Premium
The hybrid multivariate analysis method for damage detection
Author(s) -
Yang ZhiBo,
Chen XueFeng,
Xie Yong,
Zhang XingWu
Publication year - 2016
Publication title -
structural control and health monitoring
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.587
H-Index - 62
eISSN - 1545-2263
pISSN - 1545-2255
DOI - 10.1002/stc.1758
Subject(s) - robustness (evolution) , multivariate statistics , computer science , fuzzy logic , curvature , artificial intelligence , structural engineering , engineering , algorithm , pattern recognition (psychology) , mathematics , machine learning , biochemistry , chemistry , geometry , gene
Summary As the defects and superiorities of indices are mutualisms frequently, such as noise immunity and damage sensitivity, damage identification based on single damage index may hardly present the effective result all the time, so multiple indices fusion method is introduced in this paper to achieve some complementary improvements. In this paper, two kinds of no‐baseline mode shape‐based damage indices, namely, the generalized local entropy and the curvature waveform capacity fractural, are utilized to construct the basic index set for combination, and the fuzzy cluster method is introduced in order to establish fusion process. These two parts generate the hybrid multivariate analysis method finally. The multivariate analysis' superiority mainly displays on two aspects: (1) the use of no‐baseline mode shape‐based method ensures the damage detection efficiency with the absence of healthy mode shape serving as baseline and (2) the fusion conducted by cluster method provides the mutual support and complementation among indices, which can enhance the robustness of algorithm. The performances of the present method are verified via sufficient numerical examples, and then experiments are demonstrated on three typical engineering structures, namely, cantilever beam, blower wheel, and rotor, for further validations. Copyright © 2015 John Wiley & Sons, Ltd.