摘要
提出了一种结合高阶摄动和L_(1)正则化的梁结构静力随机损伤识别方法。该方法同时考虑不确定的初始模型和测量误差,建立关于单元损伤因子的随机控制方程。利用高阶摄动方法求解随机控制方程,并通过L_(1)正则化方法解决随机控制方程的不适定性问题。算例表明,该方法在提升计算效率的基础上获得了较高的损伤识别精度,且对测量噪声具有良好的鲁棒性。
An approach of stochastic static damage identification for beam structure combining high-order perturbation method and L_(1) regularization technique was proposed.The proposed approach considers not only the uncertain initial modelling errors but also the static measurement errors.Stochastic control equations about the element damage index are established,which are solved by the high-order perturbation method.The L_(1) regularization technique is used to solve the illposed problem of the stochastic control equations.The verified examples show that the proposed approach can obtain higher damage identification accuracy on the basis of improving the calculation efficiency,and the proposed approach also has certain anti-noise ability.
作者
卢博
吴志峰
黄斌
LU Bo;WU Zhi-feng;HUANG Bin(School of Civil Engineering and Architecture,Wuhan University of Technology,Wuhan 430070,China)
出处
《武汉理工大学学报》
CAS
北大核心
2020年第4期52-58,共7页
Journal of Wuhan University of Technology
基金
国家自然科学基金(51978545)。
关键词
随机损伤识别
梁结构
高阶摄动
L_(1)正则化
损伤概率
stochastic damage identification
beam structure
high-order perturbation
L_(1)regularization
damage probability
