Dielectric constant of boro silicate aluminum glasses using AI and radar sensor

Abstract

This study presents a comprehensive analysis of the dielectric constant (ε_r​) of glasses with varying chemical compositions, utilizing artificial intelligence (AI) predictions and experimental validation. An AI model, trained on over 100 glass compositions, was employed to predict ε_r based on compositional inputs such as SiO₂-Na₂O-CaO-B₂O₃-Al₂O₃. For the first time, the BGT60TR13C radar sensor was adapted for non-contact dielectric constant measurements, offering a novel methodology for material characterization. To validate the AI predictions and sensor values, two additional experimental techniques were employed: an LCR meter for capacitance-based measurements and the parallel plate capacitor method. Results showed excellent agreement among all methods, confirming the reliability of AI predictions and the accuracy of the experimental techniques. Furthermore, the dielectric constant increased with higher concentrations of network modifiers and secondary network formers. This study highlights the integration of AI and advanced sensing technologies as a powerful hybrid framework for rapid and accurate material characterization, introducing the radar sensor as an innovative tool for dielectric measurements.