Fluid Flow Turbulence Modeling

A vast majority of industrial flows are turbulent. While conditions at boundaries may be steady, shear action near walls and other velocity profiles lead to time-dependent pressure and velocity fluctuations transported by and affecting the mean path of the fluid. These "turbulent" fluctuations span a wide range of length and timescales. As a result, computational demands for exact solutions of industrial turbulent flows far exceed available computing resources.

Reynolds-Averaged Navier-Stokes (RANS) models provide an economic approach to solving these complex, turbulent flows by modeling turbulence as fluctuations within a steady, average velocity field. These models are not exact but have been tuned to provide reasonable accuracy on appropriately scaled meshes. RANS models solve additional transport equations for turbulence and introduce an eddy viscosity (also known as turbulent viscosity) to the simulation to mimic the effect of turbulence, a modeling approach that is suitable for many engineering applications.

In Discovery, the k-omega SST turbulence model is appropriate for most cases and is the default for both time-dependent and steady-state simulations. Additional turbulence models are available and can be modified based on your simulation requirements.