Evaluation of Load Carrying Capacity of Bridges Using Enhanced Methods of Analysis

In this project, enhanced methods of analysis were used to demonstrate how higher load carrying capacities can be proven through detailed assessment of bridges in the Swedish road network. Probabilistic analysis, finite element (FE) analysis and alternative design methods, and combinations of these, were described and used together with results from previously performed measurements. The focus of the project has been the application on prestressed concrete box girder bridges. Two bridges owned by the Swedish National Road Administration (Vägverket) were used as case studies: The Källösund Bridge, built with the free cantilevering method as a part of the Tjörn Bridge connection on the west coast of Sweden, and a continuous three-span bridge at a traffic interchange in Kropp close to Helsingborg in the south of Sweden. An assessment strategy for enhanced evaluation of the load carrying capacity of existing bridges is proposed. As a basis for such an evaluation, an initial structural assessment according to current practice should be performed, but with higher demands on a detailed documentation. The enhanced evaluation should thereafter be made step by step and as an integrated part of the decision process. The detailed assessment involves more advanced analysis and judgement, that are not ruled by the detailed regulations and code of practice used in common structural assessment. However, it should have the same aim, intention and safety requirements. It can be based on research results and more enhanced methods for determination of the capacity, and more advanced analysis models of the bridge can be used together with bridge-specific load and material data. Calculations and analyses are made in a continuous interaction with physical investigations of the condition of the bridge, and decisions whether to proceed with the assessment are made successively. For the Källösund Bridge, the evaluation made showed that the load carrying capacity with respect to bending moments was sufficient for a vehicle bogie load of 210 kN according to the Swedish code for road bridge assessment, see Vägverket (1998). However, the capacity with respect to shear and torsion was too small for the same load level, with the existing shear reinforcement. This was still the case when load effects determined through probabilistic evaluation were used, even though they were found to be 9% lower than the load effects determined from the deterministic code loads. An evaluation based solely on the Swedish concrete code, BBK 94 (1994), gives sufficient safety in the latter case, but, on the other hand, the other methods used (EC 2-1 and the modified compression field theory) give a too low capacity. A FE analysis performed resulted in sufficient load carrying capacity for the critical load case with respect to shear and torsion, but non-linear FE analysis of shear failures need to be studied further before the results of such an analysis can be fully adopted. Strengthening of the Källösund Bridge is, based on the evaluation performed, considered to be the safest alternative. In case of further evaluation, an interesting alternative would be to leave the bridge unstrengthened and to expose it to a test loading with successively increasing vehicle loads up to the maximum load level the bridge has been exposed to earlier. Relevant strains, deformations and possibly the support reactions could be measured in order to check if the analysis models used correctly reflects the response of the bridge. Furthermore, a confirmation of the minimum load carrying capacity would be obtained. The response would be followed during the loading and the load increase be terminated if risk of damage or cracking of the webs arises. Regarding the risk of cracking due to shear at the flange to web connection, the study showed that the bridge has sufficient resistance if the classification with respect to environmental influence is reconsidered. For the bridge at Kropp, the evaluation showed that the load A/B, i.e. the axle and bogie loads according to Vägverket (1998), can be raised from 150/255 kN to about 275/630 kN, if certain assumptions were made regarding the dynamic amplification factor and the accuracy of the traffic load. The reasonableness of these assumptions needs to be further evaluated. Further research and development are needed regarding quantification of uncertainties in the modelling of material parameters, loads and structural capacity