Buildings with torsional irregularity represent the main focus of many current investigations. However, despite this volume of research, there is no established framework that describes adequately the seismic vulnerability of reinforced concrete shear wall systems. In this study, the three-dimensional behavior of a particular shear-wall structure under earthquake effects was examined with regard to the nonlinear behavior of the reinforced concrete assembly and the parameters that characterize the structure exposed to seismic motion for damage assessment. A three story reinforced concrete shear-wall building was analyzed using the finite element method based ANSYS software. The scaled model building was subjected to shaking table tests at Saclay, France. The project was led by the Atomic Energy Agency (CEA Saclay, France) under the &ldquo / SMART 2008 Project.&rdquo / The investigation was conducted in two phases. In the first phase, the results of the finite element method and experiments were examined, and were reported in this study. For time history analysis, micro-modeling was preferred due to allowing inclusion the nonlinear effects of concrete and steel for analysis. The guiding parameters (acceleration, displacement, strain) of analytical results are compared with the corresponding values that were measured in the experiments to be able to quantify the validity of models and simulation. For the comparison of v the numerical model results with the experimental results FDE (Frequency Domain Error) method was used. After comparison of the numerical model results with the experimental results, the second phase of the SMART 2008 Project was undertaken. The second phase consisted of two parts summarized as &ldquo / Sensitivity Study&rdquo / and &ldquo / Vulnerability Analyses&rdquo / . However, in this report only the sensitivity study and fragility analyses will be reported. Sensitivity study was done to understand which parameters affect the response of the structure. Twelve parametric cases were investigated under two different ground motions. Different behavior parameters were investigated. The effective damping coefficient was found to affect the structural response at 0.2 g design level as well as at 0.6 g over-design level. At the design level, it was observed that elasticity modulus of concrete and additional masses on the specimen determined as effective on the calculated results. To derive the failure probabilities of this structure under various earthquake forces for the given limit states, fragility curves were obtained. Different seismic indicators such as PGA (Peak ground acceleration), PGV (Peak ground velocity), PGD (Peak ground displacement) and CAV (Cumulative absolute velocity) were used as seismic indicators and MISD (Maximum interstory drift) were used as damage indicator for fragility curves. In all 30 time history analyses were done. Regression analyses using least squares method were performed to determine the median capacity and its deviation. Correlation coefficients of the time history data versus fitted curves obtained from the regression analyses changes between 0.65 and 0.99. The lower cases were for PGD- MISD graphs. The scatter of the fragility curves calculated for each damage limit was slightly wider. HAZUS MH MR1 (2003) damage states were also used for the calculation of the fragility curves and compared with the SMART 2008 damage states.

Cette thèse présente trois méthodes pour l’identification des rigidités des structures d’usage commun dans l’ingénierie civile, à partir de données dynamiques expérimentales. La première méthode est développée pour des structures composées pour portiques. La deuxième méthode proposée est appliquée à des structures constituées pour des poutres isostatiques. La troisième est une méthodologie d’estimation des rigidités en flexion (EI) et au cisaillement (GA/?) pour une structure constituée de murs dont les énergies de déformation en flexion et cisaillement peuvent être soit du même ordre de grandeur, soit l’une prépondérante par rapport à l’autre. Pour chaque méthode, des simulations numériques sont effectuées pour identifier les dommages structuraux ou les variations des rigidités, en termes de localisation et de magnitude de ces dommages. L'incidence et l'impact des erreurs et bruits sur les valeurs estimées des rigidités structurales sont analysés. Les méthodologies sont également appliquées pour localiser des dommages mécaniques ou des réductions de section sur modèles de laboratoire. A partir des concepts dynamiques de base et considérant une typologie donnée de structure, la thèse développe les concepts et formulations permettant d’identifier les rigidités résiduelles des structures considérées. Les méthodes peuvent être aisément mises en oeuvre pour déterminer les éventuels dommages (localisation et intensité) qui peuvent affecter une structure, par exemple après un séisme. Peu de mesures sont requises à cet effet : des essais de vibration libre et du matériel peu onéreux de mesures sont amplement suffisants dans le cas particulier des structures étudiées / This thesis presents three methods to estimate and locate damage in framed buildings, simply-supported beams and cantilever structures, based on experimental measurements of their fundamental vibration modes. Numerical simulations and experimental essays were performed to study the effectiveness of each method. A numerical simulation of a multi-storey framed building, a real bridge and a real chimney were carried out to study the effectiveness of the methodologies in identifying damage. The influence of measurement errors and noise in the modal data was studied in all cases. To validate the experimental effectiveness of the damage estimation methods, static and dynamics tests were performed on a framed model, a simply supported beam, and a cantilever beam in order to determine the linear behavior changes due to the increase of the level of damage. The structural identification algorithms during this thesis were based on the knowledge type of the stiffness matrix or flexibility matrix to reduce the number of modal shapes and required coordinates for the structural assessment. The methods are intended to develop tools to produce a fast response and support for future decision procedures regarding to structures widely used, by excluding experimental information, thereby allowing a cost reduction of extensive and specific testing

This thesis presents three methods to estimate and locate damage in framed buildings, simply-supported beams and cantilever structures, based on experimental measurements of their fundamental vibration modes. Numerical simulations and experimental essays were performed to study the effectiveness of each method. A numerical simulation of a multi-storey framed building, a real bridge and a real chimney were carried out to study the effectiveness of the methodologies in identifying damage. The influence of measurement errors and noise in the modal data was studied in all cases. To validate the experimental effectiveness of the damage estimation methods, static and dynamics tests were performed on a framed model, a simply supported beam, and a cantilever beam in order to determine the linear behavior changes due to the increase of the level of damage. The structural identification algorithms during this thesis were based on the knowledge type of the stiffness matrix or flexibility matrix to reduce the number of modal shapes and required coordinates for the structural assessment. The methods are intended to develop tools to produce a fast response and support for future decision procedures regarding to structures widely used, by excluding experimental information, thereby allowing a cost reduction of extensive and specific testing