Hi,
I am trying to complete a buckling analysis on a tube with an external pressure. This solves itself easily by adding radial constraint on the internal edge of the tube, one axial constraint at one node and one tangential constraint on one node on each end. See pictures.
However, the radial constraint creates a artificial stiffness. Ideally, it shall represent a O-ring with some stiffness. My question is how I can easily do this. Is it possible to use radial spring elements, if so, how? Could not find anything in the manual about it.
Sincerely
David
Hi David,
The following are quotes from the Abaqus manual which might be helpful.
"You define the direction of action for SPRING1 and SPRING2 elements by giving the degree of freedom at each node of the element. This degree of freedom may be in a local coordinate system (Orientations*). The local system is assumed to be fixed".
* "You define the local coordinate system at the particular location at which it is required. You can select a rectangular, cylindrical, or spherical coordinate system. The coordinate system is defined in terms of points a, b, and c, as shown in Figure 1. You can select the method for defining points a, b, and c, as described below".
Alternatively, You could explicitly model the O-ring and define a contact between the cylinder and the O-ring. This might be preferable, because the stiffness of the spring should really depend on the numbers of springs you use (i.e. the stiffness will represent the average stiffness of the O-ring over the gap between the springs).
Regards,
Simon
Nils Wagner unfortunately I can't. However, I tried with Axial connectors that should work as springs elements where the stiffness can be described in the elasticity window. The reference length was to equal to the radius, such that only the pressure is working on the Axial connector. That being said, it does not work well. Results does not change with varying stiffness, and I get an oval shaped buckle which does not make sense with this mange "springs".
See picture for parameters I used.
Hi David,
A couple of points:
1. Why do you think that an oval shape is wrong? (It might be helpful to provide a deformed shape plot of the predicted buckle). The following link talks about the use of an initial imperfection for a buckling analysis.
https://simulation-blog.technia.com/simulation/buckling-post-buckling-imperfections-modelled-with-abaqus-fea
2. It might be good to use "non-linear" springs so that the O-ring doesn't resist the cylinder if it wants to lift off the O-ring.
Regards,
Simon
Simon Smith I have attached the buckling mode when it is radially constrained at the points where I have now attached the Axial connectors (springs).
The reason why it should not buckle as an oval shape is because of the springs that are only allowed to move radially. It should have some sort of shape relevant to the eight springs added. I understand that 8 springs does not simulate an o-ring perfectly, but it a start to try to make it work.
The cylinder is being pressed inwards due to the external pressure, so the springs or O-ring shall be compressed.
Hi David,
The "buckle shape" doesn't look unreasonable to me. Elasticity solutions look for the minimisation of potential energy (that is the competition between strain energy and external work done), so Abaqus is trying to do that for you. Clearly the springs are a source of strain energy as is the movement the cylinder. Abaqus has "decided" that this deformed shape predicted minimises the strain energy. I assume that you would get a very similar deformation pattern if the radial springs were converted to radial constraints (i.e. 8 of).
Regards,
Simon
Simon Smith Sorry, I think I confused you with that picture. I agree that the above attached buckling shape looks reasonable, and this is what I get with radial constraints.
However, the attached picture is what I get with the radial spring elements, oval shaped.
- Badges
- Science topic