US-12626027-B2 - Methods for modelling variations of repeating parts of a component

US12626027B2US 12626027 B2US12626027 B2US 12626027B2US-12626027-B2

Abstract

A computer-implemented method for modelling variations of repeating parts of a component is disclosed herein. The method includes providing parameter sets for part configurations of a plurality of different of parts, wherein the parameters of each parameter set respectively define the part configuration of a part; selecting a selection of the part configurations; providing a part model for each part configuration of the selection, wherein the part model relates forces and displacements to locations of the respective part; and building an approximator, wherein the approximator is provided for interpolating the part models to approximate a part model of a part configuration that does not belong to the selection of the part configurations.

Inventors

Tom De Weer

Assignees

SIEMENS INDUSTRY SOFTWARE NV

Dates

Publication Date: 20260512
Application Date: 20220224
Priority Date: 20210305

Claims (15)

1 . A computer-implemented method for improving an efficiency and an accuracy of a physical simulation of three-dimensional (3D) components with repeating parts by modelling variations of repeating parts of a 3D component, the method comprising: providing, by a processor, parameter sets for part configurations of a plurality of different parts, wherein parameters of each parameter set define a respective part configuration of a part of the plurality of different parts of the 3D component; selecting, by the processor, a selection of the part configurations from a defined parameter space of configurations; providing, by the processor, a part model for each part configuration of the selection, wherein the part model relates forces and displacements to locations of the respective part; building, by the processor, a computational approximator, wherein the computational approximator is provided for interpolating the part models to approximate a part model of a part configuration that does not belong to the selection of the part configurations within the defined parameter space; generating a reduced component model, using the computational approximator, for several parts of the 3D component, therein providing an improved efficiency and accuracy when a physical simulation of a 3D component is performed without the computational approximator; and manufacturing the 3D component using the reduced component model.
2 . The method of claim 1 , wherein the providing of the part model comprises: meshing a mesh of the part with solid elements; generating a preliminary part model of the meshed part by relating forces and displacements to locations defining the solid elements through a first part stiffness matrix; identifying constraints reducing a number of exterior degrees of freedom of movement by constrained degrees of freedom; and generating the part model by reducing the model order of the preliminary part model to the remaining exterior degrees of freedom by eliminating the constrained degrees of freedom of the first part stiffness matrix resulting in a reduced stiffness matrix, and wherein the building of the approximator comprises interpolating the reduced stiffness matrixes to extract the reduced stiffness matrix of a part that does not belong to the selection of the part configurations.
3 . The method of claim 2 , wherein the repeating parts are joints of a lattice structure.
4 . The method of claim 3 , wherein the selecting of the selection of the part configurations is a random selection.
5 . The method of claim 3 , wherein the reducing the model order is done using a Guyan reduction resulting in the reduced stiffness matrix.
6 . The method of claim 3 , wherein the interpolating of the reduced stiffness matrices of the part models uses Canonical Polyadic Decomposition.
7 . The method of claim 3 , further comprising: combining resulting reduced stiffness matrices of the several parts of the 3D component with other models of elements of the 3D component.
8 . The method of claim 1 , wherein the repeating parts are joints of a lattice structure.
9 . The method of claim 1 , wherein the selecting of the selection of the part configurations is a random selection.
10 . The method of claim 2 , wherein the reducing the model order is done using a Guyan reduction resulting in a reduced stiffness matrix.
11 . The method of claim 1 , wherein the interpolating of the reduced stiffness matrices of the part models uses Canonical Polyadic Decomposition.
12 . The method of claim 1 , further comprising: combining resulting reduced stiffness matrices of the several parts of the 3D component with other models of elements of the 3D component.
13 . An additive manufacturing method for generating a three-dimensional (3D) component with variations of repeating parts, the method comprising: providing, by a processor, parameter sets for part configurations of a plurality of different parts, wherein parameters of each parameter set define a respective part configuration of a part of the plurality of different parts of the 3D component; selecting, by a processor, a selection of the part configurations from a defined parameter space of configurations; providing, by a processor, a part model for each part configuration of the selection, wherein the part model relates forces and displacements to locations of the respective part; building, by the processor, a computational approximator, wherein the computational approximator is provided for interpolating the part models to approximate a part model of a part configuration that does not belong to the selection of the part configurations within the defined parameter space; generating a reduced component model, using the computational approximator, for several parts of the 3D component, therein providing an improved efficiency and accuracy when a physical simulation of a 3D component is performed without the computational approximator; and generating the 3D component using the reduced component model.
14 . A computer system for improving an efficiency and an accuracy of a physical simulation of a three-dimensional (3D) component with repeating parts by modeling variations of repeating parts of the 3D component, the computer system comprising: a processor; and a memory, wherein the memory and the processor are configured to: provide parameter sets for part configurations of a plurality of different parts, wherein parameters of each parameter set define a respective part configuration of a part of the plurality of different parts of the 3D component; select a selection of the part configurations from a defined parameter space of configurations; provide a part model for each part configuration of the selection, wherein the part model relates forces and displacements to locations of the respective part; build a computational approximator, wherein the computational approximator is provided for interpolating the part models to approximate a part model of a part configuration that does not belong to the selection of the part configurations within the defined parameter space; generate a reduced component model, using the computational approximator, for several parts of the 3D component, therein providing an improved efficiency and accuracy when a physical simulation of a 3D component is performed without the computational approximator; and generate the 3D component using the reduced component model.
15 . The method of claim 1 , further comprising: outputting, by the processor, the computational approximator for use in the manufacturing of the 3D component.

Description

The present patent document claims the benefit of European Patent Application No. 21160989.6, filed Mar. 5, 2021, which is hereby incorporated by reference in its entirety. TECHNICAL FIELD The disclosure relates to a computer-implemented method for modelling variations of repeating parts of a component. BACKGROUND Modelling variations of repeating parts of a component may need to be done in performance of engineering tasks. For example, lattice structures, (e.g., of the lightweight type), become more popular in particular due to new production techniques like additive manufacturing, which enables easier creation of, e.g., lightweight structures with variations of repeating parts of the component to be generated. Lattice structures according to a narrower definition are containing beams that are connected at joints, e.g., by repeating a unit cell. This disclosure may be applicable to lattice structures which include porous three-dimensional spatial structures formed and tessellated by unit cells with different topological geometries. This kind of lattice structure may be considered to belong to cellular structures (e.g., including foam structures, honeycomb structures, and lattice structures). If, for example, as a component a lattice structure with joints is considered, these joints may be provided as variations of repeating parts, and they may have a complex geometry. Such joints may conventionally be hard to simulate accurately without a large computational effort. Models required for industrially relevant applications may be highly detailed and may quickly become unfeasible regarding availability of computation power. Lattice structures may be simulated with the Finite Element Method (FEM). Here, the structure is subdivided (“meshed”) into a mesh of simpler geometries called Finite Elements. These elements are more easily discretized and after assembly the full geometry is captured. During the post processing, the performance of the design may be gauged by looking at the simulation results. The total stiffness and mass may be used in topology optimization frameworks, wherein stress concentrations decide the fatigue life of the structure and an eigenvalue analysis shows the dynamical behavior of the structure. There are two types of mesh that may be used for lattices. The first type of mesh is a solid mesh, wherein the lattice structure is subdivided into a large number of three-dimensional (3D) elements (e.g., Simcenter Nastran's CTETRA). Due to the geometrical complexity of lattices, a lot of these elements may be needed. This method may therefore be considered ‘accurate but slow’. The second mesh type is the one-dimensional (1D) mesh, wherein every beam in the mesh is replaced by a beam element (for example, Simcenter Nastran's CBAR). Fewer elements are thus needed, resulting in a much smaller computation time. A problem with this approach is that these beam elements are based on beam theory (e.g., Euler-Bernoulli beam theory, Timoshenko-Ehrenfest beam theory), which becomes inaccurate as the beams become thicker. Aspects of lattice modelling are known from: M. Helou and S. Kara, “Design, analysis and manufacturing of lattice structures: An overview,” International Journal of Computer Integrated Manufacturing, 31 (3):243-261, 2018. Another publication is G. Dong, Y. Tang, and Y. F. Zhao, “A survey of modeling of lattice structures fabricated by additive manufacturing,” Journal of Mechanical Design, Transactions of the ASME, 139(10), 2017. Some teachings of these publications may be summarized regarding the field of the disclosure as follows. A major problem is that by replacing the beams with one-dimensional elements, every joint is replaced by a single point. However, these nodes contain a lot of stiffness and mass. In bending-dominated lattices, they also have a large influence on the fatigue life because that is where the highest stress occurs. There are some techniques in literature that try to mitigate this problem by increasing the beam thickness close to the joints, but this is more a pragmatic but less accurate solution. Further, there is no generally accepted way to determine by how much the thickness then needs to change. This method may therefore be considered “fast but inaccurate”. Currently, there exist methods for modelling variations of repeating parts of a component in particular for modelling a joint of a lattice structure. Such methods may benefit from improvements. It is one objective of the disclosure to improve the design process of components with variations of repeating parts. SUMMARY AND DESCRIPTION Based on the prior art described above and the problems associated, the disclosure is based on the task of improving the design process of a component including variations of repeating parts, like lattice structures. The scope of the present disclosure is defined solely by the appended claims and is not affected to any degree by the statements within this summary. The present embodim