
[Sponsors] 
September 25, 2003, 23:03 
Porous Media Question

#1 
Guest
Posts: n/a

Hey all,
The quadratic resistance in CFX5.6 is stated as following a "generalized" Darcy's law, and uses the expression (for each component): dP/dx = K*rho * U * u (ignoring the linear term) Where U is the magnitude of the total velocity vector (speed), and K*rho is the R2 resistance constant. This would indicate, that the resistance in the x direction is a function of not just the u velocity, but the v, and w velocities as well (since U=sqrt(u^2+v^2+w^2)). This just doesn't feel right to me. With unidirectional flow, I've always seen: dP/dx = K*rho * u * u using the absolute value of u, instead of the vector magnitude, U. This would give a smaller resistance for nonunidirectional flow than the CFX formulation. Does someone know of a reference with a really good development of a generalized Darcy's law in 3D? One CFX paper references a "Convective Heat Transfer" by Bejan. I have this text, but the only reference I can find is a 1D formulation like the one I stated. Can anyone help? Thanks, Jeff 

September 26, 2003, 05:28 
Re: Porous Media Question

#2 
Guest
Posts: n/a

Hi jeff, I think in CFX5, the ForchheimerBrinkmanDarcy equations are in the vector form(3D) correct implemented.You can get The unidirectional flow formulation with simply setting v and w to 0. It always a Problem with thus empirical formulations and their Parameters!!In Bejan also are the equations in vector form (thik)! regards! hanni


September 28, 2003, 02:19 
Re: Porous Media Question

#3 
Guest
Posts: n/a

hi Jeff & Hannibal,i m a new CFX 5.6 user so pls dont mind if my question seems trivial to u. i have problem with specifying BRINKMAN's equation in CFX 5.6 Brinkmans equation is> dp/dl = (myu)*v/k  (myu)*grad^2(v)
how to incorporate it into momentum source in the form S = Cr1*Ui  Cr2*U*Ui + Sspec we will take Cr1 = (myu)/k but how to use the term [(myu)*grad^2(v)] in the form [Cr2*U*Ui + Sspec ] ? Waiting for ur reply. Warm Regards, Paresh Jain 

September 28, 2003, 20:32 
Re: Porous Media Question

#4 
Guest
Posts: n/a

Paresh,
In order to simulate Brinkman's equation, you'll need to set momentum sources for each equation manually (i.e. your terms will have to be calculated in CEL and put into Sspec rather than Cr2). You can get grad(u) from a CEL user FORTRAN function (see the CFX manual which shows how to do this). I suppose you can then put each of these into a user variable (one that's not solved) say ph1, phi2, phi3 and then call the grad user function again on on these new variables. Then combine everything into your Sspec term. I don't know if this will work, but it seems logical. Jeff 

October 1, 2003, 22:24 
Re: Porous Media Question

#5 
Guest
Posts: n/a

Paresh,
Also, dont forget to compute the jacobian, dS/dU, and take the maximum value as the linearisation coefficient. This will give you much better convergence behaviour. Neale 

October 5, 2003, 18:35 
Re: Porous Media Question

#6 
Guest
Posts: n/a

If each component of velocity independently generates resistance, doesn't that mean that the resistance is then a function of the coordinate system selected? If so, that doesn't seem to be the best choice of models for resistance to me.
EG assume rho and K are both = 1. If the flow is moving in a 2D flow field with velocity u,v = 0.5, 0.5 then the resistance in x and y directions would be proportional to (0.5*0.5,0.5*0.5) or (0.25,0.25) and the magnitude of resistance would be 0.353. If you then aligned the x' axis with the flow, velocity would be (0.707,0) and resistance would be (0.5,0) with a magnitude of 0.5. So depending on coordinate system, you get a different resistance to the flow. Let me know if you see something wrong with my math. 

October 6, 2003, 11:21 
Re: Porous Media Question

#7 
Guest
Posts: n/a

... to complete Martin's thought...
If you use multiply by the velocity magnitude, instead of individual components, the frame dependancy goes away. The resistance in the x and y directions are then proportional to (.707*.5, .707*.5) and the magnitude is thus .5. Similary, if you align your xaxis with the flow you get (.707*.707, .707*0) and again a magnitude of .5. Note that this also means that the pressure drop is in the right direction. If you reverse your original xaxis, you get (.707*(.5), .707*.5) and the momentum source is now negative, keeping it in the right direction. By multiplying by velocity, a negative coefficient will always ensure that the source is resistive. Regards, Robin 

October 9, 2003, 21:24 
Re: Porous Media Question

#8 
Guest
Posts: n/a

Robin/Martin,
Thanks for this. This has bugged me for a long time, but I see that the math does indeed work. I'll be able to sleep tonight. Thanks, Jeff 

November 11, 2003, 04:47 
i want join uyour org

#9 
Guest
Posts: n/a

please send informaion to me your hameed pakistan


November 11, 2003, 09:11 
Re: i want join uyour org

#10 
Guest
Posts: n/a

to whom u want to join? for what ?
Paresh Jain 

Thread Tools  
Display Modes  


Similar Threads  
Thread  Thread Starter  Forum  Replies  Last Post 
Problem with rhoPorousMRFSimpleFoam and how to define porous media  Amit Mathur  OpenFOAM  6  July 18, 2013 06:57 
Porous Media coupled with internal flow  Samuel Andrade  FLUENT  2  August 26, 2012 09:43 
Multiphase Porous Media Flow  Convergence Issues  VT_Bromley  FLUENT  5  March 23, 2011 11:38 
Discrete phase model in porous media  Pradeep  FLUENT  0  February 3, 2006 17:30 
Fluid flow in porous media question  Orlando  FLUENT  2  September 12, 2001 10:31 