Numerical prediction of boundary layer flows with the anisotropic k-.EPSILON. turbulence model.

Prediction of Three-Dimensional Developing Turbulent Flow in a Square Duct With an Anisotropic Low-Reynolds-Number k-ε Model

“…It is just an extended form of its isotropic version (Myong, 1988;Myong and Kasagi, 1990b), but broadens the range of applicability whilst maintaining most of popular features of the latter. This model has already been found capable of predicting the anisotropic normal Reynolds stresses up to the wall with the correct wall-limiting behavior, and also proven to perform satisfactorily in several flow situations, including turbulent pipe and channel flows, two-dimensional boundary layer flows with and without pressure gradients (Myong and Kasagi, 1990a, c;Myong et al, 1990). These results suggest a possibility that this model can be used to predict more complex flows influenced by the anisotropy of Reynolds stresses, e.g., secondary flows in noncircular ducts and separated flows.…”

“…Manuscript received by the Fluids Engineering Division December 5, 1990. gations have been carried out to understand in detail turbulent flows in a straight square duct as the simplest geometry in which turbulence-driven secondary flows arise (see e.g., Melling and Whitelaw, 1976;Gessner et al, 1979). In addition, since these particular flow situations provide a natural vehicle for examining the validity of existing turbulence models, turbulence-driven secondary flows in a straight square duct are of special interest to modelers, and hence numerical investigations have also been carried out to accurately predict them by Launder and Ying (1973), Gessner and Emery (1981), Nakayama et al (1983), and Demuren and Rodi (1984). Since the widely used (isotropic) k-e model has no built-in mechanism for the development of secondary flow due to its inherent isotropic characteristics, most of their methods are based on the algebraic Reynolds stress models.…”