images are processed to generate several wheel simulation objects; Perform numerical simulation calculations on each wheel simulation object to obtain mass data and natural frequency data; According to the natural frequency data and mass data, several wheel stiffness information is calculated; One or more wheel designs meeting the stiffness requirement are screened out based on the stiffness information.

[0027] A third specific embodiment of the wheel design evaluation method based on model simulation of the present invention: A wheel design evaluation method based on model simulation combines deep learning algorithms, computer-aided design (CAD), computer-aided engineering (CAE) and other technologies. The model training process is divided into the following four stages: In the first stage, various 2D wheel designs are generated using a deep generative model to produce diverse 2D wheel design images.

[0028] In the second stage, a simulation wheel generation model is constructed, which generates a wheel simulation object based on the two-dimensional wheel design image.

[0029] In the third stage, a numerical simulation model is constructed, and the wheel simulation object generated in the second stage is used to perform numerical simulation calculations to obtain the mass of the wheel simulation object and collect the natural frequency of the wheel.

[0030] In the fourth stage, the two-dimensional wheel design and its mass and natural frequency are used as labels to train a convolutional neural network (CNN). The trained convolutional neural network can take the two-dimensional wheel design image as input and output the predicted natural frequency and mass, thereby evaluating the engineering performance of the conceptual wheel design image.

[0031] In this embodiment, in the first stage, the method for generating a two-dimensional design is as follows: In order to generate various two-dimensional wheel designs, the present invention adopts a deep generative design method, which combines generative adversarial networks with topological optimization. The optimization is used to optimize the material distribution to ensure the optimal design in terms of engineering performance, while the generative adversarial network generates new designs with more diversity and aesthetic characteristics by learning the distribution of existing designs. The core of the combination of the two is that the generated design is first generated by the generative adversarial network, and then its engineering performance is further optimized by topological optimization. This iterative process ensures that the generated design is both innovative and can meet actual engineering needs.

[0032] The generative adversarial network includes a generator and a discriminator. The generator generates wheel design data according to wheel distribution characteristics. The discriminator couples the wheel design data with existing wheel design data to obtain several wheel design images.

[0033] In this embodiment, in the second stage, the method for generating a wheel simulation object is as follows: In the first step, an edge detection algorithm is used to capture the smooth and clear boundaries in the two-dimensional wheel design image and obtain the spoke edge coordinate points; The second step is to sort and group the spoke edge coordinate points obtained in the first step and form a spline curve of the wheel; The third step is to realize the automatic generation of wheel simulation object according to the spline curve of the wheel.

[0034] In this example, the method for obtaining the spoke edge coordinate points is as follows: Since the wheel image generated in the first stage has blank margins at the edge, in order to obtain a more compact image that only contains the wheel, it is necessary to remove these blank margins. First, the outer edge of the wheel is obtained by image processing technology. Coordinates and The maximum and minimum values ​​of the coordinates.

[0035] Next, a cropping region is defined in the original image using the obtained minimum value Xmin and maximum value Xmax of the x-coordinate, and the minimum value Ymin and maximum value Ymax of the y-coordinate. This cropping region is a rectangle whose boundaries are determined by these four coordinate values, that is, the coordinates of the upper left corner of the rectangle are (Xmin, Ymin), and the coordinates of the lower right corner are (Xmax, Ymax).

[0036] Finally, the original image is cropped to retain only the image content in the cropped area, and the image is scaled to This will give you a wheel image without any blank margins, which only contains the wheel itself. The overall process is as follows: In the two-dimensional wheel design image, the pixels with a large rate of change in brightness are the edges of the image. Their rate of change in brightness is mathematically defined as the gradient. In order to identify these boundaries, the gradient of each pixel is first calculated. The gradient calculation depends on the rate of change of brightness of each pixel in the image in the horizontal and vertical directions. It can be approximated by calculating the first-order derivatives in these two directions. The gradient The calculation formula is as follows:

[0037]

[0038]

[0039] in, The brightness change rate of the pixels in the wheel design image in the horizontal direction, Design the brightness change rate of the pixels in the wheel image in the vertical direction. Design any pixel in the image for the wheel The gray value at and Similarly, and They represent the horizontal and vertical coordinates of the pixel point in the wheel design image respectively.

[0040] After obtaining the gradient information of the image, the preset gradient threshold can be used to determine which pixels constitute the boundary. The choice of threshold determines the sensitivity and accuracy of edge detection, and varies depending on factors such as the specific content of the image, lighting conditions, noise level, etc. In this application, the gradient threshold is set to 100.

[0041] In this embodiment, an edge detection algorithm is used to extract the edge of the wheel image to generate a binary image, in which edge pixels are marked as white or 1, and non-edge pixels are marked as black or 0. In order to remove the edges of the rim and hub, two circular masks are defined: one for the rim and the other for the hub. The radius of the rim mask should be slightly smaller than the outer diameter of the rim, and the radius of the hub mask should be slightly larger than the inner diameter of the hub.

[0042] Apply these two circular masks to the binary image after edge detection. For the rim mask, set the pixel values ​​outside the mask area to 0 (i.e. black) to remove the edge of the rim, see Figure 2 For the wheel hub mask, set the pixel values ​​within the mask area to 0 to remove the edges of the wheel hub.

[0043] After the above processing, only the edge of the spoke area remains in the image. Finally, the processed binary image is traversed to extract the coordinates of all white pixels (i.e., spoke edges), and these coordinates are saved in a plain text file to obtain the spoke edge coordinate points. Each coordinate point can be expressed as form.

[0044] In this embodiment, the edge detection algorithm is an edge detection method with a Sobel operator, which detects the edges in the wheel design image by calculating the gradient values ​​of the wheel design image in the horizontal and vertical directions.

[0045] In this embodiment, the spoke edge coordinate points obtained in the first step are sorted and grouped, and the algorithm flow is as follows: Step 21, store all coordinate points in array A; Take the first coordinate (X0, Y0) from array A as the initial value and delete the point from array A; Create an array B to store the grouped points, and store the initial value coordinates (X0, Y0) in the i-th group of array B, where i is zero.

[0046] Step 22, perform iterative loop calculation until array A is empty, the process is as follows: Declare the initial value as a fixed point; Traverse each point in array A and use the Euclidean distance formula to calculate the distance between the current point and the fixed point. For two points on the two-dimensional plane and , the distance between them It can be calculated by the following formula:

[0047] Find the point closest to the fixed point, declare it as the new initial value, and delete it from array A; Compute the distance between the fixed point and the initial value (i.e. the closest point): If the distance is less than or equal to the threshold (here set to 5 pixels), the initial value is stored in the i-th group of array B.

[0048] Otherwise, create a new group (group i+1) and store the initial value in group i+1.

[0049] Since each time the new initial value is removed and assigned to the corresponding group, no point will belong to multiple groups at the same time. Through the above steps, the spoke edge coordinate points can be effectively sorted and grouped.

[0050] In this example, the third step is to realize the generation of a specific wheel simulation object, and its algorithm flow is as follows: Step 31, selecting an 18-inch wheel and obtaining a wheel simulation object of the wheel.

[0051] Step 32, creating a cross-sectional plane intersecting with the wheel simulation object, wherein the cross-sectional plane passes through the center of the wheel and is parallel to the axial direction of the wheel.

[0052] In step 33, the section plane is intersected with the wheel simulation object, thereby obtaining the section geometry of the spoke and the rim.

[0053] Step 34, obtaining the cross-sectional curve (the contour line of the spoke and the rim), and converting the cross-sectional curve into a coordinate point form to obtain cross-sectional information.

[0054] Perform 3D CAD modeling based on the spoke shape and cross-section information of the disk view; Step 35, loading the cross-sectional coordinate data of the spoke and rim, and drawing the cross-sectional view, which includes the following contents: Using the spline algorithm, the cross-sectional coordinate points of the spokes and rims are connected end to end to form a closed contour line.

[0055] Rotate the outline and cross section 180 degrees to form a primitive car body.

[0056] Step 36, loading the coordinate data of the spoke edge and drawing the spoke shape, which includes the following contents: Use the spline algorithm to connect the points of each group end to end to draw the spoke edge shape; Then the spoke edge shape is extruded to create a spoke-shaped object. The extrusion direction is set to the direction of the vehicle body. The spoke-shaped object is designated as the tool body, and the spoke body is designated as the target body. Subsequently, the original vehicle body is cut to remove the intersecting part of the two objects to obtain a vehicle body with a spoke shape, that is, a wheel simulation object.

[0057] In this example, in the third stage, the method for performing numerical simulation calculation using the wheel simulation object generated in the second stage is as follows: Step 301, using a meshing algorithm, processes the wheel simulation object to generate a finite element mesh, i.e., a second-order tetrahedral mesh. The maximum size of the mesh is set to 6 mm, and the material properties are set as follows: the wheel is made of aluminum, with a Young's modulus of 73500, a Poisson's ratio of 0.33, and a density of , the shear modulus is 0.001.

[0058] Step 302, calculate the mass of the wheel simulation object based on the second-order tetrahedral mesh .

[0059] Step 303: Perform modal analysis using a free modal algorithm to obtain the results of the modal analysis and the natural frequency of the lateral mode. .

[0060] Step 304, the stiffness of the wheel refers to the ability of the wheel to resist deformation when subjected to external force. Since the natural frequency is proportional to the stiffness and inversely proportional to the mass, the mass and natural frequency obtained in the above steps can be used to obtain the stiffness of the wheel using the following formula:

[0061] in, is the natural frequency, is the stiffness to be solved, It's quality.

[0062] Step 305, the shape of the spoke has a direct and significant impact on the overall engineering performance of the wheel, and this relationship can be quantitatively evaluated through stiffness analysis. The stiffness of the wheel is a key indicator to measure its ability to resist deformation due to external forces, and is directly related to the driving stability and safety of the vehicle. Specifically, if the stiffness of the wheel is too high, excessive vibration may occur during driving, affecting the riding comfort of passengers; on the contrary, if the stiffness is insufficient, the wheel is prone to excessive deformation, which in turn threatens driving safety. Therefore, the design of the spoke shape is crucial to ensure that the wheel has appropriate stiffness.

[0063] At the same time, the stiffness standards of different manufacturers involve multiple factors and are constantly changing, so the specific threshold of stiffness should be selected based on actual conditions.

[0064] In this embodiment, in the fourth stage, the two-dimensional wheel design and its mass and natural frequency are used as labels, and the method for training the convolutional neural network is as follows: A convolutional neural network is constructed, and the two-dimensional wheel design image, natural frequency, and mass are stored in pairs as training data for deep learning.

[0065] Each wheel design image generated in the first stage was rotated 72 degrees, 144 degrees, 216 degrees, 288 degrees and flipped left and right for 10 enhancements. The enhanced samples were used for training to avoid overfitting. At this time, the natural frequency and mass remain unchanged because rotation and left and right flipping do not affect the modal analysis and mass results.

[0066] The convolutional neural network model of this embodiment can quickly predict the natural frequency and mass of the wheel based on the two-dimensional design drawing, and then calculate the wheel stiffness through the relevant equations. During the design process, the manufacturer can set the lower limit of the stiffness of the mode as the design constraint condition based on the correlation between the stiffness and road noise in each mode, and use a computer equipped with a high-performance accelerator to perform evaluation and calculation. According to experiments, it takes only 0.66 seconds on average to evaluate and predict a wheel design image, which significantly shortens the iteration cycle in traditional industrial design.

[0067] Therefore, the evaluation method of the present invention enables automobile manufacturers to quickly evaluate the stiffness of the wheel and quickly screen and eliminate designs that do not meet the requirements according to the manufacturer's own stiffness standards. At the same time, a candidate solution can be efficiently selected for the subsequent detailed design stage, thereby ensuring that the design process is both efficient and meets the manufacturer's requirements.

[0068] Furthermore, the present invention can automatically generate a large amount of three-dimensional wheel design data and timely evaluate its engineering performance. In the conceptual design stage, industrial designers and engineers can use the engineering performance results of the present invention to obtain a large number of three-dimensional wheel simulation objects and discuss candidate conceptual designs suitable for the detailed design stage. The deep learning model proposed in the present invention can predict the CAE results of two-dimensional wheel design, so that industrial designers can understand the engineering performance of two-dimensional concept sketches in a timely manner.

[0069] A specific embodiment of applying the present invention to evaluate 7 wheel design drawings: The wheel design evaluation method based on model simulation of the present invention is applied to evaluate 7 wheel design images, which includes the following contents: First, select 7 wheel design images, see Figure 3 , and then conduct simulation evaluation, and the evaluation results are as follows: The true values ​​of mass are: 9.96; 10.71; 11.80; 10.83; 10.40; 10.05; 11.40; The true values ​​of the frequencies are: 814.2; 846.4; 849.6; 803.2; 795.4; 779.1; 801.6.

[0070] At the same time, the engineering performance of the above 7 wheel design images was manually evaluated by humans. The manual evaluation results are as follows: The quality predictions are: 9.90; 10.32; 11.43; 10.78; 10.20; 9.83; 11.09; The frequency predictions are: 806.5; 837.5; 824.0; 794.8; 785.2; 782.2; 796.0.

[0071] Therefore, it can be seen that the quality prediction error of the present invention is within 3%, and the frequency prediction error is within 1.5%, indicating that the scheme of the present invention is feasible and has a high accuracy.

[0072] Furthermore, through the above-mentioned mass and frequency evaluation, the stiffness of the wheel can be calculated. Therefore, in the design process, the application of the present invention can accurately and timely feedback the engineering performance of the two-dimensional concept sketch of the wheel. At the same time, the manufacturer can set the lower limit of the stiffness of the mode as a design constraint condition to quickly screen out wheels that do not meet the requirements.

[0073] An embodiment of a device applying the method of the present invention: An electronic device comprising: one or more processors; A storage device for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the above-mentioned wheel design evaluation method based on model simulation.

[0074] A computer medium embodiment using the method of the present invention: A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the above-mentioned wheel design evaluation method based on model simulation.

[0075] Those skilled in the art should understand that the embodiments of the present application can be provided as methods, systems, and computer program products. Therefore, the present application can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, optical storage, etc.) containing computer-usable program codes.

[0076] The present application is described by flowcharts or / and block diagrams of the methods, devices (systems), and computer program products of the embodiments of the present application. It should be understood that each process or / and block in the flowchart or / and block diagram and the combination of the processes or / and blocks in the flowchart or / and block diagram can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart or / and block diagram. Figure 1 Process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0077] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 Process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0078] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 Process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0079] The model in this application is an object that objectively describes the morphological structure with the help of physical or virtual representation. The object is not equal to the object and is not limited to physical and virtual. It can be a data processing function, software program, processing mode, usage method, operation method, workflow, application process, electronic hardware, circuit module, processing system, system imitation or simulation object.

[0080] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field can still modify or replace the specific implementation methods of the present invention with equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be included in the scope of protection of the claims of the present invention.

Claims

1. A wheel design evaluation method based on model simulation, characterized in that: The following steps are involved: Step 1: Obtain a two-dimensional wheel design image and collect wheel design distribution features through a wheel design acquisition model created in advance; Step 2, using the previously created simulation wheel generation model, based on the edge detection algorithm, the two-dimensional wheel design image and the wheel design distribution characteristics are processed to generate a wheel simulation object; Step 3, using the numerical simulation model created in advance, numerically calculate the wheel simulation object to obtain mass data and natural frequency data of the wheel simulation object; Step 4: Based on the wheel evaluation model created in advance, the wheel stiffness information is calculated according to the natural frequency data and mass data, and the wheel design in the two-dimensional wheel design image is judged based on the stiffness information to complete the wheel design evaluation based on model simulation.

2. A wheel design evaluation method based on model simulation as claimed in claim 1, characterized in that: Step 1: Using the wheel design acquisition model created in advance, one or more two-dimensional wheel design images are acquired, and the method for acquiring wheel design distribution features is as follows: Use deep learning algorithms to build a generative adversarial network for wheel design; Using a generative adversarial network, the wheel characteristic parameters are learned to generate one or more two-dimensional wheel design images; Based on the distribution characteristics of wheel materials, topology optimization is performed on the two-dimensional wheel design image to obtain a new two-dimensional wheel design image; Feature collection is performed on the new two-dimensional wheel design image to obtain the wheel design distribution characteristics.

3. A wheel design evaluation method based on model simulation as claimed in claim 1, characterized in that: Step 2: Using the previously created simulated wheel generation model, based on the edge detection algorithm, the two-dimensional wheel design image and the wheel design distribution characteristics are processed to generate a wheel simulation object as follows: Use edge detection algorithm to capture smooth and clear boundaries in the two-dimensional wheel design image and obtain wheel edge coordinate point data; Extract spoke and rim sections based on wheel design distribution features; According to the range distribution of the spoke cross section and the rim cross section, the wheel edge coordinate point data are sorted and grouped to obtain an edge coordinate array; The edge coordinate array includes the spoke cross-section coordinates and the rim cross-section coordinates; According to the spoke cross-section coordinates and using the spline curve algorithm, draw the cross-section of the spoke, and then rotate the cross-section of the spoke to create a spoke body; According to the rim cross-section coordinates and using a spline curve algorithm, the cross-section of the rim is drawn, and then the cross-section of the rim is rotated to create a rim body; The rim body and the spoke body are combined together to generate a wheel simulation object.

4. A wheel design evaluation method based on model simulation as claimed in claim 3, characterized in that: The method of using edge detection algorithm to capture smooth and clear boundaries in the two-dimensional wheel design image and obtain the wheel edge coordinate point data is as follows: Firstly, the image processing algorithm is used to process the two-dimensional wheel design image, and the maximum and minimum values ​​of the horizontal and vertical coordinates of the outer edge of the wheel are obtained to obtain four extreme value coordinates; The four extreme coordinates include a first minimum value and a first maximum value of the abscissa, and a second minimum value and a second maximum value of the ordinate; Based on the four extreme value coordinates, a cropping area is determined in the two-dimensional wheel design image; the cropping area is a rectangle, the coordinates of the upper left corner are the first minimum value and the second minimum value, and the coordinates of the lower right corner are the first maximum value and the second maximum value; The two-dimensional wheel design image is cropped according to the cropping area, and only the image content within the cropping area is retained to obtain a wheel image without blank margins; Using edge detection algorithm, the boundary of wheel image without blank margin is identified according to the brightness change rate, and the gradient information of all pixels is obtained; According to the preset gradient threshold, the gradient information is judged and the pixel points constituting the boundary are screened out to capture the smooth and clear boundary in the two-dimensional wheel design image; The coordinates of the pixel points constituting the boundary are extracted as wheel edge coordinate point data.

5. A wheel design evaluation method based on model simulation as claimed in claim 4, characterized in that: Using the edge detection algorithm, the method for identifying the boundary of the wheel image without blank margins according to the brightness change rate is as follows: According to the edge detection algorithm, set up a binary image and marking mechanism; The marking mechanism is that edge pixels are marked as white or 1, and non-edge pixels are marked as black or 0; Based on the brightness change rate, two circular masks are set, which include a rim mask and a hub mask; The radius of the rim mask is smaller than the outer diameter of the rim, and the radius of the hub mask is larger than the inner diameter of the hub; According to the rim mask and the hub mask, the binary image is assigned a value, which specifically includes the following contents: Set the pixel values ​​outside the rim mask area to 0, i.e. black, to remove the edge of the rim; The pixel values ​​in the hub mask area are set to 0, i.e. white, to remove the edge of the hub, so that only the edge of the spoke area remains in the wheel image, thereby completing the boundary recognition of the wheel image.

6. A wheel design evaluation method based on model simulation as claimed in claim 3, characterized in that: The method of combining the rim body and the spoke body to generate a wheel simulation object is as follows: Create a reference section passing through the wheel center and parallel to the wheel axis; Using the reference section, cut the rim body and the spoke body to obtain the wheel body section; Extracting the cross-sectional curve of the wheel body cross section, obtaining the contour lines of the spoke and the rim, and converting the cross-sectional curve into coordinate data to obtain the cross-sectional coordinate data of the spoke and the cross-sectional coordinate data of the rim; Selecting the first and last coordinate points from the cross-sectional coordinate data of the spoke and the cross-sectional coordinate data of the rim, and connecting the first and last coordinate points end to end to form a closed contour line; Rotate the outline to form an original car body; Draw the spoke edge shape based on the cross-sectional coordinate data of the spoke; Create spoke-shaped objects based on the spoke edge shape; According to the spoke-shaped object, the original vehicle body is cut to obtain a vehicle body with a spoke shape, that is, a wheel simulation object.

7. A wheel design evaluation method based on model simulation as claimed in claim 1, characterized in that: Step 3: Use the numerical simulation model created in advance to perform numerical calculations on the wheel simulation object to obtain the mass data and natural frequency data of the wheel simulation object as follows: According to the wheel simulation object, set the material property data and mesh density; Material property data include wheel material characteristics, Young’s modulus, Poisson’s ratio, density, and shear modulus; According to the mesh density, the wheel simulation object is meshed to obtain a second-order tetrahedral mesh; Based on the material property data, the second-order tetrahedral mesh is processed to obtain quality data; Combined with the mass data, modal analysis is performed on the wheel simulation object to obtain the natural frequency of the lateral mode.

8. A wheel design evaluation method based on model simulation as claimed in claim 1, characterized in that: Step 4: Based on the wheel evaluation model created in advance, the wheel stiffness information is calculated according to the natural frequency data and mass data, and the wheel design in the two-dimensional wheel design image is judged according to the stiffness information as follows: According to the correlation between wheel stiffness and road noise and the preset stiffness requirement standard, stiffness constraint conditions are set; According to the natural frequency data and mass data, the stiffness information of several wheel simulation objects is calculated; According to the stiffness constraint conditions, each stiffness information is judged to obtain stiffness information that meets the stiffness requirements; The stiffness information that meets the stiffness requirements is matched with the two-dimensional wheel design image to obtain the two-dimensional wheel design image that meets the design requirements, thereby screening out one or more wheel designs that meet the requirements.

9. A wheel design evaluation method based on model simulation, characterized in that: Includes the following: Obtain wheel design distribution characteristics; Based on the wheel design distribution characteristics, several two-dimensional wheel design images are generated to form diversified wheel designs; Using edge detection algorithms, several two-dimensional wheel design images are processed to generate several wheel simulation objects; Perform numerical simulation calculations on each wheel simulation object to obtain mass data and natural frequency data; According to the natural frequency data and mass data, several wheel stiffness information is calculated; One or more wheel designs meeting the stiffness requirement are screened out based on the stiffness information.

10. A wheel design evaluation system based on model simulation, characterized in that: It includes: one or more processors; A storage device for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement a wheel design evaluation method based on model simulation as described in any one of claims 1-9.

Images