FVM MODELING OF STRATIFIEDFLOWS USING A NON-LINEAR K-ε TURBULENCE CLOSURE

Buoyancy forces caused by stratification have substantial impacts on propagation of currents, turbulent mixing processes, production and dissipation of turbulence eddies.They also play a major role in the transportation of sediments and contaminants in these flows. In stratified flows, due to the presence of buoyancy forces, the stressescaused by turbulence are anisotropic and the implementation of this anisotropy may have a significant impact on the predicted results. The accuracy of thenumerical simulation of such flows is highly dependent on the turbulence model and the treatment of the buoyancy forces in the model. Despite the intensive research efforts to develop more complete turbulence models, k-ε models, still remain the most widely used approach by engineers and scientists for the solution of practical problems. The main advantage of a k-ε turbulence model is its reasonable computational time in comparison with the more complicated models, while reserving satisfactory results. The linear k-ε turbulence model is usually incapable of accurate prediction of turbulent flows, where the nonisotropy is considerable. The hydrodynamic model WISE (Width Integrated Stratified Environments) used herein is a 2DV numerical model developed by Hejazi (2013). WISE is a free surface numerical model based on the timedependent Reynolds averaged Navier-Stokes equations. For the development of the non-linear k-ε turbulence model, the model proposed by Speziale [3], and for the Reynolds scalar fluxes, the flux equations suggested have been implemented [4]. The new model has been simulated for several lock-release type tests, and saline water jets discharging into an ambient flow, for which the laboratory experimental values are reported in the literature. Simulated velocity and concentration profiles have been compared with the measured values, as well as two- and three-dimensional numerical simulations reported in the literature along with the predictions of the linear k-ε turbulence model.