VOF based on SIMPLE method

chinaduck · March 7, 2014, 23:30

Hello everyone,
My name is Peter. I am new student in the VOF. I am writing a VOF code based on SIMPLE method. Now, my code can compute the low density ratio flow. But when set the two phases to be water and air, the computed results are wrong. The velocity magnitude could exceed 100 m/s. I have no idea what is wrong. Could anyone help me and give me some suggestions?
Thanks.

Best,

Peter

suhas jain · May 22, 2014, 03:32

Hi Peter,

I am new to VOF too. I have written VOF on SIMPLE algorithm. I too encounter same problem as yours i.e simulation runs fine only for low density ratio flows. If you could find out the mistake and rectify it, let me know as well..

Regards,
suhas

beer · May 22, 2014, 04:14

Hi

This topic sounds interesting. Just a few thoughts of mine. First of all even commercial codes have these problem, so I think it is not trivial. I used the vof model in Fluent in the past which hilariously failed with the level-set method for water-air. The method failed because of the poor implementation of the surface force. That leads me to my second thought. Do you use surface force and if yes how is it implemented?

Regards

suhas jain · May 22, 2014, 06:55

Hi,

I have implemented CSF method with Height Function to calculate curvature and I am quiet sure that the problem is not with surface force implementation because I have validated it with static drop tests for density ratios upto 10000:1(According to paper by Francois et al.) But the validation using dynamic tests failed. I see very high velocities even when I neglect surface forces. So I guess the problem is some where in coupling between SIMPLE algorithm and VOF, or am I wrong ?

Regards,

Suhas.

beer · June 23, 2014, 23:15

Hi

Sorry for not writing so long, my mail account was not accessible.

If it works with the static test case you are probably right, that the problem is somewhere in the vof-simple coupling. I am not there yet with my own code, so I can't tell you more from experience. Maybe you can have a look at how OpenFoam solved the problem.
The fact that the dynamic case fails makes me think if you have a small enough CFL number.

rmh26 · June 25, 2014, 13:04

Quote:

Originally Posted by suhas jain

Hi,

I have implemented CSF method with Height Function to calculate curvature and I am quiet sure that the problem is not with surface force implementation because I have validated it with static drop tests for density ratios upto 10000:1(According to paper by Francois et al.) But the validation using dynamic tests failed. I see very high velocities even when I neglect surface forces. So I guess the problem is some where in coupling between SIMPLE algorithm and VOF, or am I wrong ?

Regards,

Suhas.

Did you actually validate it or are you just citing Francois et al.? The studies in that paper where the claim to eliminate the spurious currents were for drops where the exact analytical curvature of the drop was used not the values calculated from the height function. I'm solve my equations with projection methods not the SIMPLE algorithm so I'm not too familiar with what you are doing. But if you are getting errors even without surface forces something must be wrong. Are you including the density variation in you Pressure equation.

My pressure equation looks like
$\nabla \cdot \frac{1}{\rho} \nabla P = \nabla \cdot \tilde u$

when I have density variations.

leflix · July 20, 2014, 14:31

In principle somewhere in your code you have to compute the mean density and mean viscosity from the volume fraction C (which is between 0 and 1 when the interface cross a control volume) and from rho1 and rho2 and mu1 and mu2 the density and viscosity in the the water and air respectively. What kind of mean do you use? If you use an arithmetic mean as rho = C*rho1 + (1-C)*rho2 (the same for mu) it may fail for large density ratio (or large vicosity ratio)
You should better try an harmonic mean 1/rho = C*(1/rho1) + (1-C)*(1/rho2) this gives much more stable results.

March 7, 2014, 23:30	VOF based on SIMPLE method	#1
chinaduck New Member Peter Join Date: May 2012 Location: New York Posts: 18 Rep Power: 14	Hello everyone, My name is Peter. I am new student in the VOF. I am writing a VOF code based on SIMPLE method. Now, my code can compute the low density ratio flow. But when set the two phases to be water and air, the computed results are wrong. The velocity magnitude could exceed 100 m/s. I have no idea what is wrong. Could anyone help me and give me some suggestions? Thanks. Best, Peter

May 22, 2014, 03:32		#2
suhas jain New Member suhas jain suresh Join Date: Aug 2012 Location: Stanford, USA Posts: 27 Rep Power: 14	Hi Peter, I am new to VOF too. I have written VOF on SIMPLE algorithm. I too encounter same problem as yours i.e simulation runs fine only for low density ratio flows. If you could find out the mistake and rectify it, let me know as well.. Regards, suhas __________________ Suhas Jain Graduate student, Stanford University suhas.jain2000@gmail.com

May 22, 2014, 06:55		#4
suhas jain New Member suhas jain suresh Join Date: Aug 2012 Location: Stanford, USA Posts: 27 Rep Power: 14	Hi, I have implemented CSF method with Height Function to calculate curvature and I am quiet sure that the problem is not with surface force implementation because I have validated it with static drop tests for density ratios upto 10000:1(According to paper by Francois et al.) But the validation using dynamic tests failed. I see very high velocities even when I neglect surface forces. So I guess the problem is some where in coupling between SIMPLE algorithm and VOF, or am I wrong ? Regards, Suhas. __________________ Suhas Jain Graduate student, Stanford University suhas.jain2000@gmail.com

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
What is C.V. based finite element method	C-H Kuo	Main CFD Forum	4	September 19, 2022 15:06
SIMPLE Method	satyajeet13	Main CFD Forum	3	October 2, 2012 15:02
SIMPLE family and Fractional Step Method	Geon-Hong	Main CFD Forum	2	April 5, 2010 19:05
SIMPLE method for inviscid flow	abcdef123	Main CFD Forum	0	February 26, 2010 09:24
Ask help wiht a SIMPLE Method on compressible flow.	universez	Main CFD Forum	0	February 8, 2010 11:38

May 22, 2014, 04:14		#3
beer Member Join Date: Dec 2012 Posts: 92 Rep Power: 13	Hi This topic sounds interesting. Just a few thoughts of mine. First of all even commercial codes have these problem, so I think it is not trivial. I used the vof model in Fluent in the past which hilariously failed with the level-set method for water-air. The method failed because of the poor implementation of the surface force. That leads me to my second thought. Do you use surface force and if yes how is it implemented? Regards

June 23, 2014, 23:15		#5
beer Member Join Date: Dec 2012 Posts: 92 Rep Power: 13	Hi Sorry for not writing so long, my mail account was not accessible. If it works with the static test case you are probably right, that the problem is somewhere in the vof-simple coupling. I am not there yet with my own code, so I can't tell you more from experience. Maybe you can have a look at how OpenFoam solved the problem. The fact that the dynamic case fails makes me think if you have a small enough CFL number.

July 20, 2014, 14:31		#7
leflix Senior Member Join Date: Aug 2011 Posts: 272 Rep Power: 16	In principle somewhere in your code you have to compute the mean density and mean viscosity from the volume fraction C (which is between 0 and 1 when the interface cross a control volume) and from rho1 and rho2 and mu1 and mu2 the density and viscosity in the the water and air respectively. What kind of mean do you use? If you use an arithmetic mean as rho = Crho1 + (1-C)rho2 (the same for mu) it may fail for large density ratio (or large vicosity ratio) You should better try an harmonic mean 1/rho = C(1/rho1) + (1-C)(1/rho2) this gives much more stable results.