
[Sponsors] 
April 14, 2016, 11:26 
Porous domain

#1 
New Member
Join Date: Sep 2015
Posts: 13
Rep Power: 3 
Hi all,
I have got a question regarding porous media in CFX, and more generally in fluid dynamic. What does it physically mean to create a porous media with a certain porosity but with no loss (no permeability, no resistance coefficient)? Is porosity a loss on its own? Are permeability/resistance coefficient higher order losses? Thanks for your help 

April 14, 2016, 15:14 

#2 
Senior Member
Join Date: Jun 2009
Posts: 573
Rep Power: 13 
ANSYS CFX provides 3 different type of domains: fluid, solid and porous to create a model.
Defining a domain does not imply any physics until you select/activate specific model details. For example, when you define a fluid domain it is not clear yet if you have a viscous fluid, or an inviscid fluid. That is defined in the material details. Similarly, when you define a porous domain (not media) there is no loss until the parameters to describe the nature of the loss are defined. Some "porous media" behave quadratically, and others linearly. ANSYS CFX does not assume either, and it is up to you to define the characteristics of such "porous media" by setting the parameters within the "porous domain" Hope the above helps, 

April 14, 2016, 15:53 

#3 
New Member
Join Date: Sep 2015
Posts: 13
Rep Power: 3 
Thanks Opaque for your answer.
I understand what you are saying about the porous domain (and not media, sorry for the confusion). However, there is something that confuses me a bit. Physically, what would be a porous domain without losses? For example, if we model a fluidized bed ( which is a porous media) by a porous domain, then permeability is a major component of this domain. Permeability is a loss. So I would assume that every porous domain should contain some losses, shouldn't they ? Thanks again 

April 14, 2016, 16:45 

#4 
Senior Member
Join Date: Jun 2009
Posts: 573
Rep Power: 13 
When you define a porous domain, you must enter certain parameters:
Volume available for the fluid, i.e. Volume Porosity. If you do not add anything else, the loss is undefined; therefore, 0. Physical or not, it is a model defined by the user not the software. Therefore, what are you trying to model? and what data do you have available to characterize your model ? Not sure I understand where your question is coming from. The software provides you tools to model what you want, it cannot assume a behavior without input data. 

April 14, 2016, 17:20 

#5 
New Member
Join Date: Sep 2015
Posts: 13
Rep Power: 3 
It is indeed easier with input data...
What I am trying to model in effect of a spring in a valve. I don't want to include the spring geometry in the model (too complex to have an appropriate mesh) therefore I thought of including a porous domain which would represent the spring "blockage". The porosity would simply be the volume of fluide available in the domain divided by its total volume. However, where I am less sure is I should include any losses (permeability, resistance coeff...). I hope it is a bit clearer. Thanks again for your time 

November 9, 2016, 04:08 
Volume porosity

#6 
New Member

Hi,
I am trying to model a multiphase flow (resinair) in porous media with varying volume porosity in ANSYSCFX. I have 1 (atm) pressure in the inlet and 0.1 (atm) in outlet as boundary condition and volume porosity increase through thickness of the geometry from 0.5 to 1. I observe that flow front position leads where porosity is 0.5 and lags where porosity is 1. However, I expect the opposite result because when porosity is high it means more space for flow and where porosity reaches 1 I expect fluid channel behavior which comparatively has lower loss and resistance to flow. Could you please clarify this issue for me. Is there any problem in set up? or results are true. permeability: 1e10 [m^2] fluid viscosity: 0.1 [Pa s] 

November 9, 2016, 13:48 

#7 
Senior Member
Join Date: Jun 2009
Posts: 573
Rep Power: 13 
For starters, do you have a setup with static pressure specified at inlet and outlets ?
If so, such setup is illconditioned. 

November 10, 2016, 08:15 

#8 
New Member

Opaque, thanks for your reply.
I have Total pressure of 1 (atm) in the inlet and relative pressure 0.1 (atm) in the outlet. I was receiving a warning because of "Inlet" and "Outlet" Boundary conditions which disappeared after I changed boundary conditions to "Opening". but the same problem still makes me confused. I do no understand why flow leads where porosity is lower and lags where porosity is higher. As far as I am concerned, porosity deals with the relative volume available for fluid to flow. Therefore, higher porosity means better space for flow. My case study is composite manufacturing process simulation where propagation of viscous resin into a porous fiber bead continues until the full impregnation of the preform. Taking into account of "resistance loss coefficient" in the "isotropic loss model", I should admit that the value of the coefficient does not effect results. However, I expect lower loss in high porosity regions where there is more available space in comparison with densely compacted fibers which reduce porosity and increase fiber volume fraction. Physically and experimentally, flow propagation should occur in regions with high porosity values much easier. I have also checked the injection boundary condition by changing a "total pressure" BC into a "normal speed" BC of 0.02 (m/s). But the flow still leads in the regions with lower porosity values. I would be truly thankful if some one could clarify this issue for me. 

November 10, 2016, 09:52 

#9 
Senior Member
Join Date: Jun 2009
Posts: 573
Rep Power: 13 
OK.. The boundary conditions make sense..
On the flow comparisons, are you comparing fully developed flow velocity profiles, or developing flow ? For fully developed velocity profile, I agree the maximum velocity should be at the center of the channel (assuming a regular cross section, no buoyancy, etc), and larger as porosity increases. For developing flow, it is not so trivial since the developing region should be longer than for fully open space (porosity = 1). In that section, the comparisons between runs for different porosities may be misleading. My 2cents 

November 10, 2016, 10:09 

#10 
Senior Member
Join Date: Jun 2009
Posts: 573
Rep Power: 13 
Deleted repeated submit


Thread Tools  
Display Modes  


Similar Threads  
Thread  Thread Starter  Forum  Replies  Last Post 
Strange Behaviour of Porous Domain (Pictures Included)  Trev0r  CFX  0  April 20, 2015 21:29 
Radiation at interface between fluid and porous domain  Hitch8  CFX  19  April 20, 2015 06:24 
Floating point exception: Zero divide  liladhar  CFX  11  December 16, 2013 05:07 
Implementation of a porous domain  megacrout  OpenFOAM  1  January 12, 2012 08:02 
Porous domain setup from single pressure loss value  siw  CFX  1  December 8, 2011 17:36 