Simulation methods for generating reduced order models of MEMS sensors with geometric nonlinear drive motion

We compare different simulation methods for the development of reduced order models (ROM) that feature a geometric nonlinear drive motion in a MEMS device. We benchmark three different simulation methods by comparing the results to measured frequency shifts of a MEMS gyroscope. We find that quasi-static points in the dynamic trajectory of the gyroscope are suitable deflection states for generating the modal stiffness. Furthermore, we show an efficient simulation method that overcomes the burden of the transient FE simulation while exhibiting similar accuracy.