[roberthino]

hi me again...i finally managed it to run my channel flow. i put the boundary conditions for outlet and inlet both to outlet(fixed pressure and rest zero gradient) for the walls i use wall functions. now with k-epsilon and rng k-epsilon it works well and it converges after 10000 iterations with simple foam. now with the non-linear k-epsilon i made 30000 iterations and the profiles still look weird.....does it just take more time because of the more complicated turbulence model?
whatever with the rs-models like LRR notheing is working. always comes this error message:

Reading field p

Reading field U

Reading/calculating face flux field phi

Selecting incompressible transport model Newtonian
Selecting turbulence model LRR


--> FOAM FATAL IO ERROR : unexpected class name volTensorField expected volSymmTensorField
while reading object R

file: /home/maduta/OpenFOAM/maduta-1.4/run/tutorials/simpleFoam/pitzDaily/0/R at line 23.

From function regIOobject::readStream(const word&)
in file db/regIOobject/regIOobjectRead.C at line 115.

FOAM exiting


what should i do?
maybe someone can help me

[david_h]

Try the following in the "0/R" file:

1) change the class entry to "volSymmTensorField"

2) change the two occurrances of "(0 0 0 0 0 0 0 0 0)" (9 zeros) to "(0 0 0 0 0 0)" (6 zeros).

Dave

[roberthino]

well thank you dave now it worked fine.
maybe anybody has an idea why the nonlinear eddy-viscosity models are converging so bad?
i did 50000 iterations with my channelflow and still the profils for the turbulent kintetic energy and the dissipation were really bad. with the rng k-e model and the reynolds stress models i got fine results after 6000 iterations.
do i have to set up special values for any solvers?

[onurdundar]

Dear David
I want to thank you the advices about the LRR method it works fine in turbFoam.

Onur