Electronic Data Center. Delivering Direct Online Access to Scientific, Engineering, and Medical Reference Material and Information

ICHMT DIGITAL LIBRARY ONLINE

ISSN
961-91393-0-5

Print version

Year 2006

Volume 1 - Turbulence, Heat and Mass Transfer 5
Proceeding of The International Symposium on Turbulence, Heat and Mass Transfer - Dubrovnik, Croatia, September 25-29, 2006

ARTICLE:

  • Remaining Challenges in Measurements and Computations of Three-dimensional Turbulent Boundary Layers  download article

K. Chauhan
Illinois Institute of Technology, Chicago, Illinois 60616, USA

S. Gravante
Illinois Institute of Technology, Chicago, Illinois 60616, USA

I. Pelivan
University of Bremen, D-283 59 Bremen, Germany

H. Nagib
Illinois Institute of Technology, Chicago, Illinois 60616, USA


ABSTRACT

Results are presented from experiments in a cross-flow driven three-dimensional turbulent boundary layer (3DTBL) by Gravante[1] and Pelivan[2] at IIT. The direct and accurate technique of oil-film interferometry is utilized to measure wall shear stress, while three-component Stereo Particle Image Velocimetry (SPIV) and Seven Hole Probe (SHP) are used to document the flow field. Measurements from the experiment are compared to computations performed by Pelivan[2] and Chauhan[3] using RANS turbulence models available in the commercial code FLUENT. For our specially-designed test flow, we find that the variation of Cf in a 3DTBL is no longer dependent on the pressure gradient alone, and that both increasing and then decreasing behavior of the spanwise wall shear is exhibited even under a spanwise favorable pressure gradient. The streamwise wall-shear stress is under-estimated in our computations when only the test section is computed as the domain, even with proper inflow conditions. In contrast, when the flow was computed starting from the leading edge of the boundary-layer plate, the streamwise wall shear stress is over-estimated, but the spanwise shear stress is accurately computed. The computations perform well in predicting the mean flow parameters in both streamwise and spanwise directions. The Reynolds stresses are all affected by the mean flow three-dimensionality, with turbulence production and the pressure rate-of-strain tensor playing the major role in the dynamics of the Reynolds stresses. From the computations with a standard Reynolds stress model, we conclude that with a good set of boundary conditions, the current RSM can very well capture the dynamics of a 3DTBL. Parameters like Prandtl’s mixing length (l), Townsend’s structure parameter and the turbulent viscosity, which are important for turbulence modeling, are also evaluated from the experimental data and compared to the computations.

 download article

139-142 pages


« Previous article         Next article »