基于移动主成分法的桥梁损伤识别研究

Damage Detection of Bridge Based on Moving Principal Component Analysis

本文在多元统计分析方法——主成分法(PCA)局限性的基础上,提出了移动主成分法(MPCA)的桥梁损伤识别新方法,此方法借助桥梁结构实际运营过程中动力响应信号实现在线实时监测并能对其健康状况进行快速准确评估。文中对广州海珠桥模型不同工况进行数值模拟及现场实验研究,分析了结构不同损伤程度下的加速度信号及位移信号对此方法的影响规律,结果证明了其在复杂结构中的可行性,为同类型桥梁结构损伤识别提供借鉴。

Based on the limitation of principal component analysis （PCA） , i.e. a typical statistical analysis in damage identification of bridges, moving principal component analysis in this paper, which realizes online real-time monitoring and fast and accurate evaluation of method of multivariate （MPCA） is proposed the health of the bridge by colleting dynamic response signals during its the actual operation. This method can be able to achieve monitoring online and evaluation about structural health condition by means of the dynamic response of bridge structure in the work. In this paper, the numerical simulation and. field experiment of the Haizhu Bridge model in Guangzhou are carried out. The influence of the acceleration and the displacement signals under different degree of the damage is analyzed, which demonstrates the feasibility of MPCA in the complex structure and provide further reference for damage identification of similar bridges.