4.3 Excavation Supported by Diaphragm Wall with Upper Strut

4.3.1 Preparation and instrumentation

In an actual construction, propped diaphragm walls are more commonly used than the unpropped ones for excavation in soft clay. As shown in the last section, the unpropped diaphragm wall has only a limited ability to maintain stability of the wall unless a large penetration depth is employed. So it is an unsafe and uneconomical way to retain a deep excavation in soft clay. The deformation and failure mechanism of a strutted diaphragm wall during an excavation will be simulated in the centrifuge and studied.

For test Ts1, the c_u of upper soft clay was 17.5-19.0 kPa and the c_u of the lower stiff soil was 173 kPa. The thickness of the upper and lower soil layers of soils was 200 mm and 170 mm, respectively. The other details of soil and geometry of model as well as the instrumentation are given in Table 3.4 and Figure 4.14.

Similar to test To1, the bag containing the zinc chloride solution and model wall was placed before the test. The dimension of the cut replaced by the bag was 150 mm in both depth and width. The embedment of wall into the lower stiff soil was 25 mm. All these dimensions mentioned were the same as those of test To1. The excavation stages and layout of instrumentation including LVDTs, PPTs and the strain gauges on the wall were placed as displayed in Figure 4.14. A lateral strut was then fixed at the retained ground level in front of the wall. This strut was to simulate the top slab in the so called top-down construction where excavation is carried underneath a floor slab after it is cast. The same draining rate of heavy solution as that of test To1 was applied to simulate a rapid excavation.

4.3.2 Pore pressures and soil movement

The layout of PPTs for Ts1 is shown in Figure 4.14. The readings of PPTzncl in Figure 4.15 represented the excavation process and also allowed the excavation depth to be obtained. PPT6 was embedded in the soil below the bag and expected to follow a similar trend and magnitude as the reading of PPTzncl. PPTs 1 to 5 display the changes in pore pressure behind the retaining wall. PPT1 was located nearest to the wall and experienced a sharp drop within the first 120 seconds. But after that, the reading of PPT1 rebounded rapidly, as shown in Figure 4.16. At PPT3 and PPT4 which is further away from the excavation, the drop of pore pressure is later than those nearer the excavation. The sharp and early rebound of PPT 1 and PPT2 is likely to be due to local cracking which makes air and water enter the transducers located in the large deformed soil mass.

The photographs taken from video pictures at different excavation stages of test Ts1 are shown in Figure 4.17. In stages 1 and 2, it could be observed that with the strut, the diaphragm wall was able to retain the soil and some bending in the wall was noticeable. At stage 3, the bottom of the excavated side heaved and failure around the toe appeared. After this, the wall began to rotate around the ¡®hinge¡¯ at the top. Instead of the Rankine¡¯s linear slip line observed in test To1, the failure pattern was noted to be nearly circular. As excavation continued, a full failure developed. From the sequence of pictures and also the pore pressure data, the soil was deemed to fail at a depth of 7.5 m. Thus, with the presence of a top strut, failure was clearly due to a toe ¡®kick-out¡¯ with a small local failure first appearing at the toe. In contrast, for test To1, failure seems to occur almost instaneously along the entire slip line.

The variation of surface settlements with time for the four LVDTs are shown in Figure 4.18. The settlement profiles obtained are shown in Figure 4.19. By comparing Figure 4.19 with Figure 4.6 and 4.11, the differences between an unpropped excavation (test To1) and a propped excavation (test Ts1) are identified as follows. For the same excavated depth, the magnitude of the surface settlement of test Ts1 is clearly smaller than those of test To1. The shape of the surface settlement profile is also different as the maximum settlement is now located at 150 mm from the wall in test Ts1, instead of very near to the wall as in test To1. This is consistent with the failure patterns observed. In test To1, the soil failed near the top and slip occurred with the entire triangular wedge moving downwards. In test Ts1, it was the toe that failed first, and thus the soil away from the wall moved downwards.

4.3.3 Deformation and Bending Moment of Wall

The manner of the wall deflection can be approximately interpreted from Figure 4.17. In the first two stages, bending occurred in the centre while the two ends were restrained by the upper strut and the wall embedment. With further excavation, the toe began to fail and the wall started to rotate around the ¡®hinge¡¯ at the upper strut. The bending moment in the wall was estimated from the strain gauges placed along the wall and the results are shown in Figures 4.20 and 4.21. The magnitude of bending moments increased gradually with the excavation. During the time from 110 seconds to 130 seconds when the shear failure surface appeared clearly, the rate of increase in bending moments reduced. The wall bending moment continued to increase until the total collapse of the retaining system. Figure 4.21 shows the profiles of bending moment at 30 seconds, 60 seconds and 120 seconds respectively. The bending moment is at its maximum around the mid-height of the wall.

4.3.4 Equilibrium analyses for stability

Based on the conventional theory on earth pressure, a stability analysis on test Ts1 was also conducted and the results are shown in Table 4.3 in which the effect of zinc chloride solution below the excavated depth was also considered. By taking moment about the strutted point A, the critical excavation depth is determined to be about 6 m. The corresponding critical depth observed in Figure 4.17 is around 7 m. This agian confirms that for a nearly uniform soil, conventional earth pressure theory works reasonably well.