Issue with TKE and epsilon inlet profiles in ANSYS Fluent LES. Anyone could help!?

peymanoladpeymanolad Member Posts: 1

Dear all,

I am trying to simulate a wall-jet problem using ANSYS-Fluent (LES, Dynamic Smagorinsky-Lilly) and for the inlet, I have used a profile .csv file (I am imposing velocity, TKE, dissipation rate, and normal Reynolds stresses). Everything works just fine with the velocity profile, however, when I look at the TKE and epsilon profiles at the inlet, they look somewhat different that what I had imposed to the inlet. The shapes of the profiles are kind of the same, but the peak values are different e.g. for TKE, it seems that the peak values are divided by 3 (decreased from 0.07 m^2/s^2 to around 0.012 m^2/s^2). (profiles comparisons are attached. Left figures are what is being imposed and right figures are what ANSYS shows as the profiles at the inlet).

For fluent, I have calculated TKE and epsilon as follows:

TKE = resovled TKE + modeled TKE

resolved TKE = 0.5*(urms^2+vrms^2+wrms^2)

modeled TKE = 1/C*(mu_sgs/(rho*Delta)


C = C_DS^2 (C_DS being the subgrid dynamic constant)

mu_sgs (Turbulent

eddy viscosity)

rho (density)

Delta = cell-size^(1/3)

For dissipation rate:

epsilon = 2*nu*<s'ij .s'ij>,

where s'ij= Sij - <Sij>

I calculated <s'ij .s'ij> as follows:

<s'ij .s'ij> = < (Sij - <Sij>) * (Sij - <Sij>) > = <Sij^2> - <Sij>^2

I believe I have run the simulation for long enough time (around 10 washout times) to get reliable (mean) values at the inlet e.g. especially for getting epsilon through <Sij2> and <Sij>2 values as well as TKE from rms velocity components.

Why do you think I get different values of TKE and epsilon in the inlet than what I have imposed in the profile file?


  • sdebsdeb Posts: 200Forum Coordinator


    Are you plotting the profile directly in Fluent or are you using an XY plot?

    If XY plot, then are you using node values or cell centered values?

    Can you check by plotting directly the profile from Plots-->Profile Data?



Sign In or Register to comment.