ABSTRACT
Concrete box culverts facilitate drainage, railway/road crossing, or subway for transportation networks. Health assessment of box-culverts when distressed, or near end of design life, is generally conducted with non-destructive and/or partially destructive tests (NDT/PDT). The selection of NDT/PDT locations for box-culverts need to consider the inherent low redundancy. In this study, the failure probability of a box-culvert is first estimated using prior knowledge, incorporating the uncertainties in design load/strength, updated later through a Bayesian pre-posterior analysis. The testing expenses and repair costs are associated with the decisions in a value of information (VoI) framework. Proposed approach to select the best possible locations is demonstrated for a square box-culvert by minimizing overall cost. A set of parametric studies have been shown to conclude upon the parameters that may influence the life cycle cost after decision making. Further, the importance of optimizing sampling location has been emphasized with respect to cost of decision making.
