Deep-learning-based assessment of skin friction in wall-bounded turbulence

[EN] This work investigates the influence of classically coherent structures on wall-shear stress and energy dissipation in turbulent channel flow, utilizing direct numerical simulations (DNS) data and explainable deep learning (XDL). Sweeps, defined as regions of low streamwise velocity moving towa...

Descripción completa

Detalles Bibliográficos
Autores: Hoyas, Sergio|||0000-0002-8458-7288, Benedikt, Nils, Cremades, Andrés, Vinuesa, Ricardo
Tipo de recurso: artículo
Fecha de publicación:2025
País:España
Institución:Universitat Politècnica de València (UPV)
Repositorio:RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Idioma:inglés
OAI Identifier:oai:riunet.upv.es:10251/226179
Acceso en línea:https://riunet.upv.es/handle/10251/226179
Access Level:acceso abierto
Palabra clave:Turbulent channel flow
Coherent structures
Wall-shear stress
Energy dissipation
Direct numerical simulations (DNS)
Explainable deep learning (XDL)
Drag reduction
Turbulence-control strategies
Descripción
Sumario:[EN] This work investigates the influence of classically coherent structures on wall-shear stress and energy dissipation in turbulent channel flow, utilizing direct numerical simulations (DNS) data and explainable deep learning (XDL). Sweeps, defined as regions of low streamwise velocity moving toward the wall, are found to be the most influential structures for both energy dissipation and drag. Moreover, the volume of these key structures falls within a narrow range, making it easier to identify the most significant ones. Consequently, this work paves the way for the development of novel, highly efficient turbulence-control strategies for the reduction of drag.