Identifying Soil Parameters from Excavation Using MOOP

In this paper, a framework that combines the stochastic multi-objective optimization (MOOP) and the observed field data for identifying soil parameters is presented. For conducting the parameter identification, a multi-objective differential evolution algorithm is employed. Then, an elastoplastic soil model accounting for the small-strain stiffness and anisotropic elasticity behaviors of clays is developed and adopted. Finally, the proposed procedure is applied to a well-instrumented deep excavation. The observed wall deflection and ground settlement are used as objective in the optimization. The results demonstrate that the anisotropy of elasticity and small strain stiffness are two important factors influencing the ground settlement and wall deflection in the excavation. All the comparisons demonstrate that the proposed framework is effective and efficient for identifying soil parameters from excavation.