Prompt estimations of fill slope stability considering material uncertainties
An-Jui Li, Kelvin Lim
Taiwan Tech
Abstract
Slope stability analysis is one of the most critical and important topics in geotechnical engineering and thus various stability chart solutions have been developed to provide quick first assessments of slope stability. Motivated by that, this paper aims to adopt an artificial neural network with the extreme learning machine algorithm to develop a convenient and efficient tool in assessing fill slope stability. The neural network is trained using the solutions from the finite element upper and lower bound limit analysis methods. Because the conventional deterministic approach would ignore the uncertainties in soil properties, this study also uses the same technique to develop a tool that is capable of performing a reliability analysis of fill slope stability. Therefore, the tools developed in this study are capable of providing not only a quick first assessment of fill slope stability but also a reliability assessment of the slope. They can provide information for design, inspection or maintenance judgement.
Keywords: Decision-making; Limit analysis; Artificial neural network; Reliability
Topic: International Symposium of Civil, Environmental, and Infrastructure Engineering