Singular value diagnosis in dam safety monitoring effect values

• Published in: Science China Technological Sciences Vol. 54; no. 5; pp. 1169 - 1176
• Main Authors: Gu, ChongShi, Zhao, ErFeng, Jin, Yi, Su, HuaiZhi
• Format: Journal Article
• Language: English
• Published: Heidelberg SP Science China Press 01.05.2011
• Subjects: Engineering; 主成分分析; 信息矩阵; 大坝安全监测; 奇异值; 效应值; 监测诊断; 能量百分比; 融合估计; optimal estimation; singular value diagnosis; dam safety monitoring; principal component analysis; effect
• ISSN: 1674-7321; 1869-1900; 1862-281X
• DOI: 10.1007/s11431-011-4339-7

Based on the principal component analysis, principal components that have major influence on data variance are determined by the energy percentage method according to the correlation between monitoring effects. Then principal components are extracted through reconstructing multi effects. Moreover, combining with the optimal estimation theory, the method of singular value diagnosis in dam safety monitoring effect values is proposed. After dam monitoring information matrix is obtained, single effect state estimation matrix and multi effect fusion estimation matrix are constructed to make diagnosis on singular values to reduce false alarm rate. And the diagnosis index is calculated by PCA. These methods have already been applied to an actual project and the result shows the ability of the monitoring effect reflecting dam evolution behavior is improved as dam safety monitoring effect fusion estimation can take accurate identification on singular values and achieve data reduction, filter out noise and lower false alarm rate effectively.