• Registered: 05-02-2015
  • Posts: 46

Topic: Plate on Subsoil - no results

Hi everyone,
could someone please have a look on the attached input file? (Sorry its so big, I could reproduce the mistake with smaller systems...)
What I'd like to calculate is a simple cctplate laying on subsoil and a beam on one edge. There are several nodal loads on the plate.
In reality: the concrete bottom plate of a building with a concrete wall on one edge (simulated as the beam). The nodal loads are the loads of the building.

All I get are NaN results, and I just cannot find any mistake... I guess there is a typo somewhere, could someone please help me?
Thanks a lot!

Dennis

Post's attachments

oofeminput.txt 434.72 kb, 10 downloads since 2019-01-16 

You don't have the permssions to download the attachments of this post.
  • Registered: 05-02-2015
  • Posts: 46

Re: Plate on Subsoil - no results

Anyone got an idea?

  • Registered: 25-06-2013
  • Posts: 241

Re: Plate on Subsoil - no results

Hi Dennis,

I looked at the file and it seems to me that it is not preperly supported. Unfortunately, the file is very big and therefore it is very difficult to work with it if you can make a smaller example it would be helpful.

Martin

  • Registered: 05-02-2015
  • Posts: 46

Re: Plate on Subsoil - no results

Thank you for ur answer!

Unfortunately, I wasnt able to reproduce this error in a smaller system. I use a self written programm to produce the input file, and usually everything works perfect, except for this example, and I just couldnt find a mistake in the input file.
What I am actually trying to simulate is a plate on subsoil, using cctplate and trialplatesubsoil elements. I'm producing a mesh and fill it with cctplate elements combined with the same mesh with the same nodes for the subsoil elements. So, actually I have two times the same mesh on two layers. There is no other support needed.
What might chause the problem is that I connect the beam3D element to the cctplate elements on one edge as well. Although I already did that a couple of times without any problems, I think this might be the problem here...

Maybe u could check that beam-plate connection again? Although I already did that like a hundred times already.... ;-)

Thanks for ur help!
Dennis

5 Reply by johnnyontheweb (edited by johnnyontheweb 30-01-2019 17:02:34)

  • Registered: 16-10-2010
  • Posts: 407

Re: Plate on Subsoil - no results

Hi Dennis,
I has a look at the file, in my copy of oofem it seems that, in the node records, the instructions:

dofidmask 6 1 2 3 4 5 6 bc 6 1 1 0 0 0 1

made some problems.
Anyway, I can suggest to review this, because the domain is 2dMindlinPlate and you have specified 6 dofs in the dofidmasks.

ps.
I'm not really confident with 2d models - why don't you make it in 3d with the proper BCs? you can use quad1mindlinshell and tria1subsoil.

  • Registered: 05-02-2015
  • Posts: 46

Re: Plate on Subsoil - no results

Hi,
thanks for ur help.
I think that part of the input file should be correct, since I have used that already without any problems... I clamp the dofs I dont need respectively the dofs which dont appear in my 2D model. U are right, Im using the 2Dmindlinplate model with only 3 dofs.
Actually I dont want to switch to 3D, since my models are 2D plates and my whole setup is designed for the 2D problems (dofs u_z, d_u, d_v)

Could there be something wrong with the connection beam-plate?

Thank u!

Dennis

7 Reply by bp

  • bp
  • Administrator
  • Offline
  • From: Prague, Czech Rep.
  • Registered: 02-03-2007
  • Posts: 1,026

Re: Plate on Subsoil - no results

Hi,
the smaller input with a ~10 elements would help a lot.
Anyway, I have spotted following:
- node 1526 is not connected to any element (used only as reference node) and thus it is free, should be supported.
- some loads (1,3,17) define only 3 components, while applied to nodes with 6 DOFs. 
- the input for nodes 1510-1525 is not correct. You set master nodes for DOFs Dw,Rx,Ry but you mastermask says only Dw is linked to master. What the remaining rotational DOFs? I believe, you should use doftype 6 0 0 1 1 1 0 in all these nodes.

Borek

bp's Website

  • Registered: 05-02-2015
  • Posts: 46

Re: Plate on Subsoil - no results

Hi Borek,
thx for ur reply.
I finally figuered it out by myself, but ur hints are helpful as well.
-node 1526: ur right, its only a refnode. I can support this node, altough I think it this didnt cause any problems so far
-the loads 1,3,17 were wrong, this is correct. Of course they need 6dofs to be clearly defined
- nodes 1510-1525: actually I only want to connect the Dw, so I actually did that on purpose. The nodes are supposed to (physicaly) only push in one direction and not to connect the rotations.

What actually causes the NaN results is connected to ur last hint: the nodes 1510-1525 are the nodes of beam3D elements (elements 6029-6043) and since I only connected the Dw and supported Du, Dv and Rw, the beam elements were loose around their local x-axis (torsion). So what I needed to do was to support the 4th dof (Ru) on one of the 1510-1525 nodes to be statically defined. I dont want to support the 5th dof (Rv) since I want to have bending around local y-axis in my beam.

Thank u for ur help!
Dennis