Abstract
Excessive midspan deflection is often observed in large-span prestressed concrete girder bridges. In particular, the deformation caused by shrinkage and creep of concrete is an important part of deflection that often exceeds expectations. To achieve creep deformation control, extending the curing time and postponing the loading age are frequently adopted during construction. This article proposed an approach based on creep experiments with prismatic and beam specimens and viscoelastic model-based finite element analyses to evaluate the effect of the concrete strength development factor on the ultimate creep coefficient in the Comité Euro-International du Béton-Fédération internationale de la précontrainte (CEB-FIP) model. The concrete strength development factor ξcc(t/tu) is defined as the ratio of the mean compressive strength at the loading time to the maximum mean compressive strength obtained in the test. The strategies for targeted regulation and control of ξcc(t/tu) were discussed by establishing an artificial neural networks model for strength prediction from a database provided in our previous work. Uniaxial compressive prismatic specimens and pure bending beam specimens with a span of 5.0 m were used in the experiment for creep behavior observation. A viscoelastic finite element analysis (FEA) model was established based on solidification theory. The FEA model results were verified with measured data for capturing the creep behavior under ambient conditions to obtain reliable long-term creep deformation predictions. Finally, the proposed method was validated with the Xincheng bridge construction for the optimistic loading age determination, and the results indicated good feasibility in girder deflection control.