Multi-scale optimization of selectively magnetized isotropic carbon fiber fabrics for microwave absorption using machine learning

A synergistic collaboration between microscopic analysis and macroscopic engineering is essential to attain optimal scientific properties. Traditional material design methods are ineffective in navigating the vast and infinite design space of cross-scale systems. Fortunately, machine learning provides an efficient and time-saving scientific solution for exploring the boundless possibilities offered by multi-component and structural materials systems. Crucially, the development of theoretical predictions, supported by theoretical calculations and finite element simulations, will facilitate the global optimization of materials. In this study, the structure and component of selective magnetic surface @ CF fabric are screened and balanced simultaneously using target-driven optimization frame. In order to achieve dynamic adjustment of micro surface and macro structure parameters, the relative factors are translated into codes and introduced into the design. And simulation, experimental feedback and corrections are established. The flexible CCGF shows a broad effective absorption bandwidth with 8.2 GHz with a thickness of only 1.76 mm and a density of 8.8 × 10−4 kg cm−3. Step-by-step optimizations and summaries are performed for data analysis, together with finite element simulation-based electromagnetic and loss field analysis and theoretical analysis to accelerate understanding, paving the way for development of novel materials for both functional and structural applications.