[anne]

Hello,

I am running a LES to simulate the dispersion of
a passive contamimant.

I have tried linearLimited01 schemes for the scalar transport
equation but I still have negative values.

Could you please inform me about the best
schemes available in OpenFoam to chose to ensure a positive scalar?


Thanks You

Anne

[hjasak]

Use Gamma01 with the coefficient of 0.5 or 1. You also need to limit the Laplace scheme for that scalar, e.g.

laplacian(DkEff,k) Gauss linear limited 1;

On bad meshes you may need to use a lower coefficient, say 0.5

That should do it,

Hrv

[hjasak]

Forgot to say: temporal discretisation may also give you unboundedness. A good way forward is Euler implicit (first order); if you really need second order you can try my new boundedBackward scheme.

Hrv

[anne]

Dear Hrvoje,

Thank you,

I will test your suggestions.

Anne

[borrbyper]

I want to try different ddt-schemes.
In order to see the choices I wrote "x":
------------------------------------------
Creating field DpDt
--> FOAM FATAL IO ERROR : unknown ddt scheme x
Valid ddt schemes are :
6
(
CoEuler
boundedBackward
CrankNicholson
steadyState
Euler
backward
)
file: /home/workdisc/OpenFOAM/pen-1.3/run/boundp3/system/fvSchemes::default at line 27.
From function ddtScheme<type>::New(const fvMesh&, Istream&)
in file /home/dm2/henry/OpenFOAM/OpenFOAM-1.3/src/finiteVolume/lnInclude/ddtScheme.C at line 80.
FOAM exiting
------------------------------------------

When I try boundedBackward I get:
------------------------------------------
Creating field DpDt
Starting time loop
Time = 2e-06
Mean and max Courant Numbers = 0 0
diagonal: Solving for rho, Initial residual = 0, Final residual = 0, No Iterations 0
--> FOAM FATAL IO ERROR : unknown ddt scheme boundedBackward
Valid ddt schemes are :
5
(
CoEuler
CrankNicholson
steadyState
Euler
backward
)
file: /home/workdisc/OpenFOAM/pen-1.3/run/boundp3/system/fvSchemes::default at line 27.
From function ddtScheme<type>::New(const fvMesh&, Istream&)
in file /home/dm2/henry/OpenFOAM/OpenFOAM-1.3/src/finiteVolume/lnInclude/ddtScheme.C at line 80.
FOAM exiting
------------------------------------------

Why is boundedBackward no longer on the list?

[hjasak]

Easy: a bounded backward scheme I wrote is there for scalars only: you cannot define what "bounded" means in terms of a vector or a tensor. Therefore, I cannot define a bounded second-order scheme for a vector.

In order to do this well, you will need to specify a scheme on a per-ddt-term basis. Set default to none and the code will ask you for a scheme one ddt term at a time.

Hrv