Systematic update of a deep excavation model using field performance data

Abstract

Numerical simulation of construction staging of deep urban excavations is commonly used to estimate induced ground deformations. A novel method to extract the constitutive model of the soil behavior directly from field measurements of excavation response is introduced. This methodology develops a neural network based constitutive model of the soil, given the lateral wall deformation and surface settlement profile measurements. The resulting soil model, used in a numerical analysis, provides correct ground deformations and can be used in the forward prediction of future excavations or later excavation stages. The soil model can continuously evolve using additional field information.