Shed vortex structure and phase-averaged velocity statistics in symmetric/asymmetric turbulent flat plate wakes

Abstract

The near wake of a flat plate is investigated via direct numerical simulations. Many earlier experimental investigations have used thin plates with sharp trailing edges and turbulent boundary layers to create the wake. This results in large θ/DTE values (θ is the boundary layer momentum thickness toward the end of the plate and DTE is the trailing edge thickness). In the present study, the emphasis is on relatively thick plates with circular trailing edges (CTEs) resulting in θ/D values less than one (D is the plate thickness and the diameter of the CTE) and vigorous vortex shedding. The Reynolds numbers based on the plate length and D are 1.255 × 106 and 10 000, respectively. Two cases are computed: one with turbulent boundary layers on both the upper and lower surfaces of the plate (statistically the same, symmetric wake, Case TT) and the other with turbulent and laminar boundary layers on the upper and lower surfaces, respectively (asymmetric case, Case TL). The data and understanding obtained are of considerable engineering interest, particularly in turbomachinery where the pressure side of an airfoil can remain laminar or transitional because of a favorable pressure gradient and the suction side is turbulent. Shed-vortex structure and phase-averaged velocity statistics obtained in the two cases are compared here. The upper negative shed vortices in Case TL (turbulent separating boundary layer) are weaker than the lower positive ones (laminar separating boundary layer) at inception (a factor of 1.27 weaker in terms of peak phase-averaged spanwise vorticity at the first appearance of a peak). The upper vortices weaken rapidly as they travel downstream. A second feature of interest in Case TL is a considerable increase in the peak phase-averaged, streamwise normal intensity (random component) with increasing streamwise distance (x/D) that occurs near the positive vortex cores. This behavior is observed for a few diameters in the near wake. This is counter to Case TT where the peak value essentially decreases with increasing x/D. Both these effects are examined in detail, and the important contributors are identified.