Turbulent boundary layers and sediment suspension absent mean flow-induced shear

• Main Author: Johnson, Blair Anne
• Format: Dissertation
• Language: English
• Published: ProQuest Dissertations & Theses 01-01-2016
• Subjects: Civil engineering; Mechanics; Plasma physics
• ISBN: 9781369254433; 1369254431

We perform an experimental study to investigate turbulent boundary layers in the absence of mean shear at both stationary solid and mobile sediment boundaries. High Reynolds number (Reλ ∼ 300) horizontally homogeneous isotropic turbulence is generated via randomly actuated synthetic jet arrays (RASJA - Variano & Cowen 2008). Each of the arrays is controlled by a spatio-temporally varying algorithm, which in turn minimizes the formation of secondary flows or mean shear. One array consists of an 8 x 8 grid of jets, while the other is a 16 x 16 array. By varying the operational parameters of the RASJA, we also find that we are able to control the turbulence levels, including integral length scales and dissipation rates, by changing the mean on-times in the jet algorithm. Acoustic Doppler velocimetry (ADV) and particle image velocimetry (PIV) measurements are used to study the isotropic turbulent region and the boundary layer formed beneath it as the turbulence encounters the bottom boundary. Time-lapsed photography is used to monitor large-scale bed morphology of the sediment. The flow is characterized by statistical metrics including the mean flow and turbulent velocities, turbulent kinetic energy, temporal spectra, integral scales of the turbulence, and terms in the turbulent kinetic energy transport equation including energy dissipation rates, production, and turbulent transport. We evaluate the implications of assuming isotropy in computing dissipation by comparing several methods commonly used in measurements, including second-order structure functions, spatial spectra, scaling arguments, and direct computations. With our dissipation results, we calculate the empirical constant in the Tennekes (1975) model of Eulerian frequency spectra. This model allows for the determination of dissipation from temporally resolved single-point velocity measurements when there is no mean flow. We compare our boundary layer characterizations to prior literature that addresses mean shear free turbulent boundary layers via grid-stirred tank (GST) experiments, moving bed experiments, rapid distortion theory (RDT), and direct numerical simulations (DNS) in a forced turbulent box. We draw comparisons between an impermeable flat boundary, a flat permeable sediment boundary, and a rippled sediment boundary. In experiments examining turbulence above a sediment boundary, we observe sediment suspension primarily via vortical pick-up and splats. Additionally, we observe the development of ripple patterns in the sediment, which is unexpected in a facility absent mean shear or oscillations. We find a relationship between the integral length scale of the turbulent flow with the ripple spacing, suggesting a link between the turbulence levels and sediment transport. Because traditional viscous stresses due to mean velocity gradients suggest no bed friction or sediment transport, we develop a method for considering Reynolds stresses over short time periods as a surrogate for understanding the importance of bed stress in a zero mean shear environment.