[tobiasand]

Hello,

I have a problem where I am studying very small particles (micro meter scale) with lagrangian particle tracking and want to study the deposition of particles on the wall in a geometry. But due to isotropic decomposition of the turbulence kinetic energy (fluctuations) I get an overdepositon of particles on the wall. This can be fixed by multiplying the turbulent kinetic energy k with a function of y+ which goes to zero when y+ goes to zero and this will damp out the fluctuations normal to the surface. I am using a SST turbulence model.

The question is how to implement this in CFX and as I understand you have to do this with a particle source routine and this is something completely new to me.

Is there any tutorials of how to get started with Fortran and using it in CFX to get me started?

Thanks in advance,

Tobias

[Gert-Jan]

Why not use Fluent? Where you have anisotropic turbulence incorporated in the turbulent dispersion.........

[ghorrocks]

Please note y+ only exists on the control volume adjacent to the wall. It does not exist in the volume mesh. This means your y+ function will not be smooth but a function of mesh size - and therefore it will be impossible to achieve a mesh independant solution. And secondly it will stuff up the boundary layer model as you are modifying k. So I don't think your idea of modifying k with y+ is a good idea.

I suspect that a model like this would require a very fine mesh in the boundary layer to adequate capture the physics of the particles in the boundary layer. If you use a mesh which has a good boundary layer resolution, and fine particle time steps then you will get the effect you are looking for with no modifications, just accurate modelling of the real physics at work.

[Gert-Jan]

Tobias, Maybe you should look at the help

paragraph: 16.2.12. Variables and Predefined Expressions Available in CEL Expressions

You can find that a variable is available called "Wall Distance".
I think that variable makes more sence than Y plus.
Not sure under which conditions it is available, though.....

[tobiasand]

How this overdeposition is solved is described in section 9 in article [1].

The thing is how to implement it in CFX. I am pretty new to CFD and have only used CFX so far, I will try to model the problem in Fluent and see if it works better.

I will also try refine the mesh in the boundary layer region and use a finer time step and see how this will change the result.

Thank you very much for your answers.

[1] Matida, E. A., et al. "Improved numerical simulation of aerosol deposition in an idealized mouth–throat." Journal of Aerosol Science 35.1 (2004): 1-19.

[ghorrocks]

I do not have access to that article, so cannot see the model you are proposing to implement.

Before you do implement it make sure you are aware of the particle tracking options available in CFX already, especially things like turbulent dispersion.