Accurate prediction of wall-bounded flows with Reynolds-averaged Navier-Stokes turbulence models requires the improved modeling of physical processes that occur in turbulent flows such as the interaction of turbulent velocity and pressure fields, turbulent diffusion, and dissipative processes. Recently, novel models for the velocity/pressuregradient correlations through the fourth order were proposed by the authors and successfully validated against direct numerical simulation (DNS) data in wall-bounded flows without separation in a priori testing. In the current work, these models are implemented in transport equations for the Reynolds stresses to evaluate the model performance in a fullydeveloped planar channel. Other terms that need modeling in the Reynolds stress transport equations are represented in simulations by their DNS profiles to eliminate uncertainty due to modeling these terms. Problems with using DNS data as an input in simulations are highlighted. An approach to mitigate some of those problems is proposed. The computed results are compared against DNS data.
NomenclatureU i = mean velocity component in the i-direction U = mean velocity in the streamwise direction ij Π = velocity -pressure-gradient tensor, ( ) , ,