|
[Sponsors] |
October 18, 2006, 09:51 |
aspect ratio
|
#1 |
Guest
Posts: n/a
|
How can I control the aspect ratio of my mesh in CFX-mesher?I need of an inflated boundary to obtain y+ values for k-epsilon turbolence model: how high aspect ratio must be to optimize my analisys? Thanks!
|
|
October 18, 2006, 18:01 |
Re: aspect ratio
|
#2 |
Guest
Posts: n/a
|
Hi,
If you are referring to the inflation layers at the boundaries: You can control them with the properties tab. Have a look at the documentation. Optimum aspect ratio: There is no general optimum, it problem dependant. You should do a sensitivity analysis and determine what the best aspect ratio to use for your simulation. Glenn Horrocks |
|
October 19, 2006, 03:22 |
Re: aspect ratio
|
#3 |
Guest
Posts: n/a
|
Thank you Glenn. I tried several discretization by forcing either y+ value and first element high.The problem is that I obtained very small elements that are not suitable for k epsilon turbulence model.Calculation fails with a "fatal overflow error".Now I'm trying with k omega and double precision: calculation goes, but it is very very slow! Do you think this problems could be related? Do you haev some hint about this? Thanks again!
|
|
October 19, 2006, 17:54 |
Re: aspect ratio
|
#4 |
Guest
Posts: n/a
|
Hi,
Your error does not mean your elements are too small for the k-epsilon turbulence model. It means the solution has diverged. You would have to describe what you are modelling better if you want us to help you. Glenn Horrocks |
|
October 20, 2006, 01:14 |
Re: aspect ratio
|
#5 |
Guest
Posts: n/a
|
Carrying out a mesh dependency test will give you lots of insight about the discretisation error. A balancing approach is always needed. Is it worth the effort to gain the 1% accuracy if it takes a month to produce a solution? Errors such as assumptions made in turbulence model could easily overwhelm the discretisation error, as the mesh is refining.
|
|
October 20, 2006, 03:13 |
Re: aspect ratio
|
#6 |
Guest
Posts: n/a
|
Ok I'll describe my problem. I have to calculate a multiphase flow field. My domain is entirely closed and contains 60% in volume fraction of liquid and 40% of compressible fluid (Air Ideal Gas). Domain oscillates around a stable position so calculation is unsteady calculation with a time step of about 10E-5. How can I set the problem?How can I fix the pressure level for the compressible fluid since the domain is closed? Last calculation failed and gave a fatal bounds error (Absolute Pressure). Have you some hint?
Thank you! |
|
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
pressure eq. "converges" after few time steps | maddalena | OpenFOAM Running, Solving & CFD | 69 | July 21, 2011 08:42 |
[ICEM] Mesh along axis for a wedge gives elements with extreme aspect ratio | sangrampp | ANSYS Meshing & Geometry | 0 | November 17, 2010 07:57 |
aspect ratio vs. y+ | hammam | CFX | 3 | August 6, 2007 11:41 |
High Aspect Ratio elements | Flavio | CFX | 2 | November 24, 2006 13:01 |
Aspect Ratio at high turbulent and swirled flows | zago | FLUENT | 0 | May 19, 2004 03:37 |