Stochastic output error vibration-based damage detection and assessment in structures under earthquake excitation

• Published in: Journal of sound and vibration 2006-11, Vol.297 (3), p.1048-1067
• Main Authors: Sakellariou, J.S., Fassois, S.D.
• Format: Article
• Language: English
• Subjects: Acoustics; Computer simulation; Damage detection; Earthquake damage; Exact sciences and technology; Excitation; Fracture mechanics (crack, fatigue, damage...); Fundamental areas of phenomenology (including applications); Measurement and testing methods; Methodology; Physics; Seismic phenomena; Solid mechanics; Stochasticity; Structural acoustics and vibration; Structural and continuum mechanics; Vibration; Vibration, mechanical wave, dynamic stability (aeroelasticity, vibration control...)
• ISSN: 0022-460X; 1095-8568
• DOI: 10.1016/j.jsv.2006.05.009

A stochastic output error (OE) vibration-based methodology for damage detection and assessment (localization and quantification) in structures under earthquake excitation is introduced. The methodology is intended for assessing the state of a structure following potential damage occurrence by exploiting vibration signal measurements produced by low-level earthquake excitations. It is based upon (a) stochastic OE model identification, (b) statistical hypothesis testing procedures for damage detection, and (c) a geometric method (GM) for damage assessment. The methodology's advantages include the effective use of the non-stationary and limited duration earthquake excitation, the handling of stochastic uncertainties, the tackling of the damage localization and quantification subproblems, the use of “small” size, simple and partial (in both the spatial and frequency bandwidth senses) identified OE-type models, and the use of a minimal number of measured vibration signals. Its feasibility and effectiveness are assessed via Monte Carlo experiments employing a simple simulation model of a 6 storey building. It is demonstrated that damage levels of 5 % and 20 % reduction in a storey's stiffness characteristics may be properly detected and assessed using noise-corrupted vibration signals.