Computational Simulation of Chloride-Induced Corrosion Damage in Prestressed Concrete Bridge Girders

Prestressed concrete is a popular construction material for highway bridges. A variety of girder span values, cross-sectional shapes, and prestressing strand layouts has been used in bridges across the United States. A major concern for such bridges is the possibility of corrosion damage in the prestressing strands or reinforcing bars, which is commonly caused by the use of deicing salts on the deck or saltwater spray in coastal regions. The present study aims at establishing analytical tools for the accurate simulation of chloride ingress, corrosion and mechanical damage (cracking) in the concrete, and for the evaluation of the impact of corrosion on the flexural and shear strength of bridge girders. First, an efficient and accurate analytical scheme is formulated to enable the calculation of the load-carrying capacity of corrosion-damaged girders. The analyses rely on two types of models, namely, beam models and nonlinear truss models. The latter are deemed necessary to obtain reliable estimates of the shear capacity, as beam models are not well-tailored for capturing shear failures. A procedure to account for the reduction in area and deformability of corroded strands, based on visually observed corrosion damage, is proposed and implemented. The models are calibrated and validated with the results of experimental tests on prestressed girders which exhibited varying levels of corrosion damage. Further analyses allow the comparison of the capacity of corrosion-damaged girders to that of their undamaged counterparts. The accuracy of a simplified procedure, using equations in the AASHTO code to determine the flexural and shear capacity of the damaged girders, is also determined. Subsequently, a computation scheme was proposed to describe the intrusion of chloride ions in prestressed bridge girder sections. The approach accounts for multiple, coupled processes, i.e., heat transfer, moisture transport, and chloride advective and diffusive transport. The constitutive models for moisture and chloride transport rely on previous pertinent work, with several necessary enhancements. The modeling scheme is calibrated with data from previous experimental tests on concrete cylindrical and prismatic specimens. The calibrated models are then validated using data from chloride titration tests conducted on girders removed from two bridges in Virginia after 34 and 49 years of service. The results indicate that the proposed framework can accurately reproduce the experimentally measured chloride content. The modeling approach also allows the evaluation of the accuracy of simplified, design-oriented tools for estimating the evolution of chloride content with time. The multi-physics simulation scheme is further refined to account for the corrosion-induced mechanical damage (cracking), by incorporating a phenomenological description of the electrochemical reaction kinetics, generation of expansive corrosion products, and subsequent development of tensile stresses and cracking in the surrounding concrete. The impact of cracking on the chloride and moisture transport mechanisms is also taken into account. The last part of this dissertation pursues the quantification of the uncertainty governing the chloride ingress in bridge girders, through the use of a stochastic collocation approach. The focus is on understanding how the inherent uncertainty in the value of input parameters (e.g., material transport parameters, ambient conditions etc.) is propagated, leading to uncertainty in the evolution of chloride content and the expected corrosion initiation time for a given bridge.