This study is an evaluation on the seismic behavior of the Nahrain Dam, which is located in Khorasan province in the central part of Iran. The finite element method is used for the dynamic analysis of the dam. Using a slope stability computer program, the critical slip surface near the crest of the dam and the yield acceleration corresponding to this slip surface are determined. Static analysis was completed by using the finite element computer program SAP90 and SAP2000 in order to determine the stress conditions within the body of the dam prior to earthquake. The stresses obtained from this analysis were used in the assessment of the dynamic material properties of the dam. Two near by actual earthquakes were modified and used as input motions of different magnitudes. At the next step, the harvested data was used as input data for the program TELDYN to perform dynamic analysis. Permanent displacements under the scenario base motions were calculated by using the Newmark&rsquo / s method.

A new reliability-based probabilistic model is developed for the estimation of liquefaction-induced lateral ground spreading, taking into consideration the uncertainties within the model functional form and the descriptive variables as well. The new model is also introduced as performance-based probabilistic engineering tool.

Seafloor hydrothermal systems play a key role in Earth s energy and geochemical budgets. They also support the existence and development of complex chemosynthetic biological ecosystems that use the mineral-laden fluids as a source of energy and nutrients. This dissertation focuses on two inter-related topics: (1) the heat output at mid-ocean ridge (MOR) hydrothermal sites, and (2) the hydrothermal response to earthquakes at mid-ocean ridges. The response of mid-ocean ridge hydrothermal systems to seismic and magmatic activity could provide a means of using seafloor observations to assess processes occurring at crustal depths. Three decades of study of the ocean floor have made clear that tectonic/volcanic perturbations and changes in the hydrothermal activity are correlated. Yet, this relationship is still highly misunderstood. In this dissertation, we developed a mathematical model of hydrothermal circulation affected by magmatic and/or tectonic events occurring at depth. We showed that the perturbations they generate need not propagate through the entire upflow zone, as previously suggested, but need only affect the movement of the boundary layers. In this case, the fluid residence time in the discharge zone is on the order of years. The determination of heat output at MOR hydrothermal sites provides important constraints on the physics of these processes. Yet, such measurements are still very limited, available only for ~30 sites out of the ~10^3 estimated worldwide. This dissertation reports the first measurements of hydrothermal heat output at 9°50' N on the East Pacific Rise (EPR), which has been the most intensively studied site for the past 20 years (~10^2 expeditions). The values we obtained helped constraining the afore-mentioned mathematical model, which was tested based on the March 1995 microearthquake swarm near 9°50' N, EPR. Our results showed that the current interpretation based on a thermal cracking episode at the bottom of the upflow zone, may be incorrect. We argue that a diking scenario better explains the seismic and temperature data, while being consistent with the mechanics of inflation of the axial magma chamber in the 9°50' N, EPR area between two major eruptions, in 1991 and 2006.

Le terme variabilité spatiale des mouvements sismiques (SVGM en anglais pour Spatial Variability of Ground Motion) désigne les différences entre deux mesures du mouvement du sol effectuées à différents endroits, généralement en surface. La modélisation de SVGM ainsi que son effet sur la réponse dynamique des barrages est nécessaire pour l'intégration du phénomène dans les codes parasismiques. L'étude actuelle présente une mesure et une enquête approfondie sur SVGM à l'interface barrage voûte - fondation rocher. Des mesures in situ sont utilisées pour sa quantification et des simulations numériques pour la compréhension plus approfondie des phénomènes physiques qui contribuent à SVGM notamment à l'interface, c-à-d la topographie de la voûte et l'interaction sol-structure. Les données in situ provient d'une campagne sismologique qui a eu lieu sur et autour le barrage voûte du Saint Guérin pendant six mois. Le sous-ensemble d'événements consiste des événements de faible à moyenne magnitude, locaux et régionaux. Ainsi, des analyses linéaires sont permises. Tout d'abord, l'analyse dynamique du barrage voûte est réalisée; les fréquences de vibration, le coefficient d'amortissement et l'amplification de crête sont estimées profitant des mesures continues de bruit ambiant et des enregistrements sismiques. Ensuite, le SVGM est quantifiée au moyen de la phase et de la variabilité d'amplitude en utilisant des estimations de cohérence et de l'écart type de la différence des amplitudes des spectres de Fourier respectivement. Forte variabilité est observée à la fois en phase et en amplitude à l'interface de barrage-fondation rocher. Une fois que les estimateurs de la variabilité sont obtenus à partir des données, les modèles paramétriques existants sont comparés avec eux. Accentuation est mis sur deux observations: 1) les mouvements sismiques au champ libre semblent être légèrement moins variable par rapport aux mouvements à l'interface barrage-fondation rocher et 2) à l'interface barrage-fondation rocher, il y a une variabilité plus forte autour des fréquences de vibration du barrage. Ces observations suggèrent que la présence de la structure ainsi que la topographie du canyon augmentent SVGM. Cette hausse semble cependant être faible étant donné que les observations sont montés par des modèles paramétriques satisfaisants basés sur des données provenant de réseaux sismiques plates (sans topographie et structure). Des simulations numériques dans le code SPECFEM3D, basé sur les éléments spectraux, sont utilisées pour étudier plus profondément les observations in situ par le découplage des différentes causes de SVGM et évaluer l'impact de chacune. Une étude paramétrique en utilisant une topographie du canyon simplifié tente d'identifier l'effet de la topographie du canyon local sur SVGM tandis qu'un modèle géométrique précis du barrage voûte à Saint Guérin et sa topographie du canyon nous donne une idée plus précise sur l'impact de l'interaction barrage-fondation rocher sur SVGM. Les résultats de cette recherche devraient contribuer à l'amélioration de notre compréhension de SVGM à l'interface barrage-fondation rocher et de proposer des modèles de variabilité utilisés dans la conception de barrages voûtes. / Spatial variability of seismic ground motions (SVGM) denotes the differences between two time histories of the ground motion recorded at different locations, generally at the ground surface. The modeling of SVGM and the understanding of its influence on the dam's response are necessary so as design codes start to incorporate its effects in their provisions. This study presents a measure and a profound investigation of SVGM at the dam-foundation rock interface of an arch dam. In-situ measurements are used to quantify SVGM and numerical simulations to deeper understand the particular physical phenomena that contribute to SVGM at the interface, i.e. local canyon topography and rock-structure interaction. The in-situ data comes from a seismological experimental campaign that has taken place on and around Saint Guérin arch dam over the period of six months. The campaign was held in the framework of the present thesis. The subset of events consists of low to moderate magnitude local and regional earthquakes. Thus, analysis is allowed in the linear range. Firstly, dynamic analysis of the arch dam is conducted; the frequencies of vibration, the damping coefficient and the crest amplification are estimated based on ambient noise and seismic records. Then, SVGM is quantified by means of phase and amplitude variability using coherency estimates and standard deviation of difference of Fourier amplitudes respectively. High variability is observed both in phase and amplitude at the dam-foundation rock interface. Once the estimators of variability are obtained from the data, parametric models are fitted to them. Focus is given on two observations : 1) the ground motions in the free field appear to be slightly less variable with respect to the motions at the dam-foundation rock interface and 2) at the dam-foundation rock interface, there is higher variability around the frequencies of vibration of the dam. These observations suggest that the presence of the structure along with the canyon topography increase SVGM. This increase though seems to be small given that the observations are satisfactory fitted by parametric models based on data coming from flat seismic arrays. Numerical simulations in the SPECFEM3D code, based on the spectral element method are used to deeper investigate the in-situ observations by decoupling the various causes of SVGM and evaluating the impact of each one. A parametric study using a simplified canyon topography attempts to identify the effect of local canyon topography on SVGM while a geometrically accurate model of the Saint Guérin arch dam and its canyon topography gives us a better insight on the dam-foundation rock interaction impact on SVGM. Although both features are found to increase SVGM, their impact remains secondary. The findings of the present research are expected to contribute in enhancing our understanding of SVGM at the dam-foundation rock interface and proposing variability models used in arch dams' design.

Les séismes intermédiaires (40-325 km de profondeur) ont été largement documentés dans les plaques océanique en subduction mais leur mécanismes déclencheurs restent énigmatiques et très peu compris en raison (I) des incertitudes instrumentales sur l'acquisition des données géophysiques et (II) de la rareté des exemples de roches métamorphisées dans des conditions du faciès éclogites et préservant (sans ambiguïté) les structures produites par des séismes intermédiaires. Bien que toujours limité, il y a de plus en plus de preuves que les lambeaux « fossiles » de lithosphère océanique exhumée de taille pluri-kilométrique peuvent enregistrer les processus chimiques et mécaniques caractéristiques de la sismicité de profondeur intermédiaire. Ce projet de doctorat étudie l'impact de la libération et de l'infiltration des fluides métamorphiques sur la génération des brèches éclogitiques disséminées dans une zone de cisaillement de 15 km de long, exposée dans un fragment presqu’intact de lithosphère océanique Tethysienne, subduite jusqu'à 80 km de profondeur (2.6 GPa - 550 ° C, unité du Lago Superiore): le complexe métaophiolitique du Mont Viso (W Alps, Italie). Trois zones de cisaillement majeures ont été étudiées, et en particulier la Lower Shear Zone (LSZ), dans laquelle des blocs de métagabbros mylonitiques éclogitisés et brechifiés (potentiellement lors d’une phase sismique) sont dispersés, avec des blocs métasédimentaires, dans la matrice serpentineuse de la zone de cisaillement.Dans cette dernière zone de cisaillement, une attention particulière a été portée à la caractérisation pétrologique et structurale de 196 blocs exhibant ces brèches (dans lesquels plus de 100 échantillons ont été récoltés), notamment à travers l’étude (i) de leur répartition dans la zone de cisaillement et (ii) de leurs caractéristiques morphologiques (longueur, largeur, hauteur, volumétrie relative de la matrice par rapport aux clastes et nature des clastes). Ces données ont été synthétisées dans une nouvelle carte géologique détaillée des méta-ophiolites du Mont Viso à l’échelle 1 : 20.000. L’étude pétro-structurale a permis de montrer que la formation des brèches résulte de passages transitoires de la déformation ductile à cassante dans le faciès éclogitique, comme le montre la foliation mylonitique dues métagabbros (composée de l’assemblage omphacite + rutile ± grenat et quartz) recoupé par des plans de brèches cimentées par des matrices riches en omphacite ± grenat et lawsonite. Elle montre également que la formation des brèches n'est pas liée à des événements pré-alpins (brèches sédimentaires ou tectoniques superficielles) comme d’autres auteurs l’ont proposé (Balestro et al. 2013; Festa et al. 2015). L'analyse des éléments en trace (in-situ et dans la roche totale), en lien avec une caractérisation rigoureuse des microstructures dans les blocs de brèches, a permis de mettre en évidence un changement progressif des fluides circulant pendant la bréchification. Les premières ruptures fragiles (locales, M1) ont été déclenchées par des fluides dérivés localement (e.g. metagabbros) avec une injection progressive de fluides «exotiques» (dérivés de la déshydratation de serpentinite) provoquant la bréchification dans l'intégralité des LSZ and ISZ (avec la cristallisation de M2 et M3). En raison de l'extension limitée des affleurements, l'origine sismique des brèches eclogitiques reste spéculative. Néanmoins, plusieurs preuves (par exemple, les minéraux fracturés et décalés le long des niveaux micrométriques riches en omphacite) suggèrent que les brèches d'eclogite du Monviso ont été générées par rupture fragile instantanée. / Intermediate-depth earthquakes (40-325 km depth) have been extensively documented within subducting oceanic slab but their triggering mechanisms remain enigmatic due to (I) the instrumental uncertainties on geophysical data acquisition and (II) the scarcity of examples of exhumed (and unambiguously recognized) eclogite-facies earthquake-derived rocks. Although still limited, there is growing evidence that large-scale, “fossil” exhumed portions of subducted lithosphere may record both chemical and mechanical processes operating in the depth range of intermediate-depth seismicity. This thesis project investigates the role of metamorphic fluids in the formation of the eclogitic breccias (resulting from potentially seismic deformation) found in a 15 km-long shear zone developed in an almost intact fragment of Tethyan oceanic lithosphere metamorphosed to eclogite-facies peak metamorphic conditions (2.6 GPa - 550 °C) during Alpine subduction: the Monviso metaophiolite complex (W. Alps). Three major shear zones cutting across the complex at low angle were studied, with a major focus on the Lower Shear Zone -LSZ-, where blocks of variably brecciated (and potentially seismically-derived) Fe-Ti and Mg-Al metagabbros are embedded, together with metasedimentary blocks, in a talc and tremolite-rich serpentinite matrix. Particular attention was paid to the petrological and structural characterization of 196 breccia blocks (in which more than 100 samples were collected), with (i) detailed analysis and mapping of their distribution in the LSZ and (ii) morphological characterization (block size, relative volume of matrix with respect to clasts and matrix-clast composition). These data have been synthesized in a new detailed geological map of Monviso meta-ophiolite at a scale of 1: 20,000. The mylonitic foliation of intact Mg-Al-rich metagabbros (composed of omphacite + rutile ± ex-lawsonite ± quartz and locally garnet) cut by breccia planes (cemented by omphacite + garnet ± ex-lawsonite) univocally indicates brecciation at eclogite facies conditions. In the breccias the occurrence of a first omphacite-rich matrix (M1) cut by a second matrix rich in garnet + lawsonite pseudomorphs (M2) witnesses multiple brittle rupture events (probably shortly spaced in time) prior to a stage of massive eclogite facies fluid ingression (matrix M3). Trace elements analysis (in-situ and bulk) coupled to rigorous microstructural characterization of samples suggest a progressive change of fluids circulating during the brecciation. First brittle event M1 was triggered by locally-buffered fluids (e.g., from metagabbros) with later ingression of “exotic” fluids (e.g., from serpentinites) triggering the brecciation events M2 and M3. Due to the limited extension of outcrops, the coseismic origin of the eclogitic breccia remains somehow speculative; nevertheless several evidences (e.g., minerals fractured and offset along omphacite-bearing planes) suggest that Monviso eclogite breccias were generated by instantaneous brittle rupture.

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / A presente dissertação apresenta um estudo do comportamento de um muro de gravidade flexível submetido a carregamentos sísmicos. A influência do carregamento sísmico, e mais precisamente da variação da aceleração horizontal de pico é avaliada num estudo paramétrico, através da análise dos resultados obtidos para os deslocamentos e acelerações horizontais que ocorreram no paramento do muro de gravidade. Optou-se pela análise ao longo do tempo, introduzindo o carregamento sísmico ao modelo numérico do qual fazia parte o muro de gravidade através de 5 históricos temporais de acelerogramas horizontais normalizados em 0,05g, 0,10g, 0,15g, 0,20g e 0,25g. O evento sísmico de origem dos acelerogramas normalizados foi o terremoto ocorrido no Chile em 27 de fevereiro de 2010 e medido na estação em San Jose de Maipo, a 332,7km do epicentro do terremoto. Utiliza-se um software computacional aplicado à Geotecnia, o Plaxis, para a calibração de um modelo numérico em situação estática e posterior desenvolvimento das análises paramétricas em condições sísmicas. As análises realizadas para avaliação do comportamento do muro de gravidade em condição sísmica foram a verificação da influência da variação do acelerograma horizontal normalizada introduzido na base do modelo nos valores das acelerações e deslocamentos horizontais atuantes no muro e a verificação da influência do carregamento sísmico em comparação com a situação estática, comparando-se os valores dos deslocamentos horizontais obtidos na análise estática e nas análises dinâmicas. Os fatores que influenciam potencialmente nos resultados numéricos em condições sísmicas foram apresentados, destacando-se os aspectos relacionados à modelagem numérica em condição de sismo. Os modelos constitutivos oferecidos, a possibilidade de incorporação do carregamento sísmico na base do modelo e a possibilidade de consideração do amortecimento visco-elástico de Rayleigh nos materiais são as principais qualidades do Plaxis, utilizado na modelagem numérica. Avaliou-se positivamente a potencialidade do Plaxis, visto que o programa mostrou-se uma ferramenta capaz de simular o comportamento de muros de gravidade sujeitos a carregamentos sísmicos.

A Costa Peruana apresenta alta atividade sísmica, tornando imprescindível a execução de análises de estabilidade que considerem os eventos sísmicos. Com o desenvolvimento de novas ferramentas numéricas, a análise dinâmica está se tornando cada vez mais importante e usual na fase de projeto, não se justificando mais a execução de análises estáticas em locais tão vulneráveis a sismos. A presente dissertação tem como objetivo realizar a análise da estabilidade dos Penhascos de Lima, Peru, considerando quatro taludes distintos, em uma região bastante afetada por abalos sísmicos. As análises foram executadas pelos métodos de equilíbrio limite, elementos finitos e pseudo-estático, buscando-se uma comparação entre os diferentes métodos. O trabalho apresenta uma descrição completa dos taludes em estudo, uma avaliação da condição sismológica da região e finalmente a estabilidade dos Penhascos de Lima, fazendo uso dos programas computacionais Slide (método do equilíbrio limite) e Plaxis (método dos elementos finitos) em 2D. Os resultados mostraram que os três métodos adotados forneceram fatores de segurança compatíveis, principalmente quando se considera perfis menos estratificados. Para perfis homogêneos, as diferenças obtidas foram da ordem de 0,5 a 1,0 %. As análises ressaltaram a importância de considerar a condição dinâmica, e mostraram-se bastante sensíveis aos valores de carga sísmica adotado. / The Peruvian coast has high seismic activity, making essential the implementation of stability analyzes that consider the seismic events. With the development of new numerical tools, dynamic analysis is becoming increasingly important and usual in the design phase is not justified over the execution of static analysis in places as vulnerable to earthquakes. This thesis aims to conduct the analysis of the stability of the cliffs of Lima, Peru, considering for different slopes, in a region highly affected by earthquakes. The analyzes were performed by equilibrium methods limit finite elements and pseudo-static, seeking a comparison between the different methods. The dissertation presents a complete description of the slopes in the study, an evaluation of the seismic condition of the region and finally the stability of Lima cliffs, making use of computer programs Slide (limit equilibrium method) and Plaxis (finite element method). The results showed that all three methods provided adopted gave safety factors, especially when considering less laminated profiles. For homogeneous profiles, obtained differences were of the order of 1.0 at 0.5%. The analysis highlighted the importance of considering the dynamic condition, and were quite sensitive to seismic load values adopted.

Le Liban n'a pas souffert de grands tremblements de terre destructeurs depuis près de deux siècles. Il est toutefois traversé par la faille transformante majeure du Levant, séparant sur1000 km de longueur la plaque Arabique à l’est de la plaque Africaine à l’ouest. Ses principales branches au Liban sont la faille de Yammouneh qui traverse le pays du sud au nord, les failles de Serghaya et Rachaya dans sa partie Est, la faille de Roum et les failles inverses du Mont Liban dans la partie Ouest. Ces failles ont généré des séismes destructeurs dans la longue histoire connue de la région, parmi lesquels les plus importants sont ceux deJuillet 551 sur la faille du Mont-Liban, de mai 1202 sur la faille de Yammouneh, d’Octobre1759 sur la faille de Rachaya et de Novembre 1759 sur la faille de Serghaya. L'évaluation del'aléa et du risque sismique local est donc de première importance pour l'ensemble du pays.L'objectif du travail effectué dans cette thèse est d'appliquer au Liban les avancées réalisées ces dernières années dans le développement de nouveaux outils à la fois fiables et économiquement abordables pour l’évaluation de l’aléa sismique, en commençant par les grandes villes et en particulier la capitale Beyrouth. L'objectif est de mieux appréhender et comprendre le risque sismique sur le territoire libanais, pour pouvoir ensuite commencer à élaborer des politiques de prévention et des codes parasismiques qui puissent le réduire à terme.Un réseau sismologique temporaire composé de 10 stations a été installé dans Beyrouth etune partie de sa banlieue sur des sites représentatifs des principales unités géologiques présentes. Plusieurs dizaines de séismes locaux et régionaux ont pu y être enregistrés, et leur réponse sismique a été évaluée par la méthode du rapport spectral site sur référence (SSR),comparé au rapport spectral de la composante horizontale sur la composante verticale (H/V)calculé sur les tremblements de terre et sur le bruit ambiant. Les mêmes enregistrements ont également été utilisés pour prédire, par la technique des Fonctions de Green empiriques(FGE), le mouvement sismique correspondant à un événement majeur (Mw7.5) sur la faille de Yammouneh. Cet exercice de prédiction a toutefois été réalisé en deux étapes en raison des limitations dans l'application de la technique FGE en champ proche, avec deux techniques complémentaires: l'enregistrement d’un petit événement a été d'abord utilisé pour simuler un séisme de Mw6.5 sur la faille de Yammouneh, en parallèle à l'utilisation de plusieurs équations de prédiction de mouvement du sol (GMPE), soigneusement sélectionnées, pour effectuer une prédiction similaire. La comparaison FGE/GMPE a alors permis de calibrer la prédiction du mouvement du sol par GMPE à différents sites de Beyrouth pour l'événement cible de Mw7.5. Ces résultats ponctuels ont ensuite été étendus à l'ensemble de la municipalité de Beyrouth et de sa banlieue proche, en vue de mieux cerner les contours d'une future carte de microzonage, au travers d'une vaste campagne de mesures de bruit ambiant sur615 sites. Leur traitement H/V a permis d’obtenir une carte de la fréquence de résonance pour l'ensemble de la zone, carte dont la robustesse a été testée et prouvée. Des mesures sismiques actives et passives ont en outre été menées sur les principales unités géologiques à proximité des 10 sites préalablement sélectionnés et instrumentés, permettant ainsi d'obtenir les premières estimations directes de la vitesse des ondes de cisaillement (via les courbes de dispersion des ondes de Rayleigh). La comparaison de ces mesures avec les estimations -très dispersées - issues de la compilation des paramètres géologiques/géotechniques disponibles et des équations de corrélation existantes avec les valeurs N des SPT, montre tout l'intérêt de ces mesures simples et fiables. / Lebanon is one of the countries that have not suffered from large destructive earthquakes foralmost two centuries. It is however lying on the 1000 km long, left lateral Levant fault thatseparates the Arabic plate in the east from the African plate in the west. Its main branches inLebanon are the Yammouneh fault that crosses the country from south to north, the Serghayaand Rachaya faults in its Eastern part, the Roum and Mount Lebanon Thrust faults in itsWestern part. These faults have generated destructive earthquakes in the long known historyof the area. The largest events are: The July 551 earthquake on the Mount Lebanon Thrustfault, the May 1202 earthquake on the Yammouneh fault, the October 1759 on the Rachayafault and the November 1759 on the Serghaya fault. From all above, one can conclude thatLebanon is exposed to a significant seismic hazard. Assessing the local seismic hazard andrisk is therefore of primary importance for the whole country.The objective of the work undergone in this PhD is to take advantage of the latest advancesachieved worldwide to promote rather inexpensive, though reliable, seismic hazardassessment tools, to try to apply them in Lebanon starting with the big cities and specificallythe capital Beirut. These studies will help to understand the Lebanese seismic risk andsubsequently to start to elaborate seismic policies and codes that may help reducing this risk.A temporary seismological network consisting of 10 stations has been installed in Beirut anda part of its suburbs. Several tens of local and regional earthquakes could be recorded, andallowed to estimate the site response at selected sites in Beirut through the standard site toreference spectral ratio method ("SSR") on earthquakes, compared to the horizontal tovertical ratio ("H/V") calculated on earthquakes and on ambient noise. The same recordingscould also be used via the empirical Green Function’s technique ("EGF") to predict theseismic ground motion corresponding to a Mw7.5 on the Yammouneh fault. However, due tolimitations in near-field applications of the EGF technique, this prediction exercise wasperformed in two steps and with two complementary techniques: a weak event recording wasfirst used to simulate a Mw6.5 earthquake on the Yammouneh fault, while several, carefullyselected ground motion prediction equations (GMPE) were used to perform a comparativeprediction for the same earthquake. This EGF/GMPE comparison then allowed tuning theGMPE prediction of ground motion at various sites within Beirut for the target Mw7.5 event.The results were then extended in view of proposing a framework for a future microzonationviimap. A comprehensive campaign of ambient noise measurements was achieved for 615 sitesof Beirut municipality and close suburbs, the H/V processing of which allowed to derive arobust map of resonance frequency for the whole area. In addition, active and passive seismicmeasurements were conducted on different geological units near the 10 formerlyinstrumented sites, which provided quantitative estimates of the shallow S-wave velocitythrough the Rayleigh wave dispersion curves. These geophysical measurements permitted toprovide direct estimates of the shear waves velocity, which prove much more reliable thanthe highly scattered estimates derived from the compilation of the availablegeological/geotechnical parameters and the use of existing correlations equations betweenSPT N-value and S-wave velocity Vs.

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / A presente dissertação apresenta um estudo do comportamento de um muro de gravidade flexível submetido a carregamentos sísmicos. A influência do carregamento sísmico, e mais precisamente da variação da aceleração horizontal de pico é avaliada num estudo paramétrico, através da análise dos resultados obtidos para os deslocamentos e acelerações horizontais que ocorreram no paramento do muro de gravidade. Optou-se pela análise ao longo do tempo, introduzindo o carregamento sísmico ao modelo numérico do qual fazia parte o muro de gravidade através de 5 históricos temporais de acelerogramas horizontais normalizados em 0,05g, 0,10g, 0,15g, 0,20g e 0,25g. O evento sísmico de origem dos acelerogramas normalizados foi o terremoto ocorrido no Chile em 27 de fevereiro de 2010 e medido na estação em San Jose de Maipo, a 332,7km do epicentro do terremoto. Utiliza-se um software computacional aplicado à Geotecnia, o Plaxis, para a calibração de um modelo numérico em situação estática e posterior desenvolvimento das análises paramétricas em condições sísmicas. As análises realizadas para avaliação do comportamento do muro de gravidade em condição sísmica foram a verificação da influência da variação do acelerograma horizontal normalizada introduzido na base do modelo nos valores das acelerações e deslocamentos horizontais atuantes no muro e a verificação da influência do carregamento sísmico em comparação com a situação estática, comparando-se os valores dos deslocamentos horizontais obtidos na análise estática e nas análises dinâmicas. Os fatores que influenciam potencialmente nos resultados numéricos em condições sísmicas foram apresentados, destacando-se os aspectos relacionados à modelagem numérica em condição de sismo. Os modelos constitutivos oferecidos, a possibilidade de incorporação do carregamento sísmico na base do modelo e a possibilidade de consideração do amortecimento visco-elástico de Rayleigh nos materiais são as principais qualidades do Plaxis, utilizado na modelagem numérica. Avaliou-se positivamente a potencialidade do Plaxis, visto que o programa mostrou-se uma ferramenta capaz de simular o comportamento de muros de gravidade sujeitos a carregamentos sísmicos.

A Costa Peruana apresenta alta atividade sísmica, tornando imprescindível a execução de análises de estabilidade que considerem os eventos sísmicos. Com o desenvolvimento de novas ferramentas numéricas, a análise dinâmica está se tornando cada vez mais importante e usual na fase de projeto, não se justificando mais a execução de análises estáticas em locais tão vulneráveis a sismos. A presente dissertação tem como objetivo realizar a análise da estabilidade dos Penhascos de Lima, Peru, considerando quatro taludes distintos, em uma região bastante afetada por abalos sísmicos. As análises foram executadas pelos métodos de equilíbrio limite, elementos finitos e pseudo-estático, buscando-se uma comparação entre os diferentes métodos. O trabalho apresenta uma descrição completa dos taludes em estudo, uma avaliação da condição sismológica da região e finalmente a estabilidade dos Penhascos de Lima, fazendo uso dos programas computacionais Slide (método do equilíbrio limite) e Plaxis (método dos elementos finitos) em 2D. Os resultados mostraram que os três métodos adotados forneceram fatores de segurança compatíveis, principalmente quando se considera perfis menos estratificados. Para perfis homogêneos, as diferenças obtidas foram da ordem de 0,5 a 1,0 %. As análises ressaltaram a importância de considerar a condição dinâmica, e mostraram-se bastante sensíveis aos valores de carga sísmica adotado. / The Peruvian coast has high seismic activity, making essential the implementation of stability analyzes that consider the seismic events. With the development of new numerical tools, dynamic analysis is becoming increasingly important and usual in the design phase is not justified over the execution of static analysis in places as vulnerable to earthquakes. This thesis aims to conduct the analysis of the stability of the cliffs of Lima, Peru, considering for different slopes, in a region highly affected by earthquakes. The analyzes were performed by equilibrium methods limit finite elements and pseudo-static, seeking a comparison between the different methods. The dissertation presents a complete description of the slopes in the study, an evaluation of the seismic condition of the region and finally the stability of Lima cliffs, making use of computer programs Slide (limit equilibrium method) and Plaxis (finite element method). The results showed that all three methods provided adopted gave safety factors, especially when considering less laminated profiles. For homogeneous profiles, obtained differences were of the order of 1.0 at 0.5%. The analysis highlighted the importance of considering the dynamic condition, and were quite sensitive to seismic load values adopted.