Turbulence & YPlus
 

Hello, I have three questions, can any body please answer it.

1. How to select the inlet turbulence parameters like dissipation length scale & turbulence intensity in a standard Turbulence modal and, How to verify after the solution that whatever values we have chosen are correct.

2. In case of external flows, what are the parameters to be checked to say that the solution has converged?

3. What is the physical meaning of YPlus and its significance? If it is the input parameter for the software then how it should be chosen

Your answers are most welcome.

Irfan

Re: Turbulence & YPlus
 

inlet tke = 1.5*(TI*U0)^2 inlet epsl = cmu^(3/4)*(tke)^(3/2)/(kappa*l) where TI - turbulent intensity U0 = Inlet velocity (mean) l = length scale = generally 0.07*diameter

yplus decides whether the pt is in log region or viscous sublayer its not input parameter at all . it will be calculated by the code .

Re: Turbulence & YPlus
 

1. In k-epsilon and many other models freestream k must be small compared with the one generated by the walls and epsilon does not matter much. In k-w be aware that the selection of w infinite can give problems (see papers by Menter about it)

2. some residual (change of variable with time) properly averaged

3. Yplus is the distance to the wall nondimensionalized by the characteristic scale of variation of the turbulent mean velocity profile close to the wall. It means that, close to the wall, where viscosity dominates, the scaling of distance and velocity should come given only by the local wall friction tau, the density of the fluid rho and the kinematic viscosity v. On dimensional grounds a characteristic distance using these three physical quantities is d = v(rho/tau)^1/2 and a characteristic velocity is ut = (tau/rho)^1/2, so the proper nondimensionalization is y+ = y / d and ut = u / ut. Once done this, it can be demonstrated that for almost any situation in terms of external pressure gradient, external velocity...etc... you know that very close to the wall the velocity varies with the distance as

u+ = y+

This only holds for y+ < 10 and only works when boundary layer is far from separation. This is important since it means you will have to put a few points in the segment y+ < 10 if you want to resolve this layer (called viscous sublayer) and it is important to resolve this sublayer if you want to get an accurate value of the friction.

(read book by Tennekes & Lumley or more recent book of Pope about turbulence and especially wall turbulence)