Performance Optimization Engineering of Multicomponent Absorbing Materials Based on Machine Learning.

Area of Science:

• Materials Science
• Electromagnetics
• Machine Learning Applications

Background:

• Multicomponent materials offer superior microwave-absorbing (MA) properties compared to single-component alternatives.
• Traditional design of multicomponent MA materials is hindered by complex, high-dimensional design spaces and reliance on semi-empirical rules.
• Existing methods struggle to efficiently explore the vast possibilities for optimal MA material performance.

Purpose of the Study:

• To develop a performance optimization engineering approach for accelerating the design of multicomponent MA materials.
• To overcome limitations of conventional design rules in high-dimensional spaces using sparse data.
• To identify novel Ni-based multicomponent MA materials with targeted performance characteristics.

Main Methods:

• Implemented a closed-loop system integrating machine learning with the expanded Maxwell-Garnett model and electromagnetic calculations.
• Utilized experimental feedback to refine the design process and validate material performance.
• Screened and identified materials from a near-infinite design space based on desired MA performance.

Main Results:

• Successfully designed Ni surface@carbon fiber (NiF) and NiF-based multicomponent (NMC) materials meeting specific MA performance requirements.
• Achieved target MA performance for X- and Ku-bands with ultrathin thicknesses of 2.0 mm and 1.78 mm, respectively.
• Demonstrated achievement of MA targets across S, C, and all bands (2.0-18.0 GHz) as predicted.

Conclusions:

• Performance optimization engineering provides an effective strategy for designing advanced multicomponent MA materials.
• The proposed method enables efficient exploration of vast material design spaces for tailored MA properties.
• This approach facilitates the practical application of high-performance microwave-absorbing materials.