Turbulent structures in planar gravity currents and their influence on the flow dynamics

Abstract

[1] Direct numerical simulations (DNS) of planar gravity current in the Boussinesq limit have been conducted with the objective of identifying, visualizing, and describing turbulent structures and their influence on the flow dynamics. The simulations are performed for Reynolds numbers of Re = 8950 and Re = 15,000 with 31-and 131-million grid point resolutions, respectively. This range of Reynolds numbers ensures fully developed turbulent gravity currents, which have never been simulated before using DNS. The flow develops zones with different turbulence characteristics, which eventually interact with each other. The near-wall bottom flow resembles boundary layer flow with several hairpin-like vortices oriented in the direction of the flow and preferential patterns of low-and high-speed streaks. The separation between low-speed streaks at the front scales with the lobe size, which is about 200 wall units for Re = 15,000. Upstream of the front, the separation between low-speed streaks scales with the well-accepted value of 100 wall units for turbulent boundary layers. These patterns have associated regions of low and high bottom shear stresses with implications for sediment erosion and bed load transport. Most of the erosive power of the flow is found in the gravity current front. The interface between heavy and light fluids rolls up by baroclinic generation of KelvinHelmholtz vortices, which undergo sudden breakup and decay to small-scale turbulence. The effect of turbulence and three-dimensionality on the flow dynamics is addressed by comparing two-and three-dimensional simulations. Three-dimensional simulations present active mechanisms that undermine the strong flow coherence, comparing well with experimental observations.

Particulate gravity currents along V-shaped valleys

“…This difference in the dynamics of the interface between the gravity current and the ambient fluid may be the cause of the apparent difference in the turbulence in the flat and V-shaped-valley experiments. Finally, we note that the numerical simulations of Cantero et al (2008) reports that the speed of the front is predicted equally well by two and three-dimensional simulations during the initial and slumping phases but begin to diverge in the inertial and viscous phases.…”

“…The analog three-dimensional direct numerical simulations (DNS) of planar gravity current in the lock-exchange configuration and finite volume release (Cantero et al 2008) show this in detail for Reynolds numbers in the range 8950 to 15 000 using between 31×10 6 and 131×10 6 grid points. In particular the numerical calculations show that the downstream velocity is partly determined by the complex interaction between Kelvin-Helmholtz vortices.…”

Caracterización experimental del campo lejano de los vertidos de salmuera al mar

“…Muchos de estos estudios numéricos y experimentales se han centrado en el frente de la corriente y su esparcimiento [16][17], habitualmente generados por una descarga puntual controlada [10], mientras que otros analizan efectos tangibles en estas corrientes de amplio rango de escalas espaciales y temporales, como el efecto Coriolis [18]. Finalmente, existen trabajos focalizados en el análisis del cuerpo estable de estas corrientes de gravedad comúnmente generadas por descargas constantes de caudal lateralmente confinadas, es decir, bidimensionales [19][20][21].…”

A two-dimensional inviscid model of the gravity-current head by Lagrangian block simulation

“…The experimental results were correlated mostly in terms of depthaveraged internal hydraulic models. The Direct Numerical Simulations (DNS) model and the Large Eddy Simulations (LES) model have been used by Hartel et al [9,10], Thomas et al [25], Cantero et al [3,7] and Ooi et al [19,20] to study some aspect of the gravity current in detail. Well-resolved DNS can provide all turbulence-length scales down to the dissipative range.…”

“…LES with correct modelling of the sub-grid scales can be an efficient alternative, covering a wider range of parameter space utilizing affordable computation resources. DNS were conducted by Cantero et al [3,7] up to a Reynolds number of 8950. High Reynolds number LES simulations were carried out by Ooi et al [20] up to a Reynolds number of 12,600.…”