Three-dimensional performance and analyses of deep excavations

Congestion in urban areas has led to an increased emphasis on the accurate calculation of expected ground movements that develop due to a deep excavation. Increased accuracy of the predicted movements is needed to enable the design of an excavation support system that limits damage to surrounding structures and utilities. Numerous case studies of deep excavations in urban areas document the decrease of ground movements near the corners of an excavation due to the stiffening effect of the corners. However, this three-dimensional nature of the ground movement is often not considered in design. In current practice, calculation of expected movements by empirical methods or by two-dimensional finite element analysis have assumed either plane strain or axisymmetric conditions. The two-dimensional representation of a three-dimensional condition requires numerous oversimplifications. Results of plane strain analyses yield larger than actual ground movements near a corner because the three-dimensional effects are not considered in the analysis. While 3D finite element analyses are now possible, implementation is far from routine with uncertainties inherent in the process.; This thesis focuses on the measurement and analysis of three-dimensional ground movements caused by excavation. The correlation of the measured excavation induced ground response with the construction activity of the Robert H. Lurie Center excavation provided performance-based ground movement prediction charts and a baseline to validate three-dimensional finite element analyses of the excavation. In addition, the optical survey ground movement data were used to develop a complementary error function (erfc) based fitting procedure to determine the distribution of movements and distortions parallel to the Lurie excavation, as well as to other published studies where sufficient ground movement data were presented.; Results of a three-dimensional finite element parametric study of typically-sized excavations showed the dependence of the three-dimensional effects on the excavation geometry, height of excavation, and support system stiffness, provided guidelines as to when three-dimensional analyses should be performed, quantified reduction factors that allow a two-dimensional approximation of the three-dimensional responses, and confirmed the validity of the erfc function to calculate the distribution of ground movements parallel to an excavation.