CFD Online Logo CFD Online URL
www.cfd-online.com
[Sponsors]
Home > Forums > OpenFOAM

Bug in the PilchErdman Breakup Model

Register Blogs Members List Search Today's Posts Mark Forums Read

Reply
 
LinkBack Thread Tools Display Modes
Old   August 23, 2011, 04:57
Default Bug in the PilchErdman Breakup Model
  #1
Member
 
chenkai
Join Date: May 2010
Location: munich
Posts: 44
Rep Power: 7
xck1986 is on a distinguished road
Hi everyone,

I have notice, there is a bug in the PilchErdman Breakup Model in OpenFOAM.

In the source code of the Model, the following two equ.,

scalar Vd = Urmag*rho12*(B1_*taubBar * B2_*taubBar*taubBar);
......
scalar Ds = 2.0*Wec*sigma*Vd1/(Vd1*rhoc*sqr(Urmag));

which are not the same as in the paper: Use of breakup time
data and velocity history date to predict the maximun size of
stable fragments for acceleration
induced breakup of a liquid
drop
.

I think it should be rewritten like following:

scalar Vd = Urmag*rho12*(B1_*taubBar + B2_*taubBar*taubBar);
......
scalar Ds = 2.0*Wec*sigma/(Vd1*rhoc*sqr(Urmag));

Does anyone also notice this error.

Thanks

Chenkai
xck1986 is offline   Reply With Quote

Old   August 30, 2011, 04:59
Default
  #2
Member
 
chenkai
Join Date: May 2010
Location: munich
Posts: 44
Rep Power: 7
xck1986 is on a distinguished road
Nobody has ever used this model?

The problem is that, when I change the code of this model as in the paper. My Simulation results are even worse than before. I don't if here is really a bug or not.
Hope someone can help me?

Thanks a lot!
xck1986 is offline   Reply With Quote

Old   January 16, 2013, 08:08
Default
  #3
New Member
 
Join Date: Jun 2012
Posts: 14
Rep Power: 5
bigeddy is on a distinguished road
A year late but this post already exits so im updating it here.

Im looking at this model currently for my own purposes. You are correct, the model should be what you suggest as a fix i.e.

scalar Vd = Urmag*rho12*(B1_*taubBar + B2_*taubBar*taubBar);
......
scalar Ds = 2.0*Wec*sigma/(Vd1*rhoc*sqr(Urmag));

Also,

// update the droplet diameter according to the rate eq. (implicitly)
d = (d + frac*Ds)/(1.0 + frac);

makes no sense whatsoever. It should be

d = d*(1.0-frac)+Ds*frac

when im less busy ill update the git repository if someone else don't get to it first.
bigeddy is offline   Reply With Quote

Old   January 16, 2013, 14:28
Default
  #4
Super Moderator
 
niklas's Avatar
 
Niklas Nordin
Join Date: Mar 2009
Location: Stockholm, Sweden
Posts: 693
Rep Power: 19
niklas will become famous soon enough
Quote:
Originally Posted by bigeddy View Post
A year late but this post already exits so im updating it here.

Im looking at this model currently for my own purposes. You are correct, the model should be what you suggest as a fix i.e.

scalar Vd = Urmag*rho12*(B1_*taubBar + B2_*taubBar*taubBar);
......
scalar Ds = 2.0*Wec*sigma/(Vd1*rhoc*sqr(Urmag));

Also,

// update the droplet diameter according to the rate eq. (implicitly)
d = (d + frac*Ds)/(1.0 + frac);

makes no sense whatsoever. It should be

d = d*(1.0-frac)+Ds*frac

when im less busy ill update the git repository if someone else don't get to it first.
BUG!!! Inconcievable
I cant belieave I've missed this one.
I will talk to the guys and correct it.
About the 'makes no sense whatsoever' part. You could do it like that. Its not wrong, but neither is the implemented update. (as shown below)

f = dt/tau

we're trying to solve this eq:
dD/dt = -(D-Ds)/tau

1) explicit, approximate D with D0 in the RHS
(D1-D0)/dt = -(D0-Ds)/tau
D1 – D0 = -(D0 – Ds)*f
D1 = D0 - D0*f + Ds*f
D1 = (1-f)*D0 + Ds*f


2) implicit, approximate D with D1 in the RHS
(D1-D0)/dt = -(D1 – Ds)/tau
(D1-D0)= -D1*f + Ds*f
(1+f)*D1 - D0 = Ds*f
(1+f)*D1 = D0 + Ds*f
D1 = (D0 + f*Ds)/(1+f)
niklas is offline   Reply With Quote

Old   January 16, 2013, 15:29
Default
  #5
Super Moderator
 
niklas's Avatar
 
Niklas Nordin
Join Date: Mar 2009
Location: Stockholm, Sweden
Posts: 693
Rep Power: 19
niklas will become famous soon enough
now...you may wonder 'why is method 2 to prefer'?

think of large timesteps (f >> 1)

first method yields

D1 -> f*(Ds-D0) (makes no physical sense)

second method yields

D1 -> f*Ds/f = Ds (makes perfect sense)
niklas is offline   Reply With Quote

Old   January 16, 2013, 18:07
Default
  #6
New Member
 
Join Date: Jun 2012
Posts: 14
Rep Power: 5
bigeddy is on a distinguished road
Ahhh, I see, thanks for reminding me of the two methods!
bigeddy is offline   Reply With Quote

Reply

Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are On
Pingbacks are On
Refbacks are On


Similar Threads
Thread Thread Starter Forum Replies Last Post
multiphaseInterFoam for RAS turbulence model chiven OpenFOAM Bugs 7 August 23, 2011 02:52
help for different between les model (subgrid-scale model) liuyuxuan FLUENT 1 October 2, 2009 15:25
2 stage axial turbine model convergence issues sherifkadry CFX 2 September 7, 2009 20:51
multi fluid mixture model issue rystokes CFX 3 August 9, 2009 19:13
Advanced Turbulence Modeling in Fluent, Realizable k-epsilon Model Jonas Larsson FLUENT 5 March 13, 2000 04:27


All times are GMT -4. The time now is 20:21.