(577b) Passive Scalar Mixing in Anisotropic Turbulence from Line Sources

Mixing in turbulent flows is important both in nature and in engineering. Convection effects along with the motion of three-dimensional coherent structures in turbulent flow disperse a substance more efficiently than molecular diffusion does on its own. While the first effect increases the length scale of mixing and reduces its time scale, the latter works locally and enhances mixing at molecular level, which is of significant importance in chemical reactions and turbulent combustion applications. Several studies have been devoted to molecular mixing in homogeneous, isotropic turbulence. In this study, however, we explore the effects of anisotropic turbulence, in addition to molecular diffusion, on passive scalar mixing. Simulation of a turbulent channel flow is conducted by direct numerical simulation (DNS), followed by the Lagrangian Scalar Tracking (LST) method, where the motion of particles that correspond to markers of passive scalars with different Schmidt numbers (Sc) is calculated in the flow field [1-3]. Mass markers are released from several instantaneous line sources, starting at the channel wall and elevated to different distances from the channel wall to the center of the channel. The combination of diffusion, differences in the mean fluid velocity, and the effects of the coherent flow structures in the viscous wall region lead to the observation of different concentrations of particles downstream from the source. We then explore in details the effects of molecular diffusion and turbulence advection on mixing, which is observed to reach an upper limit at long time. We quantify mixing by introducing a "mixing factor" that is an indication of the quality of mixing and it represents the number of particles from one source that are present with particles from the other source at a specific location in the channel. Results from numerical simulation at friction Reynolds number of 300 will be discussed for Schmidt number scalars ranging from 0.7 to 2,400.

ACKNOWLEDGEMENTS

The computational support of XSEDE (CTS090017) is gratefully acknowledged.

REFERENCES

1. Nguyen, Q. and D.V. Papavassiliou, A statistical model to predict streamwise turbulent dispersion from the wall at small times, Physics of Fluids, 2016, 28(12), Art. 125103
2. Nguyen, Q., Srinivasan, C., and D.V. Papavassiliou, Flow induced separation in wall turbulence, Phys Rev E,  2015, 91, Art. 033019
3. Srinivasan, C., and D.V. Papavassiliou, Heat transfer scaling for wall bounded turbulent flows, Applied Mechanics Review, 2013, 65(3), Art. 031002