Simulating Arched Piece on Ground

[bucknast]

Hi,

I'm trying to simulate a part that has a slight but constant and symmetrical arch to it, that's meant to be stood on by a person, somewhat like a skateboard with a constant bend, that's touching the ground at the tangential point in the middle. How can I best simulate the interface with the ground? Getting some weird solver errors.

Thanks!

 

[jvian]

You need to describe your situation a little better than that and "weird solver errors" doesn't help much as problems which seem simple static analyses can quickly escalate into non-linear animals depending on the solution requirements. What have you tried and what are you trying to accomplish?

- J -

 

[bucknast]

I have a board, approx 32" long and 8" wide, bent so the ends are ~3" off the ground. The curve is constant so it's tangent to the ground in the middle.

I'm loading it using two surfaces on the top of the board as an analog to a person's feet positioned at the ends of the board.

The issue I'm having is with providing the simulation a constraint to represent the ground. I've tried putting a surface under the board and get a "one or more surfaces are constrained but are not part of any elements" error. I've also tried making a solid piece under the board and am getting a "the model is insufficiently constrained for the analysis" error.

The only way I've been able to get reasonable results is remove half the board and constrain the split face.

 

[brep]

If you need to look at the contact stresses between the board and the ground, then certainly a contact analysis will be required. Sounds like this is the direct you're going from your last post. But remember that you'll need to constrain the ground too... Fully fixing it underneath might be a good approach to start with.

If you don't want to compute the contact stresses, but rather are interested in the displacements/stresses in the board itself, then you can just use constraints. In any case, you should start with this analysis so you know you have the modeling of the board done correctly, before adding the nonlinearity of the contact. If the geometry/loading is symmetric then it would make a lot of sense to cut the board in the middle and use a symmetry constraint there too. This will save you some time in solving. If you use solid elements it's pretty easy, since they only have translational degrees of freedom. Just make sure that the displacement through the mirror plane is fixed. Be careful not to overconstrain (i.e. think about how the material might want to squish out in other directions, and let it do so.) You''ll probably want to add two other point constraints to stop the model from flying around. If you get the "the model is insufficiently constrained for the analysis" error, then you know you need to add these constraints correctly. There are plenty of tutorials on this type of constraint approach in the help system.