Research On System-Level Simulation Of MEMS Based On Reduced-Order Model

The focus of MEMS production has changed of from the development of a single MEMS device to the development of a whole MEMS system and the development of integrated testing methods' research required MEMS can model and simulate for MEMS system speedily. And the structural design coming into being spurred MEMS design to develop. System-level simulation methods based on macro-model have become a famous research field while can get behavior characteristic of MEMS fast and accurately even if most of MEMS are across multiple coupled energy domains.Aiming at the requirements of fast system-level modeling and simulation and the characteristics of multiple energy domains interaction, this paper has been investigated researched the Krylov subspace based project method and automated macro-modeling software framework. Obtaining full math model of MEMS device through FEM software ANSYS, and then obtaining model order reduction (MOR) through reduction algorithm based on Krylov. Lastly the model can be simulated by system simulation tool. To finish the whole process of system-level simulation through constructed various MEMS device using numerical algorithm. Enrich structure design method of MEMS CAD. Firstly, based on structure design method, the gap of finite element method (FEM) constructing model to system-level simulate has been analyzed. By obtaining device macro-model can recuperate the gap. The whole flow was built in VC++ 6.0, through the interface between ANSYS software and C++ program to build FEM math model, to obtain MOR through the reduce order arithmetic which was built in MATLAB and encapsulated, and lastly to view dynamic characteristic of MEMS system by SIMULINK which was integrated in MEMS_DUT.At last, taking a structural macro-beam,a fixed-fixed beam with multiple coupled energy domains,a macro-integrated MEMS drivers with more devices and multiple coupled energy domains as examples to show the efficiency and reliability of the MEMS_DUT. And several advices about improving model order reduction algorithm and optimizing the system-level simulation tool has been proposed.