Probabilistic Analysis of Performance of Three-Dimensional Supported Excavations Considering Soil Spatial Variability

Supported excavation systems are common geotechnical structures and play a crucial role in satisfying increasing requirements for highly efficient usage of underground space with the trend of urbanization around the world. However, current supported excavation designs mainly focus on two-dimensional deterministic analysis. Ignoring the three-dimensional impacts and uncertainties such as soil spatial variability could result in an inaccurate evaluation of deep excavation performance. This paper examines the impact of soil spatial variability on excavation-induced responses through three-dimensional finite-element modeling. An automated procedure for Monte Carlo simulation was developed, and three-dimensional stochastic finite-element modeling was carried out using random field theory. The effects of soil vertical spatial variability on maximum lateral wall deflection, maximum bending moment, maximum shear force, and strut force are analyzed. In addition, reliability analyses were performed to assess the probability of failure of the support structure. The results demonstrate a substantial impact of soil spatial variability on maximum wall deflection, maximum wall bending moment, maximum wall shear, and internal strut force in deep excavations. The results presented in this study can provide a useful reference for the reliability-based design of supported excavations in the urban environment.