An Approach for Model Updating of a Multiphysics MEMS Micromirror

Abstract

In this work, a general approach is formulated for updating the parameters of systems governed by multiphysics equations. The optimization technique is based on Genetic Algorithms (GAs). GAs represent a class of probabilistic optimization strategies loosely patterned after a simplified evolutionary scheme and Darwin’s “survival of the fittest” concepts. The GA is coupled to a commercially available multiphysics finite element program. In the context of this, issues of mode tracking, eigenvector comparisons, and approximate function evaluations are discussed. The approach is demonstrated on a micro-electro-mechanical (MEMS) micromirror which is governed by both structural and electrostatic physics. The MEMS mirror is characterized dynamically using a laser vibrometer. These experimental measurements are then used in the model updating of the finite element structural model and the electrostatic model. Several interesting observations that were encountered in this development will also be discussed.