A novel model order reduction approach for generating efficient nonlinear verilog-a models of mems gyroscopes

Abstract

This paper reports a novel model-order reduction (MOR) approach for creating fast-running, nonlinear, multiphysics models in Verilog-A. This new approach differs from previous work by creating the reduced order model (ROM) directly from an accurate, nonlinear, multi-physics representation. The mechanical and electrical nonlinearities of the MEMS structure are persevered to capture effects such as quadrature, amplitude-dependent frequency shifting and electrostatic softening. The reduction algorithm has been implemented in the commercial MEMS/IC co-design tool MEMS+. The approach's effectiveness is validated for a state-ofthe-art, three-axis, capacitive gyroscope from Murata Electronics by comparing simulations of the created Verilog-A model with experimental data.

Linear and nonlinear dynamics of micro and nano-resonators: Review of recent advances

“…Hopf bifurcations [27], synchronization [28][29][30], internal resonances [31][32][33], and veering [34][35][36] are examples of physical phenomena that were extensively investigated at the micro and nano scale. Also, significant research has been conducted recently on nonlinear mechanics of microsystems, such as dynamic integrity of electrostatic MEMS [37][38][39] and AFM [40,41], control of MEMS [42][43][44], reduced-order modeling and gyroscopes [45][46][47].…”

Efficient nonlinear simulink models of MEMS gyroscopes generated with a novel model order reduction method

“…The resulting linear reduced system is ̃̇= ̃+ + , which can be rewritten as ̃̃̇=̃̃+̃+̃, with ̃= ,̃= ,̃= and ̃= . For details about how the matrix is constructed, please refer to [1].…”

“…The basic motivation for applying model order reduction to capacitive MEMS sensors is to extract fast dynamic models that can be inserted into an analog/mixedsignal circuit design environment such as Cadence Virtuoso ® [1] or a Simulink® block diagram for multidomain simulation. ROMs are used for MEMS/IC cosimulation and are expected to run with simulation speed comparable to lumped-parameter behavioral models while preserving the accuracy of detailed finite-element or boundary-element (FEM/BEM) models.…”

“…This work follows a previously published paper [1] in which the authors described a model order reduction approach. This approach starts from an accurate model, containing all the coupled multi-physics nonlinearities.…”

Development of Methodology of the Heat Sink Thermal Aspect for Low Cost Effective CFD Solutions for Reasonably High Quality Design

“…That is, it's lots quicker to chop the initial instance and perform a simulation with the reduced model than to perform a dynamic simulation of the initial high dimensional model. Parent and perineum [4] studied the novel model-order reduction (MOR) so as to create a quick -running, nonlinear, Multi-physics models. From the survey it had been reasoned that a multi-physics reference model not solely simplifies the creation of the memory, however additionally facilitates the user to create educated decisions if Model Order Reduction is one potential answer to our problem.…”