Effects of deep excavations in soft clay on the immediate surroundings

Paper in proceeding, 2009

When excavating in an urban environment, evaluation of the magnitude and distribution of ground movements is an important part of the design process, since excessive movement can damage adjacent buildings and utilities. In order to minimize the movement of the surrounding soil, a retaining wall support system is used to provide lateral support. This article is a brief summary of the dissertation "Effects of Deep Excavations in Soft Clay on the Immediate Surroundings: Analysis of the Possibility to Predict Deformations and Reactions Against the Retaining System" presented at Chalmers in 2007, (Kullingsjö, 2007). The dissertation describes different methods for the evaluation of ground movements adjacent to a deep excavation in soft clay as well as how to estimate the lateral earth pressure that acts on the retaining system. It presents a review of: - Soil characteristics of importance for the evaluation of deformations and earth pressure. - Current empirical methods for estimating ground surface settlements. - Different classical methods for calculating lateral earth pressure. - Various soil modelling methods, with focus on the theory of elasto-plasticity. The review is followed by an extensive case study performed at the Göta tunnel project in the centre of Gothenburg, Sweden. Back analyses were performed in order to predict and interpret ground deformations and the development of stress changes against the retaining wall system. These analyses took the form of non-linear finite element analyses with three different constitutive models (an isotropic linear elastic Mohr-Coulomb model, the e-ADP, which is a total stress based model capable of modelling anisotropic undrained shear strength as well as non-linearity in shear, and MIT-S1, a bounding surface model based on effective stresses). The different outcomes of these three models are compared and discussed. Special focus has been placed on evaluating the parameters of the MIT-S1 model and its response compared to advanced laboratory tests.