Coupled multiphysics-RF reduced models for MEMS

This paper proposes a fast and accurate order reduction algorithm for the extraction of compact macromodels for electrostatic actuated micro-electro-mechanical (MEMS) switches used in radio-frequency (RF) applications. The algorithm uses as input the information obtained from field simulations, in which coupled equations of electrostatic, structural-dynamic, air flow and full wave electromagnetic fields are used. The reduced macromodel is described as a circuit netlist which includes both the electro-structural-fluid flow multiphysics and the RF behavior of the switch. For the benchmark considered, the reduction is impressive, from more than 400000 degrees of freedom of the field models, to a macromodel of order 2 both for the multiphysics part and the RF part. The accuracy is checked by comparing relevant quantities (e.g. pull-in voltage for the multiphysics part and S parameters for the RF part) of the original field model with the same quantities obtained from the reduced macro-model. Relative errors less than 2% are obtained.