Transient thermal shock modeling of layered TBC/CFRP composites using physics-informed neural networks

Thermal barrier coatings (TBCs) are applied to mitigate the thermal degradation of substrate materials in high-temperature environments and have seen limited application in polymer matrix composites (PMCs) due to the latter’s poor thermal stability. This study presents a novel TBC design for carbon fiber reinforced polymer (CFRP) composites using a physics-informed neural network (PINN). A PINN framework is developed to simulate transient heat conduction in both unprotected and TBC-protected CFRP composites, with validated with experimental data. The effects of TBC layering and interfacial thermal resistance on its overall thermal resistance performance are also analyzed. The proposed PINN model accurately predicts temperature distributions in CFRP and TBC-protected CFRP composites. It estimates CFRP backside equilibrium temperatures of predicts CFRP backside equilibrium temperatures of 239.9°C for the uncoated laminate and 216–230.5°C for TBC/CFRP systems, depending on coating architecture and interfacial resistance. Excellent agreement (within ~1%) with experimental observations is achieved when temperatures are evaluated over the representative averaging domain, while larger deviations for other averaging domains reflect spatially non-uniform heating and experimental uncertainties. This work provides the foundation for designing TBCs with superior thermal barrier performance. The PINN-based TBC design presents a potentially useful approach for thermal management in aerospace composite applications under extreme conditions.