I am at the moment analyzing a structure made up of many hexagonal shell elements. I have them connected via line-joints. I have edited the meshes, so that the corners, where three elements would meet, are cut out. This way there are only joints connecting two elements.
I am struggling with evaluating, however, to which elements I should orient/connect my joints to. Depending on the direction of the Y-Axis for each joint, the entire calculation results change, principal stresses make no sense. However, the deformation looks right. Most importantly: The spring stiffness of the joint is impacting the deformation in the expected manner.
I have now extended DAlpha of the connecting joints over 180°. In my understanding, this would reflect the actual construction. And the forces in the shells look very accurate to me like this. However, in this version, the deformation is very small and not impacted by the spring stiffness at all. I’m talking a range of 1 to 6000 in all three vector axis for Ct, which doesn’t effect the displacement in any way.
Hello @luisbarthelemy,
it is hard to say what is going wrong without seeing the Grasshopper definition.
Did you do a test with only two or three hexagonal elements?
If this does not give a decisive clue: would it be possible that you send over this minimum definition?
– Clemens
thanks for your reply. Building a smaller scale model is a good idea.
In the end, my question boils down to: What exactly is happening “behind the scenes” when you attach a line joint to both elements as opposed to one? Is it valid to increase DAlpha to/over 180°?
Since I am connecting my elements at their edges, from my understanding this would be the case - but is it?
I have realized that Karamba apparently can’t compute Line-Joints which are connected in both directions for my model (DAlpha >= 180°). Is that generally the case or do I need to do something different? The top image has the joints included but gives exactly the same result as when I leave them out.
The thing is, that a joint only connected to one element doesn’t reflect the construction (in my understanding).
How would you approach this case? I am thinking about trimming the elements and adding shells inbetween to reflect the connection.
Line joints in Karamba3D introduce additional DOFs between the element and its boundary. If there are joints defined on two sides of a line the DOFs that are connected to that line can become kinematic.
Setting DAlpha to a very large angle (e.g. 180°) makes the selection of the side on which the line joint shall be attached undefined - it does not add two hinges.
In order to experiment with the line-joints it could be helpful to add a small stiffness to the joint DOFs.
Trimming the elements and adding small shell-pieces in between is possible in theory but requires quite some additional modeling and analysis effort as compared to the line-hinge option. I would not recommend that.
