RANS models (Reynolds-Averaged Navier-Stokes) are widely used in the field of wind engineering and model turbulence across all length scales. The basic approach consists of splitting the velocity into an average and a turbulent fluctuation. The resulting additional unknowns are "closed" through averaging and supplementary equations. Within the RANS family, a distinction is made between simple algebraic models, which treat turbulence as local eddy viscosity, and the more commonly used one- or two-equation models. The latter solve additional transport equations for kinetic energy and dissipation rate. More complex approaches, such as anisotropic Reynolds stress methods, are less commonly used in practice.
Two-equation models, particularly the k-ε model and its variants, as well as the k-ω or SST (Shear Stress Transport) method, are the most widely used due to their balanced trade-off between computational effort, result quality, and calibration complexity. Classical RANS models seek the steady-state equilibrium of the turbulent problem and can also be applied in two spatial dimensions, unlike LES methods, if the problem allows it.
To account for temporal changes, URANS (Unsteady RANS) variants were developed, introducing a transient term with variable time steps. This approach, however, requires special caution, as the implicit averaging over all time scales makes it challenging to assess temporal accuracy and may suppress unsteady effects.