(126e) Turbulent Flow Structures That Mainly Contribute to Turbulent Transport From the Wall

The goal is to investigate the correlation between the velocity and the temperature field in wall turbulence, and to identify velocity structures that contribute most to the transfer of heat. This work involves numerical experiments that were conducted in a turbulent channel flow using Direct Numerical Simulation (DNS) in conjunction with the tracking of passive thermal walkers released from a line source at the wall of the channel in this flow field.

The simulations were conducted in a computational box of dimensions 3800x600x1900 (streamwise x normal x spanwise) in viscous wall units, for different Prandtl number fluids (Pr = 0.1, 0.7, 6, 20, and 50). The half channel height in viscous wall units was h = 300. The total number of the heat markers was 260,100, and they were released from a single line source at the wall. The trajectories of the heat markers were then monitored in space and time as they moved through the channel. This is a Lagrangian approach to studying the dispersion process. Each computation was repeated for 5 different initial velocities of the flow field, in order to simulate heat release in independent velocity fields. The behavior of these heat markers was observed in 12 successive locations downstream from the source in multiples of h, ranging from h to 12h. In each of these locations the flow structures that carried thermal markers towards the channel center (i.e., those that transfer heat away from the wall) and flow structures that carried thermal markers towards the channel walls (i.e., those that transfer heat towards the wall) were studied as a function of the Prandtl number.

The time at which the number of thermal walkers in each of these downstream locations was maximum was found, and then used to determine the characteristic time and length scales for transport. These were obtained based on the calculation of correlation coefficients for the markers that were captured in each of these downstream locations. The average width and height of the particle plumes that have positive vertical velocity and negative vertical velocity were also computed. These plumes were then visually compared with the velocity vector profiles of the flow field (obtained from the DNS at time of the particle capture) in order to provide a physical picture of the mechanism of heat transfer from the wall. Differences in these quantities as a function of the Prandtl number will be discussed.