Artificial intelligence in fluid dynamics and thermal transport: A comprehensive review of methods, challenges, and emerging applications

Artificial intelligence (AI) is a transformative tool in fluid dynamics and thermal transport, unlocking new possibilities for modeling, prediction, diagnostics, and system optimization. AI-driven approaches, ranging from deep learning and machine learning to physics-informed neural networks, are increasingly combined with traditional methods to address challenges that are difficult to solve using solely physics-based models. Notably, recent developments have demonstrated the potential of AI in reconstructing turbulent flows, enhancing heat transfer performance, and enabling real-time simulations across a wide range of thermal-fluid systems. This includes complex configurations, such as multiphase flows and compact heat exchangers, where conventional modeling techniques often have limitations due to nonlinear interactions, multiscale behaviors, and geometric complexities. This review provides a structured synthesis of current advances for AI thermal-fluid sciences. Contributions are categorized by flow regimes, such as laminar, turbulent, and multiphase, and transport phenomena, including conduction, convection, radiation, and phase change. Key technical challenges, such as the scarcity of high-fidelity datasets, robust generalization across varying flow and boundary conditions, and integration of physical laws into data-driven frameworks are considered. Finally, emerging research directions with strong potential to accelerate innovation in the field, including AI-assisted turbulence modeling, flow reconstruction from sparse measurements, data-driven design of high-efficiency heat exchangers, and intelligent control of multiphase and reactive flow systems are presented.