Spring Elements
I'm trying to setup spring elements but I am not getting the results I expect.
I've set a CBUSH element with k1 = 10 N/mm, and no constraints. I applied a 100 N along the axis of the element, expecting an infinite displacement (because there's no constraint).. Instead I get a displacement of 20 mm.
I applied a dof 123 = 0 constraint at the base, expecting a displacement of 10 mm, but I got a displacement of 20 mm again.
If the force is oriented transversely to the spring axis, the displacement is 0 mm.
When I have a CELAS2 element with k=10 N/mm, and the force oriented along the element, the displacement is 7.07 in the horizontal direction and 0 in the vertical direction.
What's going on?
Answers
-
In 1st case, your model is underconstrained and will not work correctly, broadly speaking an implicit model should be in equilibrium, you should have a warning as below in your out file, the results are meaningless if you have this
In 2nd case, set up should work, I copied it and got 10mm as expected (BushTest.fem and .h3d attached)
For the 3rd (CELAS) case, the DOF component is in coordinate system of the node(s) (not GA-GB) so if the element is at 45 degrees as you have it, you need to define a coordinate system aligned with the CELAS and reference it as CD on the node (see CelasTest.fem and .h3d)
0 -
Paul Sharp_21301 said:
In 1st case, your model is underconstrained and will not work correctly, broadly speaking an implicit model should be in equilibrium, you should have a warning as below in your out file, the results are meaningless if you have this
In 2nd case, set up should work, I copied it and got 10mm as expected (BushTest.fem and .h3d attached)
For the 3rd (CELAS) case, the DOF component is in coordinate system of the node(s) (not GA-GB) so if the element is at 45 degrees as you have it, you need to define a coordinate system aligned with the CELAS and reference it as CD on the node (see CelasTest.fem and .h3d)
Thank you.
I redid case 1 with the "-out" option and the warning appeared. That type of warning should really show up even when "-out" is not specified. A very large displacement result would give a clue that the model is not constrained and flying off to infinity. A displacement of 20 mm is close enough to be believable.
I redid case 2 and it worked properly. I reoriented the load horizontally (Case 3) and ran it again. There was no warning and the displacement was 9e-08. I've attached the file.
For the CELAS case , that makes sense. But what if I don't know the direction that the load will act in? How do I configure a spring that, if a load is applied at an angle to its axis, it rotates/bends and aligns its axis along the direction of the load, like a real spring would?
0 -
AM_20700 said:
Thank you.
I redid case 1 with the "-out" option and the warning appeared. That type of warning should really show up even when "-out" is not specified. A very large displacement result would give a clue that the model is not constrained and flying off to infinity. A displacement of 20 mm is close enough to be believable.
I redid case 2 and it worked properly. I reoriented the load horizontally (Case 3) and ran it again. There was no warning and the displacement was 9e-08. I've attached the file.
For the CELAS case , that makes sense. But what if I don't know the direction that the load will act in? How do I configure a spring that, if a load is applied at an angle to its axis, it rotates/bends and aligns its axis along the direction of the load, like a real spring would?
The warning is always written to the output (.out file), -out is just a toggle for echoing output to the screen/console display (equivalent to using -screen in the deck).
For your sideways load, attached version works, but you need to modify constraints (to prevent rotation), stiffness in Y, set a vector for Y orientation (I used X1,X2,X3 here, but you could use CID instead) and set spring/damper location (S) to 'top' of spring element to get 10mm (since the actual 'spring' in this case remains 0D and is rigidly coupled to the end point locations)
The CELAS cannot 'bend' as such in a linear analysis, so it will take the component of loading in the direction it applies in (this is what you were doing originally in effect, giving the 7.07mm displacement), in order to have a bending spring that behaved as you describe, you would need to do a non-linear analysis I think
0 -
Paul Sharp_21301 said:
The warning is always written to the output (.out file), -out is just a toggle for echoing output to the screen/console display (equivalent to using -screen in the deck).
For your sideways load, attached version works, but you need to modify constraints (to prevent rotation), stiffness in Y, set a vector for Y orientation (I used X1,X2,X3 here, but you could use CID instead) and set spring/damper location (S) to 'top' of spring element to get 10mm (since the actual 'spring' in this case remains 0D and is rigidly coupled to the end point locations)
The CELAS cannot 'bend' as such in a linear analysis, so it will take the component of loading in the direction it applies in (this is what you were doing originally in effect, giving the 7.07mm displacement), in order to have a bending spring that behaved as you describe, you would need to do a non-linear analysis I think
Thanks again. I feel like an idiot because I still don't understand the behavior.
I didn't understand this at all: "set spring/damper location (S) to 'top' of spring element to get 10mm (since the actual 'spring' in this case remains 0D and is rigidly coupled to the end point locations)". Can you explain that to me like I'm 5?
If I have a spring that is on a base that cannot resist rotation, and the spring has any transverse stiffness, then I expect the spring to rotate horizontally and then elongate to resist the horizontal force:
I don't understand how a spring with no transverse stiffness would behave. I'd think it should just fly off horizontally to infinity:
But that would cause the spring to elongate so it would create tension along the diagonal axis which would again force the spring to rotate completely horizontal.
If the base can resist rotation, then there can't be any tension in the spring.
I'm not sure what the hell is going on.
0
Altair Employee