Scholarlink

<p>Additive manufacturing of soft elastomers such as liquid silicone rubber (LSR) offers promising opportunities for personalized medical models, soft robotics, and wearable devices. However, their low modulus and slow curing often lead to material sagging, delamination, or collapse during the printing process, particularly in unsupported or high-curvature regions. To overcome these limitations, we present a lightweight closed-loop optimization framework that couples a geometric fidelity metric, the three-dimensional Intersection-over-Union (IoU), with a Bayesian optimization algorithm for automated process control. This system enables real-time feedback-guided adjustment of the feed rate without relying on manual trial-and-error or hardware modification. Within six iterations, the system autonomously converges to an optimal feed rate (19&#xa0;mm/s), increasing the IoU from 0.2199 to 0.8414. The IoU metric demonstrates high sensitivity to geometric deviations and serves as a robust performance indicator for guiding optimization. Our results confirm the feasibility of support-free, high-fidelity printing of complex hollow LSR structures. With fewer than 100 lines of code, the framework is easily embedded into existing direct ink writing (DIW) platforms and is extendable to multi-parameter optimization scenarios. This work offers a generalizable, data-driven paradigm for intelligent fabrication of soft structures, with potential applications in digital prototyping, biomedical engineering, and AI-assisted manufacturing.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Closed-Loop Bayesian Optimization for High-Fidelity 3D Printing of Liquid Silicone Rubber Structures

  • 期刊论文

  • 发表日期：

    2025年11月5日

查看全文
  • Zhihan Liu,
  • Shihao Wang,
  • Jiyuan Han,
  • Rui Yang,
  • Chenlu Wang,
  • Chuanjian Zhou,
  • Teng Long

摘要

Additive manufacturing of soft elastomers such as liquid silicone rubber (LSR) offers promising opportunities for personalized medical models, soft robotics, and wearable devices. However, their low modulus and slow curing often lead to material sagging, delamination, or collapse during the printing process, particularly in unsupported or high-curvature regions. To overcome these limitations, we present a lightweight closed-loop optimization framework that couples a geometric fidelity metric, the three-dimensional Intersection-over-Union (IoU), with a Bayesian optimization algorithm for automated process control. This system enables real-time feedback-guided adjustment of the feed rate without relying on manual trial-and-error or hardware modification. Within six iterations, the system autonomously converges to an optimal feed rate (19 mm/s), increasing the IoU from 0.2199 to 0.8414. The IoU metric demonstrates high sensitivity to geometric deviations and serves as a robust performance indicator for guiding optimization. Our results confirm the feasibility of support-free, high-fidelity printing of complex hollow LSR structures. With fewer than 100 lines of code, the framework is easily embedded into existing direct ink writing (DIW) platforms and is extendable to multi-parameter optimization scenarios. This work offers a generalizable, data-driven paradigm for intelligent fabrication of soft structures, with potential applications in digital prototyping, biomedical engineering, and AI-assisted manufacturing.