FEM model of bonded laminated glazing

About the development of the FEM module/workbench.

Moderator: bernd

FVerbaas
Posts: 13
Joined: Fri Mar 03, 2023 8:52 am

FEM model of bonded laminated glazing

Post by FVerbaas »

I have been trying to build an FEM model to evaluate glazing bonded onto a frame.
It is a simple model for validation of the FEM route. It consists of a glass pane of 600x1200x8 mm bonded to a substructure with a bond of 20 mm wide and 8 mm thick.
I tried to model the glass and the adhesive as separate objects and define connectivity between them. This lead to 5 geometry meshes and the solver reported there was no mesh. Then I merged the components and the solver complained it could not pair some surfaces.

I obviously am missing something. I did some search but unfortunately the term 'glass' triggers a lot of links referring to screen representation.

Is there a good tutorial on how to do this sort of jobs? Intended method of evaluation is quasi-static geometric non-linear.
The aim in the end is to implement this sort of computations in a Python script so any pointers in that direction are welcome also.
chrisb
Veteran
Posts: 53930
Joined: Tue Mar 17, 2015 9:14 am

Re: FEM model of bonded laminated glazing

Post by chrisb »

Hi and welcome to the forum!

I moved your post to the FEM forum. Although you ask just for a tutorial, it may be sensible to attach your file to get direct help concerning your issue.
A Sketcher Lecture with in-depth information is available in English, auf Deutsch, en français, en español.
User avatar
NewJoker
Veteran
Posts: 3018
Joined: Sun Oct 11, 2020 7:49 pm

Re: FEM model of bonded laminated glazing

Post by NewJoker »

To solve such a problem properly you would have to use so-called cohesive elements or cohesive contact - they include the damage behavior of the adhesive connection. Unfortunately, they are not available in CalculiX and thus also in FreeCAD. So you will have to use a simplified approach with an adhesive layer meshed with regular elements and merge the layers. Compound creation should be sufficient if there are no gaps between them.
FVerbaas
Posts: 13
Joined: Fri Mar 03, 2023 8:52 am

Re: FEM model of bonded laminated glazing

Post by FVerbaas »

chrisb wrote: Fri Mar 03, 2023 9:30 am Hi and welcome to the forum!

I moved your post to the FEM forum. Although you ask just for a tutorial, it may be sensible to attach your file to get direct help concerning your issue.
Yes, thanks.
I did not add my file because I thought it would be better to restart from clean slate based on the advice received than carry on with a file that probably carries the legacy of my earlier attempts.
FVerbaas
Posts: 13
Joined: Fri Mar 03, 2023 8:52 am

Re: FEM model of bonded laminated glazing

Post by FVerbaas »

NewJoker wrote: Fri Mar 03, 2023 12:42 pm To solve such a problem properly you would have to use so-called cohesive elements or cohesive contact - they include the damage behavior of the adhesive connection. Unfortunately, they are not available in CalculiX and thus also in FreeCAD. So you will have to use a simplified approach with an adhesive layer meshed with regular elements and merge the layers. Compound creation should be sufficient if there are no gaps between them.
Thanks.
Failure of the adhesive connection is not part of the investigation. The adhesive connection normally is in compression. Tensile stress may occur in the outside edge of the adhesive but that can be monitored via the local elongation. For the adhesive a simple linear model will do.

I understand that indeed I need to merge the layers before meshing them.
Steps would be:
1 - define glass plate part (A cube properly scaled and moved up by the thickness of the adhesive)
2 - define adhesive part (difference of two cubes to generate a ring section)
3 - define FEM container object
4 - define materials for glass part and adhesive part
6 - define lower surface of adhesive as clamped constraint
5 - merge the glass and the adhesive parts in compound and mesh the resulting compound part. There will be element faces where the two parts met and elements will have the material properties assigned to the material of the parts the volume came from.
6 - add the load constraint (UDL on glass top surface)
7 - run the calculation

Am I correct?
FVerbaas
Posts: 13
Joined: Fri Mar 03, 2023 8:52 am

Re: FEM model of bonded laminated glazing

Post by FVerbaas »

I have built the model with a meshed compound but I now run into a message that
"Calculix can not be started due to missing perrequisites:

Some material objects do not have the same reference shape type (all material objects must have the same reference shape type, at the moment). "

The glass plate is a transformed cube. The adhesive is a 'cut' of a similar transformed cube with the clear view opening. Will it help if I make the glass plate also a 'cut'? by making the cube slightly larger and clip it off to desired size? I would have assumed that the meshing has solved the differences in shape type.

Can someone help me please? The file is attached.
Attachments
ISO11336-1_ed2_TableJ1_Case2.FCStd
(353.08 KiB) Downloaded 27 times
User avatar
NewJoker
Veteran
Posts: 3018
Joined: Sun Oct 11, 2020 7:49 pm

Re: FEM model of bonded laminated glazing

Post by NewJoker »

The problem is that you assigned one of the materials to a face while the other one is assigned to a volume.
FVerbaas
Posts: 13
Joined: Fri Mar 03, 2023 8:52 am

Re: FEM model of bonded laminated glazing

Post by FVerbaas »

Thanks. That did it. I had set to select solids and there was a message saying the solid would be selected if you did choose one of its surfaces.
Now corrected. I attached the updated file.
CalculiX now stated but returns new problem: There is an 'invalid subelement name':

Code: Select all

Run of CalxuliX ccx tools solver started.
17:40:24  Running the Python command 'FEM_SolverRun' failed:
Traceback (most recent call last):
  File "C:\Program Files\FreeCAD 0.20\Mod\Fem\femcommands\commands.py", line 1053, in Activated
    run_fem_solver(self.selobj)
  File "C:\Program Files\FreeCAD 0.20\Mod\Fem\femsolver\run.py", line 108, in run_fem_solver
    fea.run()  # standard, no working dir is given in solver
  File "C:\Program Files\FreeCAD 0.20\Mod\Fem\femtools\ccxtools.py", line 634, in run
    self.update_objects()
  File "C:\Program Files\FreeCAD 0.20\Mod\Fem\femtools\ccxtools.py", line 248, in update_objects
    self.member = membertools.AnalysisMember(self.analysis)
  File "C:\Program Files\FreeCAD 0.20\Mod\Fem\femtools\membertools.py", line 255, in __init__
    std_mats = self.get_several_member(
  File "C:\Program Files\FreeCAD 0.20\Mod\Fem\femtools\membertools.py", line 326, in get_several_member
    return get_several_member(self.analysis, t)
  File "C:\Program Files\FreeCAD 0.20\Mod\Fem\femtools\membertools.py", line 126, in get_several_member
    obj_dict["RefShapeType"] = femutils.get_refshape_type(m)
  File "C:\Program Files\FreeCAD 0.20\Mod\Fem\femtools\femutils.py", line 365, in get_refshape_type
    first_ref_shape = get_element(first_ref_obj[0], first_ref_obj[1][0])
  File "C:\Program Files\FreeCAD 0.20\Mod\Fem\femtools\geomtools.py", line 189, in get_element
    return part.Shape.getElement(element)  # Face, Edge, Vertex

Invalid subelement name
Attachments
ISO11336-1_ed2_TableJ1_Case2.FCStd
(353.06 KiB) Downloaded 28 times
User avatar
NewJoker
Veteran
Posts: 3018
Joined: Sun Oct 11, 2020 7:49 pm

Re: FEM model of bonded laminated glazing

Post by NewJoker »

Something is wrong with the assignment of the second material. Try redefining it and applying to the top box again.
FVerbaas
Posts: 13
Joined: Fri Mar 03, 2023 8:52 am

Re: FEM model of bonded laminated glazing

Post by FVerbaas »

I tried that but no avail. Probably there is already some old crap in the file, so i started again. (the model is not that difficult).

I decided to drop the adhesive thing and to try to build a laminate first. A laminate, in this sense, is two glass plates bonded to the opposite sides of a sheet of interlayer material Glass has E-modulus of 70GPa, interlayer has E-modulus of 3.6 MPa. Glass plates are 12 mm thick, interlayer is 1.5 mm thick.

I understand that when the mesh is built from a compound I need not define a tie between the glass plies and the interlayer if there is no gap. Attempts to build the laminate from 3 boxes, properly translated to form a stack, did not lead to an 'automatic' tie. Attempts to define a tie on the surfaces lead to failure because CalciliX concluded there were too many nodes that had no partner.
Maybe it is better to define the laminate as a big box sized the full size of the laminate and use difference operations to cut it in three?
Also does it matter which mesher I use?

Apologize for the many questions
.
Post Reply