Probabilistic assessment of tunneling-induced building damage

This paper presents a rational and efficient probabilistic framework to address the uncertainties of geotechnical properties and the stratigraphic profile in a tunneling-induced building damage assessment. The probabilistic approach is based on the application of a hybrid point estimation method that employs a reduced number of computations without losing accuracy. The framework also relies on sensitivity analyses using tornado diagrams. Analyses of ground movements and building displacements are considered prior to the building damage assessment. The proposed approach was applied to the recently constructed extension of Line 5 of the São Paulo Metro. Two-dimensional (2D) and three-dimensional (3D) numerical analyses were carried out using a commercial finite element code by considering the Mohr-Coulomb and linear elastic constitutive models. Three probabilistic scenarios based on the lower, mean, and upper bound values of the coefficient of variation of input variables were considered. The results obtained using the proposed framework allowed estimating the probabilistic scenario that better provides an ideal coefficient of variation to adopt for input variables. The ground elastic properties and upper layer thickness had the greatest influence on ground movements and building displacements. Finally, the building damage analysis showed good agreement with the settlement data measured in the field.