The long-term camber of prestressed bridge girders is typically over-estimated by current Iowa Department of Transportation (IA DOT) methods at erection (typically 3 month after production of girders), especially for long-span bulb tee girders. This often leads to increased costs due to the haunch modifications in the field, and unnecessary delay of construction. Creep and shrinkage of concrete play an important role in the long-term camber of a prestressed bridge girder. The current models used to predict the creep and shrinkage yield large disparties with the actual behavior of concrete in prestressed girders cast using local materials in Iowa. In order to improve the accuracy of prediction of the camber of prestressed bridge girders, creep and shrinkage tests of concrete using local materials were performed. Seven mixes from three precast plants were investigated in this study, in which four of them were high performance concrete (HPC) that are currently used to cast prestressed bridge girders, and three of them were normal concrete (NC) that were utilized to produce prestressed bridge girders previously. Mineral admixtures including slag and fly ash are typically added into HPC. Half of the creep and shrinkage specimens were sealed with Sikagard 62 to minimize the evaporation of water, and the rest were unsealed. All creep and shrinkage specimens with 4 in. diameter and 8 in. height were monitored in an environmentally controlled chamber for one year. In addition, twenty-six prestressed bridge girders produced using HPC from three precast plants were monitored and the corresponding long-term camber was measured.

It was observed that due to the early age of loading and the use of slag and fly ash HPC had higher average creep coefficient and average shrinkage strain than NC for both sealed and unsealed specimens during 1 year. It was also found that sealed specimens represent the creep and shrinkage behavior of a full scale prestressed bridge girder much better than unsealed specimens, in agreement with some of the previous literature. It was also observed that the sealed creep coefficient and sealed shrinkage strain measured from the four HPC mixes were similar, and it was acceptable to use the average sealed creep coefficient and average sealed shrinkage strain of the four HPC mixes tested to predict long-term camber of prestressed bridge girders produced in Iowa.

Three simplified methods were applied to predict long-term camber of the prestressed bridge girders, including Tadros's Method (2011), Naaman's Method (2004) and an incremental method. Naamans' Method and the incremental method yielded similar results, and both methods yielded 25% errors relative to measured camber of 26 prestressed bridge girders, but Tadros's Method yielded up to 50% errors. The calculation of Naaman's Method was simpler than for the incremental method. Therefore, Naaman's Method was the recommended method to predict the long-term camber of prestressed bridge girders produced in Iowa.