Turbulence modeling

The turbulence model selected for the simulations is the standard $k-\epsilon$. In the form implemented in StarCD, the high Reynolds number forms of the $k$ and $\epsilon$ equations are used in conjuction with algebraic `law of the wall' to represent the flow up to the near-wall regions. The $k-\epsilon$ model is among the simplest turbulence models, since turbulence modelling is obtained by solution of two scalar equations. Equations account for transport of kinetic energy and dissipation rate within the domain and, provided that the grid refinement near to the boundaries is accurate, the results obtained are reliable as those obtained with more complex turbulence models (Renormalization Group $k-\epsilon$ models, Large Eddy Simulation models, etc.)[Bartels et al., 2000, Ranade et al., 1989]. Selection of this less computationally intensive turbulence model allowed to speed up simulations.