Structural behaviour of deteriorated and retrofitted concrete structures
MetadataShow full item record
Deterioration and ageing of concrete structures and the increased traffic intensities and loads are some of the major problems facing civil engineers, industry and researchers. Corrosion of embedded steel reinforcement is one of the main causes of deterioration. Deterioration of reinforced concrete may lead to a number of undesirable consequences such as loss of serviceability, loss of load carrying capacity and reduction in safety of structures and traffics. Steel bar corrosion affects the reinforcement itself, the surrounding concrete and the composite action between steel and concrete. The most critical effect is probably the reduction in cross sectional area of the affected bar. If corrosion is allowed to propagate over a sufficiently long period of time and proper interventions not are made, loss of bond may also lead to impaired stiffness and strength of affected structural components. The main objective of this PhD project is to develop methodology for assessment of deteriorating concrete structures based on nonlinear finite element simulations. For the numerical simulations to be accurate it is required that constitutive and kinematical models represent realistic approximations to the true behaviour of reinforced concrete structures attacked by reinforcement corrosion. Hence, these models must be based on existing experimental evidence of the structural consequences of reinforcement corrosion. However the available experimental data on steel bar corrosion have been obtained on small scale specimens corroded and tested in the laboratory and showed considerable scatter in the published results. The first part of the thesis deals with presentation of reliable information about change in the maximum bond strength between reinforcement and surrounding concrete and bond-slip behaviour due to corrosion. Available data from published laboratory investigations are collected and compared. Some analytical models for calculation of the residual bond strength and some bond-slip models for uncorroded and corroded bars used in the numerical simulations are presented and discussed in the thesis. An attempt to analyse the effect of the current density on the bond strength and on the residual load-carrying capacity of the concrete beams with corroded reinforcement is also performed. The second part of the PhD project is focused on validation and verification of the nonlinear finite element analysis of deteriorated and repaired concrete structures through a numerical test program. As a first step, simulations of laboratory experiments on small scale beams are performed to verify the finite element model proposed in this project. Accordingly, loss of steel bar section, and reduced bond between deteriorated concrete and corroded rebar are accounted for in the present work. The load-deflection behaviour obtained from the finite element simulations are generally in good agreement with the test data. As a second step in the numerical program, analyses of a full scale beam are carried out. A close approximation to the true sequence of loading, deterioration, unloading, repair and reloading until failure, a so-called service life cycle of concrete beams, are simulated using the phase analysis approach in the finite element software DIANA. The results of the phased finite element simulation showed that the failure load of the repaired beam was larger than the experimental and estimated probabilistic values. The central deflection of the concrete beam from the finite element analysis was significantly lower than in the experimental study at both the corrosion and repair stages. Next, numerical simulations of the service life cycle of a concrete beam without strengthening, where time-dependent effects such as creep and shrinkage were included in the constitutive models for concrete and repair material, were carried out. The results of the simulations showed that the failure load is in good agreement with the failure load corresponding to the upper bound of the 95 % confidence interval obtained in a previous probabilistic study. Initial strains due to shrinkage of the concrete and the repair material were implemented in the numerical analyses according to a cross section type model, and a model for development of free shrinkage strain in beams exposed to one-side drying. Total strain variations across the cross-section of the beam calculated from the finite element analyses are compared to data available from the experimental study. Finally, effect of some parameters influencing the results of the numerical simulations and the agreement between the laboratory and numerical results are presented.