Análise probabilística da estabilidade de taludes via teoria da confiabilidade
Carregando...
Data
2018-01-26
Autores
Giacon Junior, Admir José
Título da Revista
ISSN da Revista
Título de Volume
Editor
Universidade Estadual Paulista (Unesp)
Resumo
Devido às consequências negativas associadas à ruptura e deslizamentos de taludes, a análise de estabilidade desses assume uma parte primordial na área da engenharia geotécnica. A maioria dessas análises é realizada de maneira determinística, não incorporando as incertezas inerentes ao sistema. A constatação dessas incertezas em projetos de taludes associa-se a previsão de uma probabilidade de falha no desempenho de taludes quanto à segurança. Nesse contexto, esta dissertação busca avaliar a segurança de taludes através de uma abordagem probabilística, simulando taludes com geometrias e solos variados, verificando ainda a influência da elevação da posição do nível d´água. Para essa finalidade foram realizadas análises probabilísticas utilizando First Order Reliability Method (FORM) com base nos métodos de Equilíbrio Limite, através dos programas computacionais acadêmicos RASS (análises determinísticas) e RELgen (análises probabilísticas). Com base nos resultados, verificou-se que as superfícies críticas obtidas pelos métodos determinísticos podem não coincidir com as superfícies que apresentam a maior probabilidade de falha. Isto é verificado ainda na análise de inclinação da superfície do talude, em que inclinações apresentando fatores de segurança admissíveis, apresentaram probabilidades de falha não aceitáveis. Estas divergências são atribuídas às incertezas inerentes ao sistema. Sendo assim, as análises de estabilidade de talude baseadas apenas nos resultados de análises determinísticas não apresentam todas as variáveis presentes no contexto do projeto. Por fim, através dos resultados obtidos nesta pesquisa é possível inferir que a adoção da análise probabilística em projeto de taludes auxilia de uma forma consistente a avaliação da segurança dos taludes, contribuindo de maneira satisfatória na prática de projetos de engenharia geotécnica.
Due to the negative consequences associated with slope rupture and landslides, the stability analysis of these slopes plays an important role in geotechnical engineering. Most of these analyzes are carried out in a deterministic manner, not incorporating the uncertainties inherent to the system. The observation of these uncertainties in slope projects is associated with the forecast of a probability of failure in slope performance in terms of safety. In this context, this paper present research seeks to evaluate slope safety in a probabilistic manner, simulating slopes with varied geometries and soils, also verifying the influence of elevation of the water level position. For this purpose, were performed probabilistic analyzes using First Order Reliability Method (FORM), based on the Limit Equilibrium methods, RASS (deterministic analysis) and RELgen (probabilistic analyzes). Based on the results, it was found that the critical surfaces obtained by the deterministic methods may not coincide with the highest most probability surfaces to fail. This is still verified in surface slope analysis, where inclinations with acceptable safety factors presented probabilities of failure unacceptable. These divergences are attributed to the uncertainties inherent in the system. Therefore, slope stability analyzes based only on the results of deterministic analyzes do not present all the variables present in the project context. Finally, through the results obtained in this research it is possible to infer that the adoption of probabilistic analysis in slope design assists in a consistent way the assessment of slope safety, contributing in a satisfactory way in the practice of geotechnical engineering projects.
Due to the negative consequences associated with slope rupture and landslides, the stability analysis of these slopes plays an important role in geotechnical engineering. Most of these analyzes are carried out in a deterministic manner, not incorporating the uncertainties inherent to the system. The observation of these uncertainties in slope projects is associated with the forecast of a probability of failure in slope performance in terms of safety. In this context, this paper present research seeks to evaluate slope safety in a probabilistic manner, simulating slopes with varied geometries and soils, also verifying the influence of elevation of the water level position. For this purpose, were performed probabilistic analyzes using First Order Reliability Method (FORM), based on the Limit Equilibrium methods, RASS (deterministic analysis) and RELgen (probabilistic analyzes). Based on the results, it was found that the critical surfaces obtained by the deterministic methods may not coincide with the highest most probability surfaces to fail. This is still verified in surface slope analysis, where inclinations with acceptable safety factors presented probabilities of failure unacceptable. These divergences are attributed to the uncertainties inherent in the system. Therefore, slope stability analyzes based only on the results of deterministic analyzes do not present all the variables present in the project context. Finally, through the results obtained in this research it is possible to infer that the adoption of probabilistic analysis in slope design assists in a consistent way the assessment of slope safety, contributing in a satisfactory way in the practice of geotechnical engineering projects.
Descrição
Palavras-chave
Estabilidade de taludes, Teoria da confiabilidade, Equilíbrio limite, Slope stability, Reliability theory, Limit equilibrium