Electromagnetic-mechanical collaborative design of high-performance electromagnetic sandwich metastructure by machine learning based genetic optimization

Abstract

Electromagnetic sandwich metastructure (ESM) consisting of different functional layers, has gained increasing attention in radiation prevention and radar stealth. However, the current ESM design is primarily based on the separation design method, ignoring electromagnetic-mechanical interactions between layers. Thus, subject to thin thickness constraint of ESM, it is a great challenge to achieve broadband microwave absorption (MA) and excellent mechanical performance simultaneously. To address this issue, an electromagnetic-mechanical collaborative design approach was proposed for ESM. The relations of geometric-electromagnetic and geometric-mechanical of ESM were first identified by machine learning. They were then integrated with the heuristic genetic optimization algorithm to perform the highly efficient design. The designed ESM can achieve 36.4 GHz effective absorption bandwidth (EAB, RL ≤ −10 dB), 334.3 MPa equivalent bending strength and 83 MPa compressive strength with a thickness of 9.3 mm, possessing the widest EAB and highest bending strength within the current available MA structures (thickness less than 9.5 mm). The proposed approach provides an efficient tool for the design of electromagnetic-mechanical optimal ESM.