Explainable Neural Network Framework for Strength Prediction of Epoxy/Graphene Oxide Composites for UAV Structures Using Optimisable ANN and XAI

Abstract

This study presents an explainable and optimisable neural network framework for strength assessment of epoxy/graphene oxide (E/GO) composites intended for unmanned aerial vehicle (UAV) structural applications. While earlier investigations have demonstrated the feasibility of artificial neural networks (ANNs) for predicting composite strength, such approaches largely remain black-box models with limited physical interpretability. The present work addresses this limitation by integrating an optimisable ANN architecture with Explainable Artificial Intelligence (XAI) techniques to move beyond prediction accuracy toward mechanism-aware inference.