Extended Kalman Filter for the Inverse Analysis of a Supported Excavation Based on Field Monitoring Data for Improving Predictions of Ground Responses

This paper presents a procedure for inverse analysis of braced excavations based on the extended Kalman Filter formulation. Prior estimation of soil parameters for analysis and design of braced excavation is first made using limited laboratory data obtained from site investigation. Then during the staged excavation, an extended Kalman Filter formulation is used to update the estimated soil parameters based on monitoring data of the maximum ground settlement and wall deflection. The updated soil parameters are then used to refine the predictions of the soil-wall-support system responses in the subsequent stages of excavation, as well as to evaluate the damage potential of adjacent buildings at the final stage of excavation. A case study is utilized to demonstrate the effectiveness of the developed procedure for improving the predictions of the excavation-induced wall and ground responses. Furthermore, a more rational assessment of the damage potential of the adjacent buildings is yielded through this updating procedure. The proposed procedure is demonstrated to be an effective tool for improving the predictions of the ground responses and building damage potential in a staged excavation.