An Experimental Study on Damage Detection of Concrete Structures Using Decentralized Algorithms

Abstract

In this paper, an experimental study has been carried out to detect damage on a simply supported two-span reinforced concrete slab. Different crack damages are created by static loads on the slab and impact tests are carried out before and after removing the static loads. Two decentralized damage detection methods – Auto Correlation Function-Cross Correlation Function (ACF-CCF) method and Auto Regressive-Auto Regressive with exogenous input (AR-ARX) method, are used to localize damage from measured responses. The accuracy and sensitivity as well as the effect of sensor location and loading status of the structure were analysed with these two methods. The results show that the ACF-CCF method is more effective in detecting and locating damage than the AR-ARX method. The Novelty Index value of the ACF-CCF method could be a reliable indicator of damage in concrete structures.