摘要
数据驱动随机子空间法是一种常用的基于环境激励的结构模态参数识别方法。该方法通过QR分解计算“未来”输出数据对“过去”输出数据的投影矩阵,并利用奇异值分解(SVD)对投影矩阵进行进一步处理,以获取包含模态参数信息的状态矩阵。然而,QR分解和SVD分解的计算量较大,导致计算时间较长。为提高计算效率,分别引入了经济型QR分解和经济型特征分解,替代了传统的QR分解和SVD分解。以三自由度质量-弹簧-阻尼模型为例进行数值分析,结果表明:采用本文方法与传统方法得到的计算结果一致,但计算时间大幅减少。本文方法有助于增强数据驱动随机子空间法在结构模态参数识别中的实时性。
The data-driven stochastic subspace method is a commonly used ambient excitation-based structural modal parameter identification method.This method calculates the projection matrix of future output data onto past output data through QR decomposition,and further processes the projection matrix using singular value decomposition(SVD)to obtain the state matrix containing modal parameter information.However,the computational cost of QR and SVD decomposition is relatively high,resulting in long computation time.To improve computational efficiency,this paper introduces the economical QR decomposition and economical eigenvalue decomposition to replace the traditional QR and SVD decomposition.Numerical analysis is conducted using a three-degree-of-freedom mass-spring-damper model.The results indicate that the computed results obtained using the proposed method are consistent with those by the traditional method,but with significantly reduced computation time.The proposed method helps enhance the real-time performance of the data-driven stochastic subspace method in structural modal parameter identification.
作者
金宇文
伍彩
马俊
陈剑毅
朱道佩
JIN Yuwen;WU Cai;MA Jun;CHEN Jianyi;ZHU Daopei(School of Software Engineering,Jiangxi University of Science and Technology,Nanchang 330013,China;School of Civil Engineering,Hubei Engineering University,Xiaogan 432000,China)
出处
《贵州科学》
2025年第5期81-85,共5页
Guizhou Science
基金
江西省自然科学基金资助项目(20232BAB214074)
湖北省自然科学基金项目(2024AFB434)。
关键词
结构健康监测
模态识别
随机子空间法
模态参数
环境激励
structure health monitoring
modal identification
stochastic subspace method
modal parameters
ambient excitations
