Probabilistic Assessment of Deteriorating Prestressed Concrete Box-Girder Bridges under Increased Vehicle Loads and Aggressive Environment

Abstract

A probabilistic procedure for the assessment of the time-dependent reliability of existing prestressed concrete (PSC) box-girder bridges is presented. These bridges are subject to increased traffic loads and an aggressive environment, which result in structural deterioration such as cracking and corrosion. In this paper, a multiple-peak vehicle load model is proposed based on acquired traffic load data, which cannot be properly described by previous single-peak probability distributions. With the proposed vehicle load model, the maximal vehicle loads during the remaining life of bridges are obtained for evaluating the influence of the increase in traffic loads on bridge reliability. Time-dependent corrosion models are adopted to account for pitting corrosion because of chloride attack as well as uniform corrosion because of concrete carbonation. A degenerated shell element is used for accurate and efficient modeling of the PSC box-girder. Postcracking behaviors of reinforced concrete box-girders, simulated by using different crack models and elements, are compared with laboratory test results. To increase the efficiency of probabilistic analyses, an adaptive importance sampling method in conjunction with the truncated Latin hypercube sampling is used. Application of the proposed methodologies is demonstrated by a case study, in which the time-dependent reliabilities of a deteriorating PSC box-girder bridge are obtained.