Document Type

Article

Source Publication Title

Proceedings of The American Society for Composites; Thirty-seventh Technical Conference

Abstract

Due to continual exposure to moist environment, environmental degradation is a major threat to structural composites throughout their service life. Impedance Spectroscopy (IS)/Broadband Dielectric Spectroscopy (BbDS) is a reliable nondestructive method that has been used in polymer industries for dielectric characterization of material. While moisture absorption causes irreversible changes in the polymer matrix composites, it also alters the electrical properties of the system which is detectable using BbDS. Since the physics that drives the change in electrical characteristics is driven by the modifications imposed by water molecules (whether free or bound), both events can be linked. A dielectric spectrum over a broad frequency range can give an accurate estimation of moisture absorption in structural composites. In this work, machine learning (ML) models (Discriminant Analysis (DA) and Support Vector Machine (SVM) with Principal Component Analysis (PCA)) were developed that incorporate the dielectric data from BbDS to predict the current material state due to hygrothermal ageing. The models not only can detect the presence of moisture in composites, but also can predict the current saturation state of the material based on their dielectric properties. It uses the dielectric spectra and composite’s geometry as features to train supervised learning classification algorithms. The ML models developed in this study shows high accuracy and efficiency in predicting the current moisture state of the composite test samples due to hygrothermal ageing. [https://doi.org/10.12783/asc37/36451]

Disciplines

Engineering | Materials Science and Engineering

Publication Date

1-1-2022

Language

English

Download

Share

COinS