Computation speedup of macromodels for dynamic simulation of microsystem

Abstract

A doubly clamped microbeam actuated by electrostatic force with squeezed gas film damping is a well-known micro device for many researchers to demonstrate how reduced-order dynamic macromodels are an effective way to faithfully capture the device behaviors. However it still takes long time to directly recompute the time-dependant nonlinear terms in macromodels, which are generated by proper orthogonal decomposition (POD) method with Galerkin procedure, at every time step during the macromodel simulation. This paper proposes two methods to speed up the computation of macromodel simulations. In the first method, the computation speedup is achieved based on the concept of precomputation upon the basis functions are available while the second method scales down the size of spatial integration by using Gaussian quadrature and generalized differential quadrature. Numerical results show both methods could enhance the efficiency of the macromodel simulation.