• jacky
  • New member
  • Offline
  • Registered: 2010-03-14
  • Posts: 5

Topic: Generation of shell element mesh

Dear Sir,

I came across some problems related to the mesh generation when modeling a shear wall.

The information of the ATENA 3D:
Program version: 1.2.4  Product version: 3.3.0
Licence: full,   User number: 254

The modelling was operated as the following steps:
1. Definte materials (Concrete and reinforcement)
2. Geometry definition: 'Macroelement' > add > simple objecet > 'New object, Vx=0.6m, Vy=3m, Vz=4.5' > 'Generate' > Select 'macroelement type: shell/plate' > 'Azimuth: 0, Zenith: -90' > 'calculate: surface 2, positive' > Generate > 'Properties' > defenition of 'shell/plate layers' and 'shell/plate reinf' > select the bottom right 'add' > 'Quit' to the main operating window > FE mesh, macroelement > add, 'mesh type: brick' and select the macroelement 1> Generation > Global element size '0.1', Element: quadratic > push 'Generate'  > error exists 'Catastrophic failure'

I have tried many times to model the different dimensions of  the shear wall (shell element), but the errors exist consistently. Please help me to figure out this issue.

Thank you very much!

Best wishes,
Jacky

Quote Post 1

  • dpryl
  • CC software developer
  • Offline
  • Registered: 2007-07-13
  • Posts: 1,107

Re: Generation of shell element mesh

I recommend to upgrade to the latest ATENA version, 4.2.7. I am afraid we can not support 3.x.x versions any more.

Please understand that for shell elements, there should only be a single layer of elements in the thickness direction (defined as local Z) - the internal shell layers are there instead.

dpryl's Website

Quote Post 2

3 Reply by daescuc (edited by daescuc 2012-09-13 10:46:29)

  • Registered: 2011-01-04
  • Posts: 16

Re: Generation of shell element mesh

Hello,

I am trying to model a consolidation system, using the shell elements as you suggested.

Unfortunatelly, when I try to run the analysis, I get the "Illegal element type specified for the shell/plate macroelement x (quadratic brick)".
This error is given for  all the shell elements I used.

I tried to see in your model Slab_shell.cc3 how is defined the mesh, and when I tried to run it, I got the same error, unfortunatelly. I am using only plate elements (rectangular) with a reduced thickness. The perfect conection is assumed at the contact surfaces.

Thank you,

edit1> Any suggestions?

Quote Post 3

  • dpryl
  • CC software developer
  • Offline
  • Registered: 2007-07-13
  • Posts: 1,107

Re: Generation of shell element mesh

Hello, please set the shell macroelements to be meshed using quadratic brick elements: 
1. Make sure the macroelement geometry has six sides with four edges each (replace complicated MEs with several simple ones)
2. In Mesh - Generation, set the Elements to quadratic
3. In mesh - Macroelements, set the Mesh type to Brick

dpryl's Website

Quote Post 4

  • Registered: 2011-01-04
  • Posts: 16

Re: Generation of shell element mesh

Hello again,

Indeed, I was able to go to the solving module using your recommendations. Thank you for that.

In order to introduce you to my model, I will make a short description:
- on a reinforced concrete element I am applying two layers of composite materials (carbon fibers). These are two consecutive layers (on top of the other), but are arranged at 45 degrees angles one with respect to the other, on the same side of the element.

The next problem is with the contacts, but I am not sure why. I will write down the error/warning I got (for a large number of points) in order for you to have an idea. The thing is that, at some point, the program runs out of memory and stops, without making a single iteration. (The computer I work on is a dell with 200Gb HDD free, 12Gb ram, 2xIntelXeon 3.2GHz quad, using windows Vista professional x64 - I am specifying all these in order to see if you suggest anything with respect to the configuration - hardware or software).

Error/Warning:
Invalid master-slave BC, selections for master(s) $N$M402653202S603979902, for slave(s) $N$M402653200S603979895 : decorated dofs 7 of slave node 75175  could not be constrained,  specified in support id: 1, load case id 738197527, name: VOLUME 402653200 SURFACE 603979895 to VOLUME 402653202 SURFACE 603979902; BC ignored.

I am looking forward to have you opinion about it!

Regards,
Alex

Quote Post 5

  • dpryl
  • CC software developer
  • Offline
  • Registered: 2007-07-13
  • Posts: 1,107

Re: Generation of shell element mesh

Dear Alex,
with respect to the ""out of memory"" warning and "out of memory", please read the corresponding explanations in the ATENA Troubleshooting Manual.

To model the 2 fibre layers, I would recommend a single shell with 2 reinforcement layers, with reinf. directions in 45 degree angle between them.

Regards,

dpryl's Website

Quote Post 6