Answers are grounded in this paper and cite their sources.
Rapid advancement of MXene-reinforced epoxy composites has enabled new possibilities in structural health monitoring (SHM); however, existing studies remain limited by slow finite-element-based predictions, low generalization across microstructure variations, and an inability to accurately map morphology-dependent electromechanical behaviour. To address these gaps, this study proposes a Morphology-Aware Convolutional Sensing Framework driven by a novel MXene-Adaptive Convolutional Sensing Layer (MACSL), designed to capture filler-density-dependent and alignment-dependent sensitivity patterns. The framework is implemented using Python with TensorFlow, enabling efficient training, real-time execution, and reproducible experimentation. A synthetic dataset of 10,000 MXene–epoxy microstructure images with corresponding simulated strain and microdamage labels is generated through controlled parametric modelling of filler dispersion, cluster formation, and interfacial behaviour. The proposed MACSL-enhanced CNN achieves a significant improvement over conventional CNN and statistical models, yielding a 96.8% prediction accuracy, representing an increase of 1.1–3.2% compared with baseline deep models and a 3–6% gain over classical ML methods. Real-time simulation tests demonstrate stable electromechanical response tracking with a 13.4 ms latency, enabling near-instantaneous SHM prediction. The method effectively correlates microstructural heterogeneity with localized strain concentration, achieving enhanced microdamage identification even under noise-altered inputs. Experimental emulation reveals strong alignment between predicted and simulated strain fields, confirming the model's capability to generalize across filler morphologies. Overall, this study demonstrates that integrating adaptive morphology-aware feature extraction with deep learning significantly enhances predictive capability, computational efficiency, and material-level interpretability, establishing a powerful framework for next-generation SHM of MXene-based composites.
No takes yet. Share an insight, caveat, or question.
Lakshmaiya Natrayan (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: