Channels, Fall 2020

Channels • 2 020 • Volume 5 • Number 1 Page 10 A to determine if the experimental trend could be reproduced in the numerical study. Although 2D simulations do not address the 3D effect, a trend can be found in the simulations that aligns with the experimental data. 2 During the analysis, turbulent models for elliptical cylinders were showed to be less stable than for the circular models. To enhance stability and convergence, laminar model was used for elliptical cylinder simulation. To test if vortex formation occurred in elliptical grooves to provide the “bearing” effect discussed in [10], the streamwise velocity along lines normal to the ellipse’s surface were plotted. Fig. 11 shows these such lines. The depicted x-values refer to the x- direction distance from the center of the ellipse. The plots use the surface of the cylinder as the origin of the curve to show the velocity as a function of the distance away from the surface. The approach is to look for reversed flow in the grooves which correlates to the presence of vortices. Fig. 11 Locations of velocity distribution in boundary layer for (a) smooth cylinder (b) Model A Fig. 12 Velocity distribution in boundary layer for smooth elliptical cylinder and Model A at stagnation point (X = - 0.0254m) Fig. 12 compares the stagnation point boundary layer between both the smooth cylinder and Model A. At the stagnation point it can be seen that the boundary layers for both models correlate closely to each other. Fig. 13 shows that an early reversal of flow began to develop in the grooves of Model A, signaling the