Reliability-based design of an earth and concrete dam.

Detalhes bibliográficos
Ano de defesa: 2021
Autor(a) principal: Torrico Siacara, Adrian
Orientador(a): Não Informado pela instituição
Banca de defesa: Não Informado pela instituição
Tipo de documento: Tese
Tipo de acesso: Acesso aberto
Idioma: eng
Instituição de defesa: Biblioteca Digitais de Teses e Dissertações da USP
Programa de Pós-Graduação: Não Informado pela instituição
Departamento: Não Informado pela instituição
País: Não Informado pela instituição
Palavras-chave em Português:
Link de acesso: https://www.teses.usp.br/teses/disponiveis/3/3145/tde-22102021-094317/
Resumo: In geotechnical engineering, the use of classical deterministic approaches is part of the normal practice. Most laws and standards are based on engineers\' experience and deterministic assumptions. In the last four decades, the uncertainties and uncertainty inherent in geotechnical properties have caught the attention of engineers. They have found that a single factor of safety (FS) calculated by traditional deterministic analysis methods can not exactly represent the stability. Recently, to provide a more rational mathematical framework to incorporate different types of uncertainties in the stability estimation, different types of commercial software, which include basic probabilistic and deterministic methods, have been widely extended. The failure probability calculated is a kind of complement to the safety factor. In this doctoral thesis, reliability concepts have been reviewed first, and some advanced reliability methods, including Mean Value First-Order Second Moment (MVFOSM), FirstOrder Reliability Method (FORM), Second-Order Reliability Method (SORM), Monte Carlo Simulation (MCS), and importance Monte Carlo Simulation (IMCS), have been described and successfully applied to the dam stability analyses based on a direct coupling (DC) geotechnical software with a reliability solver. An advanced literature review on Reliability-Based Design Optimization (RBDO) was presented to show how accurate and efficient RBDO analyses of a concrete dam can be performed by using a Mono-level method as a Single-Loop Approach (SLA). Numerical methods help understand the behaviors of geotechnical installations. However, the computational cost may sometimes become prohibitive when structural reliability analysis is performed, due to repetitive calls to the deterministic solver. In this thesis, we show how accurate and efficient reliability analyses of geotechnical installations can be performed by directly coupling geotechnical software with a reliability solver. A dam is used as the study object under different operating conditions over time. The limit equilibrium method of Morgenstern-Price is used to calculate factors of safety and to find the critical slip surface. The commercial software packages Seep/W and Slope/W are coupled with StRAnD structural reliability software. Initial analysis in long term steady-state conditions shows close results of the FORM and SORM, as well as by ISMC simulation. By means of sensitivity analysis, the effective friction angle (?´) is found to be the most relevant uncertain geotechnical parameter. The correlations between different geotechnical properties are shown to be relevant in terms of equilibrium reliability indices. Finally, a critical slip surface, identified in terms of the minimum FS, isshown not to be the critical surface in terms of reliability index.Considerable variation between results of commercial software and DC software is found. FORM, SMC, and MVFOSM are evaluated. The computational cost is very high in commercial software. The angle of the contribution to the shearing resistance due to the matric suction (?b) has a considerable incidence in reliability analysis.In transient conditions, seepage analysis with limit equilibrium analysis is performed to investigate the safety of an earth dam over time, considering different initial conditions, rainfall intensities, rapid drawdown, and normal operating conditions (NOC). FORM is employed in reliability analysis. Sensitivity analyses reveal that saturated hydraulic conductivity (ks), friction angle (?´), and cohesion (c´) are the random parameters with the greatest contribution. The cumulative effect of random saturated hydraulic conductivity (mainly) makes critical times and critical slip surfaces different in the probabilistic and deterministic analyses.In an existing old earth dam, the FORM method is employed in reliability analysis. Sensitivityanalyses reveal the most important geotechnical parameters. A range of values of the relationship between the reliability index (?) and the factor of safety (FS) was found for all probabilistic and deterministic results. The differences in terms of pore water pressure betweendeterministic and probabilistic approaches are presented. Finally, a critical slip surface, identified in terms of minimum FS, is shown here not to be the critical surface in terms of the minimum reliability index.Finally, we show how accurate and efficient RBDO analyses of a concrete dam can be performed by using a Mono-level method as SLA. A concrete dam is used as the study object under different conditions of the target reliability index (?T) and heights of the dam. By means of sensitivity analysis, the most important geotechnical parameters are pointed out. Finally, for the lower ?T values considered herein, we found the eccentricity limit state function of the active constraint at the optimum solution. For larger ?T values, the sliding limit state is the active constraint.