An Experimental Study of Reynolds Stress Closures in Shear Free and Weakly Sheared Turbulence
Doktorsavhandling, 1995

The objective of this thesis work is to obtain data enabling evaluation of closure hypotheses used in the Reynolds stress transport (RST) equations. Special attention has been devoted to studies of flow fields with shear free and weakly sheared turbulence. Hot-wire technique was used to determine mean velocities, second and third order moments of fluctuating velocities and spatial velocity derivatives. The latter were estimated by measuring the velocity components at points in close proximity to each other. The wire separation should be in the range, 2 <=.DELTA.xi/.äta. <=4. The data obtained in the far wake of the cylinder indicate a non-isotropic dissipation rate tensor which could be predicted by the model of Hallbäck et al. (1990). However, significant differences were observed in the magnitudes of the measured and modelled profiles (Hanjalic & Launder 1972) for the triple velocity correlations. Using our experimental data we balanced an approximation to the RST equations, thus obtaining estimates of the distributions of the pressure strain rate terms. To study wall proximity effects on the turbulence field in the absence of mean velocity gradients, shear free turbulence near a wall was investigated experimentally. This was accomplished by passing decaying grid generated turbulence with a uniform mean velocity over a wall moving at the stream speed. The initial response of the turbulence field can be well described by the theory of Hunt & Graham (1978). However, this is not the case for larger downstream distances. In the latter region, the terms of the RST equation were measured or estimated through balancing the equation. We found that two different length scales are associated with the near wall damping of the Reynolds stresses and that macro scales are not suitable for describing the near wall effects on the energy redistribution. This implies that turbulence models for the streamwise pressure strain rate component, which assume that the intercomponent energy transfer scales according to K3/2/.epsilon., deviate considerably from corresponding experimental data.

turbulence modelling

dissipation measurements

Reynolds stress closures

isotropic turbulence

anisotropic dissipation rate tensor

hot-wire anemometry

wall proximity effects

pressure strain rate correlations


Dag Aronson

Chalmers, Institutionen för termo- och fluiddynamik






Doktorsavhandlingar vid Chalmers tekniska högskola. Ny serie: 1129