3D ADI-FDTD算法在分析微结构时的数值色散  

Numerical Dispersion Analysis of 3D ADI-FDTD for Micro-scale Structures

作　　者：吴银锋[1] 万江文[1] 冯仁剑[1] 

机构地区：[1]北京航空航天大学仪器科学与光电工程学院,北京100191

出　　处：《系统仿真学报》2009年第13期3889-3893,共5页Journal of System Simulation

基　　金：北京市自然科学基金资助(3083024);教育部新世纪优秀人才支持计划资助(NCET-04-0113);北航青年创新基金

摘　　要：利用三维无条件稳定交替方向隐式有限差分(ADI-FDTD)算法分析射频微机电器件时,为了保证仿真精度和提高计算效率,需要掌握微尺度条件下算法的数值色散特性。利用算法增长矩阵,讨论了多种因素对算法数值色散特性的影响。给出了分析微结构电磁特性时各参数的基本选取原则:空间导数取二阶差分;网格划分的深宽比小于50;取扫描频段内的中高频率点作为计算空间抽样率的参考点;ADI-FDTD算法的时间步长极限为T/2,时间步长的选择应满足时间抽样率与空间抽样率之比小于1。Numerical dispersion analysis of three-dimensional （3D） alternating-direction-implicit （ADI） finite difference time domain （FDTD） methods when analyzing for Radio Frequency Micro-Electro-Mechanical-System （RF MEMS） structures was proposed. It is the key issue for saving simulation time and assuring simulation accuracy. Using growth matrix, several factors which affect numerical dispersion were discussed in detail. The basic setup principles of parameters were suggested： Second order difference scheme is enough for spatial deviation. Aspect ratio should be less than 50 while meshing grid. The reference points when calculating the spatial sampling rate should be selected in medium-high frequency band. T/2 is nyquist time limit in ADI-FDTD method. Time step size should be determined only when the ratio of time sampling rate and spatial sampling rate is less than 1.

关 键 词：ADI-FDTD 数值色散 微结构 射频微系统 

分 类 号：TN252[电子电信—物理电子学]