[SOLVED] FEM output does not seem to be correct (beginner question)

About the development of the FEM module/workbench.

Moderator: bernd

Post Reply
User avatar
ceremcem
Posts: 226
Joined: Sun Jan 07, 2018 11:10 am

[SOLVED] FEM output does not seem to be correct (beginner question)

Post by ceremcem »

I started learning FEM by this video. However, my model doesn't seem to be analyzed correctly:

fem-out.png
fem-out.png (86.31 KiB) Viewed 1678 times

I would expect the upper "bar" has some stress (some red areas), just like in "Abs displacement" mode:

fem-out2.png
fem-out2.png (84.7 KiB) Viewed 1678 times

However, in "Abs displacement" mode I can see no stress on the "lower bar".

I don't know what "Abs displacement" or "Tresca" means (I'm following the French Cafe technique) but I can't obtain a result that is similar to my expectation (which is something like combination of the "Abs displacement" and "Tresca" modes).

Is it my lack of knowledge or are there some missing/wrong steps here?
Attachments
fem2-test.FCStd
(588.57 KiB) Downloaded 28 times
Last edited by ceremcem on Fri Aug 02, 2019 11:42 am, edited 2 times in total.
User avatar
PrzemoF
Veteran
Posts: 3520
Joined: Fri Jul 25, 2014 4:52 pm
Contact:

Re: FEM output does not seem to be correct (beginner question)

Post by PrzemoF »

Looks good to me. Tested on my setup - visually the same results.

My tutorial - FEM in 60 seconds - https://peertube.social/videos/watch/d7 ... 91a69a1446
User avatar
bernd
Veteran
Posts: 12849
Joined: Sun Sep 08, 2013 8:07 pm
Location: Zürich, Switzerland
Contact:

Re: FEM output does not seem to be correct (beginner question)

Post by bernd »

looks got to me too ...
screen.png
screen.png (429.3 KiB) Viewed 1660 times
ickby
Veteran
Posts: 3116
Joined: Wed Oct 05, 2011 7:36 am

Re: FEM output does not seem to be correct (beginner question)

Post by ickby »

what you see is correct. The options show what data of the result is used to color the part. "Abs displacement" means it uses the displacement vector, the absolute value of it. Hence sqrt(dx^2+dy^2+dz^2) where dx and dy is the displacement in x and y direction. So you see no color at the bottom truss as it is closer to the contraint and hence moves way less than the top truss.

So your pictire has nothing to do with the stress in the part. To see the stress you should use "Von Mises stress".
User avatar
PrzemoF
Veteran
Posts: 3520
Joined: Fri Jul 25, 2014 4:52 pm
Contact:

Re: FEM output does not seem to be correct (beginner question)

Post by PrzemoF »

ceremcem wrote: Tue Jul 30, 2019 7:06 pm[..]I would expect the upper "bar" has some stress (some red areas), just like in "Abs displacement" mode[..]
There is a strip of yellow at the top and bottom of the beam if you take a close look. Red is the top of the scale - try to play with a simple beam and different support and different forces - you'll get a feeling quickly what to expect and where in terms of stress and displacement.
User avatar
ceremcem
Posts: 226
Joined: Sun Jan 07, 2018 11:10 am

Re: FEM output does not seem to be correct (beginner question)

Post by ceremcem »

If you all say so, then it's probably my misinterpretation.

ickby wrote: Thu Aug 01, 2019 2:18 pm
So you see no color at the bottom truss as it is closer to the contraint and hence moves way less than the top truss.
So your pictire has nothing to do with the stress in the part. To see the stress you should use "Von Mises stress".

Well, now I started to get the idea. Moreover I couldn't notice any difference between "Von Mises Stress" and "Max Shear Stress" by the naked eyes. However, taking two screenshots and subtracting one from the other pixel by pixel outputs the slight difference:

Screenshot_2019-08-02_14-18-19.png
Screenshot_2019-08-02_14-18-19.png (255.37 KiB) Viewed 1574 times



I'm going to stick with "Von Mises Stress" till I understand what "Tresca" means.

PrzemoF wrote: Thu Aug 01, 2019 2:50 pm There is a strip of yellow at the top and bottom of the beam if you take a close look. Red is the top of the scale -
Is there a way to customize the scale where I would like to highlight any area whose stress (or displacement) exceeds a specific limit? (Or better, I'd like to make whole model highly transparent and make the high stress areas more opaque which would provide an X-ray like vision at any angle)
try to play with a simple beam and different support and different forces - you'll get a feeling quickly what to expect and where in terms of stress and displacement.
I got what you suggested. Here is a screenshot that looks like just I would expected:
Screenshot_2019-08-02_14-31-19.png
Screenshot_2019-08-02_14-31-19.png (126.09 KiB) Viewed 1574 times
fandaL
Posts: 440
Joined: Thu Jul 24, 2014 8:29 am

Re: [SOLVED] FEM output does not seem to be correct (beginner question)

Post by fandaL »

ceremcem wrote: Fri Aug 02, 2019 11:41 am ... Moreover I couldn't notice any difference between "Von Mises Stress" and "Max Shear Stress" by the naked eyes...
"Von Mises Stress" and "Max Shear Stress" are two approaches to answer same question: how to get equivalent stress (one simple number instead of normal and shear stresses in all directions). So they give similar value.
ceremcem wrote: Fri Aug 02, 2019 11:41 am Is there a way to customize the scale where I would like to highlight any area whose stress (or displacement) exceeds a specific limit? (Or better, I'd like to make whole model highly transparent and make the high stress areas more opaque which would provide an X-ray like vision at any angle)
Try to play with Pipeline postprocessing, e.g. with “Scalar Clip Filter”. Some hints according to this are on the end of this tutorial https://www.freecadweb.org/wiki/FEM_She ... site_Block
User avatar
PrzemoF
Veteran
Posts: 3520
Joined: Fri Jul 25, 2014 4:52 pm
Contact:

Re: FEM output does not seem to be correct (beginner question)

Post by PrzemoF »

ceremcem wrote: Fri Aug 02, 2019 11:41 am[..]Well, now I started to get the idea. Moreover I couldn't notice any difference between "Von Mises Stress" and "Max Shear Stress" by the naked eyes.[..]
Image from wikipedia

https://en.wikipedia.org/wiki/Von_Mises_yield_criterion
Attachments
Tresca_stress_2D.png
Tresca_stress_2D.png (23.6 KiB) Viewed 1490 times
User avatar
ceremcem
Posts: 226
Joined: Sun Jan 07, 2018 11:10 am

Re: [SOLVED] FEM output does not seem to be correct (beginner question)

Post by ceremcem »

I appreciate all of you for your helps and tips. I could barely understand the difference between Von Mises Stress and Tresca. I suppose choosing either one won't make much difference for me at this stage.
fandaL wrote: Sat Aug 03, 2019 11:55 am Try to play with Pipeline postprocessing, e.g. with “Scalar Clip Filter”. Some hints according to this are on the end of this tutorial https://www.freecadweb.org/wiki/FEM_She ... site_Block
I finally managed to follow the tutorial. Here is my result. Thanks!
Post Reply