This paper describes a multi-level strategy with increased complexity through four levels of structural analysis of concrete bridges. The concept was developed to provide a procedure that supports enhanced assessments with better understanding of the structure and more precise predictions of the load-carrying capacity. In order to demonstrate and examine the multi-level strategy, a continuous multi-span prestressed concrete girder bridge, tested until shear failure, was investigated. Calculations of the load-carrying capacity at the initial level of the multi-level strategy consistently resulted in underestimated capacities, with the predicted load ranging from 25% to 78% of the tested failure load, depending on the local resistance model applied. The initial assessment was also associated with issues of localising the shear failure accurately and, consequently, refined structural analysis at an enhanced level was recommended. Enhanced assessment using nonlinear finite element (FE) analysis precisely reproduced the behaviour observed in the experimental test, capturing the actual failure mechanism and the load-carrying capacity with less than 4% deviation to the test. Thus, the enhanced level of assessment, using the proposed multi-level strategy, can be considered to be accurate, but the study also shows the importance of using guidelines for nonlinear FE analysis and bridge-specific information. 

Assessing the load‐carrying capacity of existing bridges is an important infrastructure management task. In order to support the structural assessment of concrete bridges better, a procedure has been proposed, based on successively improving the bridge analysis. A multilevel strategy for structural analysis has been combined with concepts for verification of the desired safety margin, thus providing tools for engineers to model the structural behavior of bridges more accurately when necessary. This paper describes the procedure as applied to a prestressed concrete girder bridge, with use of experiences from the previous failure tests and associated evaluations of the bridge. Initial structural assessment indicated the critical failure mode to be due to shear in one of the girders; however, the enhanced analysis showed a complex failure involving both the girder and the bridge deck slab. Improving the structural analysis using nonlinear FE analysis for the loading initially identified as critical, increased the permitted axle loads on the bridge to 12 to 14 times those given by traditional and standardized assessment methods, depending on the concept used for safety verification. The model uncertainty was crucial for the verification of the structural safety and has to be properly taken into account. However, there are few recommendations, with regard to model uncertainties, on the application of nonlinear FE analysis, and detailed guidelines should be used for the modeling procedure in order to reduce analyst‐dependent variability in the results. The presented study demonstrates the applicability and the advantages of using the proposed procedure for successively improved analysis for bridge assessment.

This paper presents the finite element analysis (FEA) results of a multi-story reinforced concrete (RC) building having precast and cast-in-place load bearing walls. Door-type cut-out openings (height: 2.1 m, width: 0.9–4.4 m) were created at the first and second story of the building. Results from experimental tests on axially loaded RC panels were used to verify the modeling approach. The influence of cut-out openings on the response of individual RC panels, failure modes, and load redistribution to adjacent members was analyzed. Moreover, the wall bearing capacities obtained from FEA were compared with the values calculated from design equations. The results revealed that the robustness of multi-story buildings having RC load bearing wall systems decrease considerably with the creation of cut-out openings. However, owing to the initial robustness of the buildings, large cut-outopenings could be created under normal service conditions without strengthening of the building structure. Furthermore, design equations provided very conservative predictions of the ultimate capacity characterizing the solid walls and walls with small openings, whereas similar FEA and analytically predicted capacities were obtained for walls with large openings.

The objective of this paper is to validate a 2D nonlinear finite element (FE) model for estimating the post-cracking shear deformation of reinforced concrete (RC) beams. The proposed FE model treated the cracked concrete as an orthotropic material in the framework of the fixed-crack approach. The experimental data for both the overall response (including the total and shear-induced deflection) and the detailed response (including the mean shear strain, mean vertical strain and principal compressive strain angle) of five I-section RC beams, monitored by the main authors of this paper with the Digital Image Correlation technique, were used to verify the proposed model. In addition, 27 further test beams evaluated in independent research programs were collected to assemble a database. The proposed FE model was further verified against the database. Two additional FE models (the rotating-crack model developed in this work and Response-2000 developed by Bentz (2000)) were also evaluated by simulating the detailed responses of the beams in the database. The results obtained validate the proposed FE model for predicting the post-cracking shear deformation of RC beams and indicate that the proposed FE model is more suitable for simulating the shear behaviour of RC beams than the rotating-crack model or Response-2000.

Current codes often underestimate the capacity of existing bridges. The purpose of the tests presented here has been to assess the real behaviour and capacity of three types of bridges in order to be able to utilize them in a more efficient way.

The three studied bridges are: (1) Lautajokk – A one-span trough bridge tested in fatigue to check the shear capacity of the section between the slab and the girders; (2) Övik – A two span trough bridge strengthened with Near Surface Mounted Reinforcement (NSMR) of Carbon Fibre Reinforced Polymers (CFRP) tested in bending, shear and torsion; and (3) Kiruna – A five-span prestressed three girder bridge tested to shear-bending failures in the girders and in the slab.

The failure capacities were considerably higher than what the code methods indicated. With calibrated and stepwise refined finite element models, it was possible to capture the real behaviour of the bridges. The experiences and methods may be useful in assessment and better use of other bridges.

Five bridges of different types have been tested to failure and the results have been compared to analyses of the load-carrying capacity using standard code models and advanced numerical methods. The results may help to make accurate assessments of similar existing bridges. There it is necessary to know the real behaviour, weak points, and to be able to model the load-carrying capacity in a correct way.The five bridges were: (1) a strengthened one span concrete road bridge - Stora Höga ; (2) a one span concrete rail trough bridge loaded in fatigue – Lautajokk; (3) a two span strengthened concrete trough railway bridge - Övik; (4) a one span railway steel truss bridge -Åby; and (5) a five span prestressed concrete road bridge - Kiruna. The unique results in the paper are the experiences of the real failure types, the robustness/weakness of the bridges, and the accuracy and shortcomings/potentials of different codes and models for safety assessment of existing structures

Full-scale failure tests of bridges are important for improving understanding of bridges’ behaviour and refining assessment methods. However, such experiments are challenging, often expensive, and thus rare. This paper provides a review of failure tests on concrete bridges, focusing on lessons from them. In total, 40 tests to failure of 30 bridges have been identified. These include various types of bridges, with reinforced concrete or prestressed concrete superstructures, composed of slabs, girders and combinations thereof. Generally, the tests indicated that theoretical calculations of the load-carrying capacity based on methods traditionally used for design and assessment provide conservative estimates. It can also be concluded that almost a third of the experiments resulted in unexpected types of failures, mainly shear instead of flexure. In addition, differences between theoretical and tested capacities are often apparently due to inaccurate representation of geometry, boundary conditions and materials

For reinforced concrete (RC) slabs without shear reinforcement, shear and punching can be the governing failure mode at the ultimate limit state if subjected to large concentrated loads. Shear and punching of RC slabs without shear reinforcement has been a challenging problem in assessment based on current standards. To examine a previously developed enhanced analysis approach, this study was conducted by applying a multilevel assessment strategy to a 55-year old RC bridge deck slab subjected to concentrated loads near the main girder in a field failure test. This strategy clearly provides the engineering community a framework for using successively improved structural analysis methods for enhanced assessment in a straightforward manner. The differences between analysis methods at different levels of assessment were discussed regarding one-way shear and punching shear behavior of the slab. The influences of parameters, such as boundary conditions, location of concentrated loads, and shear force distribution, were investigated.

Tests have been carried out at service- and ultimate load levels of a 55 year-old prestressed concrete girder bridge. The bridge, located in Kiruna, Sweden, was continuous in five spans with a total length of 121.5 m. The overall aim of the study was to determinate the accuracy of assessment methods for existing structures and to provide procedures for optimized assessment. Before the tests a 2D finite element (FE) analysis was performed to predict the behavior and load-carrying capacity of the bridge. In order to more accurately assess the bridge response a 3D FE model has now been developed. The actual loading history and material properties has been considered in the model. A Life Cycle Cost Assessment of the bridge has also been performed

Three Swedish concrete bridges have been tested to failure and the results have been compared to assessment using standard code models and advanced numerical methods.

The three tested and assessed bridges were:

1. (1)

   Lautajokk, a 29 year old one span (7 m) concrete trough bridge tested in fatigue to check the concrete shear capacity.

    
2. (2)

   Ӧrnskldsvik, a 50 year old two span trough bridge (12 + 12 m) strengthened to avoid a bending failure.

    
3. (3)

   Kiruna Mine Bridge, a 55 year old five span prestressed concrete road bridge (18 + 21 + 23 + 24 + 20 m) tested in shear and bending of the beams and punching of the slab.

    

The main results in the paper are the experiences of the real failure types, the robustness/weakness of the bridges, and the accuracy of different codes and models. In all three cases the bridges had a considerable hidden capacity.