Hostname: page-component-7b7567c8c-7pwfn Total loading time: 0 Render date: 2026-11-16T02:45:25.294Z Has data issue: false hasContentIssue false

Mixing efficiency in run-down gravity currents

Published online by Cambridge University Press:  15 November 2016

G. O. Hughes*
Affiliation:
Department of Civil and Environmental Engineering, Imperial College London, South Kensington, London SW7 2AZ, UK
P. F. Linden
Affiliation:
Department of Applied Mathematics and Theoretical Physics, Centre for Mathematical Sciences, University of Cambridge, Wilberforce Road, Cambridge CB3 0WA, UK
*
†Email address for correspondence: g.hughes@imperial.ac.uk

Abstract

This paper presents measurements of mixing efficiency of the two counter-flowing gravity currents created by symmetric lock exchange in a channel. The novel feature of this work is that the buoyancy Reynolds number of the currents is higher than in previous experiments, so that the mixing is not significantly affected by viscosity. We find that the mixing efficiency asymptotes to 0.08 at high Reynolds numbers. We present a model of the mixing based on the evolution of idealized mean profiles of velocity and density at the interface between the two currents, the results of which are in good agreement with the measurements of mixing efficiency.

Information

Type
Papers
Copyright
© 2016 Cambridge University Press 
Figure 0

Figure 1.Schematic of the lock release experiment. Salty water of density $\unicode[STIX]{x1D70C}_{H}$ fills a lock of length $L_{lock}$, and fresher water of density $\unicode[STIX]{x1D70C}_{L}$ fills the remainder of the channel. The depth $H_{L}>H_{H}$ and is set so that the pressures on the two sides of the lock gate are the same at mid-depth of the channel.

Figure 1

Figure 2.Visualization of the gravity current from Exp 6 – see table 1. The current is dyed with food colouring and viewed against a translucent lined sheet.

Figure 2

Figure 3.Visualization of the impact of the gravity current from Exp 6 – see table 1 – with the end wall of the channel and the subsequent internal bore that is generated. Note that the turbulence levels are significantly reduced from those for the current shown in figure 2.

Figure 3

Table 1.Values of the dimensionless density difference $(\unicode[STIX]{x1D70C}_{H}-\unicode[STIX]{x1D70C}_{L})/\unicode[STIX]{x1D70C}_{0}$, the speeds of gravity current propagation $U_{M}$ and $U_{i}$ measured along the bottom and predicted from (4.2), respectively, the overall Richardson number $Ri_{O}\equiv (g(\unicode[STIX]{x1D70C}_{H}-\unicode[STIX]{x1D70C}_{L})H/4\unicode[STIX]{x1D70C}_{0}U_{M}^{2})$, the dimensionless mass anomaly $\unicode[STIX]{x1D6FC}$ transported from one layer into the other by mixing (see (3.1)), the Reynolds number $Re\equiv (U_{M}H/2\unicode[STIX]{x1D708})$ and the buoyancy Reynolds number $Re_{B}\equiv (C\unicode[STIX]{x1D70C}_{0}U_{M}^{3}\unicode[STIX]{x1D6FF}/g(\unicode[STIX]{x1D70C}_{H}-\unicode[STIX]{x1D70C}_{L})\unicode[STIX]{x1D708}(H/2))$ (where this expression uses an estimate of $N^{2}$ based on the interface thickness $\unicode[STIX]{x1D6FF}$ and $C=0.04$ is the constant in the assumed scaling for dissipation $\unicode[STIX]{x1D716}=CU_{M}^{3}/(H/2)$). The reference density $\unicode[STIX]{x1D70C}_{0}$ is taken to be the average of $\unicode[STIX]{x1D70C}_{L}$ and $\unicode[STIX]{x1D70C}_{H}$, and the uncertainties in $U_{M}$ and $Ri_{O}$ are based on the finite time taken to withdraw the barrier at the start of the experiment.

Figure 4

Figure 4.Final density profiles for the experiments in table 1 normalized by the initial density difference. The profiles were measured by withdrawing samples at different depths and the profiles are drawn with linear segments between the data points. Also shown is the assumed linear variation of density corresponding to an interfacial region of dimensionless thickness $r=0.33$ (see (4.3)).

Figure 5

Figure 5.Mixing efficiency results plotted as a function of the Reynolds number $Re$. Error bars are calculated for each experiment and are determined mainly by the relative accuracy with which the changes in free-surface height $H_{L}-H$ and $H_{H}-H$ in (2.5) can be measured. The error bars exceed the symbol size only for the lowest two $Re$ experiments.

Figure 6

Figure 6.Schematic diagram of the flow model: (a) indicates the overall flow structure and (b) and (c) show the assumed (piecewise linear) velocity and density profiles, respectively. Panel (a) indicates development of shear instability in the vicinity of each gravity current head, with intense turbulent eddies depicted by the swirls. Far enough behind each head, the interface between the counter-flowing currents has stabilized. The darkest grey shading denotes fluid of density $\unicode[STIX]{x1D70C}_{H}$, with fluid of intermediate density indicated by lighter shading in the interfacial region between the two currents. We decompose the volume transport in the exchange flow into a number of components denoted by the general symbol $Q$: the subscript denotes fluid that originated from the light ($\text{L}$) or heavy ($\text{H}$) reservoir (to the left and right, respectively), the first superscript character denotes fluid that is unmodified ($u$) or modified ($m$) by mixing with respect to its source reservoir density and the second and third superscript characters (if present) denote the direction of flow as either left to right ($lr$) or right to left ($rl$). Only the component transports originating from the light reservoir, $\dot{Q}_{L}^{u}$, $\dot{Q}_{L}^{mlr}$ and $\dot{Q}_{L}^{mrl}$, are indicated in (a) for the sake of clarity. However, as we assume anti-symmetry in the flow at vertical section b about $z^{\prime }=0$ (4.9), the corresponding components that have originated from the heavy reservoir, $\dot{Q}_{H}^{u}$, $\dot{Q}_{H}^{mrl}$ and $\dot{Q}_{H}^{mlr}$, respectively, are equal and oppositely directed. Note that the transport components $\dot{Q}_{L}^{mrl}$ and $\dot{Q}_{H}^{mlr}$ represent the rate at which volume is swapped with the counter-flowing layer as a result of turbulent mixing. Further note that although transport components from left to right are defined as positive, the arrows in the figure indicate the physical direction of the transport. With this decomposition, $\dot{Q}_{d}$ is seen to be equal to the exchange volume transport (of mixed fluid; equation (4.8)) from left to right through b, i.e. $\dot{Q}_{L}^{u}+\dot{Q}_{L}^{mlr}+\dot{Q}_{H}^{mlr}$, which is the sum of the horizontal and diagonal striped regions (for $0) in (b) (4.4). However, $\dot{Q}_{d}$ must also be equal to the rate at which volume leaves a specific reservoir (4.10), i.e. $\dot{Q}_{L}^{u}+\dot{Q}_{L}^{mlr}+|\dot{Q}_{L}^{mrl}|$. In (c), the dashed line indicates the density of unmixed fluid entering the current from left to right at section a and the effective current depth $h_{e}$ corresponds to the height of the rectangle in the upper layer with the same area as the horizontal striped region in the profile at section b (i.e. the same total buoyancy anomaly with respect to the mid-point density $\unicode[STIX]{x1D70C}_{0}$).