Engineering Properties of The Bukit Timah Granitic Residual Soil in Singapore
Engineering Properties of The Bukit Timah Granitic Residual Soil in Singapore
Engineering Properties of The Bukit Timah Granitic Residual Soil in Singapore
com
ScienceDirect
Underground Space 4 (2019) 98–108
www.elsevier.com/locate/undsp
Received 4 May 2018; received in revised form 17 July 2018; accepted 18 July 2018
Available online 16 August 2018
Abstract
Extensive site investigations were conducted prior to the construction of a Mass Rapid Transit project in Singapore that was predom-
inantly in the Bukit Timah Granite (BTG) formation residual soil. This paper evaluates the engineering properties of the BTG formation
residual soil based on data from 208 site investigation boreholes from four different sites. Based on the results from 2481 conventional
laboratory tests and 1192 in-situ tests, this paper summarizes the engineering properties of the highly variable BTG residual soil, includ-
ing conventional composition analysis, index and hydraulic properties, and strength and deformation parameters required for geotech-
nical analysis and design. Based on these results, the BTG formation is found to be quite heterogeneous. As the degree of weathering
decreases with depth from the top of the formation, the BTG residual soil becomes sandier, with reduced silt and clay fractions. The
coefficient of permeability and the compression index of the BTG residual soil vary significantly. In addition, the empirical equations
relating the shear strength (index) to the standard penetration test (SPT)-N, as well as the equations and charts for determining stiffness,
are proposed. These findings, together with the proposed equations or charts, can be used for design guidance of similar projects related
to granitic residual soils in Singapore.
Ó 2018 Tongji University and Tongji University Press. Production and hosting by Elsevier B.V. on behalf of Owner. This is an open access article
under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
Keywords: Bukit Timah Granite; Residual soil; Shear strength; Stiffness; Laboratory testing; In-situ testing
https://doi.org/10.1016/j.undsp.2018.07.001
2467-9674/Ó 2018 Tongji University and Tongji University Press. Production and hosting by Elsevier B.V. on behalf of Owner.
This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
W.G. Zhang et al. / Underground Space 4 (2019) 98–108 99
2 BTG formation
Fig. 1. Varying ground conditions from a geophysical survey (LTA Factual Geotechnical Report, 2008).
100 W.G. Zhang et al. / Underground Space 4 (2019) 98–108
Table 1
Bukit Timah Granite weathering grades (after BS 5930:1999).
Grade Classifier Typical characteristics from BS 5930:1999 Typical descriptions from the BTG formation
GVI Residual Soil derived by in situ weathering but retaining none Residual soil recovered as very soft to very stiff slightly gravelly fine
Soil of the original texture or fabric to coarse sandy silt or loose to dense slightly silty, slightly gravelly
fine to coarse sand
GV Completely Considerably weakened. Slakes. Original texture Completely weathered to stiff to hard fine to coarse sandy silt or dense
Weathered apparent to very dense fine to coarse sand
GIV Highly Large pieces cannot be broken by hand. Does not Highly weathered, completely discolored rock to very dense silty sand
Weathered readily disaggregate (slake) when dry sample and gravel with intact rock fragments, or highly fractured rock with
immersed in water low SCR and very low RQD — generally less than 10%, usually 0%
GIII Moderately Considerably weakened, penetrative discoloration Moderately weathered, fractured, moderately strong to extremely
Weathered Large pieces cannot be broken by hand strong rock. Noticeable discoloration. Makes a dull or slight ringing
sound when struck by hammer
GII Slightly Slight discoloration, slight weakening Slightly weathered, moderately to slightly fractured, strong to
Weathered extremely strong rock
GI Fresh Unchanged from original state Fresh very strong to extremely strong intact rock with original
fractures
35
30
25
Number
20
15
10
0
0-2 >2-4 >4-6 >6-8 >8-10 >10-12 >12-14 >14-16 >16-18 >18-20 >20-22
GVI thickness (m)
Fig. 3. Variation of the thickness of the GVI layer.
50
40
30
Number
20
10
0
0-3 >3-6 >6-9 >9-12 >12-15 >15-18 >18-21 >21-24 >24-27 >27-30 >30-31
GV thickness (m)
Fig. 4. Variation of the thickness of the GV layer.
W.G. Zhang et al. / Underground Space 4 (2019) 98–108 101
40
30
Number
20
10
0
10 >10~15 >15~20 >20~25 >25~30 >30~35 >35~40 50
silt and the clay fractions of the soil are affected, as shown particle density, bulk density, and dry density with depth.
in Fig. 6. The clay fraction decreases significantly with The plot indicates a reduction in the particle density and
depth for the GV layer, while the silt content also decreases an increase in the dry density and bulk density with increas-
with depth. The sand fraction, in general, increases with ing depth. However, these changes with depth are not sig-
increasing depth. This finding is consistent with the results nificant. The standard penetration test (SPT) results in
reported by Zhao, Broms, Zhou, and Choa (1994) and Fig. 7(c) demonstrate that SPT-N values generally increase
Rahardjo, Satyanaga, Leong, Ng, and Pang (2012). with depth. The soil varies from soft fill at the ground sur-
face to hard residual soil above the weathered rock. Fig-
4.2 Index properties ure 7(d) indicates a clear decrease of the water content
with increasing depth.
Figure 7(a) shows the liquid limit (LL), plastic limit Figure 8 shows the distribution of the bulk density, with
(PL), and plasticity index (PI) against depth, based on the distribution concentrated in the 17002000 kg/m3
the data from borehole DT2439. It is obvious that down range. It can be observed that the bulk density follows a
to the depth of 21 m, the three indices exhibit similar normal distribution, and the most probable value is in
dependence on depth. In general, PI decreases rapidly with the 18001900 kg/m3 range.
increasing depth, while PL and LL do not show any Determination of the Atterberg limits is important for
conclusive trends. Figure 7(b) shows the variation of the the investigation of soil behavior. The plot of the Atterberg
limits for the residual soil on the plasticity chart is indica-
tive of the soil’s clay mineral composition. Figure 9 shows
Distribution (%) the relationship between PI and LL, along with the Atter-
0 25 50 75 100 berg A, B, U lines. Most of the data points for the BTG
0 formation residual soil are below the A-line, indicating that
clay silt sand
they consist of mainly silts, as already shown in Fig. 6. The
5 dependence of the soil composition on depth, in Fig. 6, also
tends to show the predominance of silt. The BTG forma-
10 tion residual soil is generally described visually as a silty
clay or a clay-like silt. In addition, the Atterberg limits,
Depth (m)
Fig. 6. Variation of the soil composition with depth, for the BTG residual Both laboratory (falling head test) and field tests (single
soil (based on DT2439). packer, rising head, and falling head) were conducted to
102 W.G. Zhang et al. / Underground Space 4 (2019) 98–108
Fig. 7. Variation of index properties with depth for the BTG residual soil (based on borehole DT2439).
300
250
200
Number
150
100
50
0
1.5 >1.5~1.6 >1.6~1.7 >1.7~1.8 >1.8~1.9 >1.9~2.0 >2.0~2.1 >2.1
obtain the hydraulic properties of the residual soil. 5 Mechanical properties of the BTG formation residual soil
Figure 10(a) shows the coefficient of permeability (k) for
the GVI and GV layers, at different depths, for eight 5.1 Strength
laboratory test samples. Figure 10(b) shows the permeabil-
ity for the GVI and GV layers, for the data from 56 field Figure 11(a) shows the SPT-N result vs. the depth below
tests. There is a considerable scatter in the permeability the ground surface for this project. Figure 11(b) shows the
for both the GVI and GV layer soils, with the majority SPT-N result with respect to the top of the GVI layer. The
in the range 1 104 to 1 106 m/s. These findings are plots show that the degree of weathering decreases with
consistent with the study by Forsythe and Pearse- depth from the top of the formation. Although there is a
Hawkins (2014). considerable scatter in the two plots, a lower bound of the
W.G. Zhang et al. / Underground Space 4 (2019) 98–108 103
60 cu ¼ 3:025N ð1Þ
B-Line
50 U-Line However, as there is a considerable scatter in the data, a
A-Line spline function is proposed
40
PI = 0.9LL-7.2 cu ¼ 5:627N ðfor N 6 15Þ ð2aÞ
PI (%)
30
cu ¼ 84:4 þ 0:508ð100 N Þ ðfor N > 15Þ ð2bÞ
20
The spline-based fit provides a slightly higher estimation
10
PI = 0.73LL-14.6 of cu for SPT-N below 30 and a conservative estimate of cu
for SPT-N above 30.
0
0 20 40 60 80 100 In a previous study of UU test data from a single site,
LL (%) Wong (2014) proposed the empirical equation of (cu/N)
= 19N0.44. In this paper, based on the 392 data sets, a sim-
Fig. 9. Plot of PI vs. LL, with the Atterberg A, B, U limit lines.
ilar empirical regression equation is proposed, as shown in
Fig. 14.
SPT-N value with depth can be obtained as shown. Similar
trends were observed by Leong, Rahardjo, and Tang (2003). 5.2 Stiffness
Figure 12 presents the cohesion and friction angle
results for 769 data points, vs. the SPT-N result; it is obvi- To understand the soil stiffness, pre-bored pressuremeter
ous that no accurate empirical equations exist that could tests were conducted on the BTG formation residual soil,
relate the cohesion or the friction angle to the SPT-N using an Oyo-type pressuremeter Elasmeter 200, where
result, owing to the significant scatter. Table 2 presents radial displacements were directly measured using LVDTs
the categorical average mean and standard deviation for (Zeng and Huang, 2016). Goh, Cham, and Wen (2011) pro-
the cohesion and friction angle, represented by the red, yel- posed an alternative method to interpret the unload-reload
low, and black lines, respectively, based on the SPT-N val- portion of the pressuremeter by examining the elastic mod-
ues. With increasing N, both the average cohesion and ulus with respect to the corresponding radial strains (which
friction angle increase. Cohesion increases significantly, would be equal to half the shear strain in the cavity wall),
while the increase is less significant for the friction angle. instead of assuming a linear function to work out a single
In addition, the standard deviation of cohesion increases unloading modulus.
with increasing N, while the standard deviation of the fric- Figure 15 shows the secant modulus of a particular pres-
tion angle decreases, indicating a smaller scatter. suremeter curve (Ep) plotted against its corresponding
Unconsolidated-undrained (UU) triaxial compression radial strain (er = R/Ro). The pressuremeter modulus
tests were conducted on the BTG formation GVI and decreases, even within the range where a best-fit line is used
GV layer soils, to characterize the undrained shear strength to estimate the unload-reload modulus in soil investigation
(cu). Figure 13 shows the undrained shear strength vs. the reports. For example, as the radial strain increases from
SPT-N result, for 392 data sets. For these data, the best lin- 0.2% to 1% within the reloading curve, the elastic modulus
ear relationship between the undrained shear strength and decreases from 307 MPa to 103 MPa. This rapid degrada-
SPT-N is tion of the pressuremeter stiffness with the radial strain
Fig. 10. Coefficient of permeability vs. depth, for the BTG formation residual soil.
104 W.G. Zhang et al. / Underground Space 4 (2019) 98–108
(a) SPT-N variation with depth from ground surface (b) SPT-N variation with depthf rom top of BTG
Fig. 12. Relationship between: (a) cohesion and SPT-N, (b) friction angle and SPT-N.
Table 2
Strength of the BTG formation soil based on CU tests.
SPT-N of soil Average lc and the standard deviation rc cohesion (kPa) Average l/ and the standard deviation r/ friction angle (°)
N 10 lc = 8.8, rc = 10.1 l/ = 28.6, r/ = 8.2
10 < N 30 lc = 13.8, rc = 21.5 l/ = 29.8, r/ = 7.0
N > 30 lc = 18.4, rc = 24.9 l/ = 30.5, r/ = 6.6
W.G. Zhang et al. / Underground Space 4 (2019) 98–108 105
Fig. 15. Strain-dependent behavior of the pressuremeter modulus (based on borehole DT2207).
106 W.G. Zhang et al. / Underground Space 4 (2019) 98–108
5.3 Compressibility
Table 3
Pressuremeter modulus for the Singapore BTG formation residual soil.
Soils of BTG Pressuremeter moduli correlated to SPT-N (MPa)
0.1% radial strain 0.2% radial strain 0.5% radial strain 1.0% radial strain
Goh et al. (2012) 7.8 N 4.5 N 3.0 N 1.8 N
This study 11.3 N 7.1 N 3.8 N 2.4 N
W.G. Zhang et al. / Underground Space 4 (2019) 98–108 107
Fig. 18. Charts for choosing the pressuremeter modulus with SPT-N and radial strains.
(a) Pre-consolidation pressure vs. depth (b) Compression index vs. depth
Fig. 19. Variation of the pre-consolidation pressure and compression index with depth, based on 4 sites.
References