A new model and direct slicer for lattice structures

Abstract

This paper presents a model for generating strut-based lattice structures using topology optimization and their efficient direct slicing. These structures exhibit better physical properties and can represent the partial densities at the macro-scale level, which often appear in designs based on topology optimization. The fabrication of such large member structures with intricate geometries is possible by the additive manufacturing technologies which offer design freedom to produce the optimized parts for engineering applications. However, these structures generate millions of planer manifolds describing the strut members and result in large data files, thus making conventional procedures in additive manufacturing highly ineffective. Therefore, the design process for such structures requires efficient data manipulation and storage of the lattice topology. In the current work, a mathematical model for the strut primitive which connects two nodes in a cell is developed. Based on the proposed strut model, a structural optimization formulation is presented for lattice structures design under volume fraction constraint. A matrix-oriented compact data structure to express the lattice topology and the direct slicing algorithm which makes queries on the proposed compact data structure is presented as part of this work. The slicing kernel has been tailored for parallel implementation to handle engineering-scale applications which often consist of structures over a million struts. The article is organized into the "Introduction" section explaining the requirement and the novelty of this work. Following which, the automated design framework based on topology optimization procedure for lattice structures is given. The mathematical derivations and data structure of the strut-based lattice will be explained and the operations on model data for the direct slicing procedure are elaborated. Numerical experiments verifying the proposed method will be presented toward the end.