Does anyone out there have experience modelling hydrocyclones or other devices with tangential inlets?

I know that the turbulent anisotropy inherent with the high streamline curvature in these devices necessitates the use of second order closure models (i.e. RSM SSG), but even when I use such a model, the pressure drop predicted by the code (CFX10) seems to severly underpredict the experimentally obtained pressure drop (by a factor of 2). I am running the high resolution scheme as a steady state simulation on a tetra mesh. I am only running single phase and neglecting the air core, does anyone know what type of effect this will have on the pressure drop? I have tried running two phase (water and air core) but get very poor convergence.

Anyway, it would be great to hear of other peoples experience.

Thanks

I'm working on a hydrocyclone, too. You could try to set a max. time step, because I had to reduce it to obtain an air core. And I have to carry out transient simulations to get a good convergence. If you neglecting the air core the pressure drop will be under predicted, because water has a larger area to exhaust from the cyclone. hence, lower velocity, hence lower pressure drop.

[asifmohammad7]

hi
i am unable to predict the air core that is in my simulation air core is not forming in the hydrocyclone plz tell me how can i get the air core.
thank you

[ghorrocks]

What turbulence model are you using?

[asifmohammad7]

i am using turbilence model coz i also have to consider the effects of turbulence as the tangential inlet flow is at very high velocity 3m/sec or 4.5m/sec which causes turbulence.

[ghorrocks]

Sure, but what turbulence model are you using?

[asifmohammad7]

I have tried standard k-epsilon and rsm models but air core is not forming once i have also tried les model but the same thing happened. My overflow and underflow boundary conditions are pressure outlets and i m using steady state pressure based solver.

[asifmohammad7]

and sorry for my earlier silly reply ,at that time i was having headache and thought that u were asking that "why r u using turbulence model".....................

[ghorrocks]

No problem.

You are unlikely to get anything realistic with a k-e model. The curvature effects are not captured and you end up with something like solid body rotation.

RSM might work, but probably not. From what I know of this topic (it is not my area of expertise) you will need an LES approach to really get it to work.

I trust you are not trying to do a steady state LES model.

[asifmohammad7]

Thank you for your valuable advice, now i will try to use LES WITH AN UNSTEADY APPROACH

[ghorrocks]

I see. The name of LES done in steady state is called a "mistake". I suggest you do some reading on LES approaches before you commence so you know why LES is meaningless in steady state. If you do not understand the approach you will get nowhere.

[asifmohammad7]

hello
i am not getting any aircore using LES in unsteady state solver but i am getting an aircore with air fraction 20% using RSM.

[ghorrocks]

I repeat what I said before about LES. Have you checked that you are using a discretisation scheme with suitable levels of dissipation? Are you sure your mesh size is the correct size relative to the turbulent flow structures you aim to capture? Likewise your timestep?

If you have not done these things you have not done an LES simulation.

You might be interested in some 2-eqn, RSM and LES modelling I did years ago on vorticies in a model IC engine. It showed features like an aircore for RSM and LES approaches, and k-e gave solid body rotation. Have a look at http://hdl.handle.net/2100/248

[asifmohammad7]

ok,thanks for sending me the link............

[ghorrocks]

No problem. Chapter 6 is the section I am referring to, it shows the effect of various turbulence models onto vortex features similar to what you are doing.

[asifmohammad7]

thanks again............

[asifmohammad7]

HI GLENN,
CAN u plz tell me how a gas-liquid interface can be created using VOF.

[ghorrocks]

1. Do the tutorials. That will show you.
2. Please do not post irrelevant comments on somebody else's thread. Start a new thread.

[asifmohammad7]

thanks,i will start new thread.

[arjun3020]

Hi all,

Could you please tell me how to find pressure drop in cyclone using CFD post or fluent?

i have pressure contour which gives highest pressure of 1005 Pa. and lowest pressure of -303 Pa.

How to find actual pressure drop in cyclone?

See the attachment of contour.