Measurement of turbulence near shear-free density interfaces

Abstract

The results of an experimental study carried out to investigate the structure of turbulence near a shear-free density interface are presented. The experimental configuration consisted of a two-layer fluid medium in which the lower layer was maintained in a turbulent state by an oscillating grid. The measurements included the root-mean-square (r.m.s.) turbulent velocities, wavenumber spectra, dissipation of turbulent kinetic energy and integral lengthscales. It was found that the introduction of a density interface to a turbulent flow can strongly distort the structure of turbulence near the interface wherein the horizontal velocity components are amplified and the vertical component is damped. The modification of r.m.s velocities is essentially limited to distances smaller than about an integral lengthscale. Inspection of spectra shows that these distortions are felt only at small wavenumbers of the order of the integral scale and a range of low-wavenumbers of the inertial subrange; the distortions become pronounced as the interface is approached. Comparison of the horizontal velocity data with the rapid distortion theory (RDT) analyses of Hunt & Graham (1978) and Hunt (1984) showed a qualitative agreement near the interface and a quantitative agreement away from the interface. On the other hand, the RDT predictions for the vertical component were in general agreement with the data. The near-interface horizontal velocity data, however, showed quantitative agreement with a model proposed by Hunt (1984) based on nonlinear vortex dynamics near the interface. The effects due to interfacial waves appear to be important for distances less than about 10% of the integral lengthscale. As a consequence of the non-zero energy flux divergence, the introduction of a density interface to oscillating grid turbulence increases the rate of dissipation in the turbulent layer except near the interface, where a sharp drop occurs. The present measurements provide useful information on the structure of turbulence in shear-free boundary layers, such as atmospheric and oceanic convective boundary layers, thus improving modelling capabilities for such flows.

Experimental detection of turbulent thermaldiffusion of aerosols in non-isothermal flows

“…For stably stratified flow, the frequency of the grid oscillations varied from f =4.5 Hz to f =16.5 Hz, and for unstably stratified flow, the frequency of the grid oscillations varied from f =2.2 Hz to f =16.4 Hz. Our PIV measurements confirmed earlier results (Turner, 1968(Turner, , 1973Thompson and Turner, 1975;Hopfinger and Toly, 1976;Kit et al, 1997;Sánchez and Redondo, 1998;Medina et al, 2001) for the dependence of varies characteristics of the turbulent velocity field on the parameters of the oscillating grid turbulence generator in water flow experiments. In particular, the integral scale of turbulence is proportional to the distance from a grid, l∝Y, and the r.m.s.…”

Free-surface turbulence and air-water gas exchange

“…It is found that, on average, their steady flow velocity scale was about 40% relative to their turbulent velocity scale (i.e., u/U "" 2.5). They also • In Kit et al [71], the secondary mean flow was reported' to be no more than 25% of the RMS turbulent velocity. This is not very different from what was shown for this tank in figure 6-15.…”

“…It is not clear if this result is a consequence of a lack of flow isotropy. In similar grid tank experiments, Kit et al [71] were only able to compute L31, and the results for this lengthscale appeared to follow the relationship for LIlT, which is for the longitudinal lengthscale, i.e., L n "" L31'…”

Threshold criteria for incipient sediment motion on an inclined bedform in the presence of oscillating-grid turbulence

“…(1) and (2), where typically 0.25 ≤ C 1 ≤ 0.5 and C 2 = 0.1. 29 Moreover, previous studies have also shown that the turbulence generated by an oscillating grid is also effectively isotropic within the homogeneous region (i.e., z 2.5M), with measured values of the isotropy parameter, w/u, approximately constant with z and typically within the range 1.1 to 1.3, where w denotes the time-averaged rms vertical velocity component. The statistical properties of the turbulence produced by the oscillating-grid mechanism described here were measured and compared with (1) and (2).…”