2.3 Field Observations of Excavation System

As stated in Chapter 1, ground movement is becoming a major concern in the design of excavation support system. In built-up areas where the underlying ground contains a soft clay layer, the effects of excavation on adjacent critical structures such as historical buildings, pile foundations and MRT tunnels or stations (Figure 1.1) need to be investigated thoroughly. The interaction between the excavation area and existing structures must be studied carefully. In Singapore, considerable efforts have been devoted to the excavation monitoring and its effects on the adjacent buildings.

2.3.1 Lateral wall deflection

For the case of flexible walls, Lee et al. (1986) and Yong et al. (1989) reported that the maximum deflection of sheet piles generally occur within 1 m to 2 m below the excavation level if the soil is marine clay. Lee et al. (1986) reported that the rate of lateral wall movement during excavation ranges from 1.5 to 3.0 mm/day. Yong et al. (1989) reported the rate of increase in maximum displacement is about 1 to 3 mm/day during excavation, and 0.4 to 0.7 mm/day after excavation.

Typical diaphragm wall deflection profile in soft soil can be found in some literature, see for example Whittle et al. (1993). Some of the field studies of diaphragm walls in Singapore constructed in marine clay reported a maximum wall deflection of 110 mm for a 14.5 m deep excavation (Nicholson, 1987), and 135 mm for a 17.9 m deep excavation (Lim et al., 1991).

Prior to installing the first level of struts, a cantilever condition exists in the wall. With the installation and pre-loading of the first strut, lateral movement at the top of the wall is reduced and maintained at a fairly constant magnitude. However, excessively large movement can develop if excavation is allowed to proceed too far before a strut is installed. Insertion and pre-loading of struts serve to restrict movement of the wall until the start of the next stage of excavation (Lee et al., 1986).

Generally, the maximum lateral wall movement increases when excavation proceeds until the next level of struts has been installed (Lee et al., 1986; Tan et al., 1995). Once the struts are installed, the lateral movement is generally restricted or reduced ( Lee et al., 1986). Bono et al. (1992) reported that for an internally braced diaphragm wall in Boston, the greatest lateral displacement is at the centre of the wall. The displacements at the corners and in the short edge are significantly smaller, an illustration of the effect due to the presence of corners.

Lambe et al. (1972) reported that the lateral movements in soil supported by a diaphragm wall are much smaller than those in soil supported by sheet piles. Lateral movements near the sheet pile are 5 to 7 times as large as those near the diaphragm wall. Whittle et al. (1993) reported that the lateral movements in soil are significantly less than that measured in the wall. However, the lateral displacement profile of the soil and wall are very similar.

2.3.2 Ground surface settlements

Peck (1969) developed a settlement chart to estimate the magnitude and extent of ground settlement with respect to the soil conditions at an excavation site. This settlement chart was developed using data obtained from excavations in Chicago and Oslo clays supported by soldier piles or sheet piles with cross-lot struts or tie-backs. Another settlement chart proposed by O'Rourke et al. (1976) can be used to estimate settlement according to the type of soil retaining system shown in Figure 2.5. These empirical methods are widely used by practising engineers because of their simplicity. However, as the site geometry, supporting systems, construction sequence, dissipation of pore pressure and the effect of nearby structure are different, the empirical method may not reflect the behaviour rationally.

If wall deflection is a major cause of ground surface settlement during excavation, the total volume of ground settlement should be equal to the volume covered by the deflection of the wall under undrained conditions in saturated clays (Milligan, 1983). The relationship between maximum ground surface settlement and the maximum wall deflection is given by Mana and Clough (1981). The upper limit of the maximum ground settlement is equal to the maximum lateral wall deflection, and for most excavations, the maximum ground settlement is between 0.5 to 0.75 of the maximum wall deflection.

Ou et al. (1993) proposed an empirical formula to predict the ground surface settlement profile at the mid-section of an excavation where the behaviour may be characterised by a plane-strain condition. This formula was derived based on excavations with alternating silty sand and silty clay soil profiles. They also reported that the extent of ground settlement behind the wall increases with excavation depth. Lambe et al. (1972) observed that the settlement would continue after the completion of excavation due to general ground water lowering and other construction activities.

2.3.3 Stress state of soils near excavation

Henkel (1970) showed that there are two kinds of behaviour on each side of the retaining wall. One side retained by the excavation support wall is in the active mode, the other is in the passive mode. Clough and Tsui (1974) further reported that clays under undrained active condition demonstrated up to twice the strength of clay loaded under passive conditions, this suggests an early recognition of the anisotropy of the soil. The above findings are strictly applicable to total stress analysis under short-term situation and do not explain the time-dependent mechanism of soils in front of the supporting system.

Lambe (1970) noted that the stress and strain in soil near an excavation are related to undrained and drained conditions depending on the rate of excavation. During an actual excavation, over a finite period, the soil conditions are partially drained.

Burland and Fourie (1985) illustrated the stress path behaviour associated with drained excavation at the back of an embedded cantilever retaining wall in stiff clay and proposed that the testing of soil should be under the same conditions. The initial state of stress is on the K_o (coefficient of earth pressure at rest) line and during successive excavation, both the vertical and horizontal effective stresses reduce causing the stress state to approach the failure (K_f ) line.

The case histories of pore pressure measurements in clay behind a number of supported excavations in Oslo, Norway and Boston have been reported by Lambe et al. (1972) and DiBiagio and Roti (1972). Dissipation of excess negative pore pressure is noted to start immediately, and continue well beyond the completion of the excavation. Thus the idealised undrained condition, if it exists at all, is short lived. Many of the case histories reported by Lambe et al. (1972) observed a post excavation decrease in the factor of safety with time. They attributed this to the change in effective stress around excavation as a result of the dissipation of excess negative pore pressure which reduces the soil strength. The time-dependent behaviour of excavation observed is associated with this dissipation of negative pore water pressure.