[pt] Historicamente é sabido que muitas das rupturas ocorridas em barragens ou taludes naturais podem ser atribuídas ao fenômeno da liquefação de solos arenosos, causada pela ação de carregamentos dinâmicos gerados por explosão ou, mais freqüentemente, por terremotos. Quando liquefação ocorre, um súbito aumento da poropressão faz decrescer a resistência ao cisalhamento do solo e sua capacidade de suportar pontes e edifícios é significativamente reduzida. Solo liquefeito também pode exercer altas pressões sobre estruturas de contenção, causando inclinações da mesma e movimentos do solo que, por sua vez, originam recalques e destruição de estruturas localizadas sobre a superfície do terreno. O termo liquefação tem sido empregado para descrever fenômenos relacionados, que produzem efeitos similares, mas cujos mecanismos de formação são bastante diferentes. Estes fenômenos são modernamente descritos como fluxo por liquefação e mobilidade cíclica. Fluxo por liquefação é o fenômeno no qual o equilíbrio estático é destruído por carregamentos estáticos ou dinâmicos em um depósito de solo com baixa resistência residual. Colapsos causados por fluxo por liquefação são freqüentemente caracterizados por movimentos rápidos e de grande extensão. Mobilidade cíclica, por outro lado, é causada por carregamentos cíclicos em solos sob tensões cisalhantes estáticas inferiores à resistência ao cisalhamento do material, com as deformações desenvolvendo-se gradualmente. A execução de barragens de rejeito usando a técnica de construção à montante pode levar à ocorrência de liquefação estática se a velocidade de construção for suficientemente alta para causar o desenvolvimento de excessos de poropressão. A resposta de liquefação pode ser observada em amostras de solo fofo quando as tensões de cisalhamento atingem um pico seguido por uma fase de amolecimento aparente que, no caso de carregamento não drenado, é associado com a tendência do material em contrair de volume. Para alguns estados iniciais, a parte descendente da resposta do material pode ser seguida por uma fase crescente que se atenua à medida que o estado permanente ou crítico for atingido. Nesta dissertação, a modelagem da resposta de liquefação por carregamento estático, foi feita considerando-se modelos propostos na literatura por Juárez-Badillo (1999b) e Gutierrez e Verdugo (1995). Este último, principalmente após modificação introduzida pela dependência de alguns parâmetros em relação à tensão de confinamento, levou a resultados bastante satisfatórios nas retroanálises consideradas, apesar da relativa simplicidade da formulação. / [en] Historically it is known that many failures in earth dams and natural slopes can be attributed to the phenomenon of sand liquefaction, caused by dynamic loads generated by earthquake shaking or other rapid loading, such as blasts. When liquefaction occurs, the strength of the soil decreases and its ability to support foundations for buildings and bridges is significantly reduced. Liquefied soil can also exerts higher pressure on retaining walls, which can cause them to tilt or slide, yielding settlement of the retained soil with risks of destruction of structures on the ground surface. The term liquefaction has actually been used to describe a number of related phenomena, which produce similar effects but whose mechanisms are quite different in nature. These phenomena can be divided into two main categories: flow liquefaction and cyclic mobility. Flow liquefaction is a phenomenon in which the static equilibrium is destroyed by static or dynamic loads in a soil deposit with low residual strength. Failures caused by flow liquefaction are often characterized by large and rapid movements. Cyclic mobility, on the other hand, is a liquefaction phenomenon triggered by cyclic loading, occurring in soil deposits with static shear stresses lower than the soil strength. Deformations due to cyclic mobility develop incrementally because of static and dynamic stresses that exist during an earthquake. The rising of tailing dams using the upstream construction technique can lead to static liquefaction failure if the rate of construction is sufficiently high to cause excess pore pressure to develop in the tailings. The liquefaction response is observed for loose specimens when the shear stress exhibits a peak followed by a phase of apparent softening that, in undrained loading, is associated with the tendency of the material to contract (densify). For some initial loading states, the descending part of the response is followed by an increasing part, leveling-off eventually when the material reaches the final, critical (steady) state. In this thesis, the modeling of the phenomenon of static liquefaction is carried out considering the constitutive models proposed in the literature by Juárez-Badillo (1999b) and Gutierrez & Verdugo (1995). The latter, mainly after introducing the assumption that some material parameters are stress dependent and not simple constants, as in the original version, produced good matching between experimental and predicted results, in spite the simplicity of the mathematical formulation.

Liquefaction is a phenomenon in which soil loses a large percentage of its shear resistance due to increased pore water pressure and flows like a liquid. Undrained cyclic loading conditions during earthquakes cause liquefaction of soils, which can lead to catastrophic failures such as bearing capacity failures, slope failures and lateral spreads. The concepts and mechanisms of liquefaction were studied extensively by many researchers. Though the factors affecting the liquefaction response of soils during earthquakes are well documented in literature, there are still some gray areas in understanding the individual and combined effects of factors like frequency, gradation, fines content and surcharge pressure on the initiation of liquefaction. The objective of this thesis is to study the influence of ground motion, soil and site parameters on the initiation of liquefaction in saturated sand beds through laboratory shaking table model tests and numerical studies. Shaking table tests are carried out using a uniaxial shaking table on sand beds of 600 mm thickness. The initiation of liquefaction was observed and identified by measuring the pore water pressure developed during the sinusoidal cyclic loading. Free field liquefaction studies are carried out on sand beds to study the influence of ground motion parameters, namely, input acceleration and frequency of shaking on liquefaction. These studies revealed that acceleration is one of the important parameters that can affect the initiation of liquefaction in sands. Increase in acceleration reduces the liquefaction resistance of sand and a small increase in acceleration can trigger liquefaction. Frequency of shaking did not affect the initiation of liquefaction at lower frequencies but a threshold frequency which triggered instant increase in the excess pore pressures is observed. Liquefaction caused slight initial amplification followed by de-amplification of accelerations due to the stiffness reduction in soils during liquefaction, the effect being more pronounced in the top layers of the sand bed. Pore water pressure ratios during dynamic loading decreased with depth below the surface of the sand bed due to the low initial effective vertical stress and upward transmission of pore pressure during undrained loading. Shaking table tests are carried out to study the influence of soil parameters such as relative density, thickness of dry overlying sand layer and gradation. Relative density of sand can influence the liquefaction potential of sand to a great extent, about 10% increase in relative density bringing down the probability of liquefaction by about 50%. With the increase in height of dry overlying sand layer, liquefaction potential has decreased nonlinearly. Change in grain size altered the pattern of liquefaction and pore pressure development and it is observed that the liquefaction in finer sands is influenced by the frequency of shaking to a larger extent. Surcharge pressure from building loads increased the liquefaction potential and heavier structures got liquefied at lower pore water pressure ratios. Significant post-liquefaction de-amplification was observed in sand beds with surcharge pressure. Parametric numerical analyses are carried out using finite difference program FLAC (Fast Lagrangian Analysis of Continua) with FINN model to measure pore water pressures in the sand bed. Results from numerical analyses with change in the acceleration, surcharge pressure and thickness of dry overlying layer agreed well with the experimental results. However, effect of frequency in numerical studies did not match with the experimental observations, because of the inherent boundary effects in the experimental models. Results from this thesis provided important insights into the development of pore water pressures in sand beds during cyclic loading events, apart from enhancing the understanding towards the effect of various ground motion, site and soil parameters on the initiation of liquefaction in sand beds.