Long-term deformation of balanced cantilever bridges due to non-uniform creep and shrinkage

Abstract: Balanced cantilever bridges have historically experienced excessive deformations. Previous researchsuggeststhat the cause may be due to differential thickness in the box girder cross-section and underestimation of creep and shrinkage.In this project, the long-term deformationof balanced cantilever bridges due tonon-uniformcreep and shrinkage have been investigated. The non-uniformcreep and shrinkage arecaused by variations in drying rates for the different parts of the box-girder cross-sections.A finite element model was createdintheprogram Abaqusas a case study of the Alvik bridge.The finite element model was used to evaluate the difference betweennon-uniform and uniform creep and shrinkage with Eurocode 2.Further, a comparison between Eurocode 2 and Bažant’sB4 modelwas conductedfor non-uniform creep and shrinkage. The comparison aimedto evaluate the difference between industry and research specific calculation models, forthe effect of creep and shrinkage on deformations.A parameter study was alsoconducted to discern theeffect of parameters: ballast load, water-cementratio and conditions related to drying of concrete (relative humidity and perimeter exposed to air).Acomparison withthe deformationmeasurementsof theAlvik bridge was conductedto validate the resultsfrom the model.The results showed that there was a significant difference in the calculateddeformationof the bridge during the first ten years between analyses based onnon-uniform and uniformdistribution of creepand shrinkage,respectively.The non-uniformanalysis gave largerdeformations.However, only minor differences between the two approachescould be detected in the final deformation after 120 years. The main reason for the differences in the early behaviour is primarily caused by the differences in shrinkage rate between the top and bottom flanges. In these analyses, the top flange was assumed tonotdry out from the top. Thereby, the shrinkage rate of the top flange caused by one-way drying was similar to the bottom flange that was assumed to be exposed for two-waydrying.TheB4 model gave larger deformations compared to Eurocode2.This may be due to difference in the definition ofperimeter and surface. Eurocode 2 considers the perimeter exposed to air. The B4 model instead considers the entire surface area of the part.TheB4 model and Eurocode 2 show similar results asthe measurements. However, the B4 model gaveresults more consistent with the measurements.In the parameter study,lowerrelative humidity gave smaller deformations, since concrete shrinksquicker in dry ambient air.Varying the water-cement ratiodid not affect the deformationsnoticeably.Higher ballastheight gave significantly larger deformations. The height of the ballast was an uncertainfactor due to varying heights in the structural drawings of the case study. Accurate height of ballast is therefore important.