基于EMD处理信号的桥梁颤振导数识别  

IDENTIFICATION OF BRIDGE FLUTTER DERIVATIVES BASED ON PROCESSED SIGNAL BY EMD

作　　者：许福友[1] 陈艾荣[2] 刘剑锋[2] 王达磊[2] 

机构地区：[1]大连理工大学土木水利学院,大连116024 [2]同济大学土木工程防灾国家重点实验室,上海200092

出　　处：《振动与冲击》2008年第5期144-147,156,共5页Journal of Vibration and Shock

基　　金：国家自然科学基金资助项目(50478109,50708012);高等学校博士点专项科研基金(20040247026)

摘　　要：桥梁风洞试验采集信号都不可避免地被噪声污染,并且具有一定的非平稳性,因此导致桥梁颤振导数识别精度降低。为了提高识别精度,基于经验模态分方法分离信号固有模态函数,消除了高频噪声和低频非平稳趋势项。基于原始采集信号及由经验模态分解处理后的信号采用随机搜索方法识别了苏通大桥节段模型颤振导数,两种情况下的识别结果差异不容忽略。分析结果表明,经验模态分解是提高桥梁颤振导数识别精度的一种行之有效的方法。The signals collected from bridge wind tunnel tests are inevitably contaminated by noises, and appear certain nonstationarity, which result in precision reduction for bridge flutter derivatives identification. In order to improve the accuracy, the intrinsic mode functions of such signals are sifted using the empirical mode decomposition (EMD), by which the high-frequency noises and low-frequency nonstationary trends are eliminated. Flutter derivatives of Sutong Bridge section model are identified using stochastic search technique based on the original collected signals and the processed ones by EMD. The difference between two sets of results is not ignorable. The analytical results indicate that EMD is a valid method for improving the identification accuracy of flutter derivatives of bridges.

关 键 词：经验模态分解方法 颤振导数 参数识别 苏通大桥 

分 类 号：V215.34[航空宇航科学与技术—航空宇航推进理论与工程]