Scalable macromodels for microelectromechanical systems

A fully automatic macromodeling methodology to generate scalable reduced-order models for microelectromechanical systems is presented in this paper. Krylov subspace methods are used to generate reduced-order models from detailed higher-order models of the device under consideration. The entire methodology is implemented in a symbolic computation environment to preserve dependencies on physical device parameters. The macromodeling methodology is demonstrated using the combdrive microresonator as a representative example. The results are compared with analytical models, NODAS and finite element simulations.

Duke Scholars

Author Richard B. Fair Electrical and Computer Engineering