The concept is that u_tau is a result of the flow solution by means of the wall stress. Thus, computing the y+ distribution is actually solution-dependent and requires some adjustment from the solution.

Filippo Maria Denaro are you suggesting that I have to go with the iterative process till Y+ value matches or I am doing something wrong with the solution setup?

Y+ is defined as y*utau/nu, y being the distance to the wall, nu the viscosity (m2/s), and utau a friction velocity directly link to the friction (force per surface) tau by the relation Tau=ro_wall*utau*utau, ro_wall being the fluid density at the wall. The friction is a result of the computation, thus Y+ also, you must be sure to have a converged solution (which can be long sometimes).

Depending on the turbulence model and the wall treatment, you can get different values from the computation for Y+ and for friction, not always coherent with other theories or measurements.

Depending on the local size of the mesh and the numerical options chosen, Fluent will either use wall functions (suitable for Y+>~30, the near wall boundary layer is then typically modeled with a log law) or damping functions near the wall (suitable for lower Y+ values around 1). Both options are more or less valid depending on configurations.

For simple pipes, you can also have big numerical (unphysical?) effects in the simulation at the beginning of the pipe because of boundary layer not fully developped here, so the result is there not on the theory curves and you must look downstream.

Hope it helps

Bye

Parth Patel

the problem is more general in both pipe and channel flow ... The issue start in defining the Re_tau value for the case. You see that, again, it require the evaluation of u_tau which is a result of the computation. In general, for periodic pipe/channel flows two procedures are addressed: a) if you fix the pressure gradient, the value if u_tau is based on it and you set a value for Re_tau in your case (and hence you precribed a grid with a priori known values of y+) but you have to knwo that the mass flux is determined by the solution (that is the bulk Reynolds has to be computed). b) you fix the mass flow and compute dynamically the pressure gradient, this way the resulting Re_tau (and y+ distribution) is determined by the computation.

I generally use approach a) so that my Re_tau and y+ distribution is fixed a priori.