CFD Online Logo CFD Online URL
www.cfd-online.com
[Sponsors]
Home > Forums > Software User Forums > ANSYS > FLUENT

Why is there an under-relaxation factor for density and body forces?

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

Like Tree3Likes
  • 1 Post By LuckyTran
  • 2 Post By LuckyTran

Reply
 
LinkBack Thread Tools Search this Thread Display Modes
Old   October 25, 2018, 14:55
Default Why is there an under-relaxation factor for density and body forces?
  #1
Member
 
Mostafa
Join Date: Sep 2016
Posts: 30
Rep Power: 10
bikooo3878 is on a distinguished road
Hello everyone,

I'd be thankful if someone could explain why there's an UR for density and body forces when density is constant and there's no gravity.

In addition, I can comprehend why UR for density exists - for the case when density changes with temperature or pressure. Hence, there's an additional equation that's solved for the density alone.
But why is there an UR for body force when they're only calculated by the product of gravity (constant) and density (calculated)?

I've looked everywhere but couldn't find an answer. So I'm really grateful for anyone looking to help.

Thank you.
bikooo3878 is offline   Reply With Quote

Old   October 26, 2018, 10:15
Default
  #2
Member
 
Mostafa
Join Date: Sep 2016
Posts: 30
Rep Power: 10
bikooo3878 is on a distinguished road
Anyone?

Pretty please :'(
bikooo3878 is offline   Reply With Quote

Old   October 26, 2018, 11:03
Default
  #3
Senior Member
 
Lucky
Join Date: Apr 2011
Location: Orlando, FL USA
Posts: 5,742
Rep Power: 66
LuckyTran has a spectacular aura aboutLuckyTran has a spectacular aura aboutLuckyTran has a spectacular aura about
Why is there a urf for density for a constant density case? You can call it laziness from the Fluent developers when they created the GUI. These settings still exist as text lines in the .cas file whether you see them or not in the GUI. Having them shown regardless is being transparent. Maybe you like the black box approach better? Then it's just a matter of preference.

The body force is primarily for multi-phase problems. Body forces can be & some are treated implicitly in Fluent similar to how implicit under-relaxation of equations are used (notice the urf name is momentum and not velocity). Brief explanation of under-relaxation of equations. At each iteration, a relaxed equation is solved. This approach is very different than simply applying a urf to a variable and limiting the change in a variable from iteration or iteration. A limited amount of body force is injected into the correction equation during SIMPLE. There is a note that this implicit body force treatment is done here, but no actual example equations are given in the manual.
bikooo3878 likes this.

Last edited by LuckyTran; October 26, 2018 at 12:57.
LuckyTran is offline   Reply With Quote

Old   October 28, 2018, 02:04
Default
  #4
Member
 
Mostafa
Join Date: Sep 2016
Posts: 30
Rep Power: 10
bikooo3878 is on a distinguished road
I've captured a screenshot as soon as you replied and have been discussing it with my colleagues since then. With your edit, it makes much more sense.

Now I'm just puzzled between implicit and explicit urf.

If you'd only assure me that not all flow urf are always implicit. That is, the density urf is an explicit one, since it there's no implicit equation needed to calculate it. Turbulence quantities are similarly explicit. They are both calculated with pressure and velocity values at the same node.

Energy, on the other hand, is an equation to which we input an implicit urf.

Body forces depend on the nature of the problem.

If I'm wrong, I'd be grateful if you corrected me.

Thank you.
bikooo3878 is offline   Reply With Quote

Old   October 29, 2018, 14:09
Default
  #5
Senior Member
 
Lucky
Join Date: Apr 2011
Location: Orlando, FL USA
Posts: 5,742
Rep Power: 66
LuckyTran has a spectacular aura aboutLuckyTran has a spectacular aura aboutLuckyTran has a spectacular aura about
Explicit/implicit in relaxation has a different meaning than explicit/implicit solver.

Explicit relaxation means you relax the variable. This means you solve an equation and instead of taking the full update, you take only a partial update of that variable. Let's say you start with x=1 and you execute your algorithm (and do a bunch of calculations) and at the end of everything your algorithm says that the solution should be x=2. If my explicit urf is 0.5 then instead of taking x=2, I would take x=1.5 as the final solution for this iteration.

Implicit under-relaxation means you relax the entire equation. This means

that at any iteration, you don't actually solve the equation but a relaxed equation. Here I cannot say whether the solution should be x=2, because instead of solving the equation which would've given be x=2 as the solution, I solve a relaxed equation, which can produce any other result.

Both variables and equations can be relaxed and both can be applied simultaneously. For example, I can apply implicit relaxation to the momentum equation and apply explicit relaxation to the velocity. Another practical example (and you can find this happening in Fluent) is with the energy equation. I can apply implicit relaxation to the energy equation (and solve for enthalpy/internal energy using the relaxed energy equation) and instead of directly calculating temperature from the equation of state, I instead apply explicit relaxation onto the calculation of temperature. This guarantees that me temperature will always be inconsistent with the internal energy or enthalpy, but I do it anyway because it improves stability.

If you use the COUPLED solver in Fluent, you'll find more examples of both implicit and explicit relaxation applied one on top of the other.

Quote:
Originally Posted by bikooo3878 View Post
That is, the density urf is an explicit one, since it there's no implicit equation needed to calculate it.
Not necessarily true because you solve the continuity equation for density.

Quote:
Originally Posted by bikooo3878 View Post
Turbulence quantities are similarly explicit. They are both calculated with pressure and velocity values at the same node.
Which turbulence quantities? Quantities like k and omega or k and epsilon have transport equations and these are solved using the transport equations. k and epsilon are not calculated from pressure and velocity. These equations have implicit relaxation applied. The turbulent viscosity on the other hand does use explicit relaxation.

Body forces are like source terms in the momentum equation. You don't have to inject the full amount of the source term into the equation. It is possible to apply implicit relaxation to a single term.

So it's not correct for me to say that implicit relation always refers to an entire equation. The difference is that implicit relaxation relaxes the equation you solve (i.e. it changes the coefficient matrix). Explicit relaxation means you solve an equation and don't take the full update.


All equations can have implicit urf's. All variables can have explicit urf's. The question is, for a particular solver, how many are applied?
hnemati and Aoratos like this.
LuckyTran is offline   Reply With Quote

Old   October 29, 2018, 14:25
Default
  #6
Member
 
Mostafa
Join Date: Sep 2016
Posts: 30
Rep Power: 10
bikooo3878 is on a distinguished road
Thank you for your clear answer. It really has helped me a lot.

Quote:
Originally Posted by LuckyTran View Post
It is possible to apply implicit relaxation to a single term.
This is the only thing I couldn't understand. implicit relaxation is applied to the whole equation (it changes the factors), how can I apply an implicit relaxation to a just single term? Is it like injecting part of the old value in the source term?

Sorry if this is a dump question

Thank you.
bikooo3878 is offline   Reply With Quote

Old   October 29, 2018, 16:43
Default
  #7
Senior Member
 
Lucky
Join Date: Apr 2011
Location: Orlando, FL USA
Posts: 5,742
Rep Power: 66
LuckyTran has a spectacular aura aboutLuckyTran has a spectacular aura aboutLuckyTran has a spectacular aura about
Quote:
Originally Posted by bikooo3878 View Post
Thank you for your clear answer. It really has helped me a lot.



This is the only thing I couldn't understand. implicit relaxation is applied to the whole equation (it changes the factors), how can I apply an implicit relaxation to a just single term? Is it like injecting part of the old value in the source term?

Sorry if this is a dump question

Thank you.

Yes, implicit under-relaxation is like injecting a part of the old value into the equation for the new value. Under-relaxation of an entire equation looks like so, specifically you should look at (18.4-23). The difference is phi is more than just one thing. You can split phi up into its parts and apply a different urf for each part (different alpha's). There isn't a rule that says you have to apply the same under-relaxation (the same alpha) to every term. That is, you can group terms by physical meaning (advection, diffusion, source, body force) and apply different alpha's to each.


(18.4-2) also gives you an idea of physically what is happening when you apply implicit under-relaxation. It's as if the solution steps in time a certain amount (local cell time). When you apply different urf's to each term, it's as if each term (say the advective term, vs the diffusion term, vs the body force term) advances the solution in time by different amounts.
LuckyTran is offline   Reply With Quote

Old   October 30, 2018, 12:42
Default
  #8
Member
 
Mostafa
Join Date: Sep 2016
Posts: 30
Rep Power: 10
bikooo3878 is on a distinguished road
Well this explains a lot

Can't thank you enough
bikooo3878 is offline   Reply With Quote

Reply

Tags
body force, density, fluent, under relaxation factor

Thread Tools Search this Thread
Search this Thread:

Advanced Search
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 Off
Pingbacks are On
Refbacks are On


Similar Threads
Thread Thread Starter Forum Replies Last Post
simpleFoam body forces wedge BC interation! CFD-Lover OpenFOAM Programming & Development 8 April 19, 2017 18:44
Estimating drag and lift forces on rectangular solid body in Abaqus ismailqeshta Main CFD Forum 0 November 20, 2016 00:58
Body forces in OpenFOAM samiam1000 OpenFOAM 2 November 22, 2013 09:50
strong body forces Ankan Main CFD Forum 0 August 2, 2005 02:13
Body Forces Thomas P. Abraham Main CFD Forum 1 April 21, 1999 21:24


All times are GMT -4. The time now is 13:52.