Self-healing shape nanocomposites for structural longevity in aerospace applications based on interpretable artificial intelligence
摘要
In order to increase the structural durability of aerospace components, specifically aircraft wing panels, a multifunctional self-healing nanocomposite system has been developed and optimized in this work. To improve mechanical strength, electrical conductivity and potentially enable piezoelectric responsiveness, carbon nanotubes (CNTs), graphene oxide (GO), and zinc oxides (ZnO) were added to a dual-polymer matrix that included epoxy resin and shape memory polyurethane (SMPU). While ZnO is known for its piezoelectric properties, healing was thermally activated in this study, and stress-responsive behavior was not experimentally investigated. A novel Nano-Engineered Filament Winding with Embedded Healing Microvascular layers (NEW-HM) technique was employed to fabricate the composite, enabling precise integration of healing channels and uniform nanoparticle dispersion. Response Surface Methodology with Box-Behnken Design (RSM-BBD) was employed to evaluate the influence of key parameters nanoparticle concentration, SMPU-to-epoxy ratio, healing activation temperature, and crack width on healing efficiency, tensile strength retention, crack closure rate, and fatigue resistance. Having a healing efficiency of 91.88%, tensile strength retention of 85.99%, a crack closure rate of 1.03%, and fatigue resistance of 909,348 cycles, the results demonstrated significant improvements. Transformer-based Surrogate Modeling, interpretable AI, and Adaptive Bayesian Optimization were combined in a hybrid TSM-LIME_ABO technique to further enhance predictive accuracy and parameter sensitivity analysis. This integrated approach offers transformative potential for next-generation aircraft structures by providing a strong framework for creating smart, self-healing materials that offer better durability.