Wrong (?) results

Post processing utility for Elmer
Post Reply
fagara
Posts: 1
Joined: 28 Oct 2010, 18:39
Location: Italy

Wrong (?) results

Post by fagara » 28 Oct 2010, 18:45

Hi,
how can I associate cartersian coordinate 1 2 3
to that of the imported model 3d ?
and in the Post processing displacments are in mm ?
it seems a wrong result.....
help please.

Juha
Site Admin
Posts: 357
Joined: 21 Aug 2009, 15:11

Re: Wrong (?) results

Post by Juha » 28 Oct 2010, 18:58

HI,

the first question i don't really understand.

As for the second, ElmerPost makes no unit conversions but displays the input file values asis.
ElmerSolver does not know much about units either, but expects a consistent set of units to be
used in the mesh, material & other input parameters.

Regards, Juha

Captain Slow
Posts: 72
Joined: 10 Nov 2009, 19:17

Re: Wrong (?) results

Post by Captain Slow » 29 Oct 2010, 19:33

As to the first question, I think the answer that you're looking for is that co-ordinates 1,2 and 3 relate to x, y, and z.

richka
Posts: 3
Joined: 05 Jun 2011, 12:46
Antispam: Yes

Verification example

Post by richka » 05 Jun 2011, 14:45

Hello all,
I tried to calculate displacement on the end of beam with geometry 0,1 x 0,1 x 1 m. I defined the boundary displacement on the first end (wall) 0 0 0. On the second end is force in y direction -1000 N.

By theoretical equation is displacement y = ( F * L^3 ) / (3 * E * J ).
As a material I choose carbon steel with E = 2.00e+11 Pa, J is for rectangular profile b * h^3. In my example b = h and J is a^4. So J = 1e-04 m^4. After calculation y = 0.0002 meter.
I created model of beam in Salome (Beam of size 100 x 100 x 1000, this was reason for using menu Operation - Transformation - Scale Transform - Scale Factor 0,001 to get proper size of beam model). Mesh consists of quadrangle elements for obtain good results. Then I exported it via unv file into Elmer ( change . to , in unv file was necessary, I 'm using Ubuntu 10.04).
For check of proper size of beam I used in Elmers menu Model - Summary, where I find
BOUNDING BOX
X-coordinate: [ 0 , 0.1 ]
Y-coordinate: [ 0 , 0.1 ]
Z-coordinate: [ 0 , 1 ]
It seems to be correct in SI units. After choosing Steel Carbon as a material of beam, which has Youngs modul 200e9, poisson ratio 0.285 I defined force and displacement.
After calculation Elmer gave that maximum displacement in y direction is 1,69e-06.
I am confused. Everything what I defined to Elmer was in meters, newtons - SI units. And differences between theoretical equation and Elmer is in order of 100.
What I am doing in the wrong way?
I giving my beam as a Elmer project into attachment.
Have anyone already compared results from linear static with empiric equations?
Have a nice day.

Karel
Attachments
beam.tar.gz
beam
(263.97 KiB) Downloaded 188 times

raback
Site Admin
Posts: 3498
Joined: 22 Aug 2009, 11:57
Antispam: Yes
Location: Espoo, Finland
Contact:

Re: Wrong (?) results

Post by raback » 07 Jun 2011, 12:35

Hi Karel

The SI unit of "Force 2" should be [F]/[A]=Pa, not N. The area of the beam end is 0.01 which should give you the factor ~100 compared to teh theoretical results.

Unfortunately the keywords do not always have the most natural interpretations. Pressure as a keyword was already reserved by the Navier-Stokes equation so Force was chosen for the Navier's equation.

-Peter

richka
Posts: 3
Joined: 05 Jun 2011, 12:46
Antispam: Yes

Re: Wrong (?) results

Post by richka » 11 Jun 2011, 02:00

Than you very much Peter. It is very helpfull. Now I understand to definiction. Now it is clear. It will be more usefull to define NODAL FORCES. Menu for NODAL FORCES is not defined in standart Elmer graphical menu, isn't it? It must be defined via text script, as I seen in forum. It is big disadvantage that Elmer have in LINEAR STATIC no menu for NODAL FORCE and have no possibility to define NODAL FORCES to choosen NODE in graphical menu. Thank you one more time.
Karel

raback
Site Admin
Posts: 3498
Joined: 22 Aug 2009, 11:57
Antispam: Yes
Location: Espoo, Finland
Contact:

Re: Wrong (?) results

Post by raback » 12 Jun 2011, 01:35

Hi Karel

Elmer menu structure is fully programmable so you can add the nodal load to the menus even by yourself. Just edit the corresponding .xml file in edf directory and add "varname Load" to boundary condition menu definition.

-Peter

richka
Posts: 3
Joined: 05 Jun 2011, 12:46
Antispam: Yes

Re: Wrong (?) results

Post by richka » 14 Jun 2011, 12:32

Great, thank you. I will try it.

Karel

Post Reply