cavitating flow over a hydrofoil

Actone-Zero · January 25, 2011, 20:35

Hi everyone,
I'm simulating a cavitation flow over a hydrofoil and I have a problem when I have large shedding of vapor structures (low cavitation numbers leading to unsteady sheet/ckoud cavitation). It seems that the simulation is doing well but in some cases it completely diverges and I have vapor everywhere in the domain. It usually start in the wake where a cloud suddenly develop to the entire domain. I double checked all my parameters (I mean pressure, boundary conditions ...) I'm pretty sure it is right.

I'm not very familiar with all the coefficients of the Rayleigh Plesset model so I use the default values, also I noticed that using a first order Euler scheme for time discretization fix the problem but I need second order...

Do you guys have any idea where the problem can come from?

Thank you !!!

ghorrocks · January 26, 2011, 00:25

Cavitation models are tricky, transient ones even more so. Any clues to the divergence? Do the residuals start to grow? The first thing to try is smaller timesteps.

The default values I think are tuned to clean fresh water, so may need some tweaks for river or sea water.

Actone-Zero · February 3, 2011, 13:50

Ok, thank you very much for the answer.

Concerning the time step I think it is small enough. Do the CFL need to be 1 in all cases? My average value is around 1. My time step is 1e-4s for a velocity of about 5m/s, Re=750000. I started a computation with a lower time step but it takes time.

I'm now inversigating different possible problems:
1- I've seen that the model has been calibrated for leading edge and mid chord cavitation, what happen if I increase the condensation coefficient to avoid development of vapor in the entire domain?

2- I've read in the manual the following:
It should be emphasised that if a poor initial guess is supplied with the cavitation model active (i.e., a non-zero initial vapor volume fraction), then it is possible to arrive at a physically impossible situation where most of the flow domain is cavitating.

In the initial solution, I turn off the cavitation model, however my volume fraction of vapor is not 0, it is between 1e-15 and 3e-15 , which I guess can be considered as 0 ?
However, this is exactly what happens in my simulation so maybe it is not propely initialized.

3- What about the maximum density ratio? I set it to default which is 1000, however it is much higher in my case, whatever the temperature (say between 10 000 and 40 000). Do I have to change the max density raion in consequence?

ghorrocks · February 3, 2011, 18:37

0) Timestep - Do a sensitivity study to check the timestep size. Anything else is just guesswork.

1) No idea, never tried it. As long as you understand the model then go for it.

2) Multiphase simulations often have a lower limit on volume fraction to improve numerical accuracy. Don't adjust this unless you have problems.

3) If your density ratio is greater than 1000 then you might have to increase it. But be aware that this will cause numerical stability problems. Often cavitation models clip the density of the gas phase for numerical stability as the effect of density in the gas phase is small. This may be the case for you, in which case you can allow the density to clip on the density ratio limit.

January 25, 2011, 20:35	cavitating flow over a hydrofoil	#1
Actone-Zero New Member Join Date: Jan 2010 Posts: 8 Rep Power: 16	Hi everyone, I'm simulating a cavitation flow over a hydrofoil and I have a problem when I have large shedding of vapor structures (low cavitation numbers leading to unsteady sheet/ckoud cavitation). It seems that the simulation is doing well but in some cases it completely diverges and I have vapor everywhere in the domain. It usually start in the wake where a cloud suddenly develop to the entire domain. I double checked all my parameters (I mean pressure, boundary conditions ...) I'm pretty sure it is right. I'm not very familiar with all the coefficients of the Rayleigh Plesset model so I use the default values, also I noticed that using a first order Euler scheme for time discretization fix the problem but I need second order... Do you guys have any idea where the problem can come from? Thank you !!!

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Cavitating flow condensation problem	mange	FLUENT	3	August 30, 2011 02:49
reversed flow at velocity inlet / mass flow inlet	ib	FLUENT	1	March 26, 2007 14:11
How to change from mass flow to volume flow rate	stanley	FLUENT	1	February 2, 2007 07:44
Plug Flow	Franck	Main CFD Forum	3	September 4, 2003 06:57
Kármán vortex street in cavitating flow behind bodies in the cavitation tunnel	L. Könözsy	Main CFD Forum	0	April 17, 2000 14:16

January 26, 2011, 00:25		#2
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,870 Rep Power: 144	Cavitation models are tricky, transient ones even more so. Any clues to the divergence? Do the residuals start to grow? The first thing to try is smaller timesteps. The default values I think are tuned to clean fresh water, so may need some tweaks for river or sea water.

February 3, 2011, 13:50		#3
Actone-Zero New Member Join Date: Jan 2010 Posts: 8 Rep Power: 16	Ok, thank you very much for the answer. Concerning the time step I think it is small enough. Do the CFL need to be 1 in all cases? My average value is around 1. My time step is 1e-4s for a velocity of about 5m/s, Re=750000. I started a computation with a lower time step but it takes time. I'm now inversigating different possible problems: 1- I've seen that the model has been calibrated for leading edge and mid chord cavitation, what happen if I increase the condensation coefficient to avoid development of vapor in the entire domain? 2- I've read in the manual the following: It should be emphasised that if a poor initial guess is supplied with the cavitation model active (i.e., a non-zero initial vapor volume fraction), then it is possible to arrive at a physically impossible situation where most of the flow domain is cavitating. In the initial solution, I turn off the cavitation model, however my volume fraction of vapor is not 0, it is between 1e-15 and 3e-15 , which I guess can be considered as 0 ? However, this is exactly what happens in my simulation so maybe it is not propely initialized. 3- What about the maximum density ratio? I set it to default which is 1000, however it is much higher in my case, whatever the temperature (say between 10 000 and 40 000). Do I have to change the max density raion in consequence?

February 3, 2011, 18:37		#4
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,870 Rep Power: 144	0) Timestep - Do a sensitivity study to check the timestep size. Anything else is just guesswork. 1) No idea, never tried it. As long as you understand the model then go for it. 2) Multiphase simulations often have a lower limit on volume fraction to improve numerical accuracy. Don't adjust this unless you have problems. 3) If your density ratio is greater than 1000 then you might have to increase it. But be aware that this will cause numerical stability problems. Often cavitation models clip the density of the gas phase for numerical stability as the effect of density in the gas phase is small. This may be the case for you, in which case you can allow the density to clip on the density ratio limit.