When to set momentumPredictor to Yes

hsieh · March 14, 2006, 22:58

Hi,

Can someone tell me under what condiction should momentumPredictor be set to yes? Thanks!

Pei

pierre · March 15, 2006, 05:35

As far as I understand, in the pressure-velocity algorithm for two-phase flows in Foam does not need the momentum equation to be solved, so in general momentumPredictor should be set to no, especially for "challenging" test cases (i.e if there is problems with convergence or sharp volume fractio gradient). In this case the momentum equation is only constructed and used to obtain fluxes and solve the pressure equation. On the other hand solving the momentum eqaution might speed up convergence in non-challenging steady-state calculations.

Hope that helps

Pierre

hsieh · March 15, 2006, 14:54

Thanks Pierre!

How about turbFoam?

Pei

pierre · March 16, 2006, 14:07

TurbFoam uses the standard PISO algorithm, momentum equation is always solved. it can be used for unsteady or steady cases.

Pierre,

marc_wattenberg · January 22, 2009, 05:53

Hi together,

for me it is not clear where the velocity field is coming from when the momentum equation is not solved. perhaps it is a stupid question but for the moment a need a hint

thanks
marc

eugene · January 22, 2009, 06:32

U = U* - grad(p)/A

where U* = H/A

i.e. U is solved explicitly after the pressure equation.

hjasak · January 22, 2009, 11:25

Hi Marc,

It depends on the importance of the laplacian in the momentum equation: if you have important diffusivity in the momentum, you should solve the momentum equation implicitly.

Enjoy,

Hrv

idrama · March 16, 2010, 04:16

Hey Foamers,

I read your posts. Could anybody tell me where to find a direct statement about this, i.e. no long text. Friends referred me to Ferziger. But I cannot find the right page.

Cheers

sandy · March 16, 2010, 07:27

I want to know too. I guess, if we set no, maybe it is slow to get convergence to some cases, but they will get the same results, right? Who can give some light?

matejfor · March 17, 2010, 02:10

Dear Claus,

Ferzinger , Peric, Comput. Methods for Fluid Dynamics is very good book and I guess you should read it all anyway. But not to waste your time, have a look at chapter 7.5 as a start. Also these slides from stanford I googled my appeal:http://www.stanford.edu/class/me469b...mpressible.pdf

As a good quote from this site sais: "Two weeks of playing with a CFD code will save you one afternoon of reading". (sorry, I forgot the author).

good luck
matej

March 14, 2006, 22:58	Hi, Can someone tell me und	#1
hsieh Senior Member Pei-Ying Hsieh Join Date: Mar 2009 Posts: 317 Rep Power: 18	Hi, Can someone tell me under what condiction should momentumPredictor be set to yes? Thanks! Pei vivek05 likes this.

March 15, 2006, 05:35	As far as I understand, in the	#2
pierre Member Pierre Le Fur Join Date: Mar 2009 Location: UK Posts: 60 Rep Power: 17	As far as I understand, in the pressure-velocity algorithm for two-phase flows in Foam does not need the momentum equation to be solved, so in general momentumPredictor should be set to no, especially for "challenging" test cases (i.e if there is problems with convergence or sharp volume fractio gradient). In this case the momentum equation is only constructed and used to obtain fluxes and solve the pressure equation. On the other hand solving the momentum eqaution might speed up convergence in non-challenging steady-state calculations. Hope that helps Pierre minh khang, wayne14, Turin Turambar and 4 others like this.

March 15, 2006, 14:54	Thanks Pierre! How about tu	#3
hsieh Senior Member Pei-Ying Hsieh Join Date: Mar 2009 Posts: 317 Rep Power: 18	Thanks Pierre! How about turbFoam? Pei

March 16, 2006, 14:07	TurbFoam uses the standard PIS	#4
pierre Member Pierre Le Fur Join Date: Mar 2009 Location: UK Posts: 60 Rep Power: 17	TurbFoam uses the standard PISO algorithm, momentum equation is always solved. it can be used for unsteady or steady cases. Pierre,

January 22, 2009, 05:53	Hi together, for me it is n	#5
marc_wattenberg New Member Marc Wattenberg Join Date: Mar 2009 Posts: 3 Rep Power: 17	Hi together, for me it is not clear where the velocity field is coming from when the momentum equation is not solved. perhaps it is a stupid question but for the moment a need a hint thanks marc

January 22, 2009, 06:32	U = U* - grad(p)/A where U*	#6
eugene Senior Member Eugene de Villiers Join Date: Mar 2009 Posts: 725 Rep Power: 21	U = U* - grad(p)/A where U* = H/A i.e. U is solved explicitly after the pressure equation. minh khang and Zane like this.

January 22, 2009, 11:25	Hi Marc, It depends on the	#7
hjasak Senior Member Hrvoje Jasak Join Date: Mar 2009 Location: London, England Posts: 1,905 Rep Power: 33	Hi Marc, It depends on the importance of the laplacian in the momentum equation: if you have important diffusivity in the momentum, you should solve the momentum equation implicitly. Enjoy, Hrv minh khang, ancolli, Swift and 5 others like this. __________________ Hrvoje Jasak Providing commercial FOAM/OpenFOAM and CFD Consulting: http://wikki.co.uk

March 16, 2010, 04:16		#8
idrama Senior Member Claus Meister Join Date: Aug 2009 Location: Wiesbaden, Germany Posts: 241 Rep Power: 17	Hey Foamers, I read your posts. Could anybody tell me where to find a direct statement about this, i.e. no long text. Friends referred me to Ferziger. But I cannot find the right page. Cheers

March 16, 2010, 07:27		#9
sandy Senior Member Sandy Lee Join Date: Mar 2009 Posts: 213 Rep Power: 18	I want to know too. I guess, if we set no, maybe it is slow to get convergence to some cases, but they will get the same results, right? Who can give some light?

March 17, 2010, 02:10		#10
matejfor Senior Member matej forman Join Date: Mar 2009 Location: Brno, Czech Republic Posts: 182 Rep Power: 17	Dear Claus, Ferzinger , Peric, Comput. Methods for Fluid Dynamics is very good book and I guess you should read it all anyway. But not to waste your time, have a look at chapter 7.5 as a start. Also these slides from stanford I googled my appeal:http://www.stanford.edu/class/me469b...mpressible.pdf As a good quote from this site sais: "Two weeks of playing with a CFD code will save you one afternoon of reading". (sorry, I forgot the author). good luck matej Nicole, CFDTeufel and shizuka like this.