Page 135 - Numerical Analysis and Modelling in Geomechanics

116 C.W.W.NG AND Q.SHI
For investigating the influence of transient rainfall intensity on pore pressure
distributions in the slope and hence slope stability, three 1 in 10-year return daily
rainfall intensities were considered (see Table 4.1). Except for the rainfall
intensity, relevant mean values (e.g. k =4.8×10 −5 m/sec) from Table 4.1 were
s
adopted for the transient parametric seepage analyses using SEEP/W. Before any
transient seepage analysis is carried out, the initial groundwater condition is set
up by conducting a steady state seepage analysis (i.e. Q=0) under the specified
boundary hydraulic head conditions. This is based on the assumption that evapo-
transpiration from the ground surface is negligible. For comparing the ground
response with various rainfall intensities with an upslope specified water table at
62 mPD (9 m below ground level), critical sections regarding the stability of the
cut slope are considered (see Figure 4.4), which include Section A-A near the
toe, Section B-B at the mid-height, and Section C-C above the cut slope. The
computed initial pore pressure distribution varying with depth at these three
critical sections is shown in Figure 4.5. As expected, the pore water pressures are
negative above the main water table. The pore pressure distributions vary
linearly with depth, with suction values up to 50 kPa predicted within the top
10m from the exposed slope surface. The predicted initial main water tables at
these three critical sections, A-A, B-B and C-C, are at 39.8 mPD, 41.3 mPD and
45.5 mPD respectively (i.e., 2.3 m, 4.4 m and 6.5 m below the corresponding
ground levels). The numerical predictions are in reasonable agreements with
field measurements described previously. This implies that the realistic
equipotential values were applied at the upslope and downslope boundaries.
Since the pore water pressure responses to various rainfall intensities are
similar in nature, only a set of typical pore water distributions (along three
critical sections) with depth is shown in Figure 4.6. During the one-day rainfall of
267 mm/day, it can be seen that the pore pressure response varies from section to
section. At Section C-C, the ground surface is at 52 mPD and the main
groundwater table (45.5 mPD) is at 6.5 m below the ground surface, the deepest
among the three sections considered. By comparing with the initial conditions
shown in Figure 4.5, it is clearly that the main water table is hardly affected by
the 1-day rainfall. Based on some typical values of CDG, equation (4.1) would
predict that the advancement of 100% saturation wetting front is in the order of
15–20 m. This significant discrepancy is likely caused by the low unsaturated
water permeability that existed in the ground, which was not taken into account
by equation (4.1).
However, the pore water pressure regime above the main groundwater table at
Section C-C is substantially affected by the rainfall. The magnitude of negative
pore water pressure is reduced by a considerable amount, which depends on the
intensity of the rainfall simulated. For the rainfall intensity of 267 mm/day, a
perched water table appears at 51.8 mPD. This offers a theoretical illustration to
support the explanations of some of the observed slope failures in Lantau Island
in Hong Kong (Wong and Ho [29]).