UC San Diego

Post-tensioned cast-in-place concrete box girder bridges are the most popular type for new bridge construction in California since this class of bridges shows an increased ability to resist and dissipate seismic loads for long- span structures. However, due to the long-term behavior of the construction materials .i.e. concrete and steel, tension forces induced by prestressing decrease over time as a result of creep and shrinkage of concrete and steel relaxation, which is called long-term prestress loss. This loss is one of the most important factors to consider in designing and analyzing this type of bridges. Although the inaccurate prediction of the loss is not considered to impose severe effects on the ultimate capacity of the structures, it is known to lead to critical problems at the design and service stages. Underestimating the prestress loss can cause cracking and excessive deflection during service stage and overestimating it could lead to uneconomic design. A number of bridge specifications use varied approximations currently to predict the long-term prestress losses. However, they show significant scatter in values of their estimates even under the same environmental conditions. This diversity of prediction mainly results from two sources of error. The first source of the error is related to inaccurate material characteristics: the coefficients for creep and shrinkage of concrete. The other comes from excessive simplification in the method of analysis in the specifications. Consequently, the goals of this research are to verify which specifications can provide more accurate creep and shrinkage strains and to propose a simple yet more comprehensive analysis procedure to predict the long-term prestress losses including effects from important structural- and material-level parameters. To meet these objectives, monitoring responses from bridges in service and investigating material characteristics from cylinder specimens are in progress on four bridge spans of I5-I805 and I215-CA91 in San Diego and Riverside, California, respectively. Also, a simple analysis method which could be easily applied in the design is improved by incorporating the effects from the five parameters: transverse shear deformation, loading events, and horizontal reinforcements as structural-level parameters, compressive strength of concrete and thermal creep from accumulation of temperature variation as material-level parameters. The results from this research are validated by comparing with the predictions in current specifications and data from the monitored spans on I5- I805 and I215-CA91.