快速迭代收缩阈值正则化改进算法在桥梁移动荷载识别中的应用研究  

Application of Iterative Regularization Improved Fast Iterative Shrinkage Threshold Algorithm in Bridge Moving Force Identification

作　　者：陈震 孙梦晴 陈璐 郭丰 李晓克 CHEN Zhen;SUN Mengqing;CHEN Lu;GUO Feng;LI Xiaoke(School of Civil Engineering and Communication,North China University of Water Resources and Electric Power,Zhengzhou 450045,China;China Construction Seventh Engineering Division Co.,Ltd.,Zhengzhou 450004,China)

机构地区：[1]华北水利水电大学土木与交通学院,河南郑州450045 [2]中国建筑第七工程局有限公司,河南郑州450004

出　　处：《华北水利水电大学学报（自然科学版）》2023年第1期79-83,92,共6页Journal of North China University of Water Resources and Electric Power：Natural Science Edition

基　　金：国家重点研发计划项目(2017YFC0703900);国家自然科学基金项目(U2004184);河南省高等学校青年骨干教师培养计划项目(2021GGJS078);华北水利水电大学2021年度大学生创新创业项目(2021XB256)。

摘　　要：移动荷载识别是结构动力学反问题研究的重要组成部分,其关键在于精确求解车-桥系统方程离散化后得到的大型稀疏线性方程组。由于车-桥系统矩阵同时具有稀疏性、非对称性、非正定性等特点,常用的迭代算法通常难以准确高效地对该大型稀疏线性方程组进行求解。迭代收缩阈值算法源于经典梯度算法,被广泛应用于线性逆问题求解及压缩感知重构算法中。由于迭代收缩阈值算法自身的非渐进全局收敛差的缺陷,导致荷载识别结果与真实值存在一定程度的偏差,难以实现对动态荷载的完全稀疏表达。为提高迭代收缩阈值算法的收敛性与稳定性,提出一种快速迭代收缩阈值正则化改进算法,将其用于识别桥梁移动荷载。通过引入迭代正则化,改进算法可准确提取动态荷载的信号特征,进而提高移动荷载识别方法的识别精度和抗噪性能。Moving force identification is an important part of the inverse problem of structural dynamics,and the key of which is to accurately solve the large sparse linear equation obtained by discretization vehicle-bridge system equation.Due to the sparsity,asymmetry,non-positive definiteness characteristics of vehicle-bridge system matrix,commonly used algorithms are usually difficult to accurately and efficiently solve the sparse linear equations.Iterative shrinkage threshold algorithm is derived from classical gradient algorithm,which is widely used in linear inverse problem solving and compressed sensing reconstruction algorithm.Due to non-asymptotic global convergence defect of the iterative shrinkage threshold algorithm,identification results of the iterative shrinkage threshold algorithm often deviate from the true force,and it is difficult to achieve the complete sparse expression of dynamic load.To improve the convergence and stability of the iterative shrink threshold algorithm,an iterative regularization improved fast iterative shrinkage threshold algorithm is proposed to identify moving force on bridge.By introducing iterative regularization,the signal characteristics of dynamic loads can be accurately extracted and then improve the identification accuracy and robustness performance of the novel method.

关 键 词：移动荷载识别 迭代收缩阈值 正则化改进算法 识别精度 

分 类 号：TU311.3[建筑科学—结构工程]