Near-interface flow modeling in large-eddy simulation of two-phase turbulence
The smallest hydrodynamic length scales in two-phase turbulence are located at the interface between phases, or fluids, as a result of two-way coupling phenomena. Typically, these interface-generated scales are several times smaller than the dissipative scales in the surrounding bulk flow identified...
| Autores: | , , , |
|---|---|
| Tipo de recurso: | artículo |
| Fecha de publicación: | 2020 |
| País: | España |
| Institución: | Universitat Politècnica de Catalunya (UPC) |
| Repositorio: | UPCommons. Portal del coneixement obert de la UPC |
| Idioma: | inglés |
| OAI Identifier: | oai:upcommons.upc.edu:2117/328417 |
| Acceso en línea: | https://hdl.handle.net/2117/328417 https://dx.doi.org/10.1016/j.ijmultiphaseflow.2020.103406 |
| Access Level: | acceso abierto |
| Palabra clave: | Turbulence Large-Eddy simulation Near-interface flow modeling Two-phase flow Turbulència Àrees temàtiques de la UPC::Enginyeria mecànica::Mecànica de fluids |
| Sumario: | The smallest hydrodynamic length scales in two-phase turbulence are located at the interface between phases, or fluids, as a result of two-way coupling phenomena. Typically, these interface-generated scales are several times smaller than the dissipative scales in the surrounding bulk flow identified by Kolmogorov’s 1941 theory. Consequently, to properly capture these interface-generated small scales with sufficiently fine resolutions, the computational cost of performing large-eddy simulations of two-phase turbulent flow increases significantly from its (single-phase) theoretical optimum and toward values on the order of the direct numerical simulation of turbulence. Therefore, to maintain the cost of scale- resolving approaches linear with respect to the Reynolds number, this work investigates the modeling of the small-scale fluid motions in the vicinity of the viscous near-interface region of two-phase turbulent flows. Given the resemblance between the flow structures in the near-interface regions and those found in the boundary layers of turbulent wall-bounded flow, the modeling methodology proposed is inspired by ideas developed for turbulent flows interacting with solid walls, but modified to capture slip-velocity effects between phases. The performance of the approach is a priori assessed by utilizing data from direct numerical simulations of decaying isotropic turbulence laden with droplets of super-Kolmogorov size, demonstrating its computational feasibility and potential for reducing the cost of large-eddy simulation studies of two-phase turbulence. |
|---|