Predicting Interfacial Pull-Out Performance of Nano-B₄C/Aramid Material with Stage-Wise Physics-Guided Machine...

Area of Science:

• Materials Science
• Mechanical Engineering
• Computational Modeling

Background:

• Interfacial yarn pull-out is crucial for energy dissipation in soft ballistic materials.
• Experimental characterization of this complex, multistage process is time-consuming and challenging.
• Understanding interfacial mechanics is key to optimizing ballistic material performance.

Purpose of the Study:

• To develop a physics-guided machine learning (PG-HML) framework for predicting yarn pull-out behavior.
• To model the pull-out response of nano hexagonal boron carbide (nh-B4C)-functionalized para-aramid fabrics.
• To enable efficient virtual screening and reduce experimental effort in material design.

Main Methods:

• A stage-wise PG-HML framework was developed, decomposing pull-out into crimp extension, interlacement rupture, and stick-slip sliding.
• Physics-based constraints (continuity, admissibility) were integrated to ensure mechanical consistency and robustness.
• The model was trained and validated using an experimentally constrained dataset of nh-B4C/aramid fabrics.

Main Results:

• The PG-HML framework achieved high predictive accuracy (R² > 0.98) for crimp extension and rupture stages.
• The model effectively captured complex rupture transitions and post-peak stick-slip evolution.
• Increased nh-B4C content enhanced interfacial friction, rupture resistance, and energy dissipation.

Conclusions:

• The PG-HML framework provides a high-fidelity surrogate model for interfacial pull-out behavior.
• This approach significantly reduces experimental requirements for designing advanced soft ballistic materials.
• The study demonstrates the potential of physics-guided machine learning in materials discovery and optimization.