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

2D vertical laminar flow - Error: Divergence detected in AMG solver: x-momentum

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

Like Tree1Likes

Reply
 
LinkBack Thread Tools Display Modes
Old   October 17, 2015, 10:19
Unhappy 2D vertical laminar flow - Error: Divergence detected in AMG solver: x-momentum
  #1
New Member
 
Join Date: Oct 2015
Posts: 3
Rep Power: 3
srk999 is on a distinguished road
Getting the error: Divergence detected in AMG solver: x-momentum

Please see Fluent file in link below.

My goal is to view Velocity Contours and Profiles of a 2D vertical laminar flow in a pipe. It is a simple 10m x 1m pipe. I chose the geometry to be 10x0.5 and set one edge as center line. I've tried playing with relaxation factors, but I am very new to Fluent, and don't know where it's going wrong.

Please help.

Project file: https://www.dropbox.com/s/3tn417atl3...rtLam.zip?dl=0
srk999 is offline   Reply With Quote

Old   October 17, 2015, 20:54
Default
  #2
New Member
 
Danny
Join Date: Feb 2015
Posts: 13
Rep Power: 3
skewness abyss is on a distinguished road
Hello,

I don't seem to have the same version of Fluent as you do, so I was only able to access your mesh file. I ran the simulation assuming 2D planar flow and 5 m/s inlet velocity and I was able to get a converged solution 2nd order no problem. I then switch to axisymetric and I could not even initialize the solution. In the axisymetric simulations I've done, the centerline is on the x-axis and is treated with an axis boundary condition. Why don't you try aligning your pipe horizontally and setting your lower BC as an axis?

PS: If you want to get meaningful results, you should cluster elements near your walls to capture the sharp gradients that exist there.
skewness abyss is offline   Reply With Quote

Old   October 17, 2015, 21:21
Thumbs up
  #3
New Member
 
Join Date: Oct 2015
Posts: 3
Rep Power: 3
srk999 is on a distinguished road
Quote:
Originally Posted by skewness abyss View Post
Hello,

I don't seem to have the same version of Fluent as you do, so I was only able to access your mesh file. I ran the simulation assuming 2D planar flow and 5 m/s inlet velocity and I was able to get a converged solution 2nd order no problem. I then switch to axisymetric and I could not even initialize the solution. In the axisymetric simulations I've done, the centerline is on the x-axis and is treated with an axis boundary condition. Why don't you try aligning your pipe horizontally and setting your lower BC as an axis?

PS: If you want to get meaningful results, you should cluster elements near your walls to capture the sharp gradients that exist there.
Is there any way to save the project so that you can open it? There doesn't seem to be any option in Save As dialog box to save for other versions.

As per your reply, I am running the simulation as planar. Noticed something that could be symptomatic of the problem- seeing velocity on x-axis. Shouldn't velocity be zero on x-axis given it's a vertical pipe flow? Or, at least I intend to make the velocity on x-axis zero. How do I fix this?

I need to make it vertical since I have to take the acceleration 9.81m/s^2 into account. Do I input -9.81m/s^2 or just 9.81m/s^2 in Gravitation Acceleration "Y" field? Does Fluent intelligently puts the negative sign in the background?

If all else fails, I suppose I could make it a horizontal pipe flow and apply gravitational acceleration (positive or negative value?) on x-axis? What do you think?
srk999 is offline   Reply With Quote

Old   October 17, 2015, 21:38
Default
  #4
New Member
 
Danny
Join Date: Feb 2015
Posts: 13
Rep Power: 3
skewness abyss is on a distinguished road
I'm not sure how to save your project so I can view it since I only use Fluent as a stand alone component.

The velocity should be zero on a solid wall. If you use a symmetry or axis BC, the velocity is allowed to be nonzero there. The only place your velocity should be zero is at your pipe wall.

If you would like to include gravity in the simulation for a horizontal pipe aligned with the x-axis, you can set the x-component of gravity as -9.81 m/s^2. If you would like to stick to your vertical pipe, you can set the y component of gravity as -9.81 m/s^2

I recommend switch to a horizontal pipe unless someone else on the forum knows how to make an axisymetric simulation with the centerline oriented vertically.
skewness abyss is offline   Reply With Quote

Old   October 17, 2015, 21:44
Default
  #5
New Member
 
Join Date: Oct 2015
Posts: 3
Rep Power: 3
srk999 is on a distinguished road
Quote:
Originally Posted by skewness abyss View Post
I'm not sure how to save your project so I can view it since I only use Fluent as a stand alone component.

The velocity should be zero on a solid wall. If you use a symmetry or axis BC, the velocity is allowed to be nonzero there. The only place your velocity should be zero is at your pipe wall.

If you would like to include gravity in the simulation for a horizontal pipe aligned with the x-axis, you can set the x-component of gravity as -9.81 m/s^2. If you would like to stick to your vertical pipe, you can set the y component of gravity as -9.81 m/s^2

I recommend switch to a horizontal pipe unless someone else on the forum knows how to make an axisymetric simulation with the centerline oriented vertically.
Thank you. I'll go with the horizontal pipe flow then.
srk999 is offline   Reply With Quote

Old   April 25, 2016, 09:42
Default
  #6
Member
 
enass
Join Date: Feb 2015
Location: Alexandria-Egypt
Posts: 30
Rep Power: 3
enass massoud is on a distinguished road
Hiii, can you help with that please, I am trying to model single phase flow in vertical pipe. I have managed to do the horizontal case. For the vertical case, i just added the gravity in the y direction, is that all to be changed?
I revived the following error;
reversed flow in 6 faces on pressure-outlet 8.
339 4.4464e-11 4.3180e-08 1.1237e-06 -2.9912e-04 3.7380e-10 5.5147e-18 0:01:27 661

The problem is axisymmetric. When i make it plannar, no such error is noticed, Could you help me here.
enass massoud is offline   Reply With Quote

Old   April 26, 2016, 08:22
Default
  #7
Senior Member
 
Lucky Tran
Join Date: Apr 2011
Location: Orlando, FL USA
Posts: 1,372
Rep Power: 20
LuckyTran will become famous soon enough
Quote:
Originally Posted by enass massoud View Post
Hiii, can you help with that please, I am trying to model single phase flow in vertical pipe. I have managed to do the horizontal case. For the vertical case, i just added the gravity in the y direction, is that all to be changed?
I revived the following error;
reversed flow in 6 faces on pressure-outlet 8.
339 4.4464e-11 4.3180e-08 1.1237e-06 -2.9912e-04 3.7380e-10 5.5147e-18 0:01:27 661

The problem is axisymmetric. When i make it plannar, no such error is noticed, Could you help me here.
The reversed flow is simply a warning, not an error. It should go away with more iterations. If it doesn't go away, then check your problem setup. I've found from experience that a 2D axissymmetric problem is a bit more numerically unstable than 2D planar and that sometimes reducing the default urf's is needed.
LuckyTran is offline   Reply With Quote

Old   April 26, 2016, 09:08
Default
  #8
Member
 
enass
Join Date: Feb 2015
Location: Alexandria-Egypt
Posts: 30
Rep Power: 3
enass massoud is on a distinguished road
Yes the geometry has a problem. The symmetry is bout y axis? how can i solve this problem? i know that fluent makes axisymmetric around x axis
enass massoud is offline   Reply With Quote

Old   April 26, 2016, 13:40
Default
  #9
Senior Member
 
Lucky Tran
Join Date: Apr 2011
Location: Orlando, FL USA
Posts: 1,372
Rep Power: 20
LuckyTran will become famous soon enough
Quote:
Originally Posted by enass massoud View Post
Yes the geometry has a problem. The symmetry is bout y axis? how can i solve this problem? i know that fluent makes axisymmetric around x axis
Unfortunately you must follow Fluent's convention for 2D axissymmetric simulations. Reorient your geometry so that the ais of symmetry is about the x-axis.

Then enable gravity and put the direction vector for gravity to be the x-axis.
LuckyTran is offline   Reply With Quote

Old   April 27, 2016, 05:23
Default
  #10
Member
 
enass
Join Date: Feb 2015
Location: Alexandria-Egypt
Posts: 30
Rep Power: 3
enass massoud is on a distinguished road
Thanks for your reply. If you dont mind could you clarify that to me. I have a vertical pipe with centerline being in y direction. Then how will i solve that issue. Pleasse refer to attached photo
Attached Images
File Type: png model geometry.png (20.7 KB, 5 views)
enass massoud is offline   Reply With Quote

Old   April 27, 2016, 15:39
Default
  #11
Senior Member
 
Lucky Tran
Join Date: Apr 2011
Location: Orlando, FL USA
Posts: 1,372
Rep Power: 20
LuckyTran will become famous soon enough
Quote:
Originally Posted by enass massoud View Post
Thanks for your reply. If you dont mind could you clarify that to me. I have a vertical pipe with centerline being in y direction. Then how will i solve that issue. Pleasse refer to attached photo
Just do it in 3D and put gravity in the y+ direction, then you won't have any issues.

To do it in 2D, you must orient the pipe axis in the x-axis (because Fluent will only let you do the x-axis. Either go back to your mesher and reorient it there, or rotate the mesh in Fluent. Whatever.

You'll need to also put gravity in the x direction. You also need to mesh only from the centerline to one wall (because the mesh is implicitly rotated 360 degrees, or 2pi radians). You cannot mesh from wall to wall. There is a tutorial for 2D axissymmetric flows (it might not have gravity) but to enable gravity is straightforward. Just enable the gravity option, and then specify the direction vector for gravity.
LuckyTran is offline   Reply With Quote

Old   April 28, 2016, 05:20
Default
  #12
Member
 
enass
Join Date: Feb 2015
Location: Alexandria-Egypt
Posts: 30
Rep Power: 3
enass massoud is on a distinguished road
Thanks Lucky in advance. Let me know one more thing please, regarding the Boundary conditions for the velocity inlet it should be v=1m/s in the radial direction. Now the pipe in geometry and meshing is horizontal? shall i define the velocity axial or radial?
enass massoud is offline   Reply With Quote

Old   April 28, 2016, 10:07
Default
  #13
New Member
 
VT
Join Date: Jan 2016
Posts: 16
Rep Power: 2
chikko04 is on a distinguished road
i am following your same tutorials, so you are at first exercise, it's right?
To solve your issue you have to translate your geometry completly within the first quadrant, far away the axys (X and Y).
I don't know why, but that is... The solution converge in base of the viscosity you have choosed, in about 100, 130 steps, but have something strange.
If you put your geometry in horizontal, everything is fine.

If you want to talk about it, or in general of those tutorials or exerces and results, you can contact me.
Just to be clear, i am noob with fluent
enass massoud likes this.
chikko04 is offline   Reply With Quote

Old   April 28, 2016, 12:29
Default
  #14
Member
 
enass
Join Date: Feb 2015
Location: Alexandria-Egypt
Posts: 30
Rep Power: 3
enass massoud is on a distinguished road
Quote:
Originally Posted by chikko04 View Post
i am following your same tutorials, so you are at first exercise, it's right?
To solve your issue you have to translate your geometry completly within the first quadrant, far away the axys (X and Y).
I don't know why, but that is... The solution converge in base of the viscosity you have choosed, in about 100, 130 steps, but have something strange.
If you put your geometry in horizontal, everything is fine.

If you want to talk about it, or in general of those tutorials or exerces and results, you can contact me.
Just to be clear, i am noob with fluent
Yes i want to chat about that please send me your details so we can discuss it.
enass massoud is offline   Reply With Quote

Old   April 28, 2016, 18:26
Default
  #15
Senior Member
 
Lucky Tran
Join Date: Apr 2011
Location: Orlando, FL USA
Posts: 1,372
Rep Power: 20
LuckyTran will become famous soon enough
Quote:
Originally Posted by enass massoud View Post
Thanks Lucky in advance. Let me know one more thing please, regarding the Boundary conditions for the velocity inlet it should be v=1m/s in the radial direction. Now the pipe in geometry and meshing is horizontal? shall i define the velocity axial or radial?
Why would you not define the radial/axial coordinates to be the pipe radial/axial direction? Of course, there's are some common-sense convention.

The coordinate system can be whatever. Axial/radial is relative to the coordinate system. Just keep track of everything. Hopefully you know which way the flow actually goes. The coordinate system can be (x,y,z), (a,b,c), or (i,j,k,), nobody cares. The flow goes the way it needs to go. But make sure you keep track of it. ALWAYS.
LuckyTran is offline   Reply With Quote

Old   April 29, 2016, 02:18
Default
  #16
New Member
 
VT
Join Date: Jan 2016
Posts: 16
Rep Power: 2
chikko04 is on a distinguished road
Quote:
Originally Posted by LuckyTran View Post
Why would you not define the radial/axial coordinates to be the pipe radial/axial direction? Of course, there's are some common-sense convention.

The coordinate system can be whatever. Axial/radial is relative to the coordinate system. Just keep track of everything. Hopefully you know which way the flow actually goes. The coordinate system can be (x,y,z), (a,b,c), or (i,j,k,), nobody cares. The flow goes the way it needs to go. But make sure you keep track of it. ALWAYS.
I didn't select cylinder coordinates essentially for 2 reasons:
1 - the exescise talk about a 2d pipe, so, it can drawn like a rectangle where one of long edges is the axis;

2 - i am a noob and i don't know how to change to cylinder coordinates.
chikko04 is offline   Reply With Quote

Old   April 29, 2016, 02:44
Default
  #17
Member
 
enass
Join Date: Feb 2015
Location: Alexandria-Egypt
Posts: 30
Rep Power: 3
enass massoud is on a distinguished road
Thanks Lucky for your reply. I have managed to do the vertical case by making geometry with centerline lying on x axis and enabling gravity in -x direction. But unfortunately all results are exactly same to horizontal case??? As far as I know gravity does play role in vertical flow.
enass massoud is offline   Reply With Quote

Old   April 29, 2016, 03:00
Default
  #18
New Member
 
VT
Join Date: Jan 2016
Posts: 16
Rep Power: 2
chikko04 is on a distinguished road
Quote:
Originally Posted by enass massoud View Post
Thanks Lucky for your reply. I have managed to do the vertical case by making geometry with centerline lying on x axis and enabling gravity in -x direction. But unfortunately all results are exactly same to horizontal case??? As far as I know gravity does play role in vertical flow.
Have you tried to translate the geometry completely in the first quadrant? Axis included... The flow behavior is different from the horizontal case.
To say one, at first time velocity increase, then, in a few step go to max value and start to decrease till to converge.
I think that something in this case, considered by fluent and not by me/us. I don't think that is the gravity, because i didn't fix it.
chikko04 is offline   Reply With Quote

Old   April 29, 2016, 03:39
Default
  #19
Member
 
enass
Join Date: Feb 2015
Location: Alexandria-Egypt
Posts: 30
Rep Power: 3
enass massoud is on a distinguished road
Hi chikko, thanks for your reply. Well for the vertical case the geometry is identical to horizontal case but i just made the gravity in -x direction thats all?? Is that correct? According to fully developed flow, i am sure that the flow should differ. Don't you think so?
enass massoud is offline   Reply With Quote

Old   April 29, 2016, 04:11
Default
  #20
New Member
 
VT
Join Date: Jan 2016
Posts: 16
Rep Power: 2
chikko04 is on a distinguished road
this is my noob thoughts:

at first time i thought. why the physics in a pipe placed in horizontal could be different from a vertical pipe? At the end, we set forces and BC, and no where is written to add gravity in the paper.
But pratically it doesn't. Solved amg issue, moved the geometry entirely in the first quadrant i have seen a different flow behavior. At inlet the flow velocity increase faster to 1 to decrease in a second time, and at the outlet the fully developed region is less than the other case, and yes i think it's right, if i blow upward the velocity decrease with distance. The question now is: who have fixed an opposite radial force?

To be sure to have the right Reynolds number, i plotted the skin friction coefficient, and for Re = 100, for a pipe, it should be 0.16, and it was.
But if i go to validate results, i have errors.

So my second thought was that i am in wrong, i had a bad run and the the exercise has been written to understand differences between radial and axial velocity, possible fluent errors and how it works with axial problems (the axis must be in the first quadrant), but the first reasoning seem to be ok.

I didn't do a run with horizontal pipe and gravity, but i think that your ideas that's right too.

Now if i am in reason and you too, we must have the same results at same conditions (velocity inlet, viscosity etc), otherwise there is something wrong in mine or your thinking and i hope someone can explain where we are in wrong.

update:
thinking better about what have you written, i think that my first thought was right. Gravity act over all domain, and if it is constant nothing should be change in the solution if you consider the developed flow shape ( not values).
chikko04 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
divergence detected in AMG solver: momentum sudhakar FLUENT 37 August 29, 2016 20:00
Divergence detected in AMG solver Mina_Mbg FLUENT 2 December 21, 2015 19:11
divergence detected in amg solver temperature increasing relaxation sweeps Kashif Rashid FLUENT 3 May 3, 2014 06:23
Divergence detected in AMG solver : epsilion maheshbabu FLUENT 0 April 4, 2014 05:54
reverse flow and divergence in AMG solver MASOUD FLUENT 2 April 2, 2010 08:48


All times are GMT -4. The time now is 11:14.