Topology optimisation for design and additive manufacturing of functionally graded lattice structures using derivative-aware machine learning algorithms

Wu, Chi; Luo, Junjie; Zhong, Jingxiao; Xu, Yanan; Wan, Boyang; Huang, Wenwei; Fang, Jianguang; Steven, Grant P.; Sun, Guangyong; Li, Qing

doi:10.1016/j.addma.2023.103833

Cited by 7 publications

(1 citation statement)

References 56 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…, 2021). Wu et al. (2023) incorporated a derivative-aware machine learning algorithm into multiscale topology optimization approach for uniform strain pattern design.…”

Section: Introductionmentioning

confidence: 99%

A novel design method based onmulti–objective optimization for graded lattice structure by additive manufacturing

Li,

Zhu,

Fan

et al. 2024

RPJ

View full text Add to dashboard Cite

Purpose While performance demands in the natural world are varied, graded lattice structures reveal distinctive mechanical properties with tremendous engineering application potential. For biomechanical functions where mechanical qualities are required from supporting under external loading and permeability is crucial which affects bone tissue engineering, the geometric design in lattice structure for bone scaffolds in loading-bearing applications is necessary. However, when tweaking structural traits, these two factors frequently clash. For graded lattice structures, this study aims to develop a design-optimization strategy to attain improved attributes across different domains. Design/methodology/approach To handle diverse stress states, parametric modeling is used to produce strut-based lattice structures with spatially varied densities. The tailored initial gradients in lattice structure are subject to automatic property evaluation procedure that hinges on finite element method and computational fluid dynamics simulations. The geometric parameters of lattice structures with numerous objectives are then optimized using an iterative optimization process based on a non-dominated genetic algorithm. Findings The initial stress-based design of graded lattice structure with spatially variable densities is generated based on the stress conditions. The results from subsequent dual-objective optimization show a series of topologies with gradually improved trade-offs between mechanical properties and permeability. Originality/value In this study, a novel structural design-optimization methodology is proposed for mathematically optimizing strut-based graded lattice structures to achieve enhanced performance in multiple domains.

show abstract

“…, 2021). Wu et al. (2023) incorporated a derivative-aware machine learning algorithm into multiscale topology optimization approach for uniform strain pattern design.…”

Section: Introductionmentioning

confidence: 99%