How to set the two parameters of solver binterFoamb

luckyluke · May 15, 2005, 03:17

Hi,everyone.
1.There are two parameters needed by 'interFoam': one is 'cGamma', the other is 'pesi' used in 'div(phi,gamma) Gauss Gamma201 0.2'. What behavior is controled by the two parameters? Is the solver 'interFoam' diffusive? How many meshes will be crossed by the interface after long time computation?
2.Is the CICSAM scheme more diffusive than the current schoem of OpenFOAM?
3.I want to mention the current interFoam algorithm in my future article, where can I get the accurate introduction to it?

hjasak · May 15, 2005, 03:24

There is no perfect reference for this because the algorithm has never been published. The closest I've got (the description is a couple of years old and we have moved on a bit since) is Henrik Rusche's PhD Thesis from Imperial College:

@PhdThesis{Rusche:PhD,
author = {Rusche, H.},
title = {Computational fluid dynamics of dispersed two-phase
flows at high phase fractions},
school = {Imperial College, University of London},
year = 2003
}

Enjoy,

Hrv

henry · May 15, 2005, 05:01

My interface-capturing algorithm in interFoam has already been disussed in detail on this site. You can use the search facility to fine the threads.

luckyluke · May 31, 2005, 22:46

Unfortunately, I cannot find Rusche's thesis from the web site ProQuest Digital Dissertations.

duderino · June 3, 2005, 07:13

Does it actually matter how the phases are defined?

e.g.

phase1 = always the denser phase
phase2 = always the lighter phase

or does it matter which phase is disperse?

e.g.

bubble problems:
phase1 = gas
phase2 = liquid (water) or vice versa

spray probelms:

phase1 = liquid (water, fuel)
phase2 = gas or vice versa

henry · June 3, 2005, 07:16

No it doesn't matter which phase is which, in fact phase-inversion can be handled without problem.

duderino · June 3, 2005, 07:20

So my assumption that the compression works more on phase1 is wrong than?

henry · June 3, 2005, 07:33

The compression is symmetric so it doesn't matter if we solve for gamma or (1 - gamma). In fact in bubbleFoam both equations are coded and there is a method to combine the solution which I had hoped might improve robustness but it didn't make a difference to the overall behaviour of the algorithm.

May 15, 2005, 03:17	Hi,everyone. 1.There are two	#1
luckyluke Member Luckyluke Join Date: Mar 2009 Posts: 51 Rep Power: 17	Hi,everyone. 1.There are two parameters needed by 'interFoam': one is 'cGamma', the other is 'pesi' used in 'div(phi,gamma) Gauss Gamma201 0.2'. What behavior is controled by the two parameters? Is the solver 'interFoam' diffusive? How many meshes will be crossed by the interface after long time computation? 2.Is the CICSAM scheme more diffusive than the current schoem of OpenFOAM? 3.I want to mention the current interFoam algorithm in my future article, where can I get the accurate introduction to it?

May 15, 2005, 03:24	There is no perfect reference	#2
hjasak Senior Member Hrvoje Jasak Join Date: Mar 2009 Location: London, England Posts: 1,907 Rep Power: 33	There is no perfect reference for this because the algorithm has never been published. The closest I've got (the description is a couple of years old and we have moved on a bit since) is Henrik Rusche's PhD Thesis from Imperial College: @PhdThesis{Rusche:PhD, author = {Rusche, H.}, title = {Computational fluid dynamics of dispersed two-phase flows at high phase fractions}, school = {Imperial College, University of London}, year = 2003 } Enjoy, Hrv lourencosm likes this. __________________ Hrvoje Jasak Providing commercial FOAM/OpenFOAM and CFD Consulting: http://wikki.co.uk

May 15, 2005, 05:01	My interface-capturing algorit	#3
henry Senior Member Join Date: Mar 2009 Posts: 854 Rep Power: 22	My interface-capturing algorithm in interFoam has already been disussed in detail on this site. You can use the search facility to fine the threads.

May 31, 2005, 22:46	Unfortunately, I cannot find R	#4
luckyluke Member Luckyluke Join Date: Mar 2009 Posts: 51 Rep Power: 17	Unfortunately, I cannot find Rusche's thesis from the web site ProQuest Digital Dissertations.

June 3, 2005, 07:13	Does it actually matter how th	#5
duderino Member Duderino Join Date: Mar 2009 Posts: 40 Rep Power: 17	Does it actually matter how the phases are defined? e.g. phase1 = always the denser phase phase2 = always the lighter phase or does it matter which phase is disperse? e.g. bubble problems: phase1 = gas phase2 = liquid (water) or vice versa spray probelms: phase1 = liquid (water, fuel) phase2 = gas or vice versa

June 3, 2005, 07:16	No it doesn't matter which pha	#6
henry Senior Member Join Date: Mar 2009 Posts: 854 Rep Power: 22	No it doesn't matter which phase is which, in fact phase-inversion can be handled without problem.

June 3, 2005, 07:20	So my assumption that the comp	#7
duderino Member Duderino Join Date: Mar 2009 Posts: 40 Rep Power: 17	So my assumption that the compression works more on phase1 is wrong than?

June 3, 2005, 07:33	The compression is symmetric s	#8
henry Senior Member Join Date: Mar 2009 Posts: 854 Rep Power: 22	The compression is symmetric so it doesn't matter if we solve for gamma or (1 - gamma). In fact in bubbleFoam both equations are coded and there is a method to combine the solution which I had hoped might improve robustness but it didn't make a difference to the overall behaviour of the algorithm.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
exper parameters on solver	laura	CFX	1	January 9, 2009 08:50
Relaxation and Solver Parameters	norbinou	Siemens	3	September 25, 2007 02:38
parameters to select a solver for a model in gambt	abid	FLUENT	2	September 20, 2006 14:07
Do you know of any tweaks of the solver parameters to accelerate solution convergence	sek	OpenFOAM Running, Solving & CFD	0	September 15, 2006 14:45
Inflation parameters	Adrian Mawdsley	CFX	2	February 11, 2003 09:13