基于AR模型的分布式损伤诊断方法研究

Research on the Distributed Damage Identification based on AR Model

现今结构健康监测中采用的损伤诊断方法大多仅能在程度上对损伤进行相对判断,不能准确的判断损伤产生的数值。从适用于剪切型框架结构的AR模型损伤识别方法出发进行研究。首先,建立结构响应信号的AR模型,通过提取AR模型的自回归系数达到结构刚度识别的目的,以刚度值作为判定结构损伤程度的定量指标;最后为证明上述方法的可行性,利用ASCE Benchmark仿真模型进行了数值模拟。结果表明:该方法能够精确的定位损伤及进行损伤程度判断。

Most of the damage identification methods used in structural health monitoring today can only judge the damage in degree,which can't accurately judge the value of damage.This paper studies AR model damage identification method applicable to shear frame structure.First of all,the structure response signal of the AR model is set up,by extracting the AR model of regression coefficient to achieve the structural stiffness,which is used as the quantitative index to determine the degree of structural damage.Finally,the simulation model of ASCE Benchmark is used to prove the feasibility of the above method.The results show that this method can accurately locate damage and damage degree.