[gu1]

Hello,

I am performing a simulation in internal flow (simpleFoam - steadyState - RANS and LES). I was based on the model: tutorials/incompressible/pisoFoam/LES/motorBike

I modified the code so that when simulation in RANS reaches convergence (and there is no problem with it), the simulation would continue in LES (for that reason the log.simpleFoam below starts at time 700).

By evaluating the log, one realizes that the calculation of 'k' is not happening... why? can anybody help me?

The directory 0 contains: U, p, nut, and k.

Create time

Create mesh for time = 700


SIMPLE: convergence criteria
field p tolerance 1e-08
field U tolerance 1e-07
field k tolerance 1e-07

Reading field p

Reading field U

Reading/calculating face flux field phi

Selecting incompressible transport model Newtonian
Selecting turbulence model type LES
Selecting LES turbulence model Smagorinsky
Selecting LES delta type cubeRootVol
LES
{
LESModel Smagorinsky;
delta cubeRootVol;
printCoeffs on;
turbulence on;
cubeRootVolCoeffs
{
deltaCoeff 1;
}
}

No MRF models present

Creating finite volume options from "constant/fvOptions"

Selecting finite volume options model type meanVelocityForce
Source: momentumSource
- selecting all cells
- selected 113152 cell(s) with volume 6.89573e-06
Reading pressure gradient from file
Initial pressure gradient = 1.51353


Starting time loop

Time = 701

smoothSolver: Solving for Ux, Initial residual = 0.0094822, Final residual = 3.78246e-05, No Iterations 4
smoothSolver: Solving for Uy, Initial residual = 0.155348, Final residual = 0.000636507, No Iterations 5
smoothSolver: Solving for Uz, Initial residual = 0.0995475, Final residual = 0.00043678, No Iterations 5
Pressure gradient source: uncorrected Ubar = 1.00093, pressure gradient = -1.0431
GAMG: Solving for p, Initial residual = 0.415882, Final residual = 0.00117298, No Iterations 3
time step continuity errors : sum local = 0.000115104, global = -5.63421e-10, cumulative = -5.63421e-10
Pressure gradient source: uncorrected Ubar = 1.00069, pressure gradient = -0.398941
ExecutionTime = 1.06 s ClockTime = 2 s

Time = 702

smoothSolver: Solving for Ux, Initial residual = 0.00523343, Final residual = 3.49301e-05, No Iterations 4
smoothSolver: Solving for Uy, Initial residual = 0.133394, Final residual = 0.000435351, No Iterations 5
smoothSolver: Solving for Uz, Initial residual = 0.088513, Final residual = 0.000282018, No Iterations 5
Pressure gradient source: uncorrected Ubar = 0.998616, pressure gradient = 3.4089
GAMG: Solving for p, Initial residual = 0.446016, Final residual = 0.00251308, No Iterations 2
time step continuity errors : sum local = 0.000284106, global = -7.10671e-10, cumulative = -1.27409e-09
Pressure gradient source: uncorrected Ubar = 0.998964, pressure gradient = 2.45149
ExecutionTime = 1.41 s ClockTime = 2 s

Time = 703

smoothSolver: Solving for Ux, Initial residual = 0.00395981, Final residual = 1.89992e-05, No Iterations 5
smoothSolver: Solving for Uy, Initial residual = 0.107527, Final residual = 0.000351482, No Iterations 5
smoothSolver: Solving for Uz, Initial residual = 0.0776189, Final residual = 0.000772944, No Iterations 4
Pressure gradient source: uncorrected Ubar = 0.999597, pressure gradient = 3.55676
GAMG: Solving for p, Initial residual = 0.464837, Final residual = 0.00181112, No Iterations 3
time step continuity errors : sum local = 0.000217197, global = -6.57768e-10, cumulative = -1.93186e-09
Pressure gradient source: uncorrected Ubar = 0.999703, pressure gradient = 3.26612
ExecutionTime = 1.78 s ClockTime = 2 s

[piu58]

It si not a good idea to start a simulation in RANS and use the values for k and epsilon/omega to LES. They have another meaning there and differe remakably form the RANS values.

Is there a important reason not to start with LES?

[gu1]

Originally Posted by piu58

It si not a good idea to start a simulation in RANS and use the values for k and epsilon/omega to LES. They have another meaning there and differe remakably form the RANS values.

Is there a important reason not to start with LES?

Because of this article: Best-Practices for LES

I figured that because I was using the Smagorinsky model ... which needs the 'U', 'p', 'k' and 'nut' files. I did not think it would be important to use RANS earlier, since the tutorial also uses it. I am believing that the problem is occurring because I used steadyState (fvSchenes) for LES. I believe that this is not right.