PulseExploreJournal ClubDebatesTrendingResearchersJournals
Instagram
HomeExploreJournal ClubTrending
Synapse
⌘+K
Synapse
March 5, 2026Chinese Science Bulletin (Chinese Version)0 citationsOpen Access

Optimization for Radar-Infrared Compatible Stealth Structures with a Kriging Surrogate Model

View Full Paper
MLMeiqin lu:PGPengfei GuDDDazhi Ding

Key Points

  • The research aims to optimize multi-parameter stealth structures for radar and infrared applications using machine learning methods.
  • Introduced a variable fidelity model for initial sampling to reduce computational costs.
  • Adaptive training strategies were employed during surrogate model training.
  • Implemented a hybrid strategy for global multi-objective optimization.
  • Designed and fabricated a composite metasurface for stealth purposes.
  • Achieved over 90% efficient broadband absorption in the 2.8 to 20 GHz frequency range.
  • Controlled infrared emission to a low level of approximately 0.25 by optimizing metal patch duty cycle.
  • Demonstrated effectiveness in reducing optimization training time for stealth design.

Abstract

针对现代多频谱兼容隐身技术中多设计参数与多性能目标带来的复杂电磁优化挑战, 本文基于在线模型的机器学习辅助优化方法来实现超表面隐身结构的高效设计. 首先, 通过引入可靠的可变保真度模型进行初始采样, 有效降低全波仿真带来的计算成本. 在代理模型训练过程中, 自适应地选择训练策略来更新模型, 缓解因设计维数增加而引起的训练时间增长问题. 此外, 结合全局多目标优化的混合策略, 进一步提升算法在复杂解空间中的收敛性能. 基于上述方法, 我们成功设计并制备出一种雷达/红外兼容隐身的复合超表面. 该结构采用红外低发射层与雷达吸波层一体化架构, 通过在功能层表面印刷高阻碳浆图案并优化其尺寸, 在2.8 ~ 20 GHz频带内实现了高于90%的宽带高效吸收. 同时, 通过提高表面金属贴片占空比, 将结构在红外波段的发射率控制在约0.25的低水平. 实验结果表明, 所设计结构兼具优异的宽频雷达吸波性能与低红外发射率特征. 本研究验证了所提优化方法在雷达/红外兼容隐身设计中的高效性, 大幅缩短了优化训练时间, 为多频谱兼容隐身材料的设计提供了有效解决方案.

Ask AI
Helpful
Bookmark
Share
View Full Paper

Cite This Study

lu: et al. (2026) studied this question.

synapsesocial.com/papers/69a91e12d6127c7a504c1aaehttps://doi.org/10.1360/csb-2025-5846
Ask AI
Helpful
Bookmark
Share
View Full Paper