Numerical and Physical Modeling of Turbulent Shear Flows

This dissertation is an evaluation of popular turbulence schemes; both three dimensional and depth-averaged, and also includes an experimental study on shallow near bed jets. The three dimensional and RNG turbulent closure schemes are evaluated for free and bounded shear flows. For free shear flows (circular and plane turbulent jets), the scheme with standard coefficient performs equally well and in some cases better than the renormalized group scheme in predicting growth rate, decay of centerline velocity and longitudinal velocity profiles. For turbulent kinetic energy across the jet, the inner region is better predicted by the RNG scheme. The second case used to evaluate the three dimensional schemes was a submerged hydraulic jump. This flow included a free surface and solid boundary creating larger shearing forces than in a free jet. The results showed the longitudinal velocity profiles and their maximum values, in vertical direction, were estimated better by the RNG scheme. The turbulent kinetic energy was overestimated in both magnitude and elevation of its maximum position in the flow. The elevation of the recirculation region was also over predicted by both schemes; however, its longitudinal extent was predicted well. A two-dimensional, depth-averaged flow model with the depth-averaged parabolic eddy viscosity, mixing length, and turbulent closure schemes was used to simulate flow patterns downstream of lock and dam structures. The mixing length scheme was modified and performed as well as the scheme in predicting the location and size of the recirculation zones, as well as the velocity profiles across the channel. Experimental measurements on shallow near bed jets are performed. For low submergence, the horizontal growth rates have two distinct regions, with the downstream region having a higher growth rate. The longitudinal velocity profiles in the horizontal plane are self-similar. The centerline decay was slower than that of a free jet