A Quasi-Optimal Shape Design Method for Lattice Structure Construction Journal: Elsevier

Sifan Chen, Yuan Kong, Qiang Zou · 2025

Read A Quasi-Optimal Shape Design Method for Lattice Structure Construction Journal: Elsevier by Sifan Chen, Yuan Kong, Qiang Zou (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2502-02602.txt.

Introduction

Sifan Chen and co-authors build lattice structures with a shape design method that stays robust at the hard spots. Lattice frames earn their place in aerospace, cars, and medicine through strength at low weight. The trouble is the nodes, where many struts meet and models tend to break.

Their 2025 paper cuts the struts short and joins them to node shapes tuned by optimization. A recent meta-heuristic, the Grey Wolf method, searches for node forms that join cleanly and carry load well. Case studies of rising topological and geometric complexity test the pipeline end to end. The result is a construction route that gives simulators and fabricators the exact shape they need.

Catalog record

CREATORSifan Chen, Yuan Kong, Qiang Zou
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE50 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2502-02602

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

A Quasi-Optimal Shape Design Method for Lattice Structure Construction Journal: Elsevier Sifan Chen, Yuan Kong, Qiang Zou Abstract Lattice structures, known for their superior mechanical properties, are widely used in industries such as aerospace, automotive, and biomedical. Their advantages primarily lie in the interconnected struts at the micro-scale. The robust construction of these struts is crucial for downstream design and manufacturing applications, as it provides a detailed shape description necessary for precise simulation and fabrication. However, constructing lattice structures presents significant challenges, particularly at nodes where multiple struts intersect. The complexity of these intersections can lead to robustness issues. To address this challenge, this paper presents an optimization-based approach that simplifies the construction of lattice structures by cutting struts and connecting them to optimized node shapes. By utilizing the recent Grey Wolf optimization method—a type of meta-heuristic method—for node shape design, the approach ensures robust model construction and optimal shape design. Its effectiveness has been validated through a series of case studies with increasing topological and geometric complexity. 1 Introduction Lattice structures, composed of interconnected struts at the micro-scale, have found increasing applications in industries such as aerospace, automotive, and biomedical fields due to their superior mechanical properties and lightweight nature zou2024geometric . These structures enable efficient material use while maintaining structural integrity, making them ideal for performance- and weight-critical applications. Additionally, advancements in additive manufacturing technologies have made the production of these complex structures increasingly feasible liu2021survey , nazir2019state . Extensive research has been carried out to understand the mechanical behavior of lattice structures and report the “shape-property-process" relationship, both theoretically and empirically wu2019design . However, as recently noted by Verma et al. verma2020combinatorial and others tao2016design , wang2005hybrid , researchers have consistently encountered challenges in the construction of lattice structures while developing this relationship. Traditional solid modeling techniques in computer-aided design (CAD) zou2023variational often fail to address these challenges effectively zou2024meta . To bridge this gap, a simple and robust method for lattice structure construction is developed in this work. Specifically, given a weighted lattice graph—comprising edges, nodes, and weights that define the struts, connections, and parameters (e.g., radii)—the goal is to create a watertight boundary representation (B-rep) model of the lattice structure for use in applications such as mesh generation wang2018hex , chen2018finite , finite element analysis zou2023xvoxel , gu2021material , and tool path generation zou2013iso , zou2014iso , as shown in Fig. 1 . A key challenge in this process is ensuring robustness in generating nodal shapes, as nodes represent intersections of multiple struts and involve complex Boolean operations to evaluate their boundary shapes. These operations are notoriously unreliable verma2020combinatorial , liu2021memory . Although struts (i.e., cylinders) are seemingly simple to intersect, robustness issues still occur. As already highlighted by Hoffman hoffmann1989problems , even for such simple geometries, singular tangencies and small face fragments (see Fig. 2 ) can cause intersection failures. To address these robustness challenges, two main strategies have been developed. The first, implicit modeling, involves converting the lattice structure into implicit fields, performing Boolean operations on these fields, and then converting the results to meshes via iso-contouring pasko2011procedural , fryazinov2013multi , ma2022novel . While this approach is robust, it is computationally intensive and tends to generate large numbers of unnecessary or distorted triangles. The second strategy simplifies the nodal shapes, typically using convex hulls, to avoid Boolean operations. While promising results have been demonstrated verma2020combinatorial , wu2020chocc , current methods do not address the deviation between simplified and original nodal shapes, nor do they determine which portions of the nodal shape are best to be replaced. In this paper, we adopt the second strategy but enhance it with an optimization-based approach to reduce shape deviation. Specifically, we determine the optimal locations on each strut to cut the node from its adjacent struts, effectively separating the replaceable portion of the lattice from the parts to be retained. The goal is to preserve the original shape as much as possible. After the cuts, we apply a shape optimization method to fill the void left by the removed nodes. Given the high nonlinearity of this optimization problem, we use the Grey Wolf Optimizer (GWO) mirjalili2014grey to solve it, a meta-heuristic algorithm known for its efficiency due to simple solution update schemes and high parallelizability. This optimizer generates high-quality nodal shapes that are watertight, smooth, and self-intersection-free, and closely match the original nodal shapes. To the best of our knowledge, while meta-heuristic optimization has been widely used in engineering, this seems to be the first work of applying such techniques to lattice structure construction. The remainder of this paper is organized as follows: Sec. 2 reviews related studies, Sec. 3 details the proposed method, Sec. 4 presents validation results through case studies, and Sec. 5 concludes with a discussion of the method’s advantages and limitations. Figure 1: Illustration of the lattice graph representation, the B-rep model constructed, and the final fabricated object Figure 2: A singular example (tangency) of boundary evaluation for a lattice node 2 Related Work Existing lattice structure construction methods may be classified into two general categories: explicit and implicit. The former directly manipulates B-rep elements, while the latter constructs lattice structures using implicit fields. These methods are discussed in detail below. 2.1 Explicit Methods The conceptually straightforward approach to constructing lattice structures involves modeling struts as cylindrical solids and nodes as spherical solids, which are then combined using Boolean operations to complete the structure. The efficiency, accuracy, and robustness of this method largely depend on the representation schemes used to define these cylindrical and spherical solids. A variety of modeling techniques have been explored, including algebraic surface-based models wang2005hybrid , dumas2017modelling , parametric surface-based models 2021_spherePacking_conformal , subdivision surface-based models 2018_subdivision_2Dsurface , 2019_differentialOffsetGrading , 2023_2Dsurface_Xiong_Subdivisional-modelling , and mesh-based models chen20073d , 2017_light-weight-triangulation , 2017_prefab-cell_HGM . Algebraic and parametric surface models provide high accuracy due to their continuous nature, but their Boolean intersections can be computationally intensive and usually unreliable. Mesh-based models, by contrast, enable simpler intersection calculations; however, their accuracy and efficiency depend on the mesh resolution—higher resolution improves accuracy but reduces efficiency, and vice versa. Furthermore, ensuring robustness in mesh-based Boolean operations remains an ongoing challenge trettner2022emberm . There are also studies that resort to the “stitching" operation to construct B-reps of lattice structures 2017_light-weight-triangulation , verma2020combinatorial , wu2020chocc . The idea is to generate triangular meshes for the struts and nodes separately, then stitch the vertices at their interfaces to create a watertight mesh for the entire lattice. These methods primarily differ in how the triangular meshes for the nodes are generated. In the work by Chougrani et al. 2017_light-weight-triangulation , spherical nodes and strut-strut intersection curves are sampled, and the resulting points are connected to form a triangular mesh that approximates the nodal shape. (Notably, the intersection curves do not need to be pre-calculated; instead, sample points along the curves are directly obtained.) In contrast, Verma et al. verma2020combinatorial use convex hulls to approximate the nodal shape and stitch these hulls directly with the meshed, open struts. Similarly, Wu et al. wu2020chocc employ convex hulls of co-spherical circles to approximate the nodal shapes, resulting in improved shape quality, and this method is not restricted to triangular meshes. By shifting from

Read the complete text