Finite element-based geotechnical risk analysis for anchor-supported deep excavations

Abstract

AbstractEvaluation of the reliability of deep excavation support systems requires the inclusion of the soil variability in conjunction with probabilistic analysis. This is often considered a complicated process; therefore, the method is usually not followed in routine engineering practice. Recent developments in finite element modelling now enable probabilistic analyses to be performed more conveniently. As an illustrative example, the case history of a 20 m-deep excavation supported by anchored drilled shafts in the Sogutozu district of Ankara was investigated. Experimental results indicate that saturated sand-gravel bands, which are common within Ankara Clay, have the potential to cause a significant decrease in the anchor capacity. Ergo the excavation performance, including structural capacity and wall lateral displacements, was evaluated taking into regard the probabilistic nature of the effect of these bands and the variability of the soil parameters. Back-analysis was performed and compared with inclinometer readings to examine the capability of the numerical model to simulate the field behaviour. Later, additional finite element analyses were carried out through Python scripting software to investigate the reliability of the excavation. Geotechnical parameters of the Ankara Clay produced by Monte Carlo simulation method were used in these analyses. The results were evaluated within the context of recommended acceptable reliability levels. The reliability of the system was determined to be relatively low when taking into account the adverse impact of sand and gravel bands.