Output-only damage localization technique using time series model

In this paper, we present a technique to detect the time instant and location of damage in civil structures using scalar time series models, by handling operational variability and measurement noise. The scalar Autoregressive (AR) and Autoregressive with exogenous inputs (ARX) models are used to obtain the time instant of damage and its spatial location. The spatial damage feature to locate the damage is obtained using a metric constructed from the probability density values of the prediction errors of AR–ARX model. The proposed method does not resort to any computationally expensive vector time series models to locate the damage and so highly preferable in smart wireless online continuous SHM schemes. Numerical simulation studies are carried out by using a simply supported beam model. The results of the studies indicate that the proposed technique is capable of identifying both the time instant and location of damage accurately using the proposed PDF based damage index. In order to validate the proposed technique with experimental results, the time-history data from the three-story bookshelf benchmark structure of EI-LANL is used. Finally, the laboratory experimental studies carried out on an RCC simply supported beam with inflicted damage are also presented. The experimental studies clearly indicate the effectiveness of the proposed damage index to detect the location of damage, by handling operational variability and measurement noise.