Application of real-time adaptive identification technique on damage detection and structural health monitoring

Abstract

The final collapse of building structures under severe earthquake excitation will undergo stiffness degradation, which may be revealed by continuously monitoring and investigating the response of the specified component. Based on the on-line parameter-estimation feature of model-reference adaptive structural control algorithm, a real-time model-reference adaptive identification technique (MRAIT) is proposed in this study. The adaptation law for parameter estimation is designed based on Lyapunov's direct method. An energy-like candidate function is defined by composing of weighted response-tracking error and weighted parameter-estimation error. An artificial adaptive signal is calculated from both the simulated states of predesigned model and the adjustable parameters estimated by MRAIT. With the help of the proposed MRAIT under some minor earthquakes, the time-invariant dynamic properties of target structure can be identified step by step by modifying the predesigned model according to the estimated adjustable parameters. The onset of yielding or the degradation on stiffness of an elastoplastic model can be revealed if the proposed MRAIT is equipped and monitored continuously when subjected to severe earthquakes. The failure patterns of hysteretic-type structures undergoing inelastic behavior can be recognized either by the monitored artificial adaptive signal or by the variation of tracking error calculated on-line. Therefore, the onset of damage can be assessed based the proposed real-time MRAIT scheme. Copyright © 2009 John Wiley & Sons, Ltd.