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

Numerical problem

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

Reply
 
LinkBack Thread Tools Display Modes
Old   February 10, 2012, 05:40
Default Numerical problem
  #1
New Member
 
Join Date: Feb 2012
Posts: 10
Rep Power: 5
CrisCFD is on a distinguished road
Hello! I have inserted a constant in my model (dielectric constant), which is very low (2,21 E-10) and the simulation does not converge.
What can I do?
CrisCFD is offline   Reply With Quote

Old   February 10, 2012, 06:23
Default
  #2
Super Moderator
 
Glenn Horrocks
Join Date: Mar 2009
Location: Sydney, Australia
Posts: 10,945
Rep Power: 85
ghorrocks has a spectacular aura aboutghorrocks has a spectacular aura aboutghorrocks has a spectacular aura about
You question is meaningless. What does the constant do? How is it coupled to the simulation? What are you modelling?
ghorrocks is offline   Reply With Quote

Old   February 10, 2012, 07:27
Default
  #3
New Member
 
Join Date: Feb 2012
Posts: 10
Rep Power: 5
CrisCFD is on a distinguished road
Quote:
Originally Posted by ghorrocks View Post
You question is meaningless. What does the constant do? How is it coupled to the simulation? What are you modelling?
I previously described my case, in another thread...
I'm modelling a reverse electrodialysis system...so I have Na+ (cations) and Cl- (anions).
So I have two transport equation (cations and anions)
I have to consider the electrical interactions in the transport equation as an additional term, as a source term.
I just found an expression like this: 1) S+ = - σ / ε * (C-A) (cations)
and
2) S- = σ / ε * (C-A) (anions)
Where A and C are the concentrations of anions and cations, respectively
But the very low value of the costant "ε " (dielectric constant = 2,21 E-10) is a problem for the solver, a numerical problem, I think.
Also theese source terns which are dependent on each other could be a problem...I know it is a particular issue...
What do you think about it ?
CrisCFD is offline   Reply With Quote

Old   February 11, 2012, 06:24
Default
  #4
Super Moderator
 
Glenn Horrocks
Join Date: Mar 2009
Location: Sydney, Australia
Posts: 10,945
Rep Power: 85
ghorrocks has a spectacular aura aboutghorrocks has a spectacular aura aboutghorrocks has a spectacular aura about
Just because a constant term is small does not tell you it has a big effect or small. It is how it interacts with the flow which is important. There will be some form of non dimensional number to describe these effects, pesumably the ratio of electrical forces to fluid flow inertial forces or something like that. That sort of number starts telling you how significant your constant term is.
ghorrocks is offline   Reply With Quote

Old   February 15, 2012, 13:20
Default
  #5
New Member
 
Join Date: Feb 2012
Posts: 10
Rep Power: 5
CrisCFD is on a distinguished road
Quote:
Originally Posted by ghorrocks View Post
Just because a constant term is small does not tell you it has a big effect or small. It is how it interacts with the flow which is important. There will be some form of non dimensional number to describe these effects, pesumably the ratio of electrical forces to fluid flow inertial forces or something like that. That sort of number starts telling you how significant your constant term is.
I have no information about this sort of number...this is the reason I'm modelling this process
CrisCFD is offline   Reply With Quote

Old   February 15, 2012, 18:13
Default
  #6
Super Moderator
 
Glenn Horrocks
Join Date: Mar 2009
Location: Sydney, Australia
Posts: 10,945
Rep Power: 85
ghorrocks has a spectacular aura aboutghorrocks has a spectacular aura aboutghorrocks has a spectacular aura about
Starting simulation without knowing which effects are important is a reciepe for disaster. You need to do some back of the envelope calculations to work out what is significant and what is not before starting modelling.
ghorrocks is offline   Reply With Quote

Old   February 18, 2012, 10:06
Default
  #7
New Member
 
Join Date: Feb 2012
Posts: 10
Rep Power: 5
CrisCFD is on a distinguished road
Quote:
Originally Posted by ghorrocks View Post
Starting simulation without knowing which effects are important is a reciepe for disaster. You need to do some back of the envelope calculations to work out what is significant and what is not before starting modelling.
My supervisor ordered me () to consider these electrical interactions
CrisCFD is offline   Reply With Quote

Old   February 18, 2012, 21:07
Default
  #8
Super Moderator
 
Glenn Horrocks
Join Date: Mar 2009
Location: Sydney, Australia
Posts: 10,945
Rep Power: 85
ghorrocks has a spectacular aura aboutghorrocks has a spectacular aura aboutghorrocks has a spectacular aura about
Do you know how many times it has come back to "my supervisor told me to"...

Regardless of what your supervisor says, do some back of the envelope calculations to check it is significant. And the degree of significance will tell you what sort of approach you need to take.
ghorrocks is offline   Reply With Quote

Old   March 6, 2012, 12:06
Default
  #9
New Member
 
Join Date: Feb 2012
Posts: 10
Rep Power: 5
CrisCFD is on a distinguished road
Quote:
Originally Posted by ghorrocks View Post
Do you know how many times it has come back to "my supervisor told me to"...

Regardless of what your supervisor says, do some back of the envelope calculations to check it is significant. And the degree of significance will tell you what sort of approach you need to take.
Experimental measurements showed that these terms are relevant. What can I do?
CrisCFD is offline   Reply With Quote

Old   March 6, 2012, 17:06
Default
  #10
Super Moderator
 
Glenn Horrocks
Join Date: Mar 2009
Location: Sydney, Australia
Posts: 10,945
Rep Power: 85
ghorrocks has a spectacular aura aboutghorrocks has a spectacular aura aboutghorrocks has a spectacular aura about
My previous post explains what you must do. If the electrical terms are relevant, then what about the viscous terms? The inertial terms? If any of these are irrelevant then you can select the correct options to proceed. Otherwise you are just guessing.

But what is the physics of what you are modelling anyway? I seem to remember you asking about imposing a constant drift velocity on top the fluid velocity - is this the same issue?
ghorrocks is offline   Reply With Quote

Old   March 8, 2012, 14:26
Default
  #11
New Member
 
Join Date: Feb 2012
Posts: 10
Rep Power: 5
CrisCFD is on a distinguished road
Quote:
Originally Posted by ghorrocks View Post
My previous post explains what you must do. If the electrical terms are relevant, then what about the viscous terms? The inertial terms? If any of these are irrelevant then you can select the correct options to proceed. Otherwise you are just guessing.

But what is the physics of what you are modelling anyway? I seem to remember you asking about imposing a constant drift velocity on top the fluid velocity - is this the same issue?
I'm modelling the flux of cations and anions through electrodialysis membranes.
So I have to consider the electric interations as source terms in the transport equations (cations and anions). We have a laminar flow in the channel
CrisCFD is offline   Reply With Quote

Old   March 8, 2012, 18:26
Default
  #12
Super Moderator
 
Glenn Horrocks
Join Date: Mar 2009
Location: Sydney, Australia
Posts: 10,945
Rep Power: 85
ghorrocks has a spectacular aura aboutghorrocks has a spectacular aura aboutghorrocks has a spectacular aura about
My question is what is the physics of what you are modelling. To model it you need to know exactly how each important force works because you need to describe an equation which accurately describes it.

Questions include:
* How is charge balance maintained? Does it have a conservation equation?
* How is motion of the charges coupled to the fluid flow?
* Is the charge motion independant of the fluid flow?

But my point about working out the significant forces needs to be considered first. It will tell you what approach you need to take and what things can be ignored. It is quite likely that CFX is not appropriate for what you are modelling at all - you will only know once you know what forces are significant.
ghorrocks 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
conduction problem venkataramana OpenFOAM 3 December 1, 2013 08:30
UDF compiling problem Wouter Fluent UDF and Scheme Programming 6 June 6, 2012 04:43
Problem Importing Geometry ProE to CFX fatb0y CFX 3 January 14, 2012 20:42
Can I solve this problem by Fluent? Kai_kc FLUENT 1 October 27, 2010 05:29
Numerical solution to the rotating disk problem? johny Main CFD Forum 7 September 5, 2005 05:53


All times are GMT -4. The time now is 12:15.