1
$\begingroup$

I was wondering if we could model rigid body motion of bodies using finite element models. Particularly I'm interested to know if we can model motion of objects with no constraints or with some degrees of freedom (such as only rotation). Though I've used word 'rigid body' to convey the situations closest to what I'm considering, In reality there will be stress, considering non-uniform application of force on surface/body or because of rotation they undergo, and these are not small enough to ignore.

$\endgroup$
  • 1
    $\begingroup$ Rigid bodies - by definition - have no strain and stress. What are the equations you would like to solve with “no constraints”? $\endgroup$ – BalazsToth Mar 21 '18 at 9:05
  • $\begingroup$ Thanks. This is about equations of motion - F = m*dx^2/dt^2., and its rotational counter part. Though I've used word 'rigid body' to convey the situations closest to what I'm considering, In reality there will be stress, considering non-uniform application of force on surface/body or because of rotation they undergo, and these are not small enough to ignore. $\endgroup$ – tired and bored dev Mar 21 '18 at 10:54
  • $\begingroup$ Can you give us an example of what you want to model? $\endgroup$ – P. G. Mar 21 '18 at 11:06
  • $\begingroup$ Suppose you've a rotating solid sphere, which is has magnetic dipole in it. It's subjected to an external magnetic field. The sphere also has linear motion. $\endgroup$ – tired and bored dev Mar 21 '18 at 11:09
  • $\begingroup$ Finite element codes that have a nonlinear formulation (i.e. nonlinear strain-displacement relations) are designed to handle arbitrarily large displacements and rotations. Most commercial structural finite element codes have this capability and there are many open source codes that do as well. $\endgroup$ – Bill Greene Mar 21 '18 at 11:42
1
$\begingroup$

Finite element method is used to change the boundary value problem, with an infinite number of unknowns, into a system of algebraic equations with a finite number of degrees of freedom (DOFs).

For example, the response of the deformable body, which conveniently is described by partial diffrenetial equation (PDE), using finite element method is transformed into a system of algebraic equations with the finite number of DOFs.

If you write equations for rigid body or system of rigid bodies, you do not have PDE; you have immediately a system of algebraic equations. Thus no need for finite element method.

The problem with the system only rigid bodies is that is very often over-constrained, since only you have equations of equilibrium to work with. However, this is another story.

$\endgroup$
  • $\begingroup$ If we write equations of motion in differential equations format, we may not get systems of algebraic equations directly. Do we? $\endgroup$ – tired and bored dev Mar 25 '18 at 16:57
  • $\begingroup$ Correct, I made shortcut here; finite element method is applied to discretise problem in space, not in time. You can think that you apply the first discretisation in time, and then in space or vice versa. $\endgroup$ – likask Mar 25 '18 at 17:00

Your Answer

By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.

Not the answer you're looking for? Browse other questions tagged or ask your own question.