Hybrid AI-Physics framework for predicting failure zones in 3D-Printed protective gear for contact sports

The growing use of 3D-printed safety equipment in sports highlights the need for accurate predictive modelling to ensure strength and reliability. We present AIPIF-Net, a Hybrid AI-Physics Integrated Fusion Network that combines data-driven learning with physics-based simulations to predict failure zones in 3D-printed protective gear. The approach integrates deep neural networks with finite element analysis (FEA) to model material behavior under dynamic impacts. High-resolution digital twins of protective parts were generated, and experimental impact data were used to train and validate the model. Domain adaptation techniques enhance robustness across different printing materials, lattice structures, and stress conditions. Results demonstrate that the hybrid model outperforms purely data-driven or simulation-based approaches in predictive accuracy while reducing computational costs by over 40%. Predicted failure zones closely match experimental observations, confirming reliability for real-world applications. This framework accelerates the design of lightweight, high-impact-resistant protective gear and can be extended to other types of personalized sportswear, promoting safer and more effective equipment development.