In the Netherlands, a lot of traffic congestion occurs on motorways. This problem is most severe nearby larger cities. Utrecht is one of these cities. To improve traffic flow in this area, a huge masterplan is designed by Rijkswaterstaat called “A27/A12 Ring Utrecht”. One part of this masterplan consists of the motorway A27 at Amelisweerd. Here, the A27 is situated in a U-shaped concrete structure and must be expanded at both sides. Across the motorway a deck structure is going to be constructed with on top a public garden. This structure spans the total width of the A27 for 249 meters and is called The Green Connection.
According to the original design of Rijkswaterstaat, this deck structure should be realized with an intermediate support. It will be advantageous to omit this structure, since it has a complex execution method. Therefore, it’s investigated if The Green Connection can be realized without the use of an intermediate support.
Then, the execution aspects of the original design will be discussed more thoroughly. The first challenge is constructing the extended parts which is explicated according to 11 main tasks. These tasks seem to be relatively straightforward to execute. After realizing the extended parts, an intermediate support must be constructed. It is found that the existing foundation lacks bearing capacity by far. A new strengthened strip foundation with extra foundation piles must be realised in the middle of the motorway. Due to the boundary conditions (such as the water pressure beneath the structure and permanent drainage is prohibited), the only possibility left is to construct small building pits, compartments, within the existing structure. Such a compartment has a rough length of about 20 meters, will be about 6.5 meters wide and must be constructed 13 times.
Thereafter, the deck structure should be assembled. Three alternatives in methods of assembly are outlined and discussed with the help of the same key-words. All three methods could be realized. But, it’s important to indicate that with some extra investments, the remaining space for traffic could be maximized during assembly.
After discussing the execution aspects of the original design, the technical feasibility of the single span deck structure was investigated more thoroughly. In the Preliminary Study was deduced that two structural designs seem to be a possible solution in constructing The Green Connection. It turned out that the box beam design seems to be advantageous. Although this judgement is substantiated with preliminary calculations and an overall execution plan, it still required more research. Therefore, a reliable structural design is performed for a 75-meter span beam which can be used as a single span deck structure. It does exceed the boundary condition of 280 tons which was posed initially with 8%. However, no optimizations have been applied to this design. If the enumerated optimizations are performed, a beam can be designed according the boundary condition and possibly even less.
When the original design of Rijkswaterstaat is compared to the single span design, the differences are quite straightforward. In essence, the question arises whether the extra money of constructing a single span deck structure outweigh the money which can be saved by leaving out the intermediate support and constructing the extended parts 25% narrower.
Rijkswaterstaat has performed a design with an intermediate support. In this thesis, the feasibility of this design is investigated, and in particular this support. The only possible method of execution in realizing the support is upon condition that a temporary applied drainage system will be feasible and approved by the authority. When the authority states that draining ground water is prohibited, the original design isn’t feasible anymore. In that case, the single span design isn’t just an alternative to the original design, but is the only feasible solution.
Concludingly, providing the applied principles of this thesis, it is strongly recommended against constructing an intermediate support within the existing U-shaped concrete structure. Since the structural reliability of a 75-meter span beam is proven, an intermediate support wall is redundant. Therefore, the single span design is less risky, less time consuming and less expensive compared to the original design.

Most bridges in the Western European road networks are ageing. The vast majority of about 90% of these bridges have reinforced concrete as a building material. The traffic intensity, as well as the axle, and the average vehicle weight have increased since these structures were opened to traffic. Furthermore, the structural (design) codes have changed over the years. Therefore, there is a need to investigate if existing structures meet the safety/reliability level described by the current codes. However, a frequently faced problem in practice is that the original design calculations and technical drawings of a large percentage of the existing bridges are unknown or lost. Especially for bridges in the lower road network, often designed for the lower load classes B/45 and maintained by a local government, the documentation is missing. The national road network, designed for load classes A/60 is maintained by the national government and faces the same problem but to a lesser extent. Bridges within this scope have different detailing rules and execution practices than used nowadays. Plain reinforcement was in general used which is bend-up at a support. Therefore, the study is twofold: first, Reverse Engineering is applied to determine the reinforcement of existing reinforced concrete slab bridges, second the capacity margin of RE bridges are examined with the current assessment codes. A Reverse Engineering-tool is developed to automate the dimensioning of the required reinforcement according to the former design codes. This tool uses the year of design, load class and the geometric dimensions of the bridge as input parameters. A parametric study is performed to examine bridges from different design periods. Consequently, the Reverse Engineering-tool is used to assess the Reverse Engineered bridge according to the current assessment codes. The validation of the model shows for the majority of the Reverse Engineered bridges that the Reverse Engineered reinforcement is slightly less than the reinforcement amounts from the technical drawings. This proves a conservative approach where the actual structural capacity is underestimated. Consequently, an assessment of the Reverse Engineered bridge can be performed with sufficient robustness.The computer code ran with the input parameters having a normal distribution, showed the largest effect for the uncertainty in the design year and load class especially around 1940 and 1962. Therefore, the design year and load class are crucial in Reverse Engineering and assessment of an existing bridge. The capacity margin of the Reverse Engineered bridges is assessed according to the current Eurocode based design codes. The traffic- and permanent load including load factors according to the general assessment codes from the NEN8700/NEN8701 and the RBK-1-1, and the decentralised load model from TNO are applied. The assessment with the Eurocode including the load factors from the NEN8700 showed Unity Checks for bending moment at the mid-supports and mid-spans of larger than 1.0, where the Unity Checks for shear forces resulted below 1.0. The assessment with the Decentralised load model showed Unity Checks for bending moment at the mid-supports and mid-spans and shear force below 1.0. In case the amount of support reinforcement is based on the amount of span reinforcement, the bending capacity margin at the mid-supports is insufficient for large spans. Significant bending capacity margins are obtained in structural design of RC slab bridges in the period 1930-1970. The main contribution of this research is that bridges designed between 1940 and 1962 show the most critical Unity Checks for bending in the assessed period. In this period account the following design methods: The dynamic amplification factor introduced in the GBV1940 for concrete bridges, the traffic load class from the VOSB1933, the N-method to determine the cross-section capacity and the effective width method from the GBV1940 and from the Guyon Massonnet method.The capacity margin for shear is found to be almost independent of the design period. Here can be concluded that the slenderness of the bridge deck is the main contribution in the shear capacity. Bridges designed in the period 1940-1962 with the support reinforcement based on the span reinforcement and with a span length >10m designed for load class B/45 or with a span length of >11m designed for load class A/60, form the group with the most critical bending capacity. However, the size of the group of former bridges designed according to these conditions is unknown.From the results can be concluded that bridges designed between 1940-1962 with RE reinforcement are found to be legally unsafe for bending according to the parametric assessment with the Eurocode.

In this thesis, existing concrete municipal slab bridges are regarded. Many of these bridges were designed and built in the 1960s and 1970s. These bridges are still in service, and subjected to higher and more frequent loads than at the time of design. This leads to uncertainties in the safety of these bridges. Many municipalities become increasingly aware of these uncertainties and want to verify their bridges according to current standards. Especially the shear assessment is critical, since some former codes do not contain shear checks. Municipalities usually do not have sufficient financial resources to investigate and recalculate all of their bridges. A Quick Scan model is a cheap and fast tool to do structural checks, or to classify a number of bridges for structural safety. A literature study and Finite Element Modelling research was performed in order to make the Quick Scan more accurate. In general, existing municipal bridges are not exposed to the same heavy traffic as governmental highways. Nevertheless, municipal bridges were calculated with the same load models as governmental bridges. For the assessment of existing municipal bridges research was done to the possibility of reducing the design loads of Load Model 1 (LM1) from the Eurocode. This reduction can be done with the α- factor on variable loads. The possible reasons for this reduction is threefold. Firstly, municipal roads are significantly less exposed to heavy traffic compared to governmental highways. Therefore the chance of occurrence of the governing vehicle according to LM1 is decreasing. This assumption was supported by Weight-in-Motion measurements on a municipal road in Rotterdam. With the reliability index (β) for existing bridges in Consequence Class 2, this lead to a governing axle load of 225 kN, instead of the 300kN from LM1. The second reason for the reduction of the design load is the fact that municipal bridges usually have relatively small spans (5-20m). LM1 is designed to simulate a fully loaded bridge with a span of at least 20m. For small spans it can be useful to calculate with real occurring traffic, since axle distance is more important than axle loads only. A third (small) reduction is the fact that an existing bridge has a shorter reference period than a new bridge. This leads to smaller chances of the occurrence of the governing vehicle. New load models for small span municipal bridges were designed taking into account these reductions. The governing vehicle for small span bridges is a 5-axle vehicle with axle loads of 137,5-165 kN and axle distances of 140-175mm. Occurring shear forces due to these loads are higher than due to a long vehicle with higher loads, mora axles and greater axle distances. The new governing vehicle was used to determine the reduction factor α. This was determined by a comparative research in the Finite Element Modelling program RFEM. The resulting shear stresses and flexural stresses from the different load models were compared. For spans up to 11m the loads from LM1 can be reduced by an α factor 0,8. This factor increases linearly to a factor 1,0 for a 20m span. Besides the differences in loads and dimensions, municipal bridges also differ from governmental bridges in lay-out. In general, the distance from the carriageway to the edge of the slab is larger for municipal bridges due to the presence of a footpath or bicycle lane separated by a kerb. Bridges in 10 different municipalities were examined on lay-out. Roughly 58% of the municipal bridges in the Netherlands have significant edge distance (>1,2m) This leads to a difference in the force transmission, since the high axle loads from LM1 cannot occur near the edge of the slab. The influence of this edge distance for different spans was investigated with RFEM. Also, the variable and permanent loads were investigated separately to give insight in the resulting shear stress due to different loads. Distinction was made between slabs in cracked and uncracked state. Cracking has significant influence on the transverse force transmission of the axle loads. The transverse force transmission is influenced by the ratio between longitudinal and transverse stiffness, which was conservatively chosen as 1/3. Also, distinction was made between in-situ casted slabs and prefab slabs. The self-weight of in-situ casted slabs leads to a constant shear force on the support. A prefab slab was treated like a theoretical slab, where the self-weight leads to peak forces near the edge due to torsional moments near the edge. The critical edge distance was determined for different spans. For a cracked in-situ casted slab with an edge distance >1,5m, the middle of the slab is always governing in shear. An edge distance <0,7 leads to the edge of the support being governing in shear. Research by Eva Lantsoght to shear force in reinforced slabs under concentrated loads close to the supports was used for rules and assumptions based on experiments. These rules were used to get a better understanding of the results in RFEM, and to develop the Quick Scan Model. The Quick Scan model uses findings from literature research combined with findings from finite element modelling. The output is a Unity Check which can function as real structural shear check if the concrete compressive strength and steel yield strength are known. When only dimensions are known, the Quick Scan model can function as classification for a series of bridges in a municipality. The Quick Scan model was tested by a series of case studies, which did not yet lead to the verification of the model. Results from the Quick Scan model were compared to results from the FEM research. Occurring shear stresses in the Quick Scan model are conservative with a maximum error of 11%. With the possibility of lowering the reliability index β to 2,5 (‘disapproval’ level) and taking into account the possible error due to several uncertainties, the upper boundary of the Unity Check was found as 1,4. Bridges with a Unity Check: 1 < UC < 1,4 according to the Quick Scan model need further assessment or material testing in order to fulfil the requirements. Extensive research was done to loads according to the Eurocode (LM1) on slab bridges and the force transmission of these loads. The capacity of existing concrete slabs was investigated relatively brief. More research to the capacity side of existing concrete bridges is expected to be beneficial.