MEMS simulations using Elmer

Numerical methods and mathematical models of Elmer
Post Reply
fernando
Posts: 1
Joined: 10 Jun 2016, 05:49
Antispam: Yes

MEMS simulations using Elmer

Post by fernando »

Hello,

I am new to Elmer. I've been trying to simulate a MEMS cantilever, but have faced a convergence problem when using iterative solvers.

The cantilever is 100x20x1 mircometers. A structured hexahedral mesh, with elements 1x4x0.2 um was generating using GMSH. One end was fixed, and a 1uN load was applied on the free face. Assuming the material to be silicon, the theoretical displacement is 1.1um. If a direct method is used with umfpack solver, Elmer solves the problem after a few minutes, and shows a deflection of 1.1um. However, if I try an iterative method to speed up the simulation, the solution does not converge. I've tried many combinations of iterative solvers and pre-conditioners, but couldn't get it to work.

If I change the thickness of the cantilever to ~5um, the solution converges (Then the element aspect ratio is comparable to the beam example in the tutorial). So it seems that the problem could be the high aspect ratio of the elements. I can decrease the aspect ratio of elements, but then there will be a huge number of unnecessary elements, which will again take a long time to solve.

I would like to simulate more complicated MEMS structures using Elmer, so I'm interested to know how thin-bodies can be simulated with an iterative solver. I'll be grateful for any suggestions from the community.

Thanks and regards,

Fernando
Post Reply