Localization of damage in a bridge using measured response signals

초록

A damage detection and localization algorithm is proposed using a single-inputsingle-output (SISO) ARX model that provides a correlation between input and output signals. The input signal is the measured acceleration at the impact location and the output is also the measured acceleration at a place of the other accelerometers uniformly distributed in a structure. To detect and locate damage, it is assumed that a baseline statistical property for an intact condition is provided as a priori known information. By comparing the statistical properties of the current condition with those of the baseline, damage can be detected and localized. Damage indices using the statistical properties are proposed and their efficiencies are compared. A laboratory experiment on a steel plate girder bridge model has been carried to examine the usefulness of the proposed algorithm. The results are compared with that from an available time-domain algorithm.