In Amsterdam, more than 30 steel-concrete composite bridges were constructed from 1880-1960 without mechanical connectors and transverse reinforcement. Currently, there are no simplified analytical methods to determine the bearing capacity of these bridges. Thus, the bearing capacity is verified using NLFEM or oversimplified analytical calculations. This research proposes an analytical method to determine the bearing capacity of historic steel-concrete-composite bridges without mechanical connectors to avoid time-consuming FEM calculations and offers reasonable results.

An experimental and numerical study is performed on data from in situ and laboratory testing of samples from two different bridge decks from these Amsterdam bridges. The tests are accompanied by a numerical model that has been studied and adjusted to a more generalized loading case. This study determined that the exterior composite girders are critical due to their lower lateral stiffness.

An analytical model is proposed to examine the behaviour of the exterior composite girder. The model considers a 3-point bending load at midspan between the exterior composite and adjacent girder. The force distribution is described through a compatibility-based strut and tie model (C-STM). The concrete in compression is considered elastic compression struts, only limited by the ultimate load of the model. The concrete in tension is interpreted as a tensile tie, which fails when it exceeds the concrete tensile resistance. Following the failure of the tensile tie, it is assumed that a longitudinal crack propagates between the exterior composite girder and the adjacent girder. Additionally, vertical and lateral stiffness components are included in the model. These account for the flexural stiffness of the exterior and interior composite girder. The vertical stiffness is accounted for as elastic springs, and the lateral stiffness as spring beams. The interior lateral spring beam summarises all the interior composite girders' stiffness, whereas the exterior lateral spring beam only considers the exterior composite girder. Therefore, the configuration assumes that the interior spring beam is significantly stiffer than the exterior. Moreover, the stiffness of the exterior spring beam reduces when the longitudinal cracking occurs, assuming a part of the concrete fails. The C-STM is linked to the cross-section verification of longitudinal shear, biaxial bending and vertical shear resistance in two stages. Stage 1, at the load at longitudinal cracking, determines if the specimen fails at this moment, indicating that there possibly is a brittle failure. Stage 2 is after longitudinal cracking, where the steel-concrete contact perimeters have reduced, and the corresponding resistances accordingly reduce.

The failure modes obtained by the analytical model are comparable to the ones observed during the experimental testing. The analytical model showed that the bridges failed due to biaxial bending limited by partial shear interaction. One of the specimens from the testing yielded due to bending but with limited ductility. The other specimen also yielded due to bending with concrete crushing at the top concrete fibre. Further, the bearing capacities obtained from the analytical model are comparable to the failure loads from the experimental and numerical results.

The model predicts the failure modes and the bearing capacity and can therefore contribute to the assessment of the historic Amsterdam bridges, helping to reduce the assessment time of the bridges and understand their load-bearing behaviour better. Future work should focus on verifying the method by examining more bridges using FEM.

Geopolymer concrete, a relatively new building material has been suggested as an alternative to Ordinary Portland Cement (OPC) based concretes for within the building sector, due to it having a relatively low CO2 emission during production. An issue with geopolymer concrete however, is that high amounts of early age creep, shrinkage and decrease in elastic modulus over time has been observed which in conjunction with the existing research on geopolymer concrete having been predominantly conducted in laboratory settings, leads to questions about the applicability of these findings and thus the use of geopolymer concrete as a replacement for OPC based concretes in concrete structures. To address this issue, this study makes use of Fiber Bragg Grating (FBG) fiber optic sensors, placing them within reinforced steel which in turn are placed inside geopolymer concrete to obtain data that is more applicable to real world settings compared to prior findings observed within laboratory settings.
Four precast prestressed geopolymer concrete girders are produced and rebars containing FBG fiber optic sensors are placed prior to casting. Measurements before and after the release of prestressing are analysed to determine the apparent transmission length of the prestressing strands and compared to methods defined by Eurocode EN 1992-1-1. The applicability of these guidelines is evaluated, as these are intended for use with conventional concretes. The changes in strain over time, and elastic modulus changes during curing of the precast girder is monitored and analysed, focusing on the effect of time dependent changes of the concrete to the effectiveness of the prestressing forces. The girders are combined with a geopolymer concrete topping to form two beams and one slab. The beams are subject to a flexural test and a shear test respectively. The slab undergoes a cyclic loading test before being tested in two critical point load locations in two separate tests. The FBG measurements gained during the tests are evaluated using analytical methods.
The results show that the FBG fiber optic sensors integrated into steel rebar can deliver strain measurements in locations conventional strain sensors cannot be applied. The FBG data suggests that conventional concrete transmission length guidelines are applicable to this geopolymer concrete mixture. As the FBG-integrated rebars were all placed in the same location in the precast girder cross section, the strains measured by these FBGs are only representative of concrete near the core of the girder, where the concrete is not exposed to drying conditions. And because this geopolymer concrete is susceptible to drying, the difference between measurements and expected values such as the elastic modulus and the drying shrinkage can be explained by the low amount of drying that the concrete in the core of the girder experiences. Due to the positioning of rebar in the cross section of the precast girder in this study, crack initiation during the flexural test was not detected. The occurrence of cracks is detected in the measured strains as it diverged from the linear elastic model during the crack propagation phase. This shows that the cross sectional position of the sensors is important for the desired function and role of concrete in-situ FBG monitoring.
Overall, the data that the (rebar-integrated) FBG strain sensors can provide can be of great use towards the larger task of bringing geopolymer concrete structures to a real life application. In-situ FBG strain data of the precast prestressed girders provide additional information into the application of this SCGC mixture in larger scale structural elements. This data is a useful addition to the data gained through conventional methods, and results from lab testing. FBG monitoring can provide strain data from the core of structural elements, giving an insight of the behaviour during curing and (destructive) testing that is otherwise difficult to obtain.

Concrete half-joints are a specific support detail in reinforced concrete structures, which reduce the construction height of the total structure. The use of concrete half-joints became popular around 1950s, but decreased its interest as a result of new insights on structural behaviour and collapses of these structures. Typical issues regarding concrete half-joints are either due to inadequate reinforcement detailing or due to deterioration mechanisms. The most critical deterioration is when a crack at the re-entrant corner allows for water ingress and thus corrosion of the rebars. This thesis focuses on the negative influence of this corrosion on the load bearing capacity and proposes an assessment method, which includes the reinforcement detailing. Firstly, problematic half-joints have been categorised and studied for reinforcement detailing. Subsequently, they have been analysed using an analytical tool, in which corrosion was implemented on the rebars. The outcomes were validated numerically.

For this thesis research, a series of Dutch concrete bridges has been studied to identify general reinforcement issues and categorise concrete half-joints. It has been observed that the majority showed short transfer- and/or anchorage lengths of the rebars and that all of them showed no shear stirrups as hanger-reinforcement. In stead, only a horizontal- and hanger rebar are present, which can be accompanied with a diagonal rebar, prestressing at the top or nib (or combinations in between).

An analytical tool is designed to calculate the load bearing capacity of (un)corroded concrete half-joints. The analysis is based on a lower-bound approximation using a strut-and-tie approach and an upper-bound approximation using a kinematic approach. The analytical tool is used to determine the load bearing capacity of the series of investigated Dutch concrete half-joints. Both approximations are comparable when rebar failure is the governing failure mechanism. The strut-and-tie approach also incorporates detailing checks, which are not considered in the kinematic approach. Therefore large differences occur when detailing governs the capacity.

The nodes in the strut-and-tie model, in which two ties are connected to one concrete strut, appear to be critical in the lower-bound solutions. The capacity depends on the concrete strength and dimensions of the node. The dimensions are influenced by the mandrel diameter of the hanger-rebar and anchorage length of the horizontal rebar. In order to study the influence of corrosion on the load bearing capacity, the effect of a reduced rebar capacity due to an increasing corrosion rate was implemented in the analytical tool. The kinematic approach appears to be more sensitive to load bearing capacity loss, as this calculation depends mainly on the strength of the rebars. The strut-and-tie approach is able to redistribute forces over the struts and ties and is less sensitive.

In order to verify the analytical results, a numerical study is performed. The specimens are modelled in such a way that rupture of one of the rebars at the re-entrant corner is governing. Both analytical solutions appear to be conservative compared to the numerical results, in which the lower-bound solutions are very conservative. Different crack’s angles have been found between the upper-bound calculation and numerical results. If the same angle is applied in the analytical tool, the difference reduces from 7% to 2% for an uncorroded concrete half-joint without diagonal. The differences can be explained by the simplification in the kinematic approach, in which the concrete compression zone is not able to transfer shear stresses.

Based on the conclusions of the analytical tool and numerical verification, an assessment method is proposed in which the upper-bound solution is combined with the lower-bound solution. If the load on the concrete half-joint is lower than the calculated lower-bound solution, the concrete half-joint is safe. However, questions arise if the load is between the lower- and upper-bound solution, in which structural safety cannot be guaranteed. The analytical tool is still a useful tool to understand the behaviour and vulnerabilities of the concrete half-joint. The analytical tool is even more useful if the strut-and-tie approach is governed by rupture of the horizontal-, diagonal- or hanger-rebar. The kinematic approach can be extended by implementing the same crack’s angle, which occurs in the existing concrete half-joint.