ABSTRACT
Deep excavations in soft ground often require additional stabilization through ground improvement (GI). Some of the common ground improvement methods involve the use of Jet Grouting Piles (JGP), Deep Soil Mixing (DSM) or Wet Speed Mixing (WSM). JGP, DSM and WSM are achieved by mixing the soil with cement and water to create a structure that performs well under compressive, and to a lesser degree tensile, loadings. These treated soil blocks help to reduce wall deflections by providing enhanced passive resistance. Overall, ground improvement in deep excavations is an effective method to provide temporary excavation support, although it may significantly increase the costs for underground construction in soft clays.One of the major challenges in ground improvement works is the selection of the treated soil parameters during the design stage, when there is usually little or no data available. This paper focuses on the selection of appropriate ground improvement parameters during the design phase. Unconfined compressive strength (UCS) data from more than 400 cored samples of cement-treated soil, together with field instrumentation measurements, are used to carry out a back analysis of an existing excavation. The design parameters for the treated soil were selected following the different statistical approaches adopted in the Eurocode, Japan and US practices. The back analysis results show that, for the improved ground at this particular excavation, the EC7 adoption of characteristic strength and stiffness values based on the 50% fractile leads to the closest agreement between the measured and predicted wall deflections. Using random finite element analysis, the paper also examines how the strength variability within the improved soil block influences the overall stability of the excavation system.
