<p>Developing countries typically suffer far greater than developed countries as a result of earthquakes. Poor socioeconomic conditions often lead to poorly constructed homes that are vulnerable to damage during earthquakes. Literature review in this study highlights the lack of existing fragility curves for buildings in developing countries. Furthermore, fragility curves derived using empirical data are almost nonexistent due to the scarcity of post-earthquake damage data and insufficient ground motion recordings in developing countries. Therefore, this research proposes a methodology for developing empirical fragility curves using ground motion data in the form of USGS ShakeMaps.</p> <p>The methodology has been applied to a case study consisting of damage data collected in Bantul Regency, Indonesia in the aftermath of the May 2006 Yogyakarta earthquake in Indonesia. Fragility curves for non-engineered single-storey unreinforced masonry (URM) homes have been derived using the damage dataset for three ground motion parameters; peak ground acceleration (PGA), peak ground velocity (PGV), and pseudo-spectral acceleration (PSA). The fragility curves indicate the high seismic vulnerability of non-engineered URM homes in developing countries. There is a probability of 80% that a seismic event with a PGA of only 0.1g will induce significant cracking of the walls and reduction in the load carrying capacity of a URM home, resulting in moderate damage or collapse. Fragility curves as a function of PGA and PSA were found to reasonably represent the damage data; however, fits for several PGV fragility curves could not be obtained. The case study illustrated the extension of ShakeMaps to fragility curves, and the derived fragility curves supplement to the limited collection of empirical fragility curves for developing countries. Finally, a comparison with an existing fragility study highlights the significant influence of the derivation method used on the fragility curves. The diversity in construction techniques and material quality in developing countries, particularly for non-engineered cannot be sufficiently represented through simplified or idealized analytical models. Therefore, the empirical method is considered to be the most suitable method for deriving fragility curves for structures in developing countries.</p> / Master of Applied Science (MASc)

Chronic seismic hazards and resulting secondary impacts as natural conditions of the country, and loss of robust building and prudent settlement practices as aggravated by rapid population growth make cities the most vulnerable geographical and social entities in Turkey. In contrast, Turkish disaster policy is solely focused on post-disaster issues and no incentives or provision exist to encourage risk analysis or risk mitigation approaches, despite current international efforts. For the development of risk reduction policies an essential step is to prioritize settlements according to their vulnerability levels. This could be determined by hazard probabilities and attributes of the building stock of each settlement. Measurement of vulnerability levels allows the ordering of settlements into risk categories. Vulnerability levels of settlements are then assumed to depend on a number of attributes of cities to explore if vulnerability could be related to a set of urban properties. Results of statistical analyses indicate that total building loss is related to the ratio of population over the total number of buildings in mid-range settlements, and directly related to population in metropolitan cities. Relative loss on the other hand is related with rate of agglomeration and development index in almost every size category of settlements. Observations provide guiding principles for effective mitigation practices in Turkey by ordering settlements and offer means of differential implementation. These could contribute to improved safety measures in urban standards, building codes, building supervision procedures, insurance systems, investment priorities, and Law (6306) on Redevelopment of Areas under Disaster Risk.

The central focus of this dissertation is a seismic vulnerability assessment of retrofitted bridges. The objective of this work is to establish a methodology for the development of system level fragility curves for typical classes of retrofitted bridges using a probabilistic framework. These tools could provide valuable support for risk mitigation efforts in the region by quantifying the impact of retrofit on potential levels of damage over a range of earthquake intensities. The performance evaluation includes the development of high-fidelity three-dimensional nonlinear analytical models of bridges retrofit with a range of retrofit measures, and characterization of the response under seismic loading. Sensitivity analyses were performed to establish an understanding of the appropriate level of uncertainty treatment to model, assess, and propagate sources of uncertainty inherent to a seismic performance evaluation for portfolios of structures. Seismic fragility curves are developed to depict the impact of various retrofit devices on the seismic vulnerability of bridge systems. This work provides the first set of fragility curves for a range of bridge types and retrofit measures. Framework for their use in decision making for identification of viable retrofit measures, performance-based retrofit of bridges, and cost-benefit analyses are illustrated. The fragility curves developed as a part of this research will fill a major gap in existing seismic risk assessment software, and enable decision makers to quantify the benefits of various retrofits.

Chronic seismic hazards and resulting secondary impacts are due to the geological conditions of Turkey and the nature of current response mechanisms. Local know-how of building and settlement that evolved over centuries eroded with the growth in population, and the introduction of reinforced concrete building economics. This makes cities the most vulnerable geographical and social entities in Turkey. A basic formal reference of disaster management is the National Seismic Hazard Map indicating zones of hazard probabilities which are directly related to different measures in construction. This is hardly a sufficient disaster policy tool however, as cities may have very different risk profiles independent from the hazard probabilities. City level risk variations are not considered in the Seismic Hazard Map. This study intends to establish indicators for different risk levels in urban areas other then those implied by the National Seismic Hazard Map. Apart from local morphological and geological conditions, attributes of building stock, rates of unauthorized buildings and social conditions represent vulnerability indicators and could be effective in the determination of local risk levels. One specific description of risk levels is available in the obligatory reporting of the local authorities about the &quot / most likely level of disaster losses&quot / . This information, as an obligatory task of the governorates represents a local assessment of the most likely disaster losses and it is available from the GDDA. The city-level statistics of building stock on the other hand are available from the Turkish Statistical Institute. Correlation and Regression analyses are employed to determine what combinations of the independent variables might best denote city-level risks, and these may vary independently from their positions in the Hazard Map. The research may thus generate information for a more effective disaster policy.

The occurrence of two recent major earthquakes, 17 August 1999 Mw = 7.4 Izmit and 12 November 1999 Mw = 7.1 D&uuml / zce, in Turkey prompted seismologists and geologists to conduct studies to predict magnitude and location of a potential earthquake that can cause substantial damage in Istanbul. Many scenarios are available about the extent and size of the earthquake. Moreover, studies have recommended rough estimates of risk areas throughout the city to trigger responsible authorities to take precautions to reduce the casualties and loss for the earthquake expected. Most of these studies, however, adopt available procedure by modifying them for the building stock peculiar to Turkey. The assumptions and modifications made are too crude and thus are believed to introduce significant deviations from the actual case. To minimize these errors and use specific damage functions and capacity curves that reflect the practice in Turkey, a study was undertaken to predict damage pattern and distribution in Istanbul for a scenario earthquake proposed by Japan International Cooperation Agency (JICA). The success of these studies strongly depends on the quality and validity of building inventory and site property data. Building damage functions and capacity curves developed from the studies conducted in Middle East Technical University are used. A number of proper attenuation relations are employed. The study focuses mainly on developing a software to carry out all computations and present results. The results of this study reveal a more reliable picture of the physical seismic damage distribution expected in Istanbul.

[Truncated abstract] Perth is the capital city of Western Australia (WA) and the home of more than three quarters of the population in the state. It is located in the southwest WA (SWWA), a low to moderate seismic region but the seismically most active region in Australia. The 1968 ML6.9 Meckering earthquake, which was about 130 km from the Perth Metropolitan Area (PMA), caused only minor to moderate damage in PMA. With the rapid increase in population in PMA, compared to 1968, many new structures including some high-rise buildings have been constructed in PMA. Moreover, increased seismic activities and a few strong ground motions have been recorded in the SWWA. Therefore it is necessary to evaluate the seismic risk of PMA under the current conditions. This thesis presents results from a comprehensive study of seismic risk of PMA. This includes development of ground motion attenuation relations, ground motion time history simulation, site characterization and response analysis, and structural response analysis. As only a very limited number of earthquake strong ground motion records are available in SWWA, it is difficult to derive a reliable and unbiased strong ground motion attenuation model based on these data. To overcome this, in this study a combined approach is used to simulate ground motions. First, the stochastic approach is used to simulate ground motion time histories at various epicentral distances from small earthquake events. Then, the Green's function method, with the stochastically simulated time histories as input, is used to generate large event ground motion time histories. Comparing the Fourier spectra of the simulated motions with the recorded motions of a ML6.2 event in Cadoux in June 1979 and a ML5.5 event in Meckering in January 1990, provides good evidence in support of this method. This approach is then used to simulate a series of ground motion time histories from earthquakes of varying magnitudes and distances. ... The responses of three typical Perth structures, namely a masonry house, a middle-rise reinforced concrete frame structure, and a high-rise building of reinforced concrete frame with core wall on various soil sites subjected to the predicted earthquake ground motions of different return periods are calculated. Numerical results indicate that the one-storey unreinforced masonry wall (UMW) building is unlikely to be damaged when subjected to the 475-year return period earthquake ground motion. However, it will suffer slight damage during the 2475-return period earthquake ground motion at some sites. The six-storey RC frame with masonry infill wall is also safe under the 475-year return period ground motion. However, the infill masonry wall will suffer severe damage under the 2475-year return period earthquake ground motion at some sites. The 34-storey RC frame with core wall will not experience any damage to the 475-year return period ground motion. The building will, however, suffer light to moderate damage during the 2475-year return period ground motion, but it might not be life threatening.

Ces travaux, menés en région PACA, visent à comprendre comment amener les individus à se protéger du risque sismique. Dans ce cadre, nous avons exploré les effets comportementaux de la substitution de soi à autrui, puis observé les comportements effectifs de protection lors d’une étude de terrain. En interprétant la consigne de substitution en termes de comparaison sociale, le paradigme repose sur une distinction entre la contrôlabilité des effets d’un séisme et l’incontrôlabilité de son occurrence. Après avoir activé un élément central ou périphérique de la représentation sociale des séismes, il s’agit dans un premier temps d’inviter les individus à évaluer la probabilité des effets d’un séisme en consigne de substitution, et ensuite à évaluer la probabilité d’occurrence d’un séisme en consigne standard. Sous l’effet d’un biais d’optimisme comparatif, les individus déclarent systématiquement une probabilité des effets d’un séisme supérieure en consigne de substitution qu’en consigne standard. De même, selon un principe de consistance cognitive, ils déclarent une probabilité d’occurrence d’un séisme supérieure après avoir évalué la probabilité de ses effets selon une consigne de substitution. Les comportements de protection étaient ensuite mesurés à l’aide de la version française de l’échelle de préparation aux séismes. Les comportements de protection auto-déclarés sont significativement plus favorables suite à une réponse en consigne de substitution plutôt que standard. Toutefois, les effets du paradigme sur les comportements effectifs de protection sont plus mitigés. / The goal of this research, conducted in the PACA region, was to understand how to lead individuals to adopt earthquake protection behaviour. To this end, we explored the behavioural effects of self-other substitution, and observed actual protection behaviours in a field study. By interpreting self-other substitution in terms of social comparison, the paradigm relies on a distinction between the uncontrollability of an earthquake and the controllability of its consequences. After having activated a central or peripheral element of the social representation of earthquakes, participants are required to evaluate the probability of an earthquake’s effects in a substitutive context, and then the probability of an earthquake in a standard context. Comparative optimism leads participants to declare that the effects of an earthquake are more probable in a substitutive rather than standard context. Furthermore, a cognitive consistency principle seems to lead participants to declare that an earthquake is more likely after having evaluated the probability of its effects in a substitutive rather than standard context. Protection behaviours were measured using the French version of the Earthquake Readiness Scale. Self-declared protection behaviours were more favourable after a substitutive rather than standard answer. However, the effects on actual behaviours are mixed.

<div>Ensuring the resilience of a state’s transportation network is necessary to guarantee an acceptable quality of life for the people the network serves. A lack of resilience in the wake of a seismic event directly impacts the states’ overall safety and economic vitality. With the recent identification of the Wabash Valley Seismic Zone (WBSV), Department of Transportations (DOTs) like Indiana’s have increased awareness for the vulnerability of their bridge network. The Indiana Department of Transportation (INDOT) has been steadily working to reduce the seismic vulnerability of bridges in the state in particular in the southwest Vincennes District. In the corridor formed by I-69 built in the early 2000s the bridge design is required to consider seismic actions. However, with less recent bridges and those outside the Vincennes District being built without consideration for seismic effects, the potential for vulnerability exists. As such, the objective of this thesis is to develop a robust seismic vulnerability assessment methodology which can assess the overall vulnerability of Indiana’s critical bridge network. </div><div><br></div><div>A representative sample of structures in Indiana’s bridge inventory, which prioritized the higher seismic risk areas, covered the entire state geographically, and ensured robust superstructure details, was chosen. The sample was used to carry a deterministic seismic vulnerability assessment, applicable to all superstructure-substructure combinations. Analysis considerations, such as the calculation of critical capacity measures like moment-curvature and a pushover analysis, are leveraged to accurately account for non-linear effects like force redistribution. This effect is a result of non-simultaneous structural softening in multi-span bridges that maintain piers of varying heights and stiffnesses. These analysis components are incorporated into a dynamic analysis to allow for the more precise identification of vulnerable details in Indiana’s bridge inventory.</div><div><br></div><div>The results of this deterministic seismic assessment procedure are also leveraged to identify trends in the structural response of the sample set. These trends are used to identify limit state thresholds for the development of fragility functions. This conditional probabilistic representation of bridge damage is coupled with the probability of earthquake occurrence to predict the performance of the structure for a given return period. This probabilistic approach alongside a Monte Carlo simulation is applied to assess the vulnerability of linked bridges along key-access corridors throughout the state. With this robust seismic vulnerability methodology, DOTs will have the capability of identifying vulnerable corridors throughout the state allowing for the proactive prioritization of retrofits resulting in the improved seismic performance and resiliency of their transportation network.</div>

Dans le contexte du calcul sismique basé sur la notion de performance, lesingénieurs se trouvent confrontés à une tâche difficile pour estimer la performance etévaluer les risques des systèmes sol-structure en interaction. Afin d’accomplir cette tâcheavec succès, toutes les sources d'incertitudes aléatoires et épistémiques doivent être prisesen compte au cours du processus de conception. Ainsi, des méthodes appropriées sontnécessaires pour l'étude de la propagation de l'incertitude des paramètres du systèmedécrivant la structure, le sol, et les charges appliquées aux réponses structurelles endéfinissant des états limites de performance. L’objectif de cette thèse est de contribuer àl’étude du comportement sismique des structures en interaction avec le sol et d’offrir denouveaux outils pour le traitement de problèmes pertinents, orientés vers la nouvellephilosophie de conception parasismique des structures : la conception basée sur laperformance (performance-based design). Cet axe de recherche structure ce travail dedoctorat. La problématique s’inscrit dans le cadre de proposition de modèles simplifiés afind’aborder un problème compliqué tel que l’interaction sol-structure (ISS). Lecomportement non linéaire de la structure est déterminé par une approche capacitive baséesur la performance sismique telle que la méthode N2 proposée par P. Fajfar. Dans saversion originale, cette méthode considère que la structure est encastrée à sa base,négligeant ainsi l’ISS. Il s'agit d'une extension de la méthode N2 et que nous appelleronsN2-ISS. A notre connaissance, il n'existe pas d'études qui ont intégré l'ISS dans ce type deformalisme. Pour examiner la validité et la fiabilité du modèle présenté, une analysecomparative a été faite entre l'approche que nous proposons et trois autres méthodes: 1) laméthode introduite dans le code BSSC 1997, 2) la méthode proposée par Avilès & Perez–Rocha (2003) ainsi que 3) la méthode dynamique temporelle non linéaire. Les résultatsobtenus ont montré que la réponse en déplacement de la structure était assez proche dansles quatre méthodes.Les courbes de fragilité sont établies en tenant compte des effets de l’ISS et desincertitudes associées au chargement (mouvement du sol), aux propriétés de la structure,du sol, et de la fondation (impédances, ...). L’incertitude épistémique est égalementconsidérée de manière indirecte suite à la comparaison entre deux procédures d'évaluationdes états d’endommagements (Méthode de Park & Ang et Méthode de RISK-UE). L'étudea abouti aussi à une caractérisation préliminaire du risque sismique dans une partie de laville d'Oran dont les caractéristiques géologiques et géotechniques étaient disponibles.Cette étude a permis la cartographie du dommage et d'étudier la la vulnérabilité sismiquedes bâtiments. / In the context of performance-based earthquake engineering (PBEE), a challengingtask for structural engineers is to provide performance and risk assessment for structures orsoil-structure interaction (SSI) systems. In order to fulfill this task successfully, all relevantsources of aleatory and epistemic uncertainties must be accounted for during the designprocess. Thus, proper methods are required for the study of uncertainty propagation frommodel parameters describing the structure, the soil, and the applied loads to structuralresponses by defining some performance limit states. The objective of this thesis is tocontribute to the study of the seismic behavior of structures interacting with soil andprovide new tools for the treatment of relevant issues facing the new philosophy of seismicdesign of structures: performance-based design. The objective of this thesis is to contributeto the study of the seismic behavior of structures interacting with soil and provide newtools for the treatment of relevant issues facing the new philosophy of seismic design ofstructures: (performance-based design). This research structure this doctoral work. Theissue is part of proposed simplified models to address a complicated problem such as soilstructureinteraction (SSI). The nonlinear behavior of the structure is determined by acapacitive approach based on the seismic performance as N2 method proposed by P.Fajfar. In its original version, this method considers that the structure is fixed at its base,thus neglecting the ISS. The new proposed method called N2-SSI is an extension of the N2method. To our knowledge, there are no studies that have joined the SSI ISS in this type offormalism. The proposed approach is validated and compared with time history analysis,Building Seismic Safety Council (BSSC) method (NEHRP, 2003), and a method proposedby Aviles and Perez-Rocha (2003). The results obtained showed that the responsedisplacement of the structure was fairly close in the four methods.The fragility curves are established taking into account the effects of the ISS anddifferent uncertainty sources: the load (input ground motion), the soil, the structure and theSSI (impedances, ...). Epistemic uncertainty was investigated through comparison betweentwo different approaches in assessing damage states (Park and Ang and Risk-UE). Thestudy also resulted in a preliminary characterization of the seismic risk in a part of the Orancity, where geological and geotechnical characteristics were available. This study allowedthe mapping of the damage and the study of the seismic vulnerability of buildings.

La concentration de plus en plus importante de la population dans les milieux urbains exposés à une forte sismicité peut générer de plus en plus de dommages et de pertes. La réponse sismique en milieu urbain dépend des effets du site (direct amplification et non-linéarité du sol) et du couplage entre le sol et les structures (interaction sol-structure et site-ville). Par conséquent, la compréhension de la sismologie urbaine, c'est-à-dire le mouvement du sol intégrant l'environnement urbain, est critique pour réduire les dommages. Cela passe par la prédiction du mouvement du sol dans le milieu urbain, ingrédient fondamental à l'évaluation de l'aléa sismique. La prise en compte de l'amplification provoquée par la présence de sédiments est largement étudiée. Au contraire, la réponse non-linéarité du sol et du couplage entre le sol et la structure est rarement intégrée à la prédiction du mouvement du sol. A cause de leur complexité, ces problèmes ont toujours été abordés séparément. Dans ce contexte, cette thèse analyse la réponse non-linéaire du système sol-structure en intégrant la non-linéarité du sol et de l'interaction sol-structure. Deux travaux expérimentaux ont été conduits, avec comme but de proposer un proxy, rendant compte de la non-linéarité du sol. Le premier est l'essai en centrifugeuse qui reproduit à échelle réduite la réponse du sol et des structures. L'état de contrainte et de déformation est conservé en appliquant une accélération artificielle au modèle. Cet essai a été effectué à IFSTTAR Nantes dans le cadre de l'ANR ARVISE. Différentes configurations ont été testées, avec et sans bâtiments, sous différents niveaux de sollicitation, pour analyser la réponse du sol et des structures. Le deuxième utilise les enregistrements des réseaux accélérométriques verticaux de deux sites tests californiens : Garner Valley Downhole Arrat (GVDA) et Wildlife Liquefaction Array (WLA), gérés tout deux par l'Université de Californie, Santa Barbara (UCSB), Etats-Unis. La réponse in-situ est importante car elle décrit le comportement réel du site. Plusieurs informations décrivant les conditions de sites sont disponibles et les séismes enregistrés ont permis de tester plusieurs niveaux de déformations pour reconstruire la réponse globale de chaque site. De plus, le site GVDA est équipé d'une structure Soil-Foundation-Structure-Interaction (SFSI) qui a comme objectif d'étudier les problèmes d'interaction sol-structure. Dans les deux expériences, grace au réseau accélérométrique vertical dans le sol et la structure, on peut appliquer la méthode de propagation d'ondes 1D pour extraire la réponse de ces systèmes. Les ondes sont considérées comme des ondes SH qui se propage horizontalement dans une couche 1D. La méthode interférométrie sismique par déconvolution est appliquée pour extraire l'Impulse Response Function (IRF) du système 1D. On analyse ainsi la variation de Vs en fonction de la solliictation et à différente position dans le sol ainsi que la variation des éléments expliquant la réponse dynamique du système sol-structure. On propose au final un proxy de déformation permettant de rendre compte mais aussi de prédire la nonlinéarité des sols en fonction des niveaux sismiques subits. / The concentration of population in urban areas in seismic-prone regions can generate more and more damages and losses. Seismic response in urban areas depends on site effects (direct amplification and nonlinearity of the soil) and the coupling between the soil and structures (soil-structure and site-city interaction). Therefore, the understanding of urban seismology, that is the ground motion incorporating the urban environment, is critical to reduce the damage. This requires the prediction of ground motion in urban areas, a fundamental element in the evaluation of the seismic hazard. Taking into account the amplification caused by the presence of sediment has been widely studied. However, the non-linearity of the soil and the coupling between the ground and the structure is seldom integrated to the prediction of the ground motion. Because of their complexity, these problems have been addressed separately. In this context, this dissertation analyzes the non-linear response of the soil-structure by integrating the non-linearity of the soil and the soil-structure interaction. Two experimental studies were performed, with the aim of providing a proxy that reflects the non-linearity of the soil. The first is the centrifuge test that reproduces the response of soil and structures at reduced scale. The state of stress and strain is conserved by applying an artificial acceleration model. This test was performed at IFSTTAR Nantes in the framework of the ANR ARVISE. Different configurations were tested with and without buildings, under different stress levels, to analyze the response of the soil and structures. The second uses the vertical accelerometric networks of two sites in California: Garner Valley Downhole (GVDA) and the Wildlife Liquefaction Array (WLA), both managed by the University of California, Santa Barbara (UCSB), USA. In-situ response is important since it describes the actual behavior of the site. Information describing the conditions of sites is widely available and the earthquakes recorded were used to test several levels of shaking to reconstruct the overall response of each site. In addition, the GVDA site is equipped with a Soil-Foundation-Structure-Interaction structure (SFSI) which aims to study the problems of soil-structure interaction. In both experiments, thanks to the vertical accelerometer network in the ground and the structure we are able to apply the 1D wave propagation method to extract the response of these systems. The waves are considered as an SH wave which propagates in a 1D horizontal layer. Seismic interferometry by deconvolution method is applied to extract the Impulse Response Function (IRF) of the 1D system. Thus the analysis of the variation in function of elastic properties of the soil and the structure is done under several magnitude of shaking, including variation in depth and the elements of the total response of the structure including the soil-structure interaction. At the end, a deformation proxy to evaluate and also to predict the nonlinear response of the soil, the structure and the soil-structure interaction is proposed.