Nonlinear SpringA

disla

This is an example of how to configure a non-linear SpringA in Mecway.
Just make sure to cover all the range of possible elongation and reaction force that the spring will undergo.
Don't forget possible transverse reaction forces. They play an important role too and you could be thrown out of the curve without realizing it.

NOTE: There is no need for such a sophisticated unittriangle function, but I am a bit picky and I like to see that the first and last step force is equal to zero.

disla

One could use this Nonlinear spring element to set up an only Compression or Only Tension Truss.
Just provide the proper Reaction /Elongation Curve.

NOTE: Do not confuse displacement with elongation. Not the same. It has puzzle me for a while. Mecway do not provide Elongation of the Spring. Could be a nice Improvement.

For Example, an Only Tension Truss element could be set up like:

JohnM

My spring is in need of pre-tension, and I am not finding a way to do this. When I have a spring connected to ground, I can "pull" the ground side to put in pre-tension, but if the spring connects two parts, I am out of ideas. Is there a custom model input I can use to get what I need?

disla

Hi John,

Have you tried two springs instead of one and a solid or beam section in between?
The pretension would be applied to the that mid element.

This is the idea.

Victor

@JohnM For small displacements, you can use *PRE-TENSION SECTION on a beam in series with the spring but the only way I could find to set it up is quite fiddly, especially if the spring isn't axis-aligned.

A truss element with thermal strain would be easier if your model allows that.

disla

@Victor,

I was setting this up and abondon when realized the amount of equation required.
Colinearity should be imposed by pairs of nodes on each equation. Writing all nodes in one *equation card isn't enough.

Victor

Yea, you need 2 equations and if it's at a funny angle, you'd have to do a bit of maths to work them out. Or maybe avoid that by putting *TRANSFORM on the nodes so the DOFs are aligned with the spring. In that case, you might as well write the equations emulating *PRE-TENSION SECTION yourself while you're at it. I think it's basically just ux1-ux2-ux3=0 and ux3=length change.

Fatmac

i've been using pretension sections as shroud adjusters in yacht rigging models -its a bit of a work up and I imagine it sucks up some computation time but it has appears to work and sort of self aligns also.

JohnM

@fatmac I thought of that exact idea but dismissed it because I thought it was a dumb idea. apparently not! thanks