Evaluación de modelos de ocurrencia de licuación basados en ensayos cptu: mejoras desde una alternativa basada en el parámetro de estado.
Loading...
Date
2024
Journal Title
Journal ISSN
Volume Title
Publisher
Universidad de Concepción
Abstract
El estudio del fenómeno de licuación y sus consecuencias es probablemente uno de los tópicos más estudiados en la ingeniería geotécnica, conocido por su gran potencial de daño. Sin embargo, en las metodologías semi-empíricas del estado de la práctica aún existen temas que no están del todo resueltos. A pesar del constante surgimiento de nuevos casos históricos de licuación de terremotos no explorados, la forma de incorporar estos datos no ha generado mayores cambios en la metodología. Frente a esto, en escenarios que no se encuentran bien representados por las bases de datos disponibles, las metodologías disponibles aún generan bastante incertidumbre acerca de su desempeño. Por lo tanto, el objetivo de este estudio es evaluar el desempeño de los modelos de licuación con un origen menos ligado a las bases de datos disponibles. Para esto, se evaluó el uso de los métodos simplificados con un enfoque de parámetro de estado basado en CPT. Utilizando únicamente correlaciones existentes provenientes de cámaras de calibración y ensayos cíclicos de la literatura, se propuso un modelo de desencadenamiento de licuación, que fue evaluado en tres bases de datos de licuación con CPT de distintos tamaños: Global, Subducción y Canterbury. Los resultados indican que, a pesar de no existir un ajuste empírico a partir de casos históricos, el enfoque propuesto entrega, en la mayoría de los casos, mejores niveles de predicción que otros modelos del estado de la práctica. Además, se observó que, al utilizar un enfoque puramente basado en el parámetro de estado, las resistencias cíclicas estimadas (CRR7.5) en arenas densas eran considerablemente menores que el resto de los modelos, coincidente con otras investigaciones recientes. Finalmente, dado que el modelo propuesto presentó grandes similitudes con el cyclic strain approach, se discuten algunas de sus ventajas para complementar el modelo propuesto en un futuro.
The study of the liquefaction phenomenon and its consequences is probably one of the most studied topics in geotechnical engineering, known for its great damage potential. However, in the state-of-practice semi-empirical methodologies there are still issues that are not fully resolved. Despite the constant emergence of new case histories of liquefaction from new earthquakes, the incorporation of these data has not generated major changes in the methodology. For that reason, in scenarios that are currently not well represented by the available databases, the methodologies still generate considerable uncertainty about their performance. Therefore, the aim of this study is to evaluate the performance of liquefaction models with an origin less relied on available databases. To accomplish this, the use of simplified methods with a CPT-based state parameter approach was evaluated. Using only existing correlations from calibration chambers and cyclic tests from the literature, a liquefaction triggering model was proposed and evaluated against three CPT-based liquefaction databases of different sizes: Global, Subduction and Canterbury. The results indicate that, despite the lack of an empirical fit from case histories, the proposed approach result in better prediction performances than most state-of-the-art models. Additionally, result shows that using a purely state-parameter based approach, the estimated cyclic resistances (CRR7.5) in dense sands were considerably lower than in the rest of the models, being consistent with other recent developments. Finally, since the proposed approach showed great similarities with the cyclic strain approach, some of its advantages are discussed to complement the proposed model in the future.
The study of the liquefaction phenomenon and its consequences is probably one of the most studied topics in geotechnical engineering, known for its great damage potential. However, in the state-of-practice semi-empirical methodologies there are still issues that are not fully resolved. Despite the constant emergence of new case histories of liquefaction from new earthquakes, the incorporation of these data has not generated major changes in the methodology. For that reason, in scenarios that are currently not well represented by the available databases, the methodologies still generate considerable uncertainty about their performance. Therefore, the aim of this study is to evaluate the performance of liquefaction models with an origin less relied on available databases. To accomplish this, the use of simplified methods with a CPT-based state parameter approach was evaluated. Using only existing correlations from calibration chambers and cyclic tests from the literature, a liquefaction triggering model was proposed and evaluated against three CPT-based liquefaction databases of different sizes: Global, Subduction and Canterbury. The results indicate that, despite the lack of an empirical fit from case histories, the proposed approach result in better prediction performances than most state-of-the-art models. Additionally, result shows that using a purely state-parameter based approach, the estimated cyclic resistances (CRR7.5) in dense sands were considerably lower than in the rest of the models, being consistent with other recent developments. Finally, since the proposed approach showed great similarities with the cyclic strain approach, some of its advantages are discussed to complement the proposed model in the future.
Description
Tesis presentada para optar al grado de Magíster en Ciencias de la Ingeniería con mención en Ingeniería Civil.
Keywords
Licuación