A Scientometric Investigation of Artificial Intelligence for Fluid Mechanics: Emerging Topics and Active Groups

Abstract

Fluid mechanics, as one of the core disciplines of engineering technology, faces dual challenges of theoretical limitations and high computational costs when addressing complex flow problems, promoting the integration of artificial intelligence into the field. This paper systematically reviews the paradigm shift in fluid mechanics driven by AI technologies. Through a bibliometric analysis of 18,776 relevant publications up to May 2025 from the Web of Science Core Collection, we investigated the interdisciplinary intersection between two keyword sets: 22 artificial-intelligence-related terms (including Machine Learning, Neural Networks, Symbolic Regression, Reinforcement Learning, etc.) and 22 fluid-mechanics-related terms (including CFD, Aerodynamics, Fluid-Structure-Interaction, etc.). Our systematic investigation employs a modeling-computation-application framework to analyze this interdisciplinary field, presenting the most active countries, institutions, research groups, and leading scholars while mapping the overall research landscape. Key findings include: (1) From 2000 to 2014, the number of publications and citations grew slowly, but since 2014, both have shown a significant increase. (2) AI applications in fluid mechanics are categorized into modeling (dimensionality reduction and knowledge discovery, machine learning-based turbulence modeling), computation (reduced-order models and feature extraction, machine learning-based CFD simulations), and applications (data fusion and reconstruction, multi-physics and multi-disciplinary coupling, intelligent flow control, aerodynamic shape optimization). (3) The United States dominates foundational research, while China leads in technological applications. (4) Multi-dimensional network analysis—including author, institutional, and national collaboration networks, along with co-citation patterns—reveals a pronounced spotlight effect among leading global universities and research teams. The twenty representative teams selected from them have important value in the academic and engineering fields and deserve further attention. This study provides a clear and comprehensive framework for researchers in the field and offers guidance for future research in intelligent fluid mechanics.