Comparison on Identification and Damage Detection Method Using Output-only Measurement: Application to Bridge Monitoring During Scouring Test

C.-H. LOH, W.-T. HSU, S.-F. CHEN, C.-K. CHAN

Abstract


The objective of this study was to assess the ability of using different multivariate signal processing algorithm on structural system identification of a time-varying system through vibration-based response measurement. First, the automatic modal analysis in cooperated with covariance-driven stochastic subspace identification (automatic SSI-COV) is used. This algorithm proposed 4 stages of analysis (criteria to remove spurious modes, clustering of the identified poles, grouping of similar modes, estimate modal parameters with considering uncertainty estimation) to enhance the automatic ability to identify the modal properties and remove the spurious modes of the monitoring structure. Second, the blind source separation (BSS) technique called second-order blind identification (SOBI) is also considered to identify the modal response and mixing matrix. Finally, the multivariate AR model (MV-AR) is also used to identify the time-varying model parameters. Through the response data collected from a bridge scouring test in the laboratory, discussion of these three identification methods on modal parameter estimation and damage detection of the bridge structure is made. The advances and disadvantages on model parameter estimation of these three methods are discussed.

doi: 10.12783/SHM2015/174


Full Text:

 Subscribers Only