Software comparisons and measurements of segmentally constructed bridges

Analysis of large segmentally constructed cantilever bridges is a complex task. It runs through all subsystems in the construction history, keeping track of time, dead weight, prestressing, traveling formwork, temporary columns, jacking of cantilevers, etc., and then finally through the service life of the bridge. This complexity makes results very hard to judge by hand, especially the time dependent behavior, and thus, the designer becomes to a very large extent dependent on the computer program. In this paper, results obtained by various software are compared for two large segmentally constructed bridges. The comparisons revealed unacceptably large differences in long-term girder displacements and column forces. The computer program DARC, that generally gave the largest time dependent effects, was however unable to track the magnitude of measured deflections beyond 15 years for a third bridge, but agreed well with rebar strain recordings over a limited period of 2.5 years in the columns of one of the others. This investigation raises three interrelated issues: (1) improved benchmark testing of this kind of software; (2) calibration of code models for creep and shrinkage to outdoor environment and loading conditions; and (3) need for instrumentation and monitoring of segmentally constructed bridges.