2026 已投稿 Energy & AISCI液冷多智能体Bayesian surrogate

Autonomous CFD Optimization of Liquid Cold Plate Geometry via Multi-Agent LLM and Bayesian Surrogate

李飞团队 · Energy & AI

论文投稿编号 EAI-D-26-00245，已投稿至 Elsevier 期刊 Energy & AI。研究面向高热流密度算力基础设施散热瓶颈，构建“多阶段自主搜索 + 贝叶斯代理优化”的自主 CFD 优化框架，实现从几何参数化、网格生成、求解执行到多目标评分的端到端自动化闭环。

该论文可为 Projects 页中的“液冷散热冷板自动化研究流水线”提供关联成果。

BibTeX

@article{li2026liquidcoldplate,
  title = {Autonomous CFD Optimization of Liquid Cold Plate Geometry via Multi-Agent LLM and Bayesian Surrogate},
  author = {Li, Fei and Team},
  journal = {Energy & AI},
  year = {2026},
  note = {Submitted, EAI-D-26-00245}
}