6.7 Analyses on Test Series 2

In test series 2, the soil condition is nearly consolidated and thus is easier and more accurately simulated by FEM analysis. In this test series, PPTs were installed at locations shown in Figure 5.1. The corresponding locations in the finite element mesh are shown in Figure 6.1. Point P1 is located 25 mm behind the wall at a depth of 100 mm. The positions of the other points are also located within the potential failure zone on the active side of the retained soil.

6.7.1 Changes of excess pore water pressures by simulating Ts5

In test Ts5, zinc chloride solution was smoothly drained down by 50mm within 300 sec to simulate a 5 m deep excavation. After the end of the excavation, the test continued for 3 hours to examine the long term behaviour. Such process of excavation was modelled in the CRISP analysis by removing the preload that is equivalent to the pressure of zinc chloride solution on the model diaphragm wall and the base of excavation.

The measured and predicted negative excess pore pressures are modelled as shown in Figures 6.14 and 6.15 respectively. The measured drop in pore pressure at P1 is about 9 kPa while the corresponding result from analysis is 6.5 kPa at the same position. At P4 and P6 which are far away from excavation, the drop in pore pressure computed from FEM is larger than the measured value shown in Figure 6.15. Such higher value in numerical results is also exhibited in the results of ground settlements as presented in next section.

The pore pressure changes in the long term measured in the experiment are shown in Figure 6.14. The corresponding FEM results are shown in Figure 6.15. There is a rebound in the pore water pressure in the retained soil after excavation. In general, a reasonable agreement was obtained between the predicted and measured pore pressure changes through the predicted values tend to be high than the measured values.

In this analysis, the permeability of soil sample for upper layer of model in the test is take to be k =10^-8 m/s which is generally on the high side for Singapore marine clay. To investigate the effect of coefficient of permeability of soil, parametric studies using k = 10^-8 m/s, k = 10^-9 m/s and k = 5 x 10^-10 m/s are conducted. The differences of pore pressure changes at PPT1 in the three analyses are shown in Figure 6.16. With k = 10^-8 m/s, the highest value of permeability, the drop in pore pressure is about 12.5 kPa during excavation. After excavation, the rebound of pore pressure is evident. On contrast, with k = 5 x 10^-10 m/s, the drop in pore pressure is about 10 kPa during excavation. No rebound of pore water pressure is observed and there is still a very slow drop in pore pressure. This comparison indicates that for soil with very low permeability, the negative excess pore pressures develop for a very long time after the excavation process. The suction is not effective due to low permeability.

6.7.2 Ground surface settlements for test Ts5

The ground settlements versus time measured in test Ts5 during and after excavation are shown in Figure 6.17. By comparing the curves of LVDT2 in Figure 6.17 and the corresponding FEM predicted curve in Figure 6.18, the FEM results fit the experimental results relatively well. The differences between the FEM and experiments at other locations such as LVDT1, LVDT3 and LVDT 4 are obvious and similar trends could also be observed. The reason for such difference will be discussed in Section 6.8. The post excavation behaviour after 300 sec shown in Figure 6.19 was simulated with FEM and a similar trend can be found in Figure 6.18. The magnitude of settlement at LVDT2 was 0.18 mm at 300 sec when excavation stopped , and 0.22 mm at 3600 sec (Figure 6.19). For FEM, the corresponding results are 0.26 mm and 0.41 mm (Figure 6.20). Again, it can be seen that the analyses generally produces reasonable result, except for LVDT 1 where the trend relative to the other LVDTs is wrong. As explained in the previous section, this is due to the fact that slip actually occurs, but this is not properly captured in the analysis. Thus in the experiment, the settlement at LVDT 1 is considerably greater than in the analysis, a trend not followed by the other LVDTs.

6.7.3 Stress paths of soil for test Ts5

As shown in Figure 6.21, the initial stress condition is located at Ao on the Ko line in stress space. With the draining of zinc chloride, the wall is forced to move towards the excavated side, causing a lateral relaxation in the retained soil. The stress path at a typical position P1 is shown as curve P1 in Figure 6.21. For the rate of excavation employed in this series of tests which is much slower than in the first series, the time dependent effect is likely to be more significant during the excavation. As mentioned earlier, in the first series of tests, the draining rate is much faster and both the test and numerical data suggest a nearly undrained behaviour. As shown in Figure 6.21, the deviatoric stress during excavation increases sharply while the normal effective stress reduces by about 10 kPa. The stress moves towards the critical state line but is still a distance away from it. During the post-excavation stage, the pore pressure dissipates and as a result the soil swells and softens. As shown by curve P1, the soil now moves closer towards the critical state line. During the post excavation stage, the deviatoric stresses at another two points P4 and P6, which are away from the wall and shown by curves P4 and P6, are not increased as sharply as P1.