CFD Online Logo CFD Online URL
www.cfd-online.com
[Sponsors]
Home > Forums > Special Topics > Mesh Generation & Pre-Processing

why the finer the mesh the worse the results?

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

Reply
 
LinkBack Thread Tools Search this Thread Display Modes
Old   January 23, 2016, 05:35
Default why the finer the mesh the worse the results?
  #1
Senior Member
 
Join Date: Oct 2014
Posts: 124
Rep Power: 10
Ema40 is on a distinguished road
Hi,

I am simulating a turbine as multiphase flow VOF.
Why using a mesh of 0.01 m the results are very close to the exact solution (error of 1%), and with the mesh of 0.005 m they differer from the exact solution of about 13%? The exact solution has been achieved by experimental tests. This occur for different simulated hydraulic conditions (e.g. flowrates)

Thank you
Regards
Ema40 is offline   Reply With Quote

Old   January 24, 2016, 16:05
Default
  #2
New Member
 
Kamaladin Saqar
Join Date: Nov 2009
Location: Tehran
Posts: 4
Rep Power: 15
time is on a distinguished road
Send a message via Yahoo to time Send a message via Skype™ to time
Hi,
Decreasing size of elements has conditions. One of them is Courat Number. I think mesh of 0.005 m does not satisfy suitable Courant Number of the problem.
time is offline   Reply With Quote

Old   January 25, 2016, 06:49
Default
  #3
Senior Member
 
Join Date: Oct 2014
Posts: 124
Rep Power: 10
Ema40 is on a distinguished road
Thank you.
The Courant number should be close to 1. I have reduced the mesh size, and I used the same time discretization. The courant number with the finer mesh is hence closer to 1 than the coarse mesh, but the coarse mesh gives better result. How is it possible?

Thank you very much
Ema40 is offline   Reply With Quote

Old   January 25, 2016, 11:51
Default
  #4
New Member
 
Kamaladin Saqar
Join Date: Nov 2009
Location: Tehran
Posts: 4
Rep Power: 15
time is on a distinguished road
Send a message via Yahoo to time Send a message via Skype™ to time
Hi
Well that is a good case to state an mesh sensitivity study!
time is offline   Reply With Quote

Old   January 25, 2016, 16:32
Default
  #5
Senior Member
 
Join Date: Oct 2014
Posts: 124
Rep Power: 10
Ema40 is on a distinguished road
Thank you. But the sensitivity analysis works well when the accuracy increases with the reduction in the mesh dimension. I obtain the opposite...

Thank you
Ema40 is offline   Reply With Quote

Old   February 1, 2016, 07:02
Default
  #6
Senior Member
 
Colinda
Join Date: May 2012
Location: Brussels
Posts: 144
Rep Power: 12
colinda1 is on a distinguished road
Possibly there is another element in the simulation set-up that is causing the difference, not the mesh dependency. E.g. the boundary conditions, the turbulence model, etc.

The mesh dependency study is one (important, but not the only) element determining the accuracy of the solution.
colinda1 is offline   Reply With Quote

Old   February 1, 2016, 17:55
Default
  #7
Senior Member
 
Join Date: Oct 2014
Posts: 124
Rep Power: 10
Ema40 is on a distinguished road
Thank you.
I do not know. The case file is always the same, and I import on that case the mesh....Strange. Probably the problem can not be well simulated by VOF for example.

Thanks
Ema40 is offline   Reply With Quote

Old   March 15, 2016, 10:29
Default
  #8
New Member
 
VT
Join Date: Jan 2016
Posts: 16
Rep Power: 8
chikko04 is on a distinguished road
did you find out the answer to your question? I am affected by the same issue. If i use a finer mesh with same setup i have results totally differents. I am investigating to find out what cause it, but i am whitout any reasonable answer!
chikko04 is offline   Reply With Quote

Old   March 16, 2016, 06:17
Default
  #9
Senior Member
 
Join Date: Oct 2014
Posts: 124
Rep Power: 10
Ema40 is on a distinguished road
Hi, no answer still.
When the mesh reduces, it may be correct that the result change. Then the result should converge until an asyntotic value, and further mesh refinement should not affect the solution any more. But in my case the results change a lot, decrease, increase, ect...they are not asysntotic!
What about your problem?
Ema40 is offline   Reply With Quote

Old   March 17, 2016, 06:21
Default
  #10
New Member
 
VT
Join Date: Jan 2016
Posts: 16
Rep Power: 8
chikko04 is on a distinguished road
Quote:
Originally Posted by Ema40 View Post
Hi, no answer still.
When the mesh reduces, it may be correct that the result change. Then the result should converge until an asyntotic value, and further mesh refinement should not affect the solution any more. But in my case the results change a lot, decrease, increase, ect...they are not asysntotic!
What about your problem?
I am studing a turbulent flow with a SEM, if increase Polynomial order (that equivalent to increase mesh) or increase mesh to have a more equispaced mesh, the fluid become laminar, instead to be turbulent, how happen with a coarser mesh. And Re is very high...so it is impossible.
chikko04 is offline   Reply With Quote

Reply

Tags
mesh, mesh dimension, mesh independency, turbine, vof

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
how to set periodic boundary conditions Ganesh FLUENT 15 November 18, 2020 07:09
[snappyHexMesh] SnappyHexMesh for internal Flow vishwa OpenFOAM Meshing & Mesh Conversion 24 June 27, 2016 09:54
Mesh dependency - possible that refinement = worse results? crisgc FLUENT 4 November 14, 2012 13:16
bubbleFoam - behavior on finer mesh darai OpenFOAM Running, Solving & CFD 8 November 22, 2011 06:15
How to control Minximum mesh space? hung FLUENT 7 April 18, 2005 10:38


All times are GMT -4. The time now is 05:58.