[aquila1213]

Hi,
I am simulating a complicated geometry system which has some reactions.
The real case is a multi-phase turbulent flow.
but this turbulency only affects on diffusion. I want to change diffusion coefficients for laminar flow in a way that gives the same diffusion when flow is turbulent with normal diffusion coefficient. Any ideas?

[piu58]

I see the only way for this in modeling: Take a simple geometry like a flow over a plate or a flow past a sphere and compare laminar flow with turbulent one. Change the Diffusion coefficient of the laminar flow until it fits (more or less) the turbulent flow. You may use this Diffusion coefficient for geometries which are similar to your model geometry.
If you have to perform plenty of calculation with one geometry, you may use your real geometry as model, of course. You may get some experience how the effective diffusion coefficient changes, and may be able to use that for a serie of calculations with differen b.c. or material properties.

[aquila1213]

Originally Posted by piu58

I see the only way for this in modeling: Take a simple geometry like a flow over a plate or a flow past a sphere and compare laminar flow with turbulent one. Change the Diffusion coefficient of the laminar flow until it fits (more or less) the turbulent flow. You may use this Diffusion coefficient for geometries which are similar to your model geometry.
If you have to perform plenty of calculation with one geometry, you may use your real geometry as model, of course. You may get some experience how the effective diffusion coefficient changes, and may be able to use that for a serie of calculations with differen b.c. or material properties.

Thanks a lot. nice idea.
but i have another problem! I am working with COMSOL. in which I cant simulate this with models like mixture, Euler-Euler, bubbly flow. My case is a packed bed reactor with high aeration(15 ml/s) from down and solution (50 ml/hr) from up in a 6*22 cm cylinder.
Do you have any ideas?

[piu58]

Originally Posted by aquila1213

Thanks a lot. nice idea.
but i have another problem! I am working with COMSOL. in which I cant simulate this with models like mixture, Euler-Euler, bubbly flow. My case is a packed bed reactor with high aeration(15 ml/s) from down and solution (50 ml/hr) from up in a 6*22 cm cylinder.
Do you have any ideas?

I am afraid, I don't understand where the problem is. If you have such a high aerial flow you can assume that the aerial part of your reactor is pure air, without any reaction products and without any additional moisture. May be, that is not even necessary.

At the other end, the velocity of the solution is so low that it doesn't count how the inner surface of your reactor is.

I think your reactor can be replaced by a flat bed reactor with the surface of the packed bed. A small experiment gives you the layer thickness of the liquid film: Put slowly fluid in your reactor until the first drops appear at the downside, and wait for some extra drops. You you get the volume the reactor needs for get wet.
From the geometry of the bed you get the thickness of the air layer above the fluid in your flat bed reactor.

It counts to simplify the model as far as possible.