Method for replacing parts of a CAD model with more coarsely gridded parts

Abstract

The invention relates to a method for modifying a CAD model. A starting model of the object is provided with thickness information, and an optimized model is then generated, in the form of a shell model, from the starting model using an optimization process. The shell model has the shape of a first grid, wherein each node of the first grid is assigned a thickness information value. At least one starting model region which is modified during the optimization process is identified by comparing the thickness information values of the starting model with the corresponding values of the shell model. The identified region is represented in the form of a second grid, said second grid being coarser than the first grid. The thickness information values are transferred from the first grid to the second grid, and a section which is representative of the identified region is generated using the transferred values. Finally, the starting model is adapted by replacing the corresponding region with a section which is representative of the identified region.

Description

[0001] Method for replacing parts of a CAD model with coarser meshed parts

[0002] The invention relates to a method for adapting a three-dimensional object for optimizing and / or constructing a three-dimensional object, in particular a component to be manufactured, and to a computer program product and an object manufactured according to the method.

[0003] For example, DE 102021 003 565 A1 discloses methods in which an optimization result of a component, in particular an optimization result of a topology of the component, is used to construct a CAD geometry of the component based on the optimization result. The component, in particular the CAD geometry of the component, is manually created using geometries of design elements in such a way that the optimization result of the component and the CAD geometry of the component are almost identical. The manual creation of the CAD geometry of the component is necessary because the optimization result of the component only contains visualizable data and no geometrically usable data. The time required to manually create the CAD geometry of a sub-area of ​​a simple component is at least a few minutes, although the time required for more complex components increases rapidly with increasing complexity.In one of the known methods, the CAD geometry of the component is created using a CAD system - in particular Siemens NX.

[0004] US 2022 / 0284153 A1 discloses another method in which a computer system accesses models of an object, in particular geometries, in connection with an optimization process, in particular a topology optimization process. These geometries comprise an initial geometry representing a design space to which the topology optimization process is applied, and an optimized geometry representing a result of the topology optimization process performed on the initial geometry. The method also includes generating a final geometry from the topology-optimized geometry. This is done by adapting the topology-optimized geometry to the initial geometry at locations that correspond to fixed regions of the initial geometry. In addition, the topology-optimized geometry is smoothed at locations that do not correspond to fixed regions of the initial geometry.

[0005] The invention is based on the object of specifying an improved method for adapting three-dimensional objects, a component produced according to this method and a vehicle which comprises at least one such component.

[0006] The first-mentioned object is achieved according to the invention by a method for adapting a three-dimensional object having the features of claim 1. The second-mentioned object is achieved according to the invention by a component having the features of claim 7. The third-mentioned object is achieved according to the invention by the features of claim 8.

[0007] Advantageous further developments of the invention are the subject of the subclaims.

[0008] In a method according to the invention for adapting a three-dimensional object, an initial model of the object is first provided with information about at least one object parameter. Subsequently, an optimized model of the object in the form of a shell model is generated from the initial model of the object by means of an optimization process of the at least one object parameter. The shell model has the form of a first mesh, with each node of the first mesh being assigned a value of the at least one optimized object parameter.

[0009] By comparing the values ​​of at least one object parameter of the initial model with the corresponding values ​​of at least one optimized object parameter of the shell model, at least one region of the initial model modified during the optimization process is identified. The identified region is represented in the form of a second mesh, wherein the second mesh is coarser than the first mesh. The values ​​of the at least one optimized object parameter are transferred from the first mesh to the second mesh, and a representative cross-section of the identified region is created using the transferred values. Finally, the object is adapted by replacing the corresponding region of the initial model with the representative cross-section of the identified region.

[0010] The advantages achieved with the invention are, in particular, that optimized areas of an object, for example from finite element optimizations created in a computer-aided engineering tool (CAE tool), can be identified and precisely transferred to a computer-aided design tool (= CAD tool) and the original areas of the initial model of the object are replaced in the CAD tool, thereby adapting the object.

[0011] This enables faster and more precise adjustment of three-dimensional objects, as the process automatically transfers the optimized areas from the CAE tool to the CAD tool. The traditional, tedious, time-consuming, and inaccurate process of parametrically redesigning optimized areas for further transfer to the CAD tool is avoided.

[0012] Furthermore, the transferred optimized areas can be further modified by an engineer in the CAD tool, if necessary, in a more detailed manner, since the representation of the areas after the process in the CAD tool is more detailed. This leads to greater flexibility and adaptability in the design of three-dimensional objects.

[0013] This makes it possible to easily and quickly construct objects such as vehicle components with improved functionality, reduced weight and / or optimized topology, which is advantageous for manufacturing and component production.

[0014] The three-dimensional object can, for example, be a component to be manufactured, such as a plastic component, a cast component, a forged component, an extruded component, a 3D-printed component, an injection-molded component, a deep-drawn component, or the like. Such structural objects or components can be freely designed in terms of shape, material, and / or thickness distribution with few manufacturing restrictions. In one possible embodiment, the second mesh is coarser than the first mesh in zones of the identified area that correspond to a lower maximum gradient value of the at least one optimized object parameter. The tool used for the initial model, such as a CAD tool, has limited computing capacity and cannot work with fine meshes like the first mesh that result from the optimization process of another tool, such as a CAE tool.To reduce the information that needs to be transferred from the first mesh to the second mesh while avoiding excessive loss of precision in the optimization process, the zones of the identified area that exhibit lower parameter value variability due to optimization are represented as a coarser mesh. This helps avoid losing precision in the zones where precision is required by conserving computational capacity in the zones where the optimization results can be represented using a smaller number of cells of the second mesh.

[0015] In another possible embodiment, one cell of the second mesh corresponds to a plurality of cells of the first mesh. Similar to the reasons mentioned above, the tool used for the initial model, for example, a CAD tool, has limited computing capacity and cannot work with fine meshes like the first mesh resulting from the optimization process of another tool, for example, a CAE tool. In this way, an approximation process takes place in which the values ​​stored in the nodes of several cells of the first mesh are transferred to the node of a cell of the second mesh. In this way, the second mesh can reflect the result of the optimization process with a certain approximation while still being light enough to be transferred to the CAD tool.

[0016] In another possible embodiment, the values ​​of a plurality of nodes of the first network are transferred to a corresponding node of the second network by averaging. This embodiment demonstrates a general approximation technique that can be used during the transfer of values ​​from the first network to the second network. Averaging techniques can include simple averaging, weighted averaging, or other types of averaging. In another possible embodiment, the object parameter is thickness information. Optimizing the thickness parameter leads to a reduction in the weight of the produced object, an increase in its robustness, and other benefits. This optimization can, for example, be performed in combination with other parameters such as material and density to achieve further improvements in the object's performance.The thickness optimization process can be performed using various optimization techniques such as gradient-based optimization or evolutionary algorithms. Thickness information can be obtained from various sources such as simulation results or experimental data. The thickness optimization process can be performed iteratively, with the results of each iteration used to further refine the optimization process. The resulting optimized thickness information can be used to generate a representative cross-sectional surface of the identified region, which can be used in the manufacturing process to produce the optimized object.

[0017] In another possible embodiment, the representative section surface of the identified region is created by offsetting the nodes of the second mesh according to the thickness information. In another possible embodiment, the representative section surface of the identified region is created by offsetting the nodes of the second mesh according to the thickness information. This is one of the possible methods for preparing the second mesh for transfer to the tool in which the initial model is created. The representative section surface should reflect the optimized geometry of the identified region. The optimized shell model corresponding to the identified region stores the optimized values ​​of the parameters in the nodes of the first mesh and does not necessarily reproduce the geometry of the identified region.

[0018] The component according to the invention comprises a structure with optimized object parameters, which was manufactured by the method described above.

[0019] The vehicle according to the invention comprises at least one component comprising a structure with optimized object parameters, which was manufactured by the method described above. The vehicle according to the invention comprises at least one component comprising a structure with optimized object parameters, which was manufactured by the method described above.

[0020] Embodiments of the invention are explained in more detail below with reference to drawings.

[0021] Showing:

[0022] Fig. 1 shows schematically an optimized CAE shell model of a three-dimensional object,

[0023] Fig. 2 shows a schematic representation of an optimized CAE shell model in which the optimized object parameters are assigned to the nodes, as well as an initial CAD model with homogeneous object parameters and the generation of a deviation CAE shell model,

[0024] Fig. 3 shows schematically an example of the identified area of ​​the initial CAD model, the corresponding area of ​​the CAE shell model and the resulting subdivision surface model of the identified area,

[0025] Fig. 4 shows schematically the creation of the representative section of the identified area by offsetting the nodes of the second mesh according to the thickness information,

[0026] Fig. 5 shows schematically an adapted CAD model 2' of the object.

[0027] Corresponding parts are provided with the same reference numerals in all figures.

[0028] Figure 1 schematically shows an optimized model of the object (not shown) in the form of a shell model, in particular an optimized CAE shell model 1. The object can be, in particular, a vehicle component such as a sheet metal component or body component. It can also be a component to be manufactured, such as a plastic component, a cast component, a forged component, an extruded component, a 3D-printed component, an injection-molded component, a deep-drawn component, or the like.

[0029] Such structural objects or components can be freely designed, designed and manufactured in terms of shape, material and / or thickness distribution.

[0030] The object is a three-dimensional component or a three-dimensional body that can be described by its base surface 1.6 and ribs 1.5. The base surface 1.6 can be formed or composed of flat and / or curved surface pieces or surface sections and can be provided with elevations such as ribs 1.5, mounts, pins, cones, or the like. For clarity, the invention will be referred to the base surface 1.6 below.

[0031] The CAE shell model 1 of the object corresponds to a geometric initial model with a homogeneous object parameter distribution (hereinafter referred to as initial CAD model 2). The initial CAD model 2 is a raw geometric representation of the object. It includes model structure data or CAD geometry data and can be imported, for example, in digital form into a CAE tool.

[0032] The exported initial CAD model 2 is optimized through simulation, for example, using a modeling tool 3, in particular a CAE tool, and represented in the form of the shell model 1, in particular a finite element mesh. The result of this process is the optimized model of the object. This process is performed with the entire initial CAD model 2, including the base surface 1.6 and the ribs 1.5.

[0033] The modeling tool is, for example, a computer or a processing unit on which a corresponding computer program product, such as a graphics file and / or a computer graphics program, can be loaded and executed. In particular, the modeling tool is a design and / or construction tool.

[0034] A mesh 1.1 (also called first mesh 1.1) of the optimized CAE shell model 1 comprises a plurality of mesh cells 1.2 and a plurality of mesh lines 1.3. The first mesh 1.1 is designed, for example, as a structured quadrilateral mesh. Alternatively, the first mesh 1.1 can also be a triangular mesh or the like. Each node 1.4 of the mesh cell 1.2 of the first mesh 1.1 is associated with a scalar value of a physical and / or topological parameter of the optimized CAE shell model 1. The scalar values ​​can, for example, represent information on the thickness distribution of the object. The optimized CAE shell model 1 shown is, in particular, a thickness-optimized CAE shell model 1.

[0035] In some cases, after optimization through simulation, some scalar values ​​of physical parameters, such as thickness, may be discontinuous. This means that there are abrupt changes in thickness between adjacent mesh cells. To resolve this, the object is optionally checked for such abrupt thickness changes and / or discontinuities during the creation of the thickness-optimized CAE shell model 1. These abrupt thickness changes can be smoothed using the model tool. The optimization step (also called smoothing step) is not shown.

[0036] Optionally, a rib-free base surface mesh model (not shown) can be extracted or isolated from the thickness-optimized CAE shell model 1 using the model tool. Alternatively, the entire optimized CAE shell model 1 can be used for further steps, as shown in Figure 2.

[0037] The generated thickness-optimized CAE shell model 1 can be used to adapt the object, as described in more detail below.

[0038] Figure 2 schematically shows an optimized CAE shell model 1, in which the optimized object parameters, in particular thickness information, are assigned to the nodes 1.4, as well as an initial CAD model 2 with homogeneous object parameters, in particular homogeneous thickness, and the generation of a deviation CAE shell model T.

[0039] To identify the region of the initial CAD model whose thickness was modified as a result of the optimization process (by generating the thickness-optimized CAE shell model 1), a deviation CAE shell model T can be created, for example. The thickness values ​​associated with the nodes 1.4 of the first mesh 1.2 of the thickness-optimized CAE shell model 1 can be modified or changed, for example, by subtracting the scalar thickness value of the corresponding node 1.4 from the homogeneous thickness value of the initial CAD model 2.

[0040] If no difference is found, the thickness of the corresponding node of the first mesh 1.4' of a deviant CAE shell model T is set to zero. However, if the modeling tool detects a difference between the homogeneous thickness of the initial CAD model 2 and the thickness-optimized CAE shell model 1, the thickness of the corresponding node of the first mesh 1.4' of the deviant CAE shell model T can be set to this difference value (= delta). The deviant CAE shell model T is thus further used to identify the region of the initial CAD model whose thickness was modified as a result of the optimization process.

[0041] The generated deviation CAE shell model T allows the identification of the region in the initial CAD model 2 whose thickness was modified by the optimization process. If multiple nodes 1.4' with a non-zero thickness difference value exist in a specific region of the deviation CAE shell model T, the corresponding region of the initial CAD model 2 is designated as the identified region of the initial CAD model 3 and used for the further process as shown in Figure 3.

[0042] The identified area of ​​the initial CAD model 3 can, for example, be an isolated parametric CAD surface, in particular the front surface of the identified area of ​​the initial CAD model 3. This parametric CAD surface is converted into a subdivision surface model of the identified area 5 according to Figure 3.

[0043] The corresponding region of the CAE shell model 1 is also identified and referred to as the identified region of the CAE shell model 4, as further described in Figure 3. Both identified regions, that of the initial CAD model 3 and that of the CAE shell model 4, are associated with an identical region of the object and have the form of isolated surfaces. Figure 3 schematically shows an example of the identified region of the initial CAD model 3, the corresponding region of the CAE shell model 4, and the resulting subdivision surface model of the identified region 5.

[0044] As already mentioned above, the identified area of ​​the source CAD model 3 can, for example, be a front surface of the identified area of ​​the source CAD model 3. The identified area of ​​the source CAD model 3 is converted into a subdivision surface model 5, which can be further processed, for example, in the CAE tool.

[0045] The converted subdivision surface model 5 has the shape of the second mesh, which is coarser than the first mesh 1.1. Since the subdivision surface model 5 is transferred back to the CAD tool, and the CAD tool has limited computing capacity and cannot work with fine meshes like the first mesh 1.1, the second mesh 5.1 is coarser than the first mesh 1.1. For example, it may be that one cell of the second mesh 5.2 corresponds to several cells of the first mesh 1.2.

[0046] In addition, the second mesh 5.1 can be coarser, for example, in zones of the subdivision surface model 5 that correspond to the zones of the identified region of the CAE shell model 4 with a lower maximum gradient value of the thickness parameters. This principle makes it possible to reduce the information loss caused by transferring the thickness values ​​from the first mesh 1.1 to the coarser second mesh 5.1. In this way, the second mesh 5.2 better approximates the results of the thickness optimization process.

[0047] The thickness information stored in the nodes of the first mesh 1.4 is transferred to the nodes of the subdivision surface model 5.4. This can be done, for example, by averaging the thickness parameter stored in the nodes of several cells of the first mesh 1.2 that are associated with the nodes of a coarser cell of the second mesh 5.2. Alternatively, the parameters stored in the nodes of the first mesh 1.4 can be transferred to the nodes of the second mesh 5.4 that are positioned at the same or approximately the same position as the nodes of the first mesh 1.4. As a result, the subdivision surface model 5, in the form of a second mesh 5.1, approximates the results of the thickness optimization.

[0048] Figure 4 schematically shows the creation of the representative section of the identified area 6 by offsetting the nodes of the second mesh 5.4 according to the thickness information.

[0049] Figure 4 shows the section of the subdivision surface model 5' in which the information about the thickness at the nodes of the second mesh 5.4 is stored. The subdivision surface model 5 is converted into the representative section surface 6, the section 6' of which is shown in Figure 4.

[0050] In this example, the curvature of the representative section surface represents the thickness distribution through offset nodes of the second mesh 6.2. In this form, the representative section surface 6 is transferred back to the CAD tool, for example, to replace the front surface of the identified area of ​​the original CAD model 3. This results in the creation of the adapted CAD model 2' of the object, shown in Figure 5.

[0051] The resulting CAD model 2' of the object advantageously has an inhomogeneous thickness distribution, which leads to better physical properties of the object and material savings.

Claims

Mercedes-Benz Group AG latskina January 15, 2025 Patent claims 1. A method for generating a CAD model, comprising CAD geometric data of a three-dimensional object, wherein an initial model of the object (2) is provided with information about at least one object parameter, an optimized model of the object in the form of a shell model (1) is generated from the initial model of the object (2) by means of an optimization process of the at least one object parameter, wherein the shell model (1) has the form of a first mesh (1.1), wherein each node of the first mesh (1.4) is assigned a value of the at least one optimized object parameter, by comparing the values of the at least one object parameter of the initial model (2) with the corresponding values of the at least one optimized object parameter of the shell model (1), at least one region of the initial model (3) modified during the optimization process is identified, the identified region (3) is stored in the form of a second mesh (5.1), wherein the second mesh (5.1) is coarser than the first mesh (1.1), the values of the at least one optimized object parameter are transferred from the first mesh (1.1) to the second mesh (5.1), a representative sectional area of the identified region (6) is created by means of the transferred values, the object is adapted as part of the creation of an adapted CAD model (2') by replacing the corresponding region of the initial model (2) with the representative sectional area of the identified region (6).

2. Method according to claim 1, characterized in that the second network (5.1) is arranged in zones of the identified area which are of a lower maximum gradient value of the at least one optimized object parameter, is coarser than the first network (1.1).

3. Method according to claim 1 or 2, characterized in that a cell of the second network (5.2) corresponds to a plurality of cells of the first network (1.2).

4. Method according to claim 3, characterized in that the values of a plurality of nodes of the first network (1.4) are transferred to a corresponding node of the second network (5.4) by averaging.

5. Method according to claims 1 to 4, characterized in that the object parameter is thickness information.

6. The method according to claim 5, characterized in that the representative sectional area of the identified region (6) is created by offsetting the nodes of the second network (5.4) according to the thickness information.

7. A component, characterized by a structure with optimized object parameters, wherein the optimized object parameters correspond to the optimized object parameters of an adapted CAD model (2'), wherein the adapted CAD model (2') was created according to a method according to one of claims 1 to 6.

8. Vehicle, characterized by at least one component according to claim 7.

Images