Eigenvalue sensitivity of sampled time systems operating in closed loop
Eigenvalue sensitivity of sampled time systems operating in closed loop
复制标题
闭环运行的采样时间系统的特征值灵敏度
DOI：
10.1016/j.ymssp.2017.11.014
复制
发表时间：
2018
影响因子：
8.4
通讯作者：
Bernal, Dionisio
中科院分区：
文献类型：
--
作者：
Bernal, Dionisio
利用反馈建立具有高灵敏度的闭环特征结构在结构健康监测领域受到了一定的关注。虽然实际的实现必须是数字化的，因此在采样时间，工作迄今已集中在连续时间框架，无论是在设计和检查性能。它示出在本文中，在离散时间的性能，在典型的采样率，可以显着不同的预期在连续时间的配方和差异可以是特别大的特征值灵敏度的真实的部分;一个后果是重要的错误（线性估计）的损害程度，闭环稳定性丢失。正如人们所预料的那样，显式考虑采样率不会对闭环特征结构设计造成特殊的困难，并在本文中开发了相关的表达式，包括基于复摄动理论的矩阵指数导数的有效计算公式。本文提出了一个容易复制的数值例子，显示的误差水平，可以导致不考虑离散时间实现的控制器。
The use of feedback to create closed-loop eigenstructures with high sensitivity has received some attention in the Structural Health Monitoring field. Although practical implementation is necessarily digital, and thus in sampled time, work thus far has center on the continuous time framework, both in design and in checking performance. It is shown in this paper that the performance in discrete time, at typical sampling rates, can differ notably from that anticipated in the continuous time formulation and that discrepancies can be particularly large on the real part of the eigenvalue sensitivities; a consequence being important error on the (linear estimate) of the level of damage at which closed-loop stability is lost. As one anticipates, explicit consideration of the sampling rate poses no special difficulties in the closed-loop eigenstructure design and the relevant expressions are developed in the paper, including a formula for the efficient evaluation of the derivative of the matrix exponential based on the theory of complex perturbations. The paper presents an easily reproduced numerical example showing the level of error that can result when the discrete time implementation of the controller is not considered.
使用灵敏度增强控制和识别模型进行损伤识别
DOI：
--
发表时间：
2006
期刊：
影响因子：
--
作者：
J. A. Solbeck;L. Ray
通讯作者：
L. Ray
复杂扰动的特征值和特征向量的敏感性
影响因子：
2.1
作者：
D. Bernal
通讯作者：
D. Bernal