|
[Sponsors] |
November 25, 2010, 10:11 |
Tutorial/Guide: 3D flow around object
|
#1 |
New Member
Join Date: Nov 2010
Posts: 2
Rep Power: 0 |
Hi all,
I'm new to CFD Online, but I've been reading a lot from this forum. I am a final year engineering student, and I am working on an open-ended project. For my project, I chose to model a wing from the A350 in Catia (done), and I now want to perform CFD on it. We have Ansys Workbench 12.1 installed at school. After an entire day of trying to figure out how to do everything, I finally got it running and got results; however, I think the results are not accurate (bad pressure contours). I found many guides for 2D analyses, but nothing for 3D flow around an object (a wing in my case). Any help, would be greatly appreciated, even if you're just pointing me to a tutorial. FYI, here is more or less what I did: 1) Imported the catia geometry into ANSYS 2) Edited the geometry to add an enclosure around the wing 3) Used Ansys Mesh to mesh: Note that it meshed the enclosure AND the wing. Also, I'm really not 100% sure my mesh settings are proper for this setup. 4) SetUp: inlet velocity at the front of the enclosure set to 60m/s, with 2deg AOA. Outlet pressure at the back of the enclosure set to 0 gauge. I think my method is correct, but I think the results are flawed perhaps because I don't have the right settings in the mesh/setup/model. Thanks for your help. |
|
November 25, 2010, 10:58 |
|
#2 |
Member
Burak
Join Date: May 2009
Posts: 90
Rep Power: 16 |
I use gambit for meshing, so I can not help you using ansys mesher.
but some quick tips for you: 1)easy flow domain -Assuming you use part design, after making your 3d model in catia, insert another body(and make it part body, by right clicking and using submenus) and make your 3d domain there(sipmly padding a rectangular sketch). -then using boolean operations, subtract your 3d model body from your 3d domain body. -here you get your flow volume, no need to edit anymore. 2)3d settings are not very different than 2d settings. take it easy. 3)use smaller mesh size on the wing, and with approx. 1.1 to 1.3 growth ratio, make your mesh bigger through the outer surfaces. use prismatic mesh on the wing to get better results, especially viscous effects. search for "y+" or "turbulence" or "boundary layer" in the forum for details. 4)I use outflow for the back surface. 5)Look in the forum for laminar-turbulent transitions threads for fluent transition models, if you are specially interested in laminar flow and drag. thats all for now, I will post more as I remember some tips. good luck in your project ! |
|
November 25, 2010, 12:02 |
|
#3 |
New Member
Join Date: Nov 2010
Posts: 2
Rep Power: 0 |
Thanks a lot Burak for your quick tips!
I'm at work right now and can't wait to get home to try this: Responses to your comments: 1) Great idea. I will do just that. This method is also well explained in this tutorial, which I found after you guided me towards this approach. It also covers meshing for a similar problem. http://www.kxcad.net/ansys/ANSYS/wor...m_bb_mesh.html Ok for the outflow; but outpressure should still be ok, right? I'm not so much interested by the boundary layer itself. What I'm trying to demonstrate is that the wing is a working model (and it is: it is a NACA 5 digit series). I'd like to show the separation of flow at the leading edge of the wing which accelerates above the wing thus creating the pressure drop on top. I can show this either by a pressure plot on the wing surface, velocity steamlines (I think), or lift. But until I have good results, I can't analyze it The last area which may cause problem is in the setup phase. I'm not sure which model to run, and how to set it. My preliminary trials only had Viscous:Laminar model on, everything else off. What models should I run? Should I have Energy or Multiphase on? Once I know that, I'm sure I can find info on the models needed. Thanks again for your guidance. |
|
November 25, 2010, 14:59 |
|
#4 |
Member
Burak
Join Date: May 2009
Posts: 90
Rep Power: 16 |
as far as I know, for a fully turbulent case, the following setup works well:
*k-omega (SST) *Simple *I do not have well knowledge about cell / node based solutions, default is cell based *all other settings to second order *monitor drag and lift coefficients to ensure convergence (this generally needs residuals of e-07(default setting is e-03)). when drag and lift converges, than you get a solution. hope this helps. note: I always used outflow, but I think outlet pressure is ok. |
|
March 16, 2012, 12:23 |
|
#5 | |
Member
sheth roh
Join Date: Jul 2011
Posts: 56
Rep Power: 14 |
Quote:
Additionally Here is something that may help you, http://www.youtube.com/watch?v=4wRCkh_qW3c Its good tutorial for beginning. |
||
Thread Tools | Search this Thread |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
Pipe flow with obstacle - HELP | Min | FLUENT | 6 | January 31, 2017 15:28 |
Compressible flow, no data at the outlet | mireis | FLUENT | 6 | September 3, 2015 03:10 |
flow solution of falling object in fluid with remeshing | bionic | CFX | 2 | September 20, 2010 02:31 |
transform navier-stokes eq. to euler-eq. | pxyz | Main CFD Forum | 37 | July 7, 2006 09:42 |
Instability and Vortex/roll structure | Z. Zeng | Main CFD Forum | 3 | May 10, 2000 06:18 |