CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   CFX (http://www.cfd-online.com/Forums/cfx/)
-   -   Importance of droplet size (http://www.cfd-online.com/Forums/cfx/24905-importance-droplet-size.html)

Itchie December 10, 2007 03:25

Importance of droplet size
 
Hi guys,

I'm simulating an incompressible flow with liquid droplets in a continuous vapor flowing in a pipe. There are 2inlets located on the same plane. The inner inlet is a mixture of liquid and gas while the other inlet is pure gas(gas is 2x faster than droplet). I initially simulated a 3D case and the results have shown that the pressure recovery is dependent on droplet size. In my case, changing the droplet size from 10micromm to 20micro gives a 3.5 kPa difference in pressure recovery. The difference decreases as the droplet size increases.I found similar results when I used 2D axisymmetric case. However, for a 1D fortran program (journal reference), the results show that droplet size has little effect on pressure recovery compared to 3D and 2D. The difference in 1D is about 0.5 kPa per 10micromm increase in droplet size.I know that the other components (y&z)are the main cause of this phenomena but I find it hard to further explain it theoretically. I really need some insights guys. PLEASE HELP ME! Big Thanks!-itchie

Glenn Horrocks December 10, 2007 17:17

Re: Importance of droplet size
 
Hi,

Have a close read of the drag law models in the CFX manual. Make sure you are using the correct drag model for the droplets. Also check whether a distribution of droplet sizes is required or if a single droplet size is adequate.

Glenn Horrocks

Itchie December 13, 2007 03:01

Re: Importance of droplet size
 
Thank you so much for your reply Glenn! I used Schiller Nauman drag model. To consider your advise, I used a constant drag coefficient (for both CFX and FORTRAN) but I found similar relationship.The droplet size is still important for CFX unlike in 1D fortran.I used constant droplet size and the CFX result approaches the experimental value as the droplet size increases. Also, I made a mesh 1element thick with 3 nodes (almost "1D") along the radial direction and the result shows that droplet size is not important (similar to 1D fortran). I want to have theoretical explanation and this is just a try... I compared the velocity and pressure profiles for both cfx and fortran program.In FORTRAN, it shows that droplet velocity profile is nearly independent of the droplet size. However in CFX, there is a significant difference in droplet velocity profile as the droplet size changes. In CFX, the vapor can give more acceleration to smaller droplets.The larger kinetic energy absorbed by the droplet can raise the pressure of the gas more. Also, I only used continuity imbalance as a convergence criteria for 1D fortran. Do you I think I must also include momentum imbalance check? My explanation is just a try. I would really appreciate any insights from you guys.THANKS A LOT!


All times are GMT -4. The time now is 12:02.