项目名称: 大规模动力系统的模型降阶方法
项目编号: No.10801048
项目类型: 青年科学基金项目
立项/批准年度: 2009
项目学科: 金属学与金属工艺
项目作者: 林依勤
作者单位: 湖南科技学院
项目金额: 16万元
中文摘要: 大规模动力系统的来源非常广泛,包括电子电路,结构动力学和微电子力学系统等。大规模系统的模型降阶就是用一个相对很小的系统来近似最初的大规模系统。通过求解小系统,我们可以了解大系统的一些属性。上世纪末以来,模型降阶技术有了很大的发展。但是还有一些问题没有解决,例如基于Gramian 的模型降阶方法的实用性问题和奇异系统的HL2 范数最优模型降阶问题。在本项目中,我们准备解决这两个问题。此外,我们也考虑不精确矩阵-向量乘积和并行程序设计在模型降阶中的应用。
中文关键词: 动力系统;模型降阶;矩匹配;Gramian
英文摘要: Large-scale dynamical systems arise from various applications, including electronic circuits, structure dynamics and microelectromechanical systems. The problem of model reduction of a large-scale dynamical system is to find a system of much smaller size to approximate the original system. The properties of the large-scale dynamical system can be captured by solving the reduced system. Although the order-reduced techniques have been improved greatly since the end of the last century, we still need to resolve some problems, such as the applicability of the model-order reduction methods based on the system's Gramians and the optimal HL2 norm model-order reduction of singular systems. We will consider these problems in this project. Moreover, we will also investigate how to apply the inexact matrix-vector product technique and the theory of parallel programming to model reduction of large-scale dynamical systems.
英文关键词: Dynamical system; Model-order reduction; Moment matching; Gramian