
[Sponsors] 
April 24, 2006, 04:40 
Hi, Im doing my thesis work an

#1 
Guest
Posts: n/a

Hi, Im doing my thesis work and need to know which the governing equations are for the simpleFoam application.
If I understand the code correctly it should be: div(phi, U) + turbulence>divR(U)= grad(p)); where: phi is the potential of U (but it has the dimension m³/s?) what does turbulence>divR(U) represent? what does the following lines do? volScalarField AU = UEqn().A(); U = UEqn().H()/AU; 

April 24, 2006, 04:50 
Hi!
For the first part of y

#2 
Member
Andreas Hauffe
Join Date: Mar 2009
Location: Dresden, Germany
Posts: 36
Rep Power: 10 
Hi!
For the first part of your question I can help you. In the Programmer's Guide on page 54 you can see this equation. R is just nueff*grad(U) and nueff is nu+nut and is calculated by the turbulence model. Phi should be U in this case. And the p is pressure/density. Andreas 

April 24, 2006, 05:06 
Hi!
Ups, I think your right

#3 
Member
Andreas Hauffe
Join Date: Mar 2009
Location: Dresden, Germany
Posts: 36
Rep Power: 10 
Hi!
Ups, I think your right with phi. It is the potential of U. But part of the discretised form is multiplyed with a surface value because of the surface integrals, right? Andreas 

April 24, 2006, 05:11 
phi is in general the flux (rh

#4 
Member
stefan
Join Date: Mar 2009
Posts: 96
Rep Power: 10 
phi is in general the flux (rho*u)
>what does the following lines do? >volScalarField AU = UEqn().A(); >U = UEqn().H()/AU; This lines are need to precondition (jacobian) the U matrix. hth 

April 24, 2006, 05:22 
> what does turbulence>divR(U

#5 
Senior Member
Gavin Tabor
Join Date: Mar 2009
Posts: 181
Rep Power: 10 
> what does turbulence>divR(U) represent?
This is the turbulence modelling. When you average the NSE you get back a very similar set of equations  the Reynolds Averaged NavierStokes equations  which have one additional term, which represents the effect of the fluctuations around the average (the turbulence) on the mean flow. We usually write this as div(R) where R is a 2nd rank tensor (the Reynolds Stress). Since we don't know this term (the closure problem) we need to model this; which is usually done by assuming that the turbulence has a dissipative effect and creating a 'turbulent viscosity' (which is usually so large it swamps the physical viscosity) and write R=nu grad(u). In FOAM this all gets swallowed into the turbulence model; turbulence>divR is getting this out of the model in a convenient (and mathematically stylish) way. Gavin 

April 24, 2006, 05:42 
Hi!
As far as I understood

#6 
Member
Andreas Hauffe
Join Date: Mar 2009
Location: Dresden, Germany
Posts: 36
Rep Power: 10 
Hi!
As far as I understood these things in this case div(R) is not only the Reynolds stress tensor of the RANS equations. But it is only a question of definition I think. div(R) = laplacian(nu,U) + Reynolds stress tensor Andreas 

April 24, 2006, 05:52 
Hi!
Next mistake. It should

#7 
Member
Andreas Hauffe
Join Date: Mar 2009
Location: Dresden, Germany
Posts: 36
Rep Power: 10 
Hi!
Next mistake. It should be div(R) = laplacian(nu,U) + div(Reynolds stress tensor) There is a mistake in Hrv's PhD. Thesis on page 71. Andreas 

April 24, 2006, 05:58 
Thanks a lot!
This was big

#8 
Guest
Posts: n/a

Thanks a lot!
This was big help. I will follow my papers on Reynolds equation on average flow values. a thought isnt div(UU)=2U*div(U) which would give 1/2*div(phi, U) + turbulence>divR(U)= grad(p)); 

April 24, 2006, 06:12 
No, there is no mistake in Hrv

#9 
Senior Member
Hrvoje Jasak
Join Date: Mar 2009
Location: London, England
Posts: 1,806
Rep Power: 25 
No, there is no mistake in Hrv's Thesis on page 71.
Hrv
__________________
Hrvoje Jasak Providing commercial FOAM/OpenFOAM and CFD Consulting: http://wikki.co.uk 

April 24, 2006, 06:25 
Hi!
In this case I want to

#10 
Member
Andreas Hauffe
Join Date: Mar 2009
Location: Dresden, Germany
Posts: 36
Rep Power: 10 
Hi!
In this case I want to apologise to you for my mistake. But for my understanding I have to ask why in equation (2.30) the Reynold stress tensor is just added? So this equation consits of 5 vectors and a tensor. Where is my mistake? Andreas 

April 24, 2006, 06:32 
Hmm, my mistake in fact: it sh

#11 
Senior Member
Hrvoje Jasak
Join Date: Mar 2009
Location: London, England
Posts: 1,806
Rep Power: 25 
Hmm, my mistake in fact: it should say div{(overline{Uprime Uprime})} right? and I am missing the div.
Will fix it asap. Hrv
__________________
Hrvoje Jasak Providing commercial FOAM/OpenFOAM and CFD Consulting: http://wikki.co.uk 

April 24, 2006, 06:40 
Hi!
I think div{(overline{U

#12 
Member
Andreas Hauffe
Join Date: Mar 2009
Location: Dresden, Germany
Posts: 36
Rep Power: 10 
Hi!
I think div{(overline{Uprime Uprime})} is right. But your PhD thesis is just a great work to understand what OpenFOAM does. And in comparison to stupid things I'm telling sometimes it is without mistakes! Andreas 

April 24, 2006, 07:02 
where can I find this PhD thes

#13 
Guest
Posts: n/a

where can I find this PhD thesisk. It sounds like Ive could use a thing or two from it?


April 24, 2006, 07:30 
You can try the FOAM CFD resou

#14 
Senior Member
Hrvoje Jasak
Join Date: Mar 2009
Location: London, England
Posts: 1,806
Rep Power: 25 
You can try the FOAM CFD resource site http://www.foamcfd.org  there's quite a lot of stuff there including a bunch of PhD Theses done with OpenFOAM I managed to get in electronic form. Alternatively, you can try my personal web pages on http://www.h.jasak.dial.pipex.com.
Enjoy, Hrv
__________________
Hrvoje Jasak Providing commercial FOAM/OpenFOAM and CFD Consulting: http://wikki.co.uk 

April 2, 2008, 08:59 
hi to all,
I m a beginner i

#15 
Guest
Posts: n/a

hi to all,
I m a beginner in OpenFoam and I want to program fluctuations with tensor... for example I want the following program to do with fluctuations. magSqr(fvc::grad(U))+ sqr(gradU.component(tensor::YX) + gradU.component(tensor::XY)) does someone can help me? best regards 

April 2, 2008, 09:01 
hi to all,
I m a beginner i

#16 
Guest
Posts: n/a

hi to all,
I m a beginner in OpenFoam and I want to program fluctuations with tensor... for example I want the following program to do with fluctuations. magSqr(fvc::grad(U))+ sqr(gradU.component(tensor::YX) + gradU.component(tensor::XY)) can someone help me? best regards 

Thread Tools  
Display Modes  


Similar Threads  
Thread  Thread Starter  Forum  Replies  Last Post 
governing equation of sliding mesh  torn  FLUENT  1  August 10, 2006 03:34 
UDF to add two governing equation?? *NM*  T.K. Yang  FLUENT  0  May 9, 2006 03:03 
UDF to add two governing equation??  T.K. Yang  FLUENT  0  May 9, 2006 02:54 
Linearisation of governing equations  MärtaKarin Brandt  CFX  3  November 14, 2003 10:33 
1D governing equations  Andrei Chernousov  Main CFD Forum  5  July 13, 1999 15:06 