FLOW paper published in Nature Communications

Our colleagues Ricardo Vinuesa and Andrés Cremades and thir collaborators have published their work titled 'Identifying regions of importance in wall-bounded turbulence through explainable deep learning' in recent issue of Nature Communications.

In this study explainable deep learning is used to assess the importance of various coherent structures in wall-bounded turbulence. The so-called Q events, which are regions of intense Reynolds shear stress, are not the most important ones, as opposed to what was believed in classical turbulence knowledge. Completely new structures are identified based on an objective definition of importance. The study of these structures has tremendous implications regarding the fundamental knowledge of turbulence and the possible strategies for flow control.

Read the article here:
www.nature.com/articles/s41467-024-47954-6