# Turbulence free-stream boundary conditions

In most CFD simulations it is necessary to specify values of the turbulence variables at the inlets. For example, if you are using a $k-epsilon$ you have to specify values of $k$ and $\epsilon$ on the inlets. This is often difficult and a source of uncertainty, since the incoming turbulence is rarely known exactly. Most often you are forced to make a more or less educated guess of the incoming turbulence.

Estimating the incoming turbulence model variables, like turbulent energy, dissipation or even Reynolds stresses, directly is often difficult. Instead it is easier to think of the incoming turbulence in terms of variables like turbulence intensity and turbulent length-scale or eddy-viscosity ratio. These properties are more intuitive and can more easily be related to physical characteristics of the problem.