Probabilistic Models for the Tensile Strength of Corroding Strands in Posttensioned Segmental Concrete Bridges

The presence of air voids, moisture, and chlorides inside tendon systems on segmental posttensioned (PT) bridges has been cited as a reason for the early age corrosion and failure of strands in these bridges. This paper develops probabilistic models to predict the time-variant tension capacity of PT strands exposed to wet-dry conditions. A total of 384 unstressed and 162 stressed strand test specimens were exposed to various void, moisture, and chloride conditions for 0, 12, 16, and 21 months; the residual tension capacities of the strands were then determined. Using these experimental data, a Bayesian approach is used to develop probabilistic capacity models for unstressed and stressed strands. The tension capacities of stressed strands under potential void, wet-dry, and chloride conditions in the field are predicted using the developed models. Probabilistic time-variant models are formulated in such a way that they can be updated by other researchers using additional information from the testing of unstressed strands only, avoiding expensive and cumbersome testing of stressed strands. The mean absolute percentage errors of these models are less than 3.2%, indicating good overall model accuracy.