Grasshopper

algorithmic modeling for Rhino

Hello

The result of my Karamba calculation does not match whit the results of my final element program DIANA, i know the results in DIANA are correct. The reaction forces are the same in grasshopper+karamba and DIANA so the geometry is correct. But the principial stresses don`t match at all. See the attached pictures and gh. file

Regards,

Jeroen

 

Views: 1783

Attachments:

Replies to This Discussion

OK.

Without clemens being here I tell you the following:

i also checked it in a pro-FE-program, in my case Dlubal RFEM 4. The support reactions match, the principal stresses in the middle are the same, also the max. displacement approx. is the same - whereas the slight displacement-mismatch and the stress-deviations at the supports here come from the same problem, i think: your model - again.

you strongly should consider the issue of pointed load or support introduction into a shell/plate structure. as you should know, this is a general problem of areal FE modelling, since singularities in the results occur at these points, and blur into the surroundings. supporting a 43m spanning concrete shell on 4 infinitely small points is likely not to work consistently - and you cant blame any program for calculating mistakes different from another program.

regards,

robert

Hello Robert,

 

Thank you for your fast reaction. Yes i already thought the problem will be in the corners. I only just do`nt know how to make the corners whider/bigger, because when the geometry is changing, the points of the mass will also change, and the the supports will change. Do you know how i can easily make whider supports.

 

Regards,

 

Jeroen

You should add small edge beams near the four corners of your shell. In this way you can avoid the stress singularity that results from supporting a surface structure in a point. Otherwise the resulting deflections are mesh dependent and increase without limit as you refine the mesh near the corners.

Best,

Clemens

Hello Robert,

Is it possible that i only show the results of the top circle of the shell structure and not of the corners? So he calculates the whole structure, but i only want to see the results of the top circle. Than the problem is solved for my project.

Regards,

Jeroen

Hello Robert,

 

I think i got the solution. I split the mesh in two, a mesh on the top (id c) and a mesh on the bottem (id b). I want to view the results of the mesh on the top (id c) in the modelview and in the shellview, is this possible?

When i put the c to the id of the modelview he still view the hole shell structure. see the attachement

 

Regards

Jeroen

Attachments:

depending on the version of karamba there is a bug in the id-handling of the model view, this is being fixed at the moment.

generally the best way is to avoid those singularities. e.g. by inserting two spring-beams on each corner of your shell that have very high stiffness - so the load introduction takes place with on 3 or more nodes for each corner. beams can without problems be placed on the edges of a shell-element and distribute the forces more realistically. like this, pointwise load introduction into shells often is handled in practice.

if you just divide your shell to view just a part, the results dont get really correct with that..

best

r

Hello Robert,

Thanks for your reaction.

Is there a version of Karamba where these bug is not in the id-handeling. Otherwise i can use that version.

 

Best

Jeroen

it's easily avoided when just replacing the model-view component by a new one or alternatively playing around with the id's you feed in..

i dont really know which one didnt have the bug. a release version is not currently there.

Is it possible that i can get a lest of how the program is calculating the princ. stress in the shell structure. There is only a list of a beam structure in the manual

 

Best,

 

Jeroen

You can test the principal stress directions calculated by karamba by considering a plate under various, in-plane boundary loads.

Best,

Clemens

Hello Clemens,

Thanks for your reaction, but I don`t understand how i can make the short edge beams near the corners.

Do you have an example or can you show it in my gh. file? See the attachement

 

Best,

Jeroen

Attachments:

Sorry I can not dig into your definition for lack of time. Take the lines at the perimeter of your shell and feed them into a LineToBeam-component. You can either strengthen only the corner regions at the supports or all free shell edges. Doing the latter is probably a good idea for free edges are normally the weak parts of shell structures. Test various beam heights and see how the maximum shell displacement varies.

Best,

Clemens

RSS

About

Translate

Search

Photos

  • Add Photos
  • View All

Videos

  • Add Videos
  • View All

© 2024   Created by Scott Davidson.   Powered by

Badges  |  Report an Issue  |  Terms of Service