基于Bayesian估计的整体和局部信息融合的损伤识别

在结构损伤识别中,如何充分利用整体和局部传感器测试得到的信息来增加结果的准确性是一个值得研究的问题。该文提出基于Bayesian 理论的结构整体局部信息融合的损伤识别方法。首先根据Bayesian理论建立了关于频率、位移模态和应变模态结构损伤的概率模型,静态应变信息提供了Bayesian理论的先验信息,使该概率模型充分利用了各种传感器信息;然后为了减少计算量,采用分步损伤识别的方法,在采用模态应变能指标初步定位损伤范围的基础上,用该文提出的逐个单元消去法定位损伤单元。最后对20 跨刚桁架模型进行试验研究证明了该方法的有效性,并且比较考虑与不考虑应变传感器信息的损伤识别结果。

DAMAGE IDENTIFICATION BASED ON BAYSIAN FUSION OF GLOBAL AND LOCAL INFORMATION

The damage identification using both global and local information is a possible method to increase accuracy of damage detection results.In this paper a new method based on Bayesian fusion of global and local information is proposed.The Bayesian probability model associated with frequencies,displacement mode shapes and strain modes are set up.The static strain information also is used as the prior information in the Bayesian model.In order to reduce computational cost of a complex structure,the strain energy damage index is firstly used to find the potential damage ranges.Then the exact damage elements are detected by the sequence elimination method proposed in the paper.The experiment on a 20-bay planar rigid truss structure is carried out to illustrate the damage detection ability of the proposed approach.