Simulation and Optimization of Electromagnetic Absorption of Polycarbonate/CNT Composites Using Machine Learning.

Area of Science:

• Materials Science
• Electrical Engineering
• Computational Science

Background:

• Electronic devices generate electromagnetic interference (EMI), necessitating effective shielding materials.
• Polymer-nanofiller composites offer a promising solution for EMI shielding applications.
• Predicting the electromagnetic absorption of these materials is crucial for performance evaluation.

Purpose of the Study:

• To develop an artificial intelligence (AI) system for predicting the electromagnetic absorption of polycarbonate-carbon nanotube composite films.
• To investigate the efficacy of multilayer perceptron (MLP) neural networks in EMI shielding material characterization.
• To optimize the AI system for accurate and efficient prediction of absorption index.

Main Methods:

• Development of a novel AI system utilizing 15 specialized multilayer perceptron (MLP) neural networks.
• Automated selection of appropriate MLP networks for specific sample categories.
• Hyper-parameter optimization using hold-out validation to ensure optimal performance.
• Calculation of S-parameters to determine the electromagnetic absorption index.

Main Results:

• The AI system achieved a high average accuracy of 99.7997% in predicting electromagnetic absorption.
• The system demonstrated a fast average calculation time of 0.01295 seconds.
• Polycarbonate-5 wt.% carbon nanotube composite identified as an optimal microwave absorber based on Rozanov formalism.

Conclusions:

• AI, specifically MLP neural networks, provides a highly accurate and efficient method for predicting EMI shielding performance.
• The developed AI system can significantly accelerate the design and selection of advanced EMI shielding materials.
• Polymer-nanofiller composites, particularly polycarbonate-carbon nanotube, show great potential for effective EMI shielding applications.