Principal Component Analysis Method with Space and Time Windows for Damage Detection
Principal Component Analysis Method with Space and Time Windows for Damage Detection
复制标题
用于损伤检测的时空窗主成分分析方法
DOI：
10.3390/s19112521
复制
发表时间：
2019
期刊：
影响因子：
--
通讯作者：
Jiang Zhenyu
中科院分区：
文献类型：
--
作者：
Zhang Ge;Tang Liqun;Zhou Licheng;Liu Zejia;Liu Yiping;Jiang Zhenyu
长期结构健康监测已成为保证基础设施安全的重要手段。然而，确定从SHM系统的大量数据中提取有价值的信息以有效识别损伤的方法仍然是一个重大挑战。在传统的主成分分析（PCA）方法中引入空间和时间窗，为结构损伤检测提供了一种新的有效方法。对一个平面梁模型的数值结果表明，由于空间和时间窗的存在，所提出的双窗PCA方法（DWPCA）比以前的移动PCA（MPCA）方法（仅结合时间窗和PCA）具有更高的损伤识别灵敏度。进一步的研究表明，与MPCA方法相比，该方法具有更高的分辨率，在定位损伤的空间窗口，也在定量评估损伤的严重程度。最后，以一座实际桥梁的有限元模型为例，验证了DWPCA方法比传统方法具有更高的损伤检测灵敏度，具有一定的工程应用潜力。
Long-term structural health monitoring (SHM) has become an important tool to ensure the safety of infrastructures. However, determining methods to extract valuable information from large amounts of data from SHM systems for effective identification of damage still remains a major challenge. This paper provides a novel effective method for structural damage detection by introduction of space and time windows in the traditional principal component analysis (PCA) technique. Numerical results with a planar beam model demonstrate that, due to the presence of space and time windows, the proposed double-window PCA method (DWPCA) has a higher sensitivity for damage identification than the previous method moving PCA (MPCA), which combines only time windows with PCA. Further studies indicate that the developed approach, as compared to the MPCA method, has a higher resolution in localizing damage by space windows and also in quantitative evaluation of damage severity. Finally, a finite-element model of a practical bridge is used to prove that the proposed DWPCA method has greater sensitivity for damage detection than traditional methods and potential for applications in practical engineering.