AI and Statistical Learning for Predictive Thermal-Fluid Systems

Thermal and fluid dynamics underpin engineering systems spanning clean energy, aerospace, transport, and industrial processes. Their analysis increasingly requires the prediction of multi-scale, multi-physics, and stochastic phenomena across wide parameter spaces. Conventional high-fidelity computational fluid dynamics and empirical correlations can be computationally demanding, difficult to generalize, or insufficient for real-time decision-making. Recent advances in artificial intelligence, Physics-Informed Neural Networks, operator learning, reduced-order modeling, and statistical inference are creating new opportunities to accelerate simulation while retaining physical consistency. However, important challenges remain concerning model interpretability, extrapolation, data scarcity, uncertainty quantification, and the reliable integration of physical laws with heterogeneous observations. Further investigation is needed to establish mathematically rigorous, scalable, and trustworthy approaches for complex thermal-fluid systems.This Research Topic aims to advance AI, machine learning, and statistical methodologies for modeling, analyzing, and controlling thermal and fluid systems. It will highlight methods that combine data-driven learning with governing transport principles to improve prediction, computational efficiency, and generalization across complex flow and heat-transfer regimes. Contributions may address how physics constraints can enhance model reliability; how uncertainty, sensitivity, and stochasticity can be quantified; and how learned models can support optimization, control, and real-time forecasting. The Research Topic also seeks to promote collaboration among thermal-fluid engineers, applied mathematicians, statisticians, and data scientists, with emphasis on interpretable and practically deployable solutions for high-dimensional and multi-physics applications.To gather further insights in AI-enabled modeling and statistical analysis of thermal and fluid systems, we welcome articles addressing, but not limited to, the following themes:- Physics-Informed Neural Networks and other physics-constrained learning methods for thermal-fluid applications- Operator learning, including DeepONet and Fourier Neural Operators, for parametric and multi-scale systems- Statistical correlation, sensitivity analysis, uncertainty quantification, and Bayesian inference- Data assimilation and hybrid modeling using experimental, observational, and simulation data- Reduced-order modeling, dynamic mode decomposition, and surrogate modeling for complex transport phenomena- AI-driven prediction, control, and optimization of heat exchangers, energy-conversion systems, and turbulent flows- Explainable and interpretable AI for physical discovery and model validation- Robustness, generalization, and error estimation in data-driven thermal-fluid models- Applications to environmental flows, renewable energy, aerospace systems, and industrial transport processes Topic editors: Bhargavi D, Ramjee Repaka, Ameeya Kumar Nayak, Deepika Neela, S. Swayamjyoti Source: https://www.frontiersin.org/research-topics/85453/ai-and-statistical-learning-for-predictive-thermal-fluid-systems