Abstract
In order to limit construction conflicts in bridges with precast, prestressed concrete girders, it is essential that designers are able to predict midspan girder camber with sufficient accuracy at key stages. This paper describes a study of the effectiveness of various strategies for improving camber predictions when compared with field-measured values from actual bridge girders produced in the southeastern United States. The study also incorporated concrete materials data from nearly 2000 girder production cycles among four regional producers, in addition to a laboratory study of mechanical and time-dependent properties of representative concrete mixtures. A standard incremental time-step analysis software was developed and utilized for the parametric study included in this work. For the girder production cycles monitored in this study, the use of regionally calibrated prediction models for material properties (concrete compressive strength, modulus of elasticity, creep, and shrinkage) resulted in the elimination of approximately 80% of the prediction error associated with current camber prediction practices within the region. As compared with a mean overprediction error of 68% for current design practice, implementation of calibrated prediction models reduced the mean overprediction to approximately 10%. The most effective prediction improvement techniques were determined to be reliance on expected rather than specified concrete compressive strength, the use of an appropriate aggregate correction factor for modulus of elasticity computations, and the use of an incremental time-step analysis method incorporating AASHTO Load and Resistance Factor Design (LRFD) or International Federation for Structural Concrete (fib) Model Code 2010 creep and shrinkage prediction equations.
Get full access to this article
View all access options for this article.