基于奇异值分解及特征正交分解的结构损伤检测方法  

Structure damage detection method based on singular value decomposition and proper orthogonal decomposition

作　　者：杨斌[1] 程军圣[1] 

机构地区：[1]湖南大学汽车车身先进设计制造国家重点实验室,长沙410082

出　　处：《振动与冲击》2014年第22期164-167,共4页Journal of Vibration and Shock

基　　金：国家自然科学基金(51075131);湖南省自然科学基金(11JJ2026)

摘　　要：为直接通过结构振动响应提取损伤特征参数,将奇异值分解和特征正交分解运用到结构响应分析中。该方法首先通过对结构响应功率谱矩阵的奇异值分解,获得模态频率,然后在模态频率处计算互相关矩阵,利用特征正交分解获得收敛于结构模态向量的特征正交模态,进而构建了损伤定位向量,最后通过结构单元应力的不同分布准确定位了损伤位置。实验数据分析结果表明,该方法能有效的进行损伤检测和定位。To extract the sensitive damage features directly from the structural vibration responses,the singular value decomposition and the proper orthogonal decomposition were applied to decompose the structural vibration response matrix into a set of proper orthogonal modes,which can represent the structure natural properties.The singular value decomposition was applied on the structural vibration power spectral matrix to figure out the modal frequencies,then at each modal frequency the correlation matrix was calculated.The proper orthogonal decomposition was applied on the correlation matrix calculated to get the proper orthogonal modes which can converge to the normal structural modes.The proper orthogonal modes were used to construct a damage locating vector,and to determine the damage location through inspecting the different stress distribution in each element.The experimental results show that the proposed method can detect and locate the damage effectively.

关 键 词：特征正交分解 奇异值分解 损伤定位向量 结构损伤检测 

分 类 号：TH113.1[机械工程—机械设计及理论]