deepmodeling / jax-fem

Differentiable Finite Element Method with JAX
GNU General Public License v3.0
290 stars 45 forks source link

Discrepancy with linear_elasticity demo and analytical/commercial-FEM-software solution #37

Closed baldurbjo closed 1 month ago

baldurbjo commented 1 month ago

Hi, when I run the linear_elasticity demo on my machine, I get a max displacement value of -1.45. When I compute the same problem using analytical beam equations or using Creo Simulate, I get a result that is four times lesser, -0.36.

Why does this occur?

baldurbjo commented 1 month ago

I just noticed that the load is added as a traction load, that is why the displacement is quadruple on a 2x2 beam cross section. Problem solved.