Fracture process simulation and crack resistance behavior analysis of transition-layer ceramic coating based on real image reconstruction model

Due to the intricate nature of the structures and interface morphologies within transition-layer ceramic coatings, the investigation of microscopic dynamic fracture behavior presents a significant challenge. In this work, we combined the deep learning-based image recognization model with peridynamic method to analyze the dynamic process of crack propagation in transition-layer and dual-layer coatings, and evaluate their crack resistance. Through the establishment of high-precision discrete models based on the real coating morphology and simulation of fracture behavior under tensile loading conditions, it was found that the transition structure leads to more dispersed crack initiation positions, followed by forming a complex network-like crack propagation pathway which involves the connection of isolated cracks with each other. The crack resistance behavior of the coating was evaluated by quantifying the number of broken peridynamic bonds and calculating the fracture energy of the coating. The results showed that, compared with dual-layer coatings, transition-layer coatings exhibit lower fracture energy in the initial stage of failure, followed by an increase and exceeding that of the dual-layer coating. The transition structure facilitates crack initiation while impeding crack growth and delaying the failure process.