Discussion Groups

Simulation Mechanical and Multiphysics

13 Posts
0 Kudos
Registered: ‎04-08-2013
Accepted Solution

To smooth or not to smooth ?

278 Views, 2 Replies
05-28-2013 02:24 AM

Hi everyboy,

I read the thin on line help manual we have and the forum to find some informations about the different process used to smooth constraints fields in the result tab, but was unsuccessfull.

Gaps are so huge between different methods, I wonder if this option provides reliable results to exploit.

*Expert Elite*
595 Posts
127 Kudos
Registered: ‎08-30-2012

Re: To smooth or not to smooth ?

05-28-2013 01:38 PM in reply to: Ju_brl

Hi Ju_brl,


I believe that you are asking about the command "Results Contours > Settings > Smooth Results" and "Smoothing Options" which are described on the pages "Results Options Menu" and "Results Environment" in the Wiki Help. Most of the options were included simply because it was easy to program them (for example, "Maximum", "Minimum", "Sum" type of smoothing functions) or because they were requested by a customer ("Smooth before applying operators"). I use the defaults 99.9% of the time because those are the ones that make sense.


I have used these on rare occasions:

  1. "Smoothing Options > Smooth across part boundaries" can be used when different "parts" in the model are really just different sections of a real part but other wise are the same material, etc. For example, you might want to make the two flanges and web of an I-beam out of different parts for some reason. In reality, the stress is a smooth function from the web into the flange. This would be shown this way in the results environment with this option checked.
  2. "Smoothing Options > Smoothing function > Range" can be used to get a sense of the accuracy of the model/mesh. This will plot the difference between two (or more) results that occur at the same physical location. For example, the stresses along the face of a part should be smooth and continuous, but FEA does not provide that type of result. It has a stress result for every element connected to a node. Ideally, all of the results are close to being the same, so a contour of the Range should be close to 0 psi. If there are large jumps because the mesh is too coarse, then the Range contour will show larger values.

If you have any other questions, please clarify what type of result you are asking about, what you are trying to do with that results, the type of analysis, and the type of element.

John Holtz, PE
Mechanical Engineer
Pittsburgh, PA

Simulation Mechanical user since Dec 1997
13 Posts
0 Kudos
Registered: ‎04-08-2013

Re: To smooth or not to smooth ?

05-31-2013 07:50 AM in reply to: AstroJohnPE

This absolutely enough for.


Thank you very much for your help John.