Evaluation of Soil Variability Influence on Deep Excavation Analysis–Simplified Approach

Abstract

A simplified procedure is developed to provide a tool for assessing the variation of the computed excavation responses, including wall and ground deformations, induced by input parameter uncertainties through a finite element solution. A well-documented case history is employed for demonstration. The proposed scheme appears accurate and easy-to-use in the propagation of parameter uncertainties. Moreover, this procedure is capable of assessing the variability effect of either a single or multiple input parameters. Further, the sensitivity of the uncertainty effect on the excavation characteristics is also examined. The proposed procedure enables a valuable insight regarding how much variation of wall and ground movements that can be expected as a result of parameter uncertainties.

RIFT: Robust and interpolation-free technique for objective functions in geotechnical inverse analysis

“…In general, the soil strata included alternating layers of clay and sand. There was a thick clay layer from a depth of 8 m to a depth of 33 m. Such a layer has been demonstrated to exert a strong effect on excavation responses [28][29][30]. The gravel layer remained below a depth of 46 m. There were seven excavation stages, with a final excavation depth of 19.7 m. The excavation width was 43 m. The water level was 2 m below the ground surface.…”

“…The details of the parameter derivation for the sand layers can be found in previous publications [17,31]. In addition, for simplicity, the same parameters were used for all three clay layers because the middle layer (8.0-33.0 m) was of much greater importance to the wall and ground deformations than were the other two layers [28][29][30]. In other words, four parameters were evaluated in the inverse analysis: k, j, M, and m.…”