复杂陈述式仿真模型的指标分析  

Index Analysis of Complex Declarative Simulation Models

作　　者：丁建完[1] 陈立平[1] 周凡利[1] 

机构地区：[1]华中科技大学国家CAD支撑软件工程技术研究中心,武汉430074

出　　处：《系统仿真学报》2006年第8期2092-2096,共5页Journal of System Simulation

基　　金：国家"八六三"高技术研究发展计划(2003AA001031);国家自然科学基金(60574053);国家重点基础研究发展规划项目(2003CB716207)

摘　　要：基于方程的面向对象建模语言使用微分代数方程描述系统构件的物理现象和行为。采用此类语言的物理系统建模经常产生高指标DAE问题。众所周知,高指标问题是困扰DAE系统数值求解的一大难题。针对复杂大系统建模产生的DAE系统的指标问题进行了研究,提出了先分解后约简的指标分析策略,并给出了相应算法。采用基于二叉树的符号微分方法实现了方程求导。文中策略及算法已在多领域物理系统建模与仿真平台MWorks中实现。Many current object-oriented equation-based modeling languages use Differential Algebraic Equations （DAEs） to describe the physical phenomena and behavior of the modeled physical system. To support flexible and safe reuse of model components, high-index DAE problems are natural in object-oriented modelling. It is well known that it is numerically difficult to solve a high-index DAE problem. Index reduction methods can be used as a remedy. A new index reduction algorithm for large scale DAE systems was proposed. The method decomposed the complete DAE problem into a sequence of subproblems. For each DAE subproblem, the Pantelides＇s algorithm was used to identify the minimal subset of equations, differentiation of which is necessary for transforming them to a form that can be solved by general purpose numerical solvers. A symbolic differentiation algorithm based on binary tree is used to differentiate equations. The proposed strategies and algorithms have been successfully implemented in a modeling and simulation system, named MWorks.

关 键 词：微分代数方程 指标约简 符号微分 仿真 多领域建模 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]