Supported excavations: Observational method and inverse modeling

An inverse analysis procedure that uses construction monitoring data to update predictions of deformations for supported excavation systems is presented. The numerical procedure is used to optimize the finite element model of a 12.2-m-deep excavation through Chicago glacial clays by minimizing the errors between monitoring data and computed displacements. The field observations are obtained from inclinometer data that measured lateral movements of the soil behind the supporting walls on opposite sides of the excavation throughout construction. Five construction stages are defined for the inverse analysis. At every new construction stage, the inclinometer data relative to that stage are added to the observations already available to "recalibrate" the model of the excavation. The constitutive responses of the soils are represented by the hardening-soil (H-S) model. Of the six basic H-S input parameters, only one parameter per layer is optimized, while the other parameters are either kept constant or related to the updated value of the optimized parameter. The methodology is effectively used to recalibrate the model of the excavation at early construction stages, such that good "predictions" are made of the behavior of the soil at later stages.