Application of nonlinear optimization technique to back analyses of deep excavation

Abstract

This study presents a nonlinear optimization technique (NOT) for conducting the back analyses of geotechnical engineering problems based on the field observations. Additional auxiliary techniques are incorporated to enhance the convergence and stability of the NOT. The developed NOT and additional auxiliary techniques are incorporated into a finite element code and then applied to the back analysis of excavation-induced wall deflection. A number of hypothetical excavation cases with various scenarios of stratigraphy and two quality excavation case histories are used to validate the developed NOT, in which the dominant soil parameters are treated as target parameters. Results show that the wall deflections of all hypothetical and actual excavation cases at each stage can be accurately and efficiently back-figured. The developed NOT has a potential to be an useful tool for preventing the building damage through accurately and efficiently predicting the excavation-induced deformations at subsequent stages.