Direct numerical simulation of near-wall turbulence: Passive and active control

In this work we analyze high-resolution numerical data bases for a turbulent channel flow with near-wall turbulence control applied at the lower wall. The simulations cover a range of {dollar}Resbtau{dollar}, from 140-200 and are based on parallel spectral element-Fourier discretizations. The control mechanisms that we investigate are both passive and active in nature. Passive control is achieved via surface modification at the lower wall, where streamwise aligned grooves of various sizes are employed. Active control at the lower wall is achieved via applied electric and magnetic fields in fluid of small electrical conductivity, i.e. sea water. These applied fields generate a Lorentz body force near the wall. We have investigated two different electrode and magnet configurations for flow control. The first configuration consists of alternating streamwise electrodes and magnets which produce a Lorentz force in the streamwise direction of the flow. The second active control configuration consists of a distributed array of electro-magnetic control tiles which produced a Lorentz force with non-zero components in all three directions of the flow.; For these flow control simulations, particular attention is paid to the near-wall vorticity dynamics and structures, and the relationship between velocity-vorticity correlations and the Reynolds stresses. Specifically, we find the following: (1) As the Reynolds stress and shear stress increase from riblet valley to riblet tip, the vortex stretching term {dollar}overline{lcub}omegaspprime,omegaspprimesb2{rcub}{dollar} increases substantially as well; (2) Riblets modify the initial development of vortex structures at the wall via increases in normal vorticity at the riblet tips; (3) The application of the streamwise Lozentz force at a wall increases the drag force as well as introduces large secondary motions near the wall; (4) Both static and time-dependent forcing from the electro-magnetic control tiles increases the drag force at the wall due to the introduction of areas of concentrated streamwise vorticity near the tile; (5) Hairpin-like structures are evident in flow visualizations of vortex lines at both controlled and uncontrolled walls; (6) A set of O(100) eigenfunctions constructed using the proper orthogonal decomposition and defined on the full channel domain are not sufficient to accurately reconstruct near-wall turbulence structures of three-dimensional, complex geometry turbulent flows.