人工智能在氧化物/氧化物陶瓷基复合材料性能预测与优化研究中的应用

Oxide/oxide ceramic matrix composites (Ox/Ox CMCs), owing to high specific strength, excellent oxidation resistance, and thermal stability, hold significant potential for applications in aerospace and high-temperature energy systems. However, their in-service behavior is governed by the coupled effects of multiscale architectures and multiphysics interactions involving thermal, mechanical, and environmental factors. Traditional experimental and numerical approaches face limitations in terms of efficiency and generalizability when addressing such complexities. In recent years, artificial intelligence (AI) methods have demonstrated remarkable capabilities in nonlinear modeling, high-dimensional data analysis, and rapid optimization, offering novel strategies for the intelligent prediction and design of Ox/Ox CMCs. This review systematically examines the failure mechanisms and influencing factors associated with key performance aspects of Ox/Ox CMCs, including tensile strength, impact resistance, fatigue behavior, creep response, and thermal stability. A detailed survey of AI applications in mechanical property prediction and design workflow optimization is presented. For prediction tasks, representative modeling paradigms are summarized in relation to distinct failure modes, emphasizing problem-driven model construction strategies. For design applications, a four-stage framework- comprising design space definition, performance modeling, inverse design, and iterative validation- is adopted to classify applicable AI techniques and review representative case studies. Furthermore, critical challenges are identified in data quality, multiscale information integration, multiphysics coupling, and model interpretability. Strategies incorporating physical priors into AI frameworks are proposed, and future research opportunities are discussed, including generative modeling, multiscale constitutive behavior modeling, process-property co-optimization, and intelligent diagnostics.