Documente Academic
Documente Profesional
Documente Cultură
Keywords: liquefaction, dilatometer test, cone penetration test, shear wave velocity
ABSTRACT: Investigations at Treasure Island involving CPT, DMT, SPT and Vs measurements provide a
relatively rare opportunity to compare these methods in predicting liquefaction. All the in-situ techniques
were relatively consistent in predicting liquefaction within a depth range of 3 to 7.5 m below the ground
surface. The factors of safety predicted by the CPT and DMT were quite consistent except at shallow depths
where the DMT gave higher values owing to the higher Ko in this zone. Analysis of the CSR and KD data
points within the depth range from 3 to 7.5 m suggests that the liquefaction boundary curve is reasonable for
sand but may be somewhat unconservative for silty sands. The factor of safety predicted by the Vs correlation
was consistently lower than other methods owing to insensitivity to higher Ko values and higher silt content.
1 INTRODUCTION et al. 1998, Robertson & Wride 1998, and Idriss &
Boulanger 2004). However, it is sometimes difficult
Procedures for assessing the liquefaction potential of to determine which method to rely upon when there
sands and silty sands have been developed for a are conflicts between competing methods as has
number of in-situ tests including: the Standard been noted by Rollins et al. (1998) as well as Liu &
Penetration (SPT) test (Youd et al. 2001), the Cone Mitchell (2006).
Penetration (CPT) test (Robertson & Wride 1998), This paper provides additional field performance
and the shear wave velocity (VS) test (Andrus & data than can be added to the database for
Stokoe 2000). liquefaction assessment with the DMT for both clean
More recently, Monaco et al. (2005), proposed a and silty sands. In addition, the results from the
method for predicting liquefaction using the KD DMT based approach are compared with results
value obtained from Flat Dilatometer (DMT) testing. from SPT, CPT and VS methods at the same site. The
Research has shown that KD is more sensitive than test results were obtained from a site on Treasure
VS to factors such as stress history, aging, Island, a man-made island in San Francisco Bay
cementation, and structure, which greatly increase where liquefaction was pervasive during the M6.7
the liquefaction resistance for a given relative Loma Prieta Earthquake in 1989.
density (Maugeri & Monaco 2006). Unfortunately,
the database for assessing liquefaction from DMT
testing is relatively small and additional results from 2 SITE CHARACTERIZATION AT
field test sites are necessary to improve the TREASURE ISLAND LIQUEFACTION TEST
reliability of the procedure. This is particularly (TILT)
important for profiles containing silty sand in
addition to clean sand. Treasure Island was constructed by building a rock-
As noted by several researchers, it is often useful fill berm on native shoal sands around the perimeter
to evaluate liquefaction using more than one in-situ of the island and placing dredged sand within the
test to confirm the potential for liquefaction (Rollins berm using hydraulic filling techniques. The test site
described in this investigation was located near the could be anchored to the ground for increased
interior of the island. This site was thoroughly reactive force. Readings were taken at 0.05 m
investigated in connection with a series of lateral intervals. The soil profile interpreted from the cone
pile load tests conducted after inducing liquefaction penetration testing along with profiles of cone tip
using small explosive charges. This test program, resistance (qc), sleeve friction (Fs), friction ratio (Rf),
known as the Treasure Island Liquefaction Test and soil behavior type index (Ic) are provided in Fig.
(TILT), has been described in a number of papers 1. The Ic value provides a detailed indication of the
(Ashford & Rollins 2001, Ashford et al. 2004, variation of soil types with depth. There is a
Rollins et al. 2005, Weaver et al. 2005). Prior to decrease in cone tip resistance from the ground
lateral pile load testing, about 0.9 m of the hydraulic surface consistent with overconsolidation, then the
fill was excavated to bring the ground surface closer value oscillates between 3 and 7 MPa within the
to the water table. Test results are referenced relative interbedded sand and silty sand layers typical of
to this excavated ground surface although hydraulic fill. The friction ratio from 0 to 7 m is
corrections required by the various methods are generally about 0.5%. The cone tip resistance
based on the ground surface elevation at the time of decreases to around 1 MPa in the clayey silt/silty
testing. clay layer while the friction ratio ranges from 0 to
3%. The Ic in the clayey silt layer is typically 2.6 or
2.1 Cone Penetrometer (CPT) Testing greater indicating that this layer is not liquefiable;
A cone penetration sounding was performed using however there are some thin interbedded layers with
the Univ. of Michigan truck mounted CPT rig which Ic below 2.6.
Generalized
Cone Tip Resistance, qc (MPa) Sleeve Friction, Fs kPa) Friction Ratio, Rf (%) Soil Behavior Type Index, Ic
Soil Profile
0 5 10 15 20 0 25 50 75 0.0 1.0 2.0 3.0 1.00 1.50 2.00 2.50 3.00 3.50
0 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
0 0 0 0
Clay
1 1 1 1 1
Clayey Silt to Silty Clay
Interbedded
Sand
2 and
2 2 2 2
Silty Sand
Depth Below Excavated Surface (m)
3 3 3 3 3
Silty Sand to Sandy Silt
4 4 4 4 4
Depth (m)
Depth (m)
Depth (m)
Depth (m)
5 5 5 5 5
6 6 6 6 6
Silty Sand
7 7 7 7 7
Clean to Silty Sand
8 8 8 8 8
Clayey Silt
9 9 9 9
Gravelly Sand
10 Silty Sand 10 10 10 10
11 11 11 11 11
Fig. 1. Soil profile interpreted from Cone Penetration (CPT) testing along with profiles of cone tip resistance (qc), sleeve
friction (Fs), friction ratio (Rf) and soil behavior type index (Ic).
Generalized Dilatometer Modulus, ED Material Index, ID Horiz. Stress Index, KD Horiz. Stress Ratio, Ko
Soil Profile (MPa)
0 20 40 60 0 1 2 3 4 5 0 10 20 30 0 0.5 1 1.5 2
0 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
0
0
0 0
1 1
1
1 1
Interbedded
Sand
and
2 2
2
2
2
Silty Sand
Depth Below Excavated Surface (m)
3 3
3
3
3
4 4
4
4
4
5 5
5 5
5
Clay
6 6
6 6
6
Silty Sand
Silt
7 7
7 7
7
Silty Sand
8 8
8 8
8
Clayey Silt
9 9
9 9
9
Sand
Silty Sand
10 10
10 10
10
Fig. 2. Soil profile interpreted from flat dilatometer (DMT) testing along with profiles of dilatometer modulus, material
index, horizontal stress index, and earth pressure ratio.
3 3
located about 15 m north and south, respectively of
4 4
the DMT/CPT soundings. Test borings were
performed with drilling mud to stabilize the hole. A
5 5 safety hammer, lifted with a rope and cathead
system, was used to perform the tests. Energy
6 6 measurements indicated that the hammer was
Silty Sand providing 57% of the theoretical free-fall energy on
7 7 average. The normalized penetration resistance
CPT1-1 (N1)60 was computed using the equation:
8 8 CPT1-6
Clayey Silt
Downhole (N1)60 = CNCEN (4)
9 9
0
0
Fine Sand
w/ Shells
1 (SP) 1
1
2 2
2
Depth Below Excavated Surface (m)
Interbedded
3 3
3
Fine Sand
and
Silty Sand
4 4 4
(SP-SM)
5 5 5
6 Fine Silty 6 6
Sand
(SM)
7 7
7
SPT 2
SPT 1
8 8
8
Gray Silty
Clay (CL)
9 9
9
Silty Sand
(SM)
10 10
10
Fig. 4. Soil profile from SPT borings along with profiles of (N1)60 and fines content.
3 LIQUEFACTION ASSESSMENT of about 3 m to 7.5 m below the ground surface.
This result provides a high degree of confidence that
According to the simplified procedure developed by this layer actually liquefied during the Loma Prieta
Seed & Idriss (1971), the factor of safety (FS) earthquake.
against liquefaction is given by the equation Despite the fact that the DMT based approach
was primarily developed for clean sand, the
FS = (CRR7.5/CSR) MSF (5)
computed factor of safety tracks the factor of safety
obtained with the CPT quite well in the depth range
where CSR = calculated cyclic stress ratio generated
from 3 m to 7.5 m where a number of silty sand
by the earthquake shaking; and CRR7.5 = cyclic
layers are encountered. In contrast, the factor of
resistance ratio for magnitude 7.5 earthquakes, and
safety from the DMT is considerably higher near the
MSF = magnitude scale factor = 102.24/Mw2.56, and
ground surface than predicted by the CPT or SPT
Mw is the magnitude of the earthquake under
methods. The high DMT-based factors of safety
consideration (Youd et al. 2001). For this study,
appear to be associated with higher Ko values within
which considers performance during the 1989 M6.7
this depth range (see Fig. 2). Although the CPT
Loma Prieta earthquake, the magnitude scaling
based factor of safety also increases in this range,
factor is 1.33.
presumably owing to the higher Ko values, they are
Seed & Idriss (1971) also developed the equation
still lower than predicted by the DMT approach. The
for the cyclic stress ratio (CSR) as follows
increased factor of safety predicted by the DMT is
likely a result of the increased sensitivity of the
CSR = (av /’vo) = 0.65(amax /g)(vo/’vo)rd (6) DMT to Ko effects relative to the CPT (Maugeri and
Monaco 2006). Because the database of field
where amax = peak horizontal acceleration at the performance data relative to the DMT approach is so
ground surface generated by the earthquake; g = limited, it is not possible to determine if the
acceleration of gravity; vo and ’vo are total and increased factor of safety is justified at the present
effective vertical overburden stresses, respectively; time.
and rd = stress reduction coefficient = 1.0 - 0.00765z As noted by a number of researchers (e.g.
for z < 9.15 m. Ground motion recordings on Ishihara et al. 1977, Seed 1979), the liquefaction
Treasure Island and ground response analyses resistance of sand clearly increases as the Ko value
conducted at a number of sites around Treasure increases. However, the correlations between CRR
Island indicate that amax in the vicinity of the test site and in-situ test parameters [i.e. qc, (N1)60, Vs, KD]
was approximately 0.16g (Rollins et al. 1994) have generally been assumed to be independent of
For simplicity, the cyclic resistance ratio (CRR) Ko because an increase in Ko is expected to produce
was generally computed based on recommendations a comparable increase in the in-situ test parameter as
by Youd et al. (2001) where the resistance ratio from suggested by Seed (1979).
CPT results is determined using techniques Salgado et al. (1997) examined the effect of Ko on
formulated by Robertson & Wride (1998) and the both liquefaction resistance and cone penetration
resistance ratio based on VS is based on charts resistance separately and concluded that the CRR vs
developed by Andrus & Stokoe (2000). With respect qc1 relationship was relatively unaffected by Ko for
to DMT results, the CRR was determined using the qc1 values less than 12 MPa and slightly
equation: unconservative for higher qc1 values. In contrast,
Harada et al. (2008) also investigated the influence
CRR=0.0107KD3–0.0741KD2+0.2169KD–0.1306 (7) of Ko on both liquefaction resistance and on SPT and
CPT resistance and found that both the CRR vs qc1
proposed by Monaco et al. (2005) where KD is the and CRR vs (N1)60 curves were conservative without
horizontal stress index obtained from the DMT. consideration of Ko effects. Harada et al. (2008)
A comparison of the factor of safety against recommend a suite of CRR vs qc1 curves to properly
liquefaction computed using the CPT, DMT, SPT account for Ko effects. Therefore, there is some
and VS approaches for determining CRR is provided controversy in the literature regarding the effect of
in Fig. 5. Although there are some notable Ko on liquefaction correlations based on in-situ tests
differences, the agreement between the various which is deserving of additional research. This issue
methods is generally quite good. For example all is particularly important in evaluating liquefaction
methods predict liquefaction (FS < 1.0) from a depth
resistance after ground improvement because ground Generalized Factor of Safety Against
improvement typically increases both the soil Soil Profile Liquefaction
density and Ko. 0.0 0.5 1.0 1.5 2.0 2.5
0 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
0
Interbedded
investigated. At shallow depths where Ko is high, the Sand
lower factor of safety may be attributed to Vs being and
2 2
Silty Sand
less sensitive to Ko than the CPT qc or DMT KD as
DMT
curve for silty sands. These curves would lead to Silty Sand
CPT
higher factors of safety for the silty sand layers and 7 7
Vs
better agreement with the other methods.
Typically, the boundary between liquefaction and SPT 1
8 8
no liquefaction for a given in-situ test parameter [qc, Clayey Silt SPT 2
(N1)60, Vs, KD] is evaluated by plotting the CSR for a
given event versus the in-situ parameter for the 9 9
employed for the Treasure Island sites using DMT Fig. 5. Comparison of factor of safety against
results both the loosest silty sand and sand layers liquefaction computed using CRR obtained from CPT,
plot considerably above the CRR-KD curve. DMT, SPT and VS test results.
Although this result is consistent with field
performance, it is not particularly helpful in defining 0.7
the location of the boundary curve. Clean Sand
In lieu of waiting for additional field performance 0.6
data, the CSR and KD data pairs within the depth Liquefaction
0.5
range from 3.0 m to 7.5 m have been plotted against
CRR, CSR
the CRR vs KD curve in Figs. 6 and 7 for clean sand 0.4 Monaco et
and silty sands, respectively. As noted previously, 0.3 al (2005)
all of the in-situ tests generally predicted
liquefaction within this zone. For the points 0.2
identified as sand based on the DMT ID, the data No
0.1 Liquefaction
points in Fig. 6 plot right up to the boundary curve
but do not cross over, suggesting that the curve is 0
appropriately positioned for this CSR value. For the 0 1 2 3 4 5 6 7 8 9 10
data points identified as silty sand based on the Horizontal Stress Index, KD
DMT ID, the data points in Fig. 7 typically plot to
the left of the curve but some points cross over the Fig. 6. Comparison of CSR vs KD values for clean sand
curve indicating that the curve may be somewhat predicted to liquefy between 3.0 and 7.5 m depth in
unconservative for silty sands at this CSR value. comparison with the CRR vs KD curve for clean sand
proposed by Monaco et al. (2005).
0.7 boundary curve by Monaco et al. (2005) is
Silty Sand reasonable for clean sand but may be somewhat
0.6 unconservative for sitly sands.
Liquefaction
0.5 5. The liquefaction factor of safety from the Andrus
CRR, CSR