# Determining Ideal Mesh Size

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

 October 7, 2018, 03:50 Determining Ideal Mesh Size #1 New Member   Jacob Join Date: Sep 2018 Posts: 7 Rep Power: 7 Dear All, I'm pretty new to CFD in general and am using Ansys Fluent. I have sketched out what i need in design modeller. However, I'm quite concern on the mesh size. Is there a relation between velocity of the air and mesh size? I can't decide on what mesh size and the amount of elements to use. Thank you guys.

 October 7, 2018, 22:05 #3 New Member   Jacob Join Date: Sep 2018 Posts: 7 Rep Power: 7 Thank you for your informative insights. I will take your advice and look more into it.

 October 8, 2018, 10:45 #4 Senior Member     André Join Date: Mar 2016 Posts: 133 Rep Power: 10 If you are running a transient simulation, you can look calculate the Courant number: Co = v * dt / dx, where v is your maximum velocity in the domain, dt is your time step (defined in solver), and dx is your typical mesh size. This number should be close to 1 to ensure accuracy and numerical stability. If you are running a steady-state simulation, typically the smaller the better, but you need to have a reasonable number of total elements depending on what hardware capabilities you have. There are also some implications if you use turbulence models, which dictate the recommended cell size near walls based on the fluid velocity and properties. __________________ Sapere aude!

 October 23, 2019, 18:22 #5 New Member   Hamed Join Date: Mar 2017 Location: Iran Posts: 29 Rep Power: 9 Thanks for the information but would you let me know how I should decide about the smallest single mesh size in a simulation respective to the dimensions of body?or domain. Do we need to try on various range based upon former knowledge either experience or if there is identical equation or solution for calculating smallest cell size in the steady state simulation?

 October 23, 2019, 18:33 #6 Senior Member   Join Date: Dec 2016 Posts: 152 Rep Power: 10 I'm guessing by "single mesh size" you mean cell (aka, element), in which you have to follow exactly what I just stated above. Now, if you're running a simulation similar to what others have published, look at what they've done for a mesh sensitivity analysis and try to emulate it. There is no such thing as "best mesh", there's only the most "reliable mesh" which you determine from running a sensitivity analysis.

 October 23, 2019, 20:08 #7 New Member   Hamed Join Date: Mar 2017 Location: Iran Posts: 29 Rep Power: 9 Really appreciate that. Regards

 September 15, 2022, 00:34 reply #8 New Member   gedu Join Date: Aug 2022 Posts: 1 Rep Power: 0 in this of transient analysis. There is only one equation and two unknowns( timestep and the mesh size). Co = v * dt / dx so what to determine whether mesh size or the time step?

April 14, 2023, 07:37
#9
Member

Sakun
Join Date: Nov 2019
Location: United Kingdom
Posts: 93
Rep Power: 6
Quote:
 Originally Posted by Pema in this of transient analysis. There is only one equation and two unknowns( timestep and the mesh size). Co = v * dt / dx so what to determine whether mesh size or the time step?
Hi,

i have the same issue, did you manage to solve the issue ?

Regards,

 April 14, 2023, 11:39 #10 Senior Member   Join Date: Jun 2009 Location: Technische Universität Chemnitz Posts: 107 Rep Power: 16 Fluent is an implicit solver, so the CFL is not the speed limit that it is for an explicit method. Depending on the model, it is not uncommon to have CFL>100 for certain parts of the flow field. Sakun likes this.

 Tags mesh and grid, mesh element, velocity