-2
$\begingroup$

I'm searching for the displacement of the surface of a elastic rectangle for a given x and y and a force at a position. Like a bowling ball on a trampoline.

The equation should include a variable for the elasticity of the surface and the position and strength of the point-shaped force. The return should be a delta-z to any given xy-pair.

I tried a lot of sin cos stuff but could not figure it out. One of the problems is the surface should be fixed at the edges (like a trampoline). So the influence of the force is getting weaker (or the surface is getting stiffer) near the edge.

There is no need for a physically correct model. A function that gives me roughly such a curve would help me a lot.

Any suggestions?

Improve this question
$\endgroup$

1 Answer 1

Reset to default
5
$\begingroup$

Your problem is one of solving Laplace's equation subject to Neumann boundary conditions. This problem is equivalent to a problem in electrostatics. For small displacements, the height of the trampoline is equivalent to the electric potential, and the bowling ball is eqivalent to a point charge. The fixed boundary of the trampoline is a conducting, grounded box. You should use the method of image charges1. Your original charge maps into a two-dimensional lattice of image charges, as reflected in the sides of the rectangle. The sign of the image charge alternates in a checkerboard pattern. To find the resulting potential, You could either sum the two-dimensional infinite series analytically, or just add up the closest charges numerically. I think the analytical solution involves an elliptic integral, which I prefer to avoid, so if I had to do this I'd do it numerically.

Improve this answer
$\endgroup$
1
  • 3
    $\begingroup$ The cartoon image of a bowling ball on a trampoline is quite far from a small-displacement linear approximation, but I suppose if the trampoline can handle several jumping humans then the bowling ball is light enough. The series you suggest converges rather slowly as it stands; the elliptic integrals aren't that bad, and at least they have more rapidly converging series expansions. $\endgroup$
    – Noam D. Elkies
    Commented Jan 13, 2013 at 0:39

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .