基于随机子空间的分布式结构模态参数识别  

Distribution modal identification based on stochastic subspace method

作　　者：张敏[1] 谢慧才[2] Sung-Han Sim B.F.Spencer Jr 

机构地区：[1]汕头大学工学院,汕头515063 [2]深圳大学土木工程学院,深圳518060 [3]伊利诺伊大学香槟分校工学院,美国香槟61801

出　　处：《深圳大学学报（理工版）》2009年第4期394-399,共6页Journal of Shenzhen University(Science and Engineering)

基　　金：国家自然科学基金资助项目(50678099)

摘　　要：针对大型建筑结构集中处理庞大数据获得结构模态参数的不便性,提出适用于密集布排的传感器网络结构的分布式模态参数识别方法.以混凝土钢管拱桥为实验平台,按不同子结构划分工况,通过随机子空间方法有效地从环境激励下的响应中提取子结构模态参数,结合稳定图进行子结构系统定阶,去除虚假模态.利用粒子群优化算法和平均技术调整子结构振型,获取桥梁结构的整体振型.以模态置信度为判据对比分析该分布式算法和集中式算法的识别结果.结果表明,该方法具有良好的识别效果,可用于不同形式复杂结构的模态振型识别.For large-scale civil structures ,the rich information can be harvested trom a dense array ot sensors, This is a big challenge for centralized processing technique. A distributed method to obtain modal parameters in smart sensor networks （SSN） with a dense array of sensors is proposed. Using a concrete-filled steel arch bridge as a testing model, Stochastic subspace identification is implemented to identify local modal parameters from ambient vibration response for the different subgroups cases. Stabilization diagram is a novel approach to define the rank of the system while get rid of noisy modes. Global mode shape is combined from rescaled local mode shapes through particle swarm optimization and average method. The results show that it enjoys a better efficiency according to modal assurance criterion （MAC）. This distributed method can be applied to any other complicated structures to determine modal parameters.

关 键 词：结构健康监测 传感器网络结构 随机子空间 粒子群优化算法 模态参数 

分 类 号：TU317[建筑科学—结构工程]