Assessment and Detection of Damage in Structural Systems

摘要

This report presents a new damage detection and assessment algorithm based on a system identification method.The algorithm can be used as a non-destructive diagnostic technique to aid in decisions of repair and rehabilitationof existing structures. The developed algorithm is able to detect and assess damage in a structure whenthe measured data are sparse and polluted with noise.Damage is defined as the reduction in estimated parameters between two time separated inferences. Damageis localized by updating parameter groups hierarchically and assessed by perturbing measured data. Parametergroups are continuously subdivided until all the existing damaged parts in a structural system are completely localized.The measured data are perturbed to examine the sensitivity of each parameter with respect to the measureddata, and to thereby provide a statistical basis for assessing damage. The bias of the mean estimate from the baselinevalue and this bias divided by the standard deviation of the Monte Carlo perturbation sample are calculatedas damage indices. The upper limit values for both damage indices are determined from the simulation studyfor the undamaged baseline structure and are used to determine damage in the existing structure.Case studies are performed to examine the developed algorithm. Simulated static responses are used to identifydamaged planar bowstring truss structures. Real measured modal responses from a laboratory experiment areused to identify the cracked cantilever structure.