| 李雪艳,赖煜山,刘恩,刘荣林,赵卫.基于Volterra级数和递推最小二乘法的结构振动响应非线性特征和损伤识别[J].计算力学学报,2025,42(3):404~411,449 |
|
| 
码上扫一扫！ |
| 基于Volterra级数和递推最小二乘法的结构振动响应非线性特征和损伤识别 |
| Identification of nonlinear characteristics and structural damage from vibration response based on Volterra series and recursive least squares method |
| 投稿时间：2023-11-29 修订日期：2024-02-05 |
| DOI：10.7511/jslx20231129001 |
| 中文关键词: Volterra级数 递推最小二乘法 非线性特征识别 损伤识别 贝叶斯融合 |
| 英文关键词:Volterra series recursive least squares method nonlinear characteristic identification structural damage identification bayesian fusion |
| 基金项目:国家重点研发计划(2019YFC1511004-05);广东省重点领域研发计划(2019B111106001)资助项目. |
|
| 摘要点击次数: 8 |
| 全文下载次数: 5 |
| 中文摘要: |
| 当结构发生损伤时有可能引发结构的非线性振动,所以进行非线性识别可以检测结构损伤。而Volterra级数可以为非线性振动分析提供简明的分析模型,可以分离振动响应的线性和非线性成分,从而建立起更灵敏的非线性特征指标。在识别Volterra核时,需要输入激励响应和输出振动响应,是一个反问题求解问题。为了避免测试激励响应,本文使用不同测点的振动响应作为输入响应。为了避免矩阵求逆和病态求解问题,提出使用递推最小二乘法。研究了Volterra级数分离的各振动响应成分的特点,然后构建了非线性特征指标,进行非线性损伤位置的识别。为了消除激励位置和输入振动响应位置对非线性特征指标的影响,使用贝叶斯融合对最后的特征指标向量进行融合,最后能准确地识别出损伤位置。通过七层框架结构的多工况分析,验证了所提方法的有效性。 |
| 英文摘要: |
| When a structure is damaged,nonlinear vibration will be induced.So nonlinear identification can identify structural damage.A Volterra series can provide a concise analytical model for a nonlinear vibration analysis.Linear and nonlinear components of a vibration response can be separated by a Volterra series.Thus more sensitive nonlinear characteristic indicators can be established.When Volterra kernels are identified,it is an inverse problem that requires the input excitation response and the output vibration responses.To avoid the measurement of the excitation response,vibration response from different measurement points are proposed to be used as input responses.To avoid matrix inversion and ill-posedness,a recursive least-squares method is proposed.The characteristics of linear and nonlinear vibration response components separated by a Volterra series are studied.Nonlinear characteristic indicators are constructed for identifying nonlinear damage location.In order to eliminate the influence of the excitation location and the input vibration response location on nonlinear feature indicators,Bayesian fusion is used to fuse the final feature vectors,and the damage location can be accurately identified.The effectiveness of the proposed method is verified through several scenario analyses of a seven-story framework structure. |
| 查看全文 查看/发表评论 下载PDF阅读器 |
|
|
|
|
