基于深度学习的多波段红外隐身超表面设计
Multi-Band Infrared Stealth Metasurface Design Based on Deep Learning
王磊 1董健 1郑崇 2谢博伟 3张文杰 1翟载腾 4刘林华1
作者信息
- 1. 山东大学能源与动力工程学院,济南 250061;山东大学光-热辐射研究中心,青岛 266237
- 2. 光辐射重点实验室,北京 100854
- 3. 山东大学能源与动力工程学院,济南 250061;上海卫星工程研究所,上海 200000
- 4. 上海卫星工程研究所,上海 200000
- 折叠
摘要
现代战争中,军事目标的红外隐身对于提高其生存和突防能力具有重要作用.本文面向多目标波段的红外隐身,即近红外激光主动探测波长下(1.54 μm)具高吸收特性、两个大气窗口波段具有低发射并兼顾非大气窗口波段的辐射散热需求,结合严格耦合波分析法和深度学习方法逆向设计Ge/Ag/Ge多层膜圆孔超表面结构,针对优化得到的结构参数,通过分析潜在的物理机制和不同角度下的发射、吸收特性,达到了多波段协同隐身的目的.
Abstract
In modern warfare,infrared stealth of military objectives plays an important role in im-proving their survival and breakout capabilities.In this paper,aiming at the multispectral infrared stealth,i.e.,narrow-band high absorption at the active detection wavelength of near-infrared laser(1.54 μm),low emission in the two atmospheric window bands and taking into account the radia-tive heat dissipation requirements in the non-atmospheric window bands,we combine the rigorous coupled-wave analysis and deep learning methods to reversely design the Ge/Ag/Ge multilayer film circular hole metasurface structure.By analyzing the potential physical mechanisms and the emis-sion and absorption characteristics at different angles for the optimized structure parameters,the multi-band synergistic stealth is achieved.
关键词
红外隐身/超表面/神经网络/严格耦合波分析Key words
infrared stealth/metasurface/neural network/rigorous coupled-wave analysis引用本文复制引用
基金项目
国家自然科学基金资助项目(51906128)
上海市青年科技英才扬帆计划(20YF1446800)
出版年
2024