Vortex Structures and Mixing of Compressible Mixing Layer

Zhou Qian
2014-01-01
Abstract:Using the seventh-order upwind compact scheme for spatial discretization of the convective flux,the explicit central eighth-order scheme for viscous flux terms,the unsteady three-dimensional compressible Navier-Stokes equations are solved for the compressible spatially evolving plane mixing layer.The direct numerical simulation(DNS)is performed without resorting to any sub-grid models.The convective Mach number of the mixing layer is chosen to be 0.7for characterizing the moderate compressibility.The finite difference grid size is 4160×351×256.The spatial developing structures are shown in detail from the initial instability flow until the full development of turbulent flow.The results show that the large-scale vortex structures,such asΛvortices,hairpin vortices and flower structures,make the momentum thickness grow rapidly along the streamwise flow.The discussions are focused on the capacity of the moderate compressible shear layer to mix the fluids from the two different sides.It is found that the large-scale vortex structures significantly affect the thickness of the visual layer thickness shown by the mass fraction of the passive scalar.Particularly,strong pumping flows between the two counter-rotating vortices ofΛvortices as well as hairpin vortices injects large amounts of fluid from one side to the other side.The small vortices contained in the flower structures also help mix the fluids through their energetic twisting motion,only at a relatively small scale.Generally speaking,the dynamic evolution of these large-scale vortex structures produces very curvy interfaces between the upper side and the lower side,and the large contact area of the curvy interfaces appreciably enhances the mixing between the fluids from the two sides.The fine scale structures in the fully developed turbulence play an important role in making the mixing of the shear layer more uniform.However the contribution of the fully developed turbulence to the visual mixing layer growth is very limited.
What problem does this paper attempt to address?