Hinges between plates/shells

[thschrader]

Does anyone have an idea how I can simulate joints between the plates?
I can define the lines 1+2 as a hinge with my commercial software.
Any experience with FreeCAD?
Thanks.

concrete_plate.FCStd

(213.49 KiB) Downloaded 52 times

hinges.JPG (86.23 KiB) Viewed 2283 times

[bernd]

Would you like to define the hinges to be used with FreeCAD FEM calculix or with StructuralAnalysis IFC export or for some other reason?

I assume for the use with FreeCAD FEM Calculix. This has been in my mind and on my TODO for a long time. Even before the Sat May 20, 2017 ...

Never had the time range to get it done ...

[thschrader]

I assume for the use with FreeCAD FEM Calculix. This has been in my mind and on my TODO for a long time.

Yes, with FC-FEM, because the the modeling options are far better than with my commercial software.
Maybe you can rearrange your todo-list...

[bernd]

Are you able to compile FreeCAD on your Windows?

[thschrader]

Are you able to compile FreeCAD on your Windows?

I never tried. What must I do?

[bernd]

Maybe you can rearrange your todo-list...

If we do it together I will go for it. May be user1234 jumps in too? We could even make a new topic in German subforum if you guys do like to write in German.

You will need to compile FreeCAD to test our new object. Compiling FreeCAD is not difficault. Since I do only compile on Linux I am not of any help here. On all further steps once you are able to compile FreeCAD I will guide you. There is a huge wiki site about this and there are afew people around who will help. Make a new topic on instsll/compile for this.

First thing we need is the tweaked inp file for the example you posted. Means the inp file with hinges. Than we know what we need to generate. It is a good one, we should use it.

[fandaL]

Possible approach is to separate elements by duplicating nodes on the hinge edge, so that elements on one side does not share nodes with elements on the other side.

Next, you would generate *EQUATIONs on duplicate nodes to connect translation DOF, so that rotation DOF are independent (hinged). You can look to, example:

Code: Select all

*EQUATION
2
11,  1, 1, 22, 1, -1

which means that equation have 2 terms:
1 * u11 + -1 * u22 = 0
i.e. node number 11 DOF 1 equals to node number 22 DOF 1. So that for all translation DOF between nodes 11 and 22, equations will be

Code: Select all

*EQUATION
2
11, 1, 1, 22, 1, -1
2
11, 2, 1, 22, 2, -1
2
11, 3, 1, 22, 3, -1

Maybe the easier way would be to use *MPC instead of *EQUATION

Code: Select all

*MPC
BEAM, 11, 22

which keeps the distance between nodes 11 and 22 constant according to the manual, but I’m not sure if *MPC in this case triggers geometrically nonlinear calculation or not. That may prolong the calculation compared to *EQUATION.

[user1234]

I would say (no warranty) for the hinge is the displacement dialog the right thing. Select a line for a hinge, the hinge direction (in that case x and z direction, y if it is fixed bearing ) and let the rotation free.

0.png (267.34 KiB) Viewed 2013 times

But i can not verify that at the moment. I must calculating something, where the fixed position and consequently the tension of the length is respected. All formulas in the books i found do not do that. There is always only z fixed (it took me a while to figure that out). But hat also means, that really every little bit fitting tolerance (every 0.05mm in the hinge in 1m free length) have a very big effect on the displacement (often important for functionality). The von Mises stress and shear stress is similar.

calculation no hinge, only loose bearing:

Code: Select all

l = 1000.0   # mm
b = 1000.0   # mm
t = 5.0      # mm

rho = 7900.0 # kg/m³
g = 9.81     # m/s²
E = 210000.0 # N/mm²

m = l * b * t * ( 10.0 ** -9.0 ) * rho
F = m * g
I = b * t ** 3 / ( 12.0 )
f = 5 * F * l ** 3.0 / ( 384.0 * E * I )

print ( "m = " + str (m) + " kg" )
print ( "F = " + str (F) + " N" )
print ( "I = " + str (I) + " mm⁴" )
print ( "f = " + str (f) + " mm" 

Code: Select all

m = 39.5 kg
F = 387.495 N
I = 10416.666666666666 mm⁴
f = 2.306517857142857 mm

result in FreeCAD FEM:

nohinge.png (462.71 KiB) Viewed 2013 times

|2.3mm| vs. |2.4mm| --> OK.

with hinge:

hinge.png (457.42 KiB) Viewed 2013 times

f = |0.96mm| --> less the a half, because the sheet is strictly position fixed (not rotation) on the ends.

file with both in it:

test34_hinge_1.FCStd

(32.73 KiB) Downloaded 45 times

Greetings
user1234

[bernd]

Totally correct if you would like to have a hinged support, but Thomas would like to set the hinge between two faces ore between two beams and not on a support.

See https://de.wikipedia.org/wiki/Dreigelenkbogen for example. the bahnhof Zürich is a good example for a hinge as well. https://www.alamy.de/stockfoto-gleis-12 ... 02265.html

Ahh brand new station of Nebelhorn cable car is a perfect example of a Dreigelenkbogen. ...

nebelhornbahn_talstation.jpg (440.24 KiB) Viewed 1985 times

[user1234]

To be honest, i do not get that out of the description from the first post. But yes, that in FEM will be pretty hard.

Greetings
user1234