Deep soil mixing (DSM) is an established ground improvement technique employed in civil projects. Despite the superiority of field tests for understanding this technique, their high cost has directed researchers' focus on laboratory tests, resulting in limited attention given to operational factors. Consequently, in current research, a small-scale DSM setup was developed to investigate the influence of operational factors such as mixing time and execution procedure on strength and deformation characteristics of laboratory-scale DSM columns. For the installation of DSM columns, mixing times of 130, 190 and 250 seconds were used, together with normal and zigzag execution procedures, cement dosages (alpha) of 300, 400 and 500 kg/m(3), and total water-to-cement (W-total/C) ratios of 2.5, 3.0 and 3.5. Laboratory samples were also prepared using the same alpha values and (W-total/C) ratios for comparison with DSM columns. The sand bed was prepared with 5 % and 30 % moisture contents. Experimental observations showed that saturating the sand bed enhances the mixing quality by preventing slurry water infiltration into the soil surrounding the DSM columns. Results indicated that increasing mixing time and adopting zigzag execution procedure improved mixing quality, unconfined compressive strength (UCS), secant modulus (E-50), and strain at maximum stress (epsilon(Maximum Stress)), whilst reducing strength variability. Moreover, the outcomes showed that UCS and E-50 of samples have a direct and inverse relationship with alpha and (W-total/C), respectively, and that the nature of these relationships, not their magnitude, were not affected by mixing time and execution procedure. Additionally, findings indicated that the failure mode of DSM samples was influenced by operational factors, whereas (E-50/UCS) ratio was not.

The recent construction of an underground mass rapid transit (MRT) station in Singapore involved 21 m deep excavations within underconsolidated marine clay. The lateral earth support system comprised 1 m thick diaphragm walls socketed into the underlying Old Alluvium and 4 levels of preloaded cross-lot struts. Deep soil mixing (DSM) and jet grouting piles (JGP) were used to improve up to 15 m thickness of the marine clay formation. Field monitoring data showed that these ground improvement processes caused large outward deflections of the diaphragm wall panels at some locations prior to the excavation and may have caused yielding within the wall panels. In this paper, the impacts of these prior wall deformations on the subsequent performance of the excavation support system are investigated. The measured performance at two indicative cross sections is compared with results from simplified 2D finite element analyses. The analyses simulate the effects of ground improvement through prescribed boundary pressures and represent the yielding of the diaphragm wall panels through zones of reduced bending stiffness. We show that large outward wall deflections and curvature observed during jet grouting at one contribute to higher inward wall movements and strut loads measured during excavation, while smaller movements (and curvature) prior to excavation at a second similar cross cause negligible change in the performance of the temporary earth retaining system. The results highlight (1) the importance of controlling ground movements associated with ground modification processes such as jet grouting, (2) the uncertainties in estimating mechanical properties for the improved soil mass, and (3) the need to improve the representation of non-linear, flexural properties (M-kappa) of reinforced concrete diaphragm panels.