Assessment of existing concrete bridges : bending stiffness as a performance indicator

Sammanfattning: Optimizing the use of existing civil reinforced concrete (RC) structures could be interpreted in such a way as to say that the capacity should be used and taken care of in an effective manner, both from a technical and economical point of view, keeping the safety in mind. Achieving this requires thorough understanding of the structure and also of the tools used for assessing current and future capacity and needs. Monitoring together with finite element modelling could give relevant and important information about a structure's capacity. In a case where monitoring alone is used, it is beneficial if a quantity is monitored which is interpretable on material and geometrical level. It is further important that the measure is practically possible to capture, and that it reflects the behaviour in a theoretically well-known mode. One example of a quantity which fulfils these requirements is the bending stiffness. In the Serviceability Limit State (SLS), in particular, a high bending stiffness is beneficial as this reduces deflections, vibration amplitudes and crack widths.It is shown within the thesis that four phases are distinguished during loading of an RC member; Un-cracked phase (I), Crack forming phase (II), Crack stabilised phase (III) and Failure phase (IV). It is also shown that corrosion and flexural strengthening are possible to capture through the bending stiffness by monitoring the curvature. Linear elastic theory has in addition been concluded to give satisfactory results in terms of good agreement between measured and theoretical results. It is shown that it is possible to determine the highest load which the structure has been previously exposed to, presuming that the structural element has not reached phase (III). The stiffness is almost constant in phase (III) which implies that it is the same for a certain load interval. One limitation coupled to the stiffness plateau formed in phase (III) is that it is difficult to predict a possible failure by monitoring the bending stiffness, caused by the limited forewarning prior to the beginning of phase (IV). Other tools, such as reliability-based assessment, become especially important here since active degradation, for example, is difficul to verify by curvature measurements in phase (III). Estimating the safety, and also finding the probable failure mode is important since curvature measurement is not effective in the Ultimate Limit States (ULS) and only captures the behaviour in bending. In the reliability-based assessment, the agreement between analytical results and actual capacity of the particular failure mode must be treated with special attention, since it has been shown that the model uncertainty can affect both the safety level and also probable failure mode. If it can be shown from monitoring that the structure is located within phase (I) the load effect during the past time has not affected the integrity of the structure in terms of bending cracks.It is preferrable to use the global curvature when evaluating the bending stiffness, since this property gives a more robust average curvature and also additional information about the structural member. Especially changing bond properties, during e.g. corrosion, is more likely to be detected if the global curvature is monitored. Another important conclusion is that the local and global stiffness development is very similar. This implies that a crack at a certain location is not allowed to increase without redistribution of stresses, which affects the global stiffness in an comparable extent. Two criteria are suggested for the least distance over which the global curvature should be measured, LG. The first one concerns the fact that at least four macro cracks is suggested to be covered and is based on the maximum crack spacing recommended by Eurocode (2004). The other requirement is that the distance should not be that small that the estimated deflection become less than one hundred times the in-built measurement error in the displacement gauge. A measurement error above one percent is hence not allowed.Curvature assessment could be useful from three different aspects ' Condition assessment. The monitored quantity is back-calculated to input data, such as material property or geometry. That is, solving the inverse bridge management problem. Decisions about the use of the structure are then based on the outcome of this assessment.' Refined calculations in serviceability and ultimate limit states. Use the results to refine the models used for SLS and ULS performance. For example, it might be possible to treat the structure in its actual condition. ' Optimized LCC. Time until a major repair and/or strengthening procedure is estimated using the bending stiffness development captured by curvature measurement.The approach using bending stiffness as a performance indicator is applied in two case studies in Sweden, the Panken road bridge located east of Karlstad and the railway bridge located in Örnsköldsvik. The Panken Bridge was located within phase (III) (crack stabilised phase), while the Örnsköldsviks Bridge was located within phase (I) (un-cracked phase). It is shown in these case studies that monitoring of the bending stiffness through curvature measurements can give valuable information regarding how the structure is affected by loads and/or degradation. One challenge when evaluating the bending stiffness from curvature measurements is that time dependent mechanisms, e.g. creep, could affect the curvature but not necessarily the bending stiffness. Time dependent mechanisms will thus give rise to what is here defined as a "fictitious stiffness change". Any movement or deformation which produces a fictitious stiffness change must be given extra attention to avoid misleading results. Another challenge is that monitoring is commonly performed for additional loading, which means that the curvature caused by the dead weight of the structure itself is in most cases not captured. Further research is suggested to address the effects of these phenomena for curvature assessment applications.