In the construction industry the demand for repair and maintenance of concrete structures constantly increases. Still, the performance of current concrete repairs is not satisfactory and there is an urgent need for improvement. Understanding the damage development in a repair system, and how to predict, model and prevent its failure is critical to improving performance of concrete repairs. Therefore, the aim of this thesis is to study damage development and failure modes in repair systems. For this purpose, interaction between the repair material and the existing substrate was investigated. First the dynamics of moisture exchange in repair systems is studied, because this interaction is critical for the development of properties of repair material and interface. Furthermore, the interaction between the repair material and the substrate under different exposure and loading conditions was investigated. The most common causes of repair failure were studied here: flexural effects, drying shrinkage and ongoing corrosion of reinforcing bars in the repair system. The main repair material used in this study is SHCC (strain-hardening cementitious composite). MOISTURE MOVEMENT IN THE REPAIR SYSTEM In Chapter 3 moisture movement in repair systems was investigated using X-ray absorption technique. Influence of w/c of the repair material, the duration of sealed curing, primer application and the initial moisture content of the substrate on moisture exchange and formed microstructure in the repair system were investigated. Experimental results show that the absorption rate by the dry substrate in the first 5 hours after casting of the repair material (regular OPC paste without any chemical admixtures, as used in this research) is independent of the w/c of the repair material. Furthermore, water from the repair material is absorbed by the substrate at the same rate as pure water. The substrate absorption significantly reduced the water content of the repair material and was found to be critical for development of the repair material properties. Repair material cast on the initially dry substrate achieved lower degree of hydration compared to the repair material cast on the initially saturated substrate. Wet substrate ensures more uniform water distribution and microstructure development in the repair material compared to the repair material cast on a dry substrate. At later age, water migrates back from the substrate to the repair material. The lower the w/c of the repair material, the more water goes back from the substrate to the repair material. Duration of sealed curing has an effect on development of properties of the repair material and interface but its effect is dependent on the initial saturation state of the substrate. It was found that, when the substrate was initially saturated, increasing duration of (sealed) curing both from1 to 3 days and from3 to 5 days had a beneficial influence on the hydration of the repair material. However, when the substrate was initially dry, curing samples for 3 instead of 1 day was beneficial, but curing samples for 5, instead of 3 days did not result in significant improvements. Moisture movement not only affects the moisture distribution but also the microstructure around the interface in a repair material. It was found that with higher substrate absorption significantly more voids in the repair material, close to the interface, formed. Dry substrate absorbs water from the repair material and releases air. This air stays entrapped at the interface resulting in high void content. This may affect the strength of the material in this region. MICROMECHANICAL INTERFACE PROPERTIES IN THE REPAIR SYSTEM In Chapter 4 the properties of the interface, repair material, and substrate were investigated at the microscale using nanoindentation. Measured moduli of elasticity and hardness were used indirectly as input in micromechanical modelling. Uniaxial tension test is simulated with the lattice model and interface, repair material, and substrate fracture properties were obtained. It was shown that the ratio between interface tensile strength and repair material tensile strength at the age of 28 days, is lower than 0.9 for pure Portland Cement paste. This ratio decreases further as the addition of BFS increases. FAILURE MODES IN THE REPAIR SYSTEM In Chapter 5 experimental and numerical studies were performed to investigate the influence of the interface and SHCC material properties on the fracture performance of repair systems due to mechanical loading. Three point bending test, DIC (digital image correlation) and epoxy impregnation were used for investigating the fracture behaviour of the repair system in experiments. The lattice fracture model was used as a numerical tool for studying damage development. The influence of substrate surface roughness was explained. Surface roughness of the substrate does not affect the load-bearing capacity in flexural tests, but there is a substantial difference in crack pattern and debonding tendency in the repair system. When the substrate surface is rough, cracks from the repair material are interlocked by grooves and directed to the substrate. In flexural and reflective cracking tests low interface toughness (low interface strength and smooth surface of the substrate) are beneficial. With less restraint at the interface, there is more local debonding around the cracks, resulting in more microcracking in the repair material (SHCC). As a result, higher ductility of the repair system is achieved. This is different from what we find in standard recommendations for surface preparation, which advises roughening of the substrate surface. However, very low bond strength and smooth surface might also lead to uncontrolled debonding and shift in failure mode (complete delamination). In Chapter 6 experimental and numerical studies were performed to investigate the influence of interface and SHCC material properties on the fracture performance of repair systems due to drying shrinkage of repair material. Free drying shrinkage of repair materials was measured and the effects of drying shrinkage (cracking and delamination)on repair system beams were investigated. Lattice moisture model and lattice fracture model were used for studying damage development due to restrained shrinkage. Influence of substrate surface roughness, repair material thickness, interface strength and type of repair material were investigated. It was shown that interface strength and surface roughness are more important for performance of thinner overlays compared to the thicker ones due to the higher moisture gradient. With weak bond and thinner repair material, the system is susceptible to large debonding and a big crack. If there is no continuous delamination, with the same bond and drying conditions, thin overlays result in more cracks with smaller spacing and smaller crack widths. As already shown in Chapter 5, damage development and final failure mode are sensitive to the interface strength and surface roughness. In contrast to recommendations from mechanical tests (i.e. Chapter 5), high interface strength and high surface roughness are necessary for optimal performance of the repair system with SHCC as a repair material under restrained shrinkage conditions. In Chapter 7 experimental and numerical studies were performed to investigate the influence of interface and SHCC material properties on the fracture performance of repair systems subjected to ongoing corrosion of rebars in the repair material. Experimentally, rebars in the repair material were exposed to accelerated corrosion. For the numerical study, the lattice model, used in Chapters 5 and 6, is applied. Influence of type of repair material, interface strength, substrate strength and substrate surface roughness on damage development is studied. It was shown that in case of continuing rebar corrosion, surface roughness had a similar influence as in Chapters 5 and 6. Grooves enable cracks to continue to propagate to the substrate. A high interface strength and a rough surface are beneficial to exploit the ductility of the repair material. On the contrary, a smooth substrate surface and a low interface strength cause uncontrolled failure through delamination. However, some delamination around the crack is also beneficial, especially when the substrate is stronger than the repair material, as the stresses will be partially relieved resulting in smaller cracks in the repair material (SHCC). A CASE STUDY In Chapter 8 SHCC is applied as a repair material in a trial patch repair in a deteriorated concrete tunnel (the Maastunnel). For surface preparation of the substrate, recommendations and conclusions from the previous chapters were used. Five different types of repair materials were applied in small patches in the tunnel. Laboratory tests were combined with on-site investigations in order to study the performance of different repair materials. It was shown that when concrete used as a repair material (as opposed to repair mortar), the repair system exhibited the best performance (no cracking nor delamination). However, in case that corrosion continues in the bars, use of SHCC is beneficial as it will result in smallest crack widths and more ductility of the repair system. CONCLUSIONS AND RECOMMENDATIONS In Chapter 9, conclusions of this study and practical recommendations for the repair application are given.