Reliability based assessment of prestressed concrete bridgessubject to creep using a coupling procedure

摘要

In France, the vertical deflection measurements at the cantilever free end of arnbridge have shown unexpected excessive creep deformations. In order to understand the possiblernreasons for this phenomenon, we have studied the impact of two sources of uncertainties:rnmodel ignorance and simplifications, and statistical and measurment errors or uncertainties. In arnfirst step, we have constituted a large database by collecting creep testing information from thernmain research centers in Europe. These database experimental results are then compared withrnthe long-term deformation predicted by the design codes of practice: BPEL, CEB 78, CEB 90,rnEUROCODE 2 and ACI. For most of these codes, the creep deformation is, more or less, underestimated.rnThe error is quantified by statistical analysis to show how confidence can be given torndesign codes. In the second step, a coupling between the reliability software PHIMECA and thernfinite element tool ST1 is performed in order to analyze the bridge reliability with respect tornprediction of creep deformation. It shows the role of creep parameter variability and hence, justifiesrnpartly the gap between predicted and observed creep deformations. In this reliability assessment,rnthe EUROCODE 2 and the ACI give the best results compared with other codes ofrnpractice.