Inverse analysis of horizontal coefficient of subgrade reaction modulus for embedded retaining structures

Embedded retaining structures are widely adopted for supporting deep excavations in urban areas. Their stability primarily relies on the lateral passive resistance of the ground. Usually, the ground resistance is considered as a set of elastic soil springs (i.e. Winkler's springs approach) following routine engineering design code. Such simplified subgrade reaction method is capable of simulating general soil-structure interaction and predicts retaining structure behavior largely in line with field observations. However, in practice, it’s difficult to determine the horizontal coefficient of subgrade reaction modulus, i.e. parameter m (the stiffness of soil springs k = mz) due to lack of experiments and soil uncertainties. In this study, an inverse analysis was conducted to determine the stiffness of soil springs m against the inclinometer data of lateral wall displacements for a deep excavation case in Shanghai clay. The inverse calculation of m involved an iterative computation process aiming to minimize the discrepancy between the computed wall displacements and the field measurements. In particular, sensitivity analyses were conducted to evaluate the effect of horizontal soil springs in different soil layers on the lateral wall displacements. Results show the m values of the soil layers play an important role on the lateral displacement, while the optimization of m may greatly improve computational accuracy and efficiency. The findings derived from the computed results and the comparison with the field measurements will provide some guidance for the design of embedded retaining wall in Shanghai clay.