Three-dimensional large-eddy simulations of two-stream mixing layers developing spatially from laminar boundary layers are presented, replicating wind-tunnel experiments carried out in Part 1 of this study. These simulations have been continued through the mixing transition and into the fully turbulent self-similar flow beyond. In agreement with the experiments, the simulations show that the familiar mechanism of growth by vortex amalgamation is replaced at the mixing transition by a previously unrecognised mechanism in which the spanwise-coherent large structures individually undergo continuous linear growth. In the post-transition flow it is this continuous linear growth of the individual structures that produces the self-similar growth of the mixing-layer thickness, the large-structure interactions occurring as a consequence of the growth, not its cause. New information is also presented on the topography of the organised post-transition flow and on its cyclical evolution through the lifetimes of the individual large structures. The dynamic and kinematic implications of these findings are discussed and shown to define quantitatively the growth rate of the homogeneous post-transition mixing layer in its organised state. _______________________________________________________________

No abstract

SUMMARYA series of spatially developing mixing layers are simulated using the large eddy simulation (LES) technique. A hyperbolic tangent function and data derived from boundary layer simulations are used to generate the inflow condition, and their effects on the flow are compared. The simulations are performed in both two and three dimensions. In two-dimensional simulations, both types of inflow conditions produce a layer that grows through successive pairings of Kelvin-Helmholtz (K-H) vortices, but the composition ratio is lower for the hyperbolic tangent inflow simulations. The two-dimensional simulations do not undergo a transition to turbulence. The three-dimensional simulations produce a transition to turbulence, and coherent structures are found in the post-transition region of the flow. The composition ratio of the three-dimensional layers is reduced in comparison to the counterpart two-dimensional runs. The mechanisms of growth are investigated in each type of simulation, and amalgamative pairing interactions are found in the pre-transition region of the three-dimensional simulations, and throughout the entire computational domain of those carried out in two-dimensions. The structures beyond the post-transition region of the three-dimensional simulations appear to behave in a much different manner to their pre-transition cousins, with no pairingtype interactions observed in the turbulent flow. In order to accurately simulate spatially developing mixing layers, it is postulated that the inflow conditions must closely correspond to the conditions present in the reference experiment.

Large Eddy Simulations (LES) of spatially developing turbulent mixing layers have been performed for flows of uniform density and Reynolds numbers of up to 50,000 based on the visual thickness of the layer and the velocity difference across it. On a fine LES grid, a validation simulation performed with a hyperbolic tangent inflow profile produces flow statistics that compare extremely well with reference Direct Numerical Simulation (DNS) data. An inflow profile derived from laminar Blasius profiles produces a flow that is significantly different to the reference DNS, particularly with respect to the initial development of the flow. When compared with experimental data, however, it is the boundary layer-type inflow simulation produces the better prediction of the flow statistics, including the mean transition location. It is found that the boundary layer inflow condition is more unstable than the hyperbolic tangent inlet profile. A suitably designed coarse LES grid produces good predictions of the mean transition location with boundary layer inflow conditions at a low computational cost. The results suggest that hyperbolic tangent functions may produce unreliable DNS data when used as the initial condition for studies of the transition in the mixing layer flow.

New flow-visualization experiments on mixing layers of various velocity and density ratios are reported. It is shown that, in mixing layers developing from laminar initial conditions, the familiar mechanism of growth by vortex amalgamation is replaced at the mixing transition by a previously unrecognized mechanism in which the spanwise-coherent large structures individually undergo continuous linear growth. In the organized post-transition flow it is this continuous linear growth of the individual structures that produces the self-similar growth of the mixing-layer thickness, with the occasional interactions between neighbouring structures occurring as a consequence of their growth, not its cause. It is also observed that periods during which the post-transition mixing layer comprises orderly processions of large structures alternate with periods during which no large-scale organization is apparent downstream of the transition location. These two fully turbulent flow states are characterized by different growth rates, entrainment ratios and orientations of the mixing layer relative to the free streams. The implications of these findings are discussed.