Turbulence is one of the most complicated phenomena observed in nature, making its precise definition difficult. Literature gives many definitions, as, for example, the one included in : “A fluid motion is described as turbulent if it is three-dimensional, rotational, intermittent, highly disordered, diffusive and dissipative”.
In order to fully capture the turbulence by numerical modelling, one has to solve the equations of motion for fluid flow on all spatial and temporal scales. This approach is referred to as “direct numerical simulation” (DNS). For industrial applications, the computational resources required by a DNS far exceed the capacity of the most powerful supercomputers currently available.
Instead, RWIND Simulation uses a different technique where flow variables, such as velocity or pressure, are decomposed into mean (averaged) components and fluctuating components. In other words, governing equations of fluid motion are averaged in order to remove the small scales, resulting in a modified set of equations that are computationally less laborious to solve. Those equations are referred to as “Reynolds-averaged Navier-Stokes equations” (RANS).
In order to solve RANS in RWIND Simulation, the k–ε turbulence model is used, which introduces two transport equations for the turbulence properties: the first one is the transport equation of the turbulence kinetic energy k, and the second equation governs the transport of the dissipation rate ε of k. This method represents the most widely used and tested model for CFD calculations. Robustness, economy, and reasonable accuracy for a wide range of turbulent flow applications explain its popularity in industrial flow simulations.