We attended the 3rd European Research Community on Flow, Turbulence and Combustion (ERCOFTAC) Workshop on Machine Learning for Fluid Dynamics, held at Centrum Wiskunde & Informatica (CWI), Amsterdam from 4–6 March 2026. The workshop brought together researchers working on machine learning applications in fluid dynamics, including data-driven closure modelling, reduced-order modelling, and surrogate modelling.
From our group Zhuolin, Baris, and Haochen presented our latest research. Zhuolin spoke on learning a heat-flux closure for jet-in-crossflow, Baris presented our work on turbulence generation and data assimilation in wall-bounded flows using a latent diffusion model, and Haochen introduced our recent efforts to develop a data-driven unified turbulence model through multi-objective learning. Professor Xiao also chaired the Turbulence Modeling–RANS2 session on the second day. The workshop was a valuable opportunity to share our work, exchange ideas with the community, and learn more about the state of the art in machine learning for fluid mechanics.
