摘要
提出了一种结合摄动法和L1正则化方法的随机梁式结构静力损伤识别方法。考虑初始模型误差和测量误差的影响,建立了关于随机损伤指数的控制方程,并将摄动法和L1正则化方法相结合,对随机损伤指数的控制方程进行求解,进而从概率的角度对结构的损伤进行识别。损伤试验结果表明,和传统的最小二乘求解法相比,本文方法能够更为准确地识别多处局部损伤的位置及大小,对实际结构损伤检测具有较好的参考价值。
This paper proposes a new static damage identification method of random beam structures based on a perturbation technique and L1regularization method.Considering the impact of model uncertainty and measurement errors on damage identification,the control equations with respect to the random damage index of beam structures are established.Using the perturbation technique and L1 regularization,the solutions of the random control equations are determined.In this manner,the probability of structural damage can be achieved.From the test results,it is observed that the accuracy of the proposed method is higher than the least squares method to identify the locations and magnitudes of multiple local damages.The proposed method can be used in the damage detection of actual structures in the future.
作者
黄斌
鲁溢
HUANG Bin;LU Yi(School of Civil Engineering and Architecture,Wuhan University of Technology,Wuhan 430070,China)
出处
《计算力学学报》
EI
CAS
CSCD
北大核心
2020年第1期69-74,共6页
Chinese Journal of Computational Mechanics
基金
国家自然科学基金(51578431)资助项目.
关键词
随机梁式结构
损伤识别
测量数据
L1正则化
损伤概率
random beam structure
damage identification
measurement data
L1regularization
damage probability