MEMS惯性传感器的系统级建模技术研究

A Method for System-Level Modeling of Micromachined Inertial Sensors

阐述了基于多端口组件网络的MEMS建模方法——把MEMS器件分割为多个功能结构部件,并建立其参数化多端口组件模型,通过组件端口的相互联结形成的网络表征整个MEMS器件。根据物体的相对运动理论并采用能量法表述了建立微惯性器件的组件模型的一般方法。在此研究基础上,形成了MEMS的可重用组件模型库并以此建立了一电容式微陀螺的系统级模型。仿真结果表明多端口组件网络方法是快速建立和仿真MEMS系统级模型的有效方法,而物体的相对运动理论和能量法对建立惯性MEMS的组件模型具有普遍指导意义。

We present a modeling method based on multi-port-element network (MuPEN) for rapidly performing the system-level mixed-domain simulation of micromachined inertial sensors with interface circuit. With this method, micromachined inertial sensors are divided into many functional components, such as anchors, beams, plates, electrostatic gaps and so on. All components are modeled as parameterized multi-port-elements (MPEs) whose behaviors are evaluated using the rigid body relative movement theory and energy method. Then MPEs are connected together to form a network for representing the system-level model of the micromachined inertial sensors. Using this method, we established component models for spatial beams and comb structures, and then a reusable MPEs library is formed. As an application of the MuPEN method, the system-level model of a micro-gyroscope is established using the MPEs library. Finally, simulation results show that the modal frequencies obtained by MuPEN method agree well with those by finite element analysis, and that the MuPEN method is effective for rapidly performing system-level time-domain simulation.