Reliability-based evaluation of the performance of the damage locating vector method

Abstract

Structural health monitoring (SHM) has the potential to be an effective way to maintain and improve the public safety of a nation. In recent years, numerous SHM methods have been developed by various researchers. Among these, the damage locating vector (DLV) method proposed by Bernal has received significant attention because of demonstrated efficacy when implemented on a wireless sensor network. However, the DLV method is a model-based approach requiring a numerical model of the structure. This numerical model cannot perfectly represent the physical structure due to intrinsic model idealization and environmental uncertainty. In this paper, the performance of the DLV method in the presence of model uncertainty is examined using a system reliability analysis by using the first order reliability method (FORM) and compared with Monte Carlo simulation. In addition, the validity of the threshold value, that is critical but empirically determined in previous research, is also examined.