|
[Sponsors] |
Flow Pattern Difference B/W 2D and 3D simulation - orifice |
|
LinkBack | Thread Tools | Search this Thread | Display Modes |
June 21, 2020, 08:05 |
Flow Pattern Difference B/W 2D and 3D simulation - orifice
|
#1 |
Member
kumar
Join Date: Dec 2019
Posts: 33
Rep Power: 6 |
Hello every one
I am trying to simulate turbulent flow in the orifice, I tried to solve it in 2D but the flow separates from the straight axis after the expansion, if I simulate the same in 3D the flow is straight, is there any particular reason? that flow works differently in 2D and 3D simulation. Thank You |
|
June 22, 2020, 15:24 |
|
#2 |
Senior Member
Erik
Join Date: Feb 2011
Location: Earth (Land portion)
Posts: 1,167
Rep Power: 23 |
What you are seeing is the "Coanda effect".
This can occur in both 2D and 3D flow. You have two different geometries here though? One is planar (2D), and the 3D one is a circular orifice? |
|
June 23, 2020, 16:38 |
|
#3 |
Member
kumar
Join Date: Dec 2019
Posts: 33
Rep Power: 6 |
Hello yes two geometries, with some boundary conditions
|
|
June 23, 2020, 16:45 |
|
#4 |
Senior Member
Filippo Maria Denaro
Join Date: Jul 2010
Posts: 6,769
Rep Power: 71 |
The physics of the two cases is totally different
|
|
June 23, 2020, 16:54 |
|
#5 |
Member
kumar
Join Date: Dec 2019
Posts: 33
Rep Power: 6 |
Thank You , I have solved it using Axis symmetric in the case of 2D ,I would like to ask you how to give length scale to the recirculating zones i.e streamlines of the turbulent flow in ansys .....
|
|
June 24, 2020, 13:33 |
|
#6 |
Senior Member
Erik
Join Date: Feb 2011
Location: Earth (Land portion)
Posts: 1,167
Rep Power: 23 |
That does not look like axis-symmetric for the 2D case. I can see differences on the top and bottom of your 2D figure. Looks like you did planar flow?
If you did do axisymmetric, then of course it won't show the Coanda effect, as you are constraining the flow to be axisymmetric. |
|
June 25, 2020, 07:08 |
|
#7 |
Senior Member
Gerry Kan
Join Date: May 2016
Posts: 347
Rep Power: 10 |
Dear Gkchpa:
When you use axisymmetry in Fluent, you need to set your y=0 axis at the central axis of your domain, and as such you only need to simulate half of the orifice (top or bottom, you pick), instead of the whole cross section like you have displayed here. Another unrelated note with regard to Fluent and axisymmetry: I don't know if this problem would appear in your solution, the results might look a bit funky and will be different from your 3D results. If that happens you need to adjust the k-epsilon model constants to recover the correct behavior. If this issue does not occur in your CFD solutions, then I think they have finally corrected it. Hope that helps, Gerry. |
|
|
|