Failure prediction of thermal barrier coatings on turbine blades under calcium-magnesium-alumina-silicate corrosion and thermal shock

Failure of thermal barrier coatings (TBCs) can reduce the safety of aero-engines. Predicting the lifetime of TBCs on turbine blades under real service conditions is challenging due to the complex multiscale computation required and the chemo-thermo-mechanically coupled mechanisms involved. This paper proposes a multiscale deep-learning method for TBC failure prediction under typical thermal shock conditions involving calcium-magnesium-alumina-silicate (CMAS) corrosion. A micro-scale model is used to describe local stress and damage with consideration of the TBC microstructure and CMAS infiltration and corrosion mechanisms. A deep learning network is developed to reveal the effect of microscale corrosion on TBC lifetime. The modeled spalling mechanism and area are consistent with the experimental results, with the predicted lifetime being within 20% of that observed. This work provides an effective method for predicting the lifetime of TBCs under real service conditions.