Predicting Time-Dependent Deformations in Prestressed Concrete Girders

Abstract

A prestressed concrete girder camber prediction program, which utilizes a time-step approach, was improved to a new version including modifications, new features, and the fib Model Code 2010 material prediction model. The original version of the software was developed by Claire E. Schrantz in 2008. A few crucial corrections were included by Brandon R. Johnson in 2012. In the new version, a user is able to obtain predicted strains and stresses at designated depth, define actual and design compressive strengths, import variables from and export output data to a spreadsheet for further analysis. The camber prediction software includes four concrete modulus of elasticity (MOE) development models: two-point MOE model (uses MOE test results from two ages), AASHTO LRFD, ACI 209R-92, and fib MC 2010. Creep and shrinkage prediction models include AASHTO LRFD, ACI 209R-92, and fib MC 2010. Experimental data were collected from four previous research projects at Auburn University. This part of the study was used to verify the application’s capability of predicting time-dependent strain, curvature, and deflection. Measurements of strain and camber values were obtained from full-scale prestressed girders constructed with vibrated concrete (VC) or self-consolidating concrete (SCC). Camber and strain measurements of nineteen AASHTO BT-54 girders, fourteen AASHTO BT-72 girders, and six AASHTO Type I girders were collected. iii Measured camber data from twelve 15 in. deep T-Beams were also used in this study. Compressive strength of the concrete mixtures at 28 days varied between 6300 and 13,600 psi. Time-dependent properties and deflections were predicted by using the new version of the camber prediction software. The two-point MOE model in combination with various creep and shrinkage prediction models was chosen for predictions. Predictions were compared with the collected data from the fifty one girders to evaluate the accuracy of the creep and shrinkage prediction models. The comparison of time-dependent responses showed that camber growth, which is affected by creep, is overestimated for high-strength girders on average and include both over- and underpredictions for moderate-strength girders. ACI 209 predicts creep most accurately for high-strength girders, and AASHTO LRFD predicts creep most accurately for moderate-strength girders. Shrinkage predictions, which have an effect on prestress losses, are overestimated for high-strength girders and have mixed distribution of estimations for moderate-strength girders