- D3.517 - Design of an AI-Based Decision Support System for Field-Tailored Personal Protective Equipment Using Multi-Layer Protective Composites

Occupational and environmental exposure to airborne allergens and irritant particles is a significant risk factor for allergic rhinitis, occupational asthma, and contact dermatitis. Protective clothing materials capable of reducing allergen penetration and minimizing secondary skin exposure may contribute to lowering sensitization and exacerbation risks. However, quantitative evaluation frameworks linking material properties to allergen barrier performance remain limited.

This study aimed to develop an AI-based material evaluation and decision support framework for multi-layer protective textiles designed to reduce allergen exposure. Mechanical stability, pore structure characteristics, and thermal durability indicators were structured into a unified material performance database. Barrier-related properties derived from BET surface analysis and thermal stability from TGA testing were integrated as predictive descriptors. A Random Forest-based model was implemented to estimate allergen-blocking performance under simulated exposure conditions. SHAP (Shapley Additive Explanations) analysis was applied to identify key material features influencing predicted barrier effectiveness.

The predictive model demonstrated stable performance across heterogeneous material groups and enabled quantitative scoring of allergen barrier potential. Feature importance analysis indicated that pore-size distribution and surface area characteristics were major contributors to predicted barrier efficiency. Multi-layer configurations with optimized porosity profiles showed improved modeled allergen-blocking performance compared with single-layer structures. The AI-supported evaluation reduced variability in material ranking and supported reproducible, condition-based selection.

The proposed AI-driven framework enables quantitative assessment and interpretable prediction of allergen barrier performance in multi-layer textile systems. By supporting data-informed material selection and configuration optimization, this approach may contribute to the development of protective garments designed to mitigate allergen exposure and potentially reduce allergic disease risk in high-exposure environments.