| description abstract | In the design of conventional steel and concrete highway girder bridges, the amount of live load carried by a single girder is quantified through distribution factors (DFs). However, moment DFs for the recently developed fiber-reinforced polymer (FRP) composite tub (CT) girders are not defined in current design codes, leaving questions as to these bridges’ live load distribution. To address this, two diagnostic live load tests were performed on an in-service CT girder bridge under heavy truck loads. The measured strains from each test were analyzed and compared with predictions from a calibrated, high-fidelity finite-element (FE) model, which was shown to accurately predict the live load distribution. Then, a simplified FE model was developed that was also shown to accurately predict the live load distribution at a significant computational discount. The simplified model was subsequently used in a parametric study to assess current CT girder design practices, which were found to be consistently conservative. One conclusion was that unintended end restraint of the girder significantly impacted the maximum measured flexural strains, and the high-fidelity FE model required modification and calibration to capture this effect. However, simplified FE models that were not calibrated to the field data were shown to reasonably accurately predict the live load distribution to the most heavily loaded, interior girders. Ultimately, the results of this study show that live load moment DFs should be developed that are specific to the CT girder to ensure efficient future designs with this new technology. | |
