Bolt pretension

General Questions
1

Hi all
I am very sorry for asking noob question but i am running out of ideas.
I am testing some simple case to find out what i am doing wrong - just two plates (2D) with solid 2 bolts:
two steps analysis:
first:
bolts conferted to a compound part, boundary layer applied to one half of each bolt, and bolt pretension force applied to same surface, contacts are linear with friction
second step :
the same but with bolt pretension setted up as a displacement fixed

Results when i run only first step results looks ok. But when i am turning on second step, calculation crashes. Like suimulation is diveging, with increasing loads
for only first step:

obraz
obraz1453×536 117 KB

My files are there:
https://www.dropbox.com/scl/fi/wdcl6x0rbtpsknc52ak1q/Documents_035724.zip?rlkey=28wisv498jpj5y3b08ls8gyjy&st=9ni0151s&dl=0

Thank you in advance

2

Try simplifying the model to see what causes non-convergence:

  • disable Nlgeom
  • remove friction (it’s really bad for convergence and should be avoided when not absolutely necessary)
  • replace contact with tie constraints

I would also add some load in the second step.

And it might be better to model the plates as solids (meshed with hexahedral elements) due to CalculiX’s limitations and issues with shell elements (and due to the fact that you want to analyze contact between two plates one on top of the other),

3

Thank you for help
i have not written but :
i have already checked both Nlgeom on and off
both tie constraints, and linear and hard behaviour

but i just have defined plates as 3d geometry and this time it works… But why???
the only difference between steps is bolt pretension changed to fixed so it has nothing to do with plates
so maybye it is because of two steps approach?? - prepomax is using *Boundary, op=New and calculix is extruding 2d mesh to 3d… Maybye it is trying to reextrude mesh which causes interference with constraints (including position of bolts after first step)??

4

The source code is written in a way that 2D elements are extruded and have additional constraints assigned, therefore it is better to work straight with 3D elements if you can.

5

There are known convergence issues e.g. with shell + Nlgeom + rigid body constraints. In the case of some other combinations with 1D and 2D elements, you may also experience non-convergence so it’s good to try with 3D (solid) elements.

6

i have tryied to try running using 2d elements for plate but with boundary op=MOD - i had to do this outside of prepomax (as far as i know) so i run through cmd
calculix is telling me that it finished calculations but on this computer i do not have python to be able to see frd file…

7

You can open the .frd file in PrePoMax, GraphiX (native GUI of CalculiX) or FreeCAD. There’s also converter to ParaView.

8

oh thank you i found how to do this in prepomax
well results seems fine and same as for simulation with one step:

https://www.dropbox.com/scl/fi/6dfictodccnfkshlkk6oc/Solver_111311.7z?rlkey=xi1623g87wsuuof0k8jdmv9fg&st=2aiy4kki&dl=0

so i guess it is better at least for bolt pretension with 2d planes to use boundary op=MOD
i guess maybye it is because of mesh extrusion…

9

if it is true than is it possible to change boundary op=new in prepomax or it needs to be done outside?