A Visualization Method for Digital Twin Models of Adapters

Abstract

The present invention proposes a method for visualizing a digital twin model of an adapter, which belongs to the field of visualization, and includes: S1, setting the target imaging size of the adapter model on the screen; S2, calculating the ratio of the target imaging size to the actual physical width size of the adapter model to be visualized, obtaining a scaling factor, scaling the adapter model to be visualized in proportion according to the scaling factor, obtaining the scaled adapter model to be visualized, and measuring the surface area of ​​the scaled adapter model to be visualized; S3, constructing a discrete point cloud grid of the adapter model to be visualized, and generating a digital twin model of the adapter model to be visualized according to the discrete point cloud grid; S4, writing a shader of the digital twin model of the adapter model to be visualized to render the digital twin model, and visually displaying the rendered digital twin model. This application uses point cloud grids to capture the details of physical objects, fuses the digital twin model with actual physical information, and improves the visualization effect through shader rendering, thereby improving the visual expression of the digital twin model.

Description

Technical Field

[0001] The present invention relates to the field of visualization technology, and in particular to a visualization method for a digital twin model of an adapter. Background Art

[0002] As a key component of the missile launcher, the adapter's compressive stress-strain characteristics, creep resistance and other mechanical properties have an important impact on the missile launch performance. Therefore, the study of mechanical properties is an indispensable part of the design and optimization of the adapter. The adapter digital twin model can realize multi-scale, multi-physics field coupling virtual modeling and simulation analysis. For example, combined with the adapter material parameters and load conditions, its stress, strain and other mechanical responses during the launch process can be accurately predicted.

[0003] However, the digital twin data of the adapter contains high-accuracy global strain performance data of the adapter. The amount of this data is large, including a large amount of mechanical response data such as strain and stress. If appropriate methods are not used to process it, it is difficult to understand its internal laws, and then it is difficult to know the performance status of the adapter, and it is difficult to intuitively reflect the overall physical state of the adapter, making it impossible to make judgments and decisions.

[0004] Therefore, finding a method that can not only enable the adapter data twin model to carry physical state information to quickly perceive the state of the adapter, but also process the adapter strain performance data is a technical problem that needs to be urgently solved by technical personnel in this field. Summary of the invention

[0005] In view of this, the present invention proposes a digital twin model visualization method for an adapter, which can reasonably adjust the display size of the adapter model on the screen according to actual scene requirements, integrate the digital twin model with actual physical information, and improve the visualization expression of the digital twin model.

[0006] The technical solution of the present invention is implemented as follows: The present invention provides a method for visualizing a digital twin model of an adapter, comprising the following steps:

[0007] S1. Setting the target imaging size of the adapter model on the screen;

[0008] S2, calculating the ratio of the target imaging size to the actual physical width size of the adapter model to be visualized, obtaining a scaling factor, scaling the adapter model to be visualized in equal proportion according to the scaling factor, obtaining the scaled adapter model to be visualized, and measuring the surface area of ​​the scaled adapter model to be visualized;

[0009] S3, constructing a discrete point cloud mesh of the adapter model to be visualized according to the scaled adapter model to be visualized and the surface area of ​​the adapter model to be visualized, and generating a digital twin model of the adapter model to be visualized according to the discrete point cloud mesh;

[0010] S4. Write a shader for the digital twin model of the adapter model to be visualized to render the digital twin model, and visualize the rendered digital twin model.

[0011] On the basis of the above technical solution, preferably, step S3 specifically includes:

[0012] S31, setting vertex density according to the number of vertices and the surface area of ​​the adapter model to be visualized;

[0013] S32, constructing discrete point cloud data of the scaled vertices of the adapter model to be visualized according to the vertex density;

[0014] S33, connecting adjacent vertices corresponding to the discrete point cloud data to form a triangular mesh, and setting a normal of the triangular mesh;

[0015] S34. Import the triangular mesh into the three-dimensional scene to obtain a digital twin model of the adapter model to be visualized.

[0016] Based on the above technical solution, preferably, step S32 specifically includes:

[0017] Constructing scaled three-dimensional point cloud data of the adapter model to be visualized according to vertex density;

[0018] The distribution of vertices in the three-dimensional point cloud data is adjusted, irregularly distributed vertices and occluded vertices are identified and removed, and discrete point cloud data of the vertices of the adapter model to be visualized are obtained.

[0019] On the basis of the above technical solution, preferably, step S4 specifically includes:

[0020] S41, extracting initial vertex coordinates in the discrete point cloud grid, and extracting initial strain performance data of the digital twin model according to the initial vertex coordinates;

[0021] S42, acquiring strain performance data of different load steps according to the initial strain performance data, and processing the strain performance data to form a load-strain performance data sequence;

[0022] S43, specifying strain information for vertices of the digital twin model according to the load-strain performance data sequence;

[0023] S44, creating material resources, binding the strain information with the digital twin model through the material resources, and obtaining the digital twin model with physical information attached;

[0024] S45. Write a shader for the digital twin model to which the physical information is attached, obtain the rendered digital twin model, and perform a visual display.

[0025] On the basis of the above technical solution, preferably, the processing of the strain performance data specifically includes:

[0026] The strain performance data of different load steps are input into a data fusion network for data fusion, wherein the data fusion network includes a low-fidelity sub-network, a linear sub-network and a nonlinear sub-network, the low-fidelity sub-network maps the strain performance data of different load steps, the linear sub-network learns the linear relationship between the strain performance data of different load steps, and the nonlinear sub-network learns the nonlinear relationship between the strain performance data of different load steps.

[0027] On the basis of the above technical solution, preferably, the structure of the data fusion network is:

[0028] The low-fidelity sub-network includes four first hidden layers of 128 neurons and three first non-linear activation functions, wherein the first non-linear activation function is located between two first hidden layers;

[0029] The linear sub-network includes three second hidden layers of 128 neurons;

[0030] The nonlinear sub-network includes four third hidden layers with 128 neurons and three second nonlinear activation functions, wherein the second nonlinear activation function is located between two third hidden layers.

[0031] On the basis of the above technical solution, preferably, the data fusion process specifically includes:

[0032] The strain performance data of different load steps are input into the low-fidelity sub-network, and low-fidelity output data is obtained after mapping;

[0033] The low-fidelity output data and the input of the low-fidelity sub-network are formed into a 4-dimensional vector, and the 4-dimensional vector is input into the linear sub-network and the nonlinear sub-network in parallel;

[0034] The outputs of the linear sub-network and the nonlinear sub-network are weighted summed to obtain the fused strain performance data.

[0035] On the basis of the above technical solution, preferably, step S43 specifically includes:

[0036] Set a color coding mapping relationship according to the value range of the strain performance data, and convert the strain performance data into corresponding color values;

[0037] Color values ​​are assigned to triangle mesh vertices in order according to the load step, so that each vertex has strain information.

[0038] Further preferably, the shader for the digital twin model with physical information attached includes a vertex shader for the digital twin model with physical information attached and a fragment shader for the digital twin model with physical information attached, the vertex shader transforms the vertex data of the digital twin model with physical information attached, and the fragment shader interpolates and calculates the pixel information to obtain the color of each pixel, wherein the vertex data includes the position coordinates, strain information and normal vector of the vertex.

[0039] Based on the above technical solution, preferably, the transformation processing includes geometric transformation, shading calculation and lighting calculation.

[0040] The digital twin model visualization method of the present invention has the following beneficial effects compared with the prior art:

[0041] (1) By reasonably adjusting the display size of the adapter model on the screen according to the actual scene requirements, using the point cloud grid to capture the details of the physical object, integrating the digital twin model with the actual physical information, and improving the visualization effect through shader rendering, the visual expression of the digital twin model is improved;

[0042] (2) The data fusion network is used to effectively fuse and integrate the strain data at different load steps, fully exploit the linear and nonlinear characteristics of the strain data, generate a more complete and accurate strain performance data sequence, better capture the complex relationship between the strain performance data at different load steps, and enhance the modeling and prediction capabilities of the strain performance data under different load conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0044] Figure 1 is a flow chart of the digital twin model visualization method of the present invention;

[0045] Figure 2 It is a diagram of adapter imaging change process of the digital twin model visualization method of the present invention;

[0046] Figure 3 A digital twin model rendering flow chart of the digital twin model visualization method of the present invention;

[0047] Figure 4 A rendering pipeline block diagram of the digital twin model visualization method of the present invention;

[0048] Figure 5 A structural diagram of a data fusion network of a digital twin model visualization method of the present invention;

[0049] Figure 6 A process diagram for specifying strain information of the digital twin model visualization method of the present invention;

[0050] Figure 7 A process diagram of converting strain performance data into RGB values ​​in the digital twin model visualization method of the present invention;

[0051] Figure 8 A process diagram for rendering and coloring strain information of the digital twin model visualization method of the present invention. DETAILED DESCRIPTION

[0052] The following will be combined with the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only 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.

[0053] like Figure 1 As shown, the present invention provides a method for visualizing a digital twin model of an adapter, comprising the following steps:

[0054] S1. Set the target imaging size of the adapter model on the screen.

[0055] In the embodiment of the present application, the target imaging size can be set based on the distribution of objects in the virtual three-dimensional scene, and the target imaging size of the adapter on the screen can also be set according to the observation requirements during information interaction.

[0056] S2. Calculate the ratio of the target imaging size to the actual physical width size of the adapter model to be visualized to obtain a scaling factor, scale the adapter model to be visualized in proportion according to the scaling factor to obtain the scaled adapter model to be visualized, and measure the surface area of ​​the scaled adapter model to be visualized.

[0057] It can be understood that the adapter model to be visualized is scaled proportionally according to the scaling factor so that the actual display size of the adapter model to be visualized matches the expected imaging size.

[0058] In one embodiment of the present application, the surface area of ​​the scaled adapter model to be visualized is measured in 3DsMax software.

[0059] S3. Construct a discrete point cloud mesh of the adapter model to be visualized according to the scaled adapter model to be visualized and the surface area of ​​the adapter model to be visualized, and generate a digital twin model of the adapter model to be visualized according to the discrete point cloud mesh.

[0060] like Figure 2 As shown, specifically, step S3 specifically includes:

[0061] S31, setting vertex density according to the number of vertices and the surface area of ​​the adapter model to be visualized;

[0062] S32, constructing discrete point cloud data of the scaled vertices of the adapter model to be visualized according to the vertex density;

[0063] S33, connecting adjacent vertices corresponding to the discrete point cloud data to form a triangular mesh, and setting a normal of the triangular mesh;

[0064] S34. Import the triangular mesh into the three-dimensional scene to obtain a digital twin model of the adapter model to be visualized.

[0065] The vertex density is n / s=1, where n represents the number of vertices and s represents the surface area of ​​the adapter model to be visualized after scaling. That is, on the surface of the adapter model to be visualized after scaling, each unit area will be filled with a vertex, and the vertex distribution density is 1 / unit area.

[0066] By setting the appropriate vertex density, the degree of detail of the point cloud can be controlled, thereby balancing the geometric accuracy and rendering efficiency of the digital twin model of the adapter model to be visualized. The point cloud data is discretized using a triangular mesh, and the rendering effect is enhanced by calculating the normal vectors of the mesh vertices. The constructed triangular mesh is imported into the three-dimensional scene to form a digital twin model consistent with the actual adapter model to be visualized, which not only ensures the geometric accuracy of the digital twin model, but also optimizes the rendering efficiency.

[0067] Further, step S32 specifically includes:

[0068] Constructing scaled three-dimensional point cloud data of the adapter model to be visualized according to vertex density;

[0069] The distribution of vertices in the three-dimensional point cloud data is adjusted, irregularly distributed vertices and occluded vertices are identified and removed, and discrete point cloud data of the vertices of the adapter model to be visualized are obtained.

[0070] It can be understood that by constructing the scaled three-dimensional power data of the adapter model to be visualized, it is ensured that the point cloud can fully cover the geometric shape of the adapter model to be visualized, and the vertices in the three-dimensional point cloud data are discretized to optimize the distribution and quality of the point cloud data, ensure that the point cloud data of the vertices is more uniform and orderly, improve the validity and availability of the point cloud data, and avoid geometric errors or visual defects.

[0071] In one embodiment of the present application, the Quad Remesher tool is used in 3DsMax to arrange the vertices on the adapter surface, adjust the vertex spacing to be approximately equal, and remove irregularly distributed vertices and some vertices in obscured areas.

[0072] In one embodiment of the present application, the width of the adapter model to be visualized on the screen is set to 300 pixels, and the adapter size is converted by parallel projection calculation and then scaled. The surface area value of the scaled adapter model to be visualized is measured to be approximately 1357205.

[0073] When the vertex density n / S=1 is set to construct the discrete point cloud of the adapter model to be visualized, theoretically, about 1,357,205 vertices need to be filled. After culling and adjustment, the discrete point cloud of the adapter model to be visualized actually contains 1,206,420 vertices in total.

[0074] S4. Write a shader for the digital twin model of the adapter model to be visualized to render the digital twin model, and visualize the rendered digital twin model.

[0075] Specifically, step S4 specifically includes:

[0076] S41, extracting initial vertex coordinates in the discrete point cloud grid, and extracting initial strain performance data of the digital twin model according to the initial vertex coordinates;

[0077] S42, acquiring strain performance data of different load steps according to the initial strain performance data, and processing the strain performance data to form a load-strain performance data sequence;

[0078] S43, specifying strain information for vertices of the digital twin model according to the load-strain performance data sequence;

[0079] S44, creating material resources, binding the strain information with the digital twin model through the material resources, and obtaining the digital twin model with physical information attached;

[0080] S45. Write a shader for the digital twin model to which the physical information is attached, obtain the rendered digital twin model, and perform a visual display.

[0081] As can be understood by those skilled in the art, a load step includes calculation steps for multiple different load conditions, such as static load, dynamic load, temperature load, pressure load, etc. Each load condition will produce different physical responses such as stress and strain on the structure.

[0082] The initial geometric information of the digital twin model is obtained from the discrete point cloud mesh, and the corresponding strain performance data is obtained according to different load steps to generate a strain performance data sequence. According to the strain performance data sequence, the corresponding physical information is assigned to each vertex of the digital twin model. The strain data is closely integrated with the digital twin model through the binding of material resources, and the strain information is visualized on the digital twin model using a shader. The strain distribution of the adapter model to be visualized under different loads is intuitively presented, providing engineers with valuable data analysis and design verification basis.

[0083] like Figure 3 and Figure 4 As shown, Figure 3 Rendering process diagram for Figure 4 It is a rendering pipeline block diagram. The geometric information and vertex attributes, rendering information, and rendering settings of the adapter are input into the rendering pipeline for rendering to obtain the rendered digital twin model. The rendering pipeline includes an application stage, a geometry stage, and a rasterization stage. The application stage processes vertex strain information and lighting information after receiving the vertex data of the digital twin model. At the same time, the geometry stage edits the vertex shader, tessellation shader, and geometry shader according to the vertex data of the digital twin model, and performs clipping and screen mapping. The result of the screen mapping is input into the rasterization stage, the triangle mesh of the digital twin model is set, and the triangle mesh is traversed, the fragment shader is edited, and operations are performed on a fragment-by-fragment basis to calculate the color and depth of each fragment to obtain the screen image.

[0084] Furthermore, the processing of the strain performance data specifically includes:

[0085] The strain performance data of different load steps are input into a data fusion network for data fusion, wherein the data fusion network includes a low-fidelity sub-network, a linear sub-network and a nonlinear sub-network, the low-fidelity sub-network maps the strain performance data of different load steps, the linear sub-network learns the linear relationship between the strain performance data of different load steps, and the nonlinear sub-network learns the nonlinear relationship between the strain performance data of different load steps.

[0086] The effective fusion and integration of strain data at different load steps is achieved through the data fusion network, generating a more complete and accurate strain performance data sequence, which can better capture the complex relationship between strain performance data at different load steps.

[0087] like Figure 5 As shown, where F l represents the linear part, F nl represents the nonlinear part, α represents the hyperparameter of the linearity between the low fidelity and the input data, and further, the structure of the data fusion network is:

[0088] The low-fidelity sub-network includes four first hidden layers of 128 neurons and three first non-linear activation functions, wherein the first non-linear activation function is located between two first hidden layers;

[0089] The linear sub-network includes three second hidden layers of 128 neurons;

[0090] The nonlinear sub-network includes four third hidden layers with 128 neurons and three second nonlinear activation functions, wherein the second nonlinear activation function is located between two third hidden layers.

[0091] Specifically, the data fusion process includes:

[0092] The strain performance data of different load steps are input into the low-fidelity sub-network, and low-fidelity output data is obtained after mapping;

[0093] The low-fidelity output data and the input of the low-fidelity sub-network are formed into a 4-dimensional vector, and the 4-dimensional vector is input into the linear sub-network and the nonlinear sub-network in parallel;

[0094] The outputs of the linear sub-network and the nonlinear sub-network are weighted summed to obtain the fused strain performance data.

[0095] Among them, the specific structures of the first hidden layer, the second hidden layer and the third hidden layer can be set according to actual needs, or the specific structures can be the same, and this application does not make any specific restrictions on this.

[0096] The linear and nonlinear characteristics of strain data are fully explored through linear and nonlinear subnetworks, and the comprehensive learning of complex strain relationships is achieved by using parallel linear and nonlinear subnetworks, thereby enhancing the modeling and prediction capabilities of strain performance data under different load conditions.

[0097] like Figure 6 and Figure 7 As shown, in one embodiment of the present application, step S43 specifically includes:

[0098] Set a color coding mapping relationship according to the value range of the strain performance data, and convert the strain performance data into corresponding color values;

[0099] Color values ​​are assigned to triangle mesh vertices in order according to the load step, so that each vertex has strain information.

[0100] By converting strain performance data into color codes and assigning them to triangular mesh vertices, the strain data can be visualized, allowing the digital twin model to more intuitively display the distribution characteristics of the data and reflect the changes in strain magnitude in different areas, which helps engineers quickly understand and analyze strain performance data.

[0101] In one embodiment of the present application, the initial vertex coordinates are V0={(X1, Y1, Z1), (X2, Y2, Z2), ..., (X i , Y i , Z i )}, i is a natural number;

[0102] Obtain the strain performance data of 20 load steps, process the strain performance data, and form a load-strain performance data sequence. n ={data 1n , data 2n ,……,data in}, 1≤n≤20;

[0103] Set the color bar according to the load-strain performance data sequence, and convert the load-strain performance data into corresponding RGB values ​​according to the corresponding strain performance data interval on the color bar. n is the color corresponding to each vertex of the digital twin model at the nth load step, where RGB n ={rgb 1n , rgb 2n , ..., rgb in}.

[0104] like Figure 8 As shown, in one embodiment of the present application, the shader for writing the digital twin model with physical information attached includes a vertex shader for writing the digital twin model with physical information attached and a fragment shader for writing the digital twin model with physical information attached, the vertex shader transforms the vertex data of the digital twin model with physical information attached, and the fragment shader interpolates and calculates the pixel information to obtain the color of each pixel, wherein the vertex data includes the position coordinates, strain information and normal vector of the vertex. The transformation processing includes geometric transformation, shading calculation and lighting calculation. For example, the vertex data of the digital twin model is translated, rotated, scaled, etc., and the color and lighting effects of the vertex are specified.

[0105] It is understandable that the shader is a tool to control the work of the GPU. To map the strain data to the adapter in an orderly and accurate manner and form a strain field cloud map, the Shader must be custom programmed. First, write the vertex shader to receive the vertex data (position coordinates, color attributes, etc.) processed by the CPU; then write the fragment shader to receive the data processed by the vertex shader, perform triangle rasterization, interpolate and calculate pixel information, form screen imaging, and realize the lighting model, texture mapping, shadow effects, etc., so that the final image presents a more realistic effect.

[0106] By writing vertex shaders and fragment shaders, the visual attachment of physical information (such as strain data) is realized, and the digital twin model is integrated with the actual physical information, enhancing the digital twin model's ability to express reality and improving the visual expression of the digital twin model, making it easier for engineers to analyze and make decisions.

[0107] This application reasonably adjusts the display size of the adapter model on the screen according to the actual scene requirements, uses the point cloud mesh to capture the details of the physical object, and improves the visualization effect through shader rendering. It improves the realism and expressiveness of the digital twin model as a whole, better reflects the characteristics of the actual physical objects, and provides more intuitive and effective visualization support for engineering applications.

[0108] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for visualizing a digital twin model of an adapter, characterized in that: The following steps are involved: S1. Setting the target imaging size of the adapter model on the screen; S2, calculating the ratio of the target imaging size to the actual physical width size of the adapter model to be visualized, obtaining a scaling factor, scaling the adapter model to be visualized in equal proportion according to the scaling factor, obtaining the scaled adapter model to be visualized, and measuring the surface area of ​​the scaled adapter model to be visualized; S3, constructing a discrete point cloud mesh of the adapter model to be visualized according to the scaled adapter model to be visualized and the surface area of ​​the adapter model to be visualized, and generating a digital twin model of the adapter model to be visualized according to the discrete point cloud mesh; S4. Write a shader of the digital twin model of the adapter model to be visualized to render the digital twin model, and visualize the rendered digital twin model; Step S4 specifically includes: S41, extracting initial vertex coordinates in the discrete point cloud grid, and extracting initial strain performance data of the digital twin model according to the initial vertex coordinates; S42, acquiring strain performance data of different load steps according to the initial strain performance data, and processing the strain performance data to form a load-strain performance data sequence; S43, specifying strain information for vertices of the digital twin model according to the load-strain performance data sequence; S44, creating material resources, binding the strain information with the digital twin model through the material resources, and obtaining the digital twin model with physical information attached; S45. Write a shader for the digital twin model to which the physical information is attached, obtain the rendered digital twin model, and perform a visual display.

2. The digital twin model visualization method of an adapter according to claim 1, characterized in that: Step S3 specifically includes: S31, setting vertex density according to the number of vertices and the surface area of ​​the adapter model to be visualized; S32, constructing discrete point cloud data of the scaled vertices of the adapter model to be visualized according to the vertex density; S33, connecting adjacent vertices corresponding to the discrete point cloud data to form a triangular mesh, and setting a normal of the triangular mesh; S34. Import the triangular mesh into the three-dimensional scene to obtain a digital twin model of the adapter model to be visualized.

3. The digital twin model visualization method of an adapter according to claim 2, characterized in that: Step S32 specifically includes: Constructing scaled three-dimensional point cloud data of the adapter model to be visualized according to vertex density; The distribution of vertices in the three-dimensional point cloud data is adjusted, irregularly distributed vertices and occluded vertices are identified and removed, and discrete point cloud data of the vertices of the adapter model to be visualized are obtained.

4. The method for visualizing a digital twin model of an adapter according to claim 1, characterized in that: The processing of the strain performance data specifically includes: The strain performance data of different load steps are input into a data fusion network for data fusion, wherein the data fusion network includes a low-fidelity sub-network, a linear sub-network and a nonlinear sub-network, the low-fidelity sub-network maps the strain performance data of different load steps, the linear sub-network learns the linear relationship between the strain performance data of different load steps, and the nonlinear sub-network learns the nonlinear relationship between the strain performance data of different load steps.

5. The method for visualizing a digital twin model of an adapter according to claim 4, characterized in that: The structure of the data fusion network is: The low-fidelity sub-network includes four first hidden layers of 128 neurons and three first non-linear activation functions, wherein the first non-linear activation function is located between two first hidden layers; The linear sub-network includes three second hidden layers of 128 neurons; The nonlinear sub-network includes four third hidden layers with 128 neurons and three second nonlinear activation functions, wherein the second nonlinear activation function is located between two third hidden layers.

6. The method for visualizing a digital twin model of an adapter according to claim 5, characterized in that: The data fusion process specifically includes: The strain performance data of different load steps are input into the low-fidelity sub-network, and low-fidelity output data is obtained after mapping; The low-fidelity output data and the input of the low-fidelity sub-network are formed into a 4-dimensional vector, and the 4-dimensional vector is input into the linear sub-network and the nonlinear sub-network in parallel; The outputs of the linear sub-network and the nonlinear sub-network are weighted summed to obtain the fused strain performance data.

7. The method for visualizing a digital twin model of an adapter according to claim 1, characterized in that: Step S43 specifically includes: Set a color coding mapping relationship according to the value range of the strain performance data, and convert the strain performance data into corresponding color values; Color values ​​are assigned to triangle mesh vertices in order according to the load step, so that each vertex has strain information.

8. The method for visualizing a digital twin model of an adapter according to claim 1, characterized in that: The shader for writing the digital twin model with physical information attached includes a vertex shader for writing the digital twin model with physical information attached and a fragment shader for writing the digital twin model with physical information attached. The vertex shader transforms the vertex data of the digital twin model with physical information attached, and the fragment shader interpolates and calculates the pixel information to obtain the color of each pixel, wherein the vertex data includes the position coordinates, strain information and normal vector of the vertex.

9. The method for visualizing a digital twin model of an adapter according to claim 8, characterized in that: The transformation processing includes geometric transformation, shading calculation and lighting calculation.

Images