Professional Documents
Culture Documents
net/publication/257896189
Article in Journal of Zhejiang University - Science A: Applied Physics & Engineering · November 2011
DOI: 10.1631/jzus.A1100015
CITATIONS READS
0 60
3 authors, including:
SEE PROFILE
All content following this page was uploaded by Gordon Tung-Chin Kung on 09 April 2014.
Abstract: The flat dilatometer test (DMT) has the potential to be a useful tool in the evaluation of liquefaction potential of soils.
In practice, it is necessary to carefully examine existing DMT-based methods for evaluating liquefaction potential. We con-
ducted the DMT and cone penetration test (CPT) in high liquefaction potential areas to examine the existing DMT-based
methods for liquefaction potential evaluation. Specifically, the DMT and CPT were conducted side-by-side at each of six in-situ
sites, and thus it is feasible to utilize those test results to validate the existing DMT-based methods. The DMT parameter,
horizontal stress index (KD), is used as an indicator for estimating liquefaction resistance of soils in terms of cyclic resistance
ratio (CRR). The analysis results revealed that the existing KD-based liquefaction evaluation methods would overestimate the
CRR of soils, which leads to overestimation of the factor of safety against liquefaction. Also, the estimations of DMT-KD values
by using the CPT-qc as well as the correlation between DMT-KD and CPT-qc proposed by the previous studies would be sig-
nificantly smaller than field measurements. The results reflected that further validation of the existing DMT-based methods for
liquefaction evaluation is desirable.
Key words: Liquefaction, Flat dilatometer test (DMT), Horizontal stress index, Earthquake
doi:10.1631/jzus.A1100015 Document code: A CLC number: TU43
through a direct correlation between the parameters of (2005), Grasso and Maugeri (2006), and Tsai et al.
DMT and CPT (Tsai et al., 2009). In addition, (2009) is briefly introduced herein.
Robertson (2009) has correlated main DMT parame- Monaco et al. (2005) proposed a CRR curve
ters with CPT parameters and evaluated the correlation based on a study of the correlations between cone tip
using published records and existing links to various resistance (qc) from CPT and blow count (N) from
other parameters, as well as comparison profiles. SPT and relative density (Dr), and between DMT
Fig. 1 shows that the DMT-based curve of cyclic horizontal stress index (KD) and Dr. Their DMT-based
resistance ratio (CRR) presented by Tsai et al. (2009) model is expressed as follows:
significantly differs from the ones proposed by
Monaco et al. (2005) and Grasso and Maugeri (2006). CRR 7.5 0.0107 K D3 0.0741K D2 0.2169 K D 0.1306,
This difference could confuse geotechnical engineers (1)
when attempting to select a DMT-based CRR curve to
practically evaluate the liquefaction potential of soils. where CRR7.5 is the cyclic resistance ratio of soil at a
It would be desirable to further examine the applica- magnitude of earthquake equal to 7.5.
bility of the existing DMT-based CRR curves for Specifically, the relationship between qc and Dr
liquefaction evaluation. To this end, this study col- adopted by Monaco et al. (2005) to formulate the
lected five sets of side-by-side DMT and CPT data CRR curve may be the one proposed by Jamiolkowski
presented by Tsai et al. (2009) and conducted another et al. (1985a) or Jamiolkowski et al. (1985b). Monaco
side-by-side test set to collectively examine the ex- et al. (2005) did not clearly indicate which equation
isting DMT-based CRR curves. was adopted. Subsequently, Grasso and Maugeri
(2006) followed the methodology adopted by Monaco
0.6 et al. (2005), in which the relationships between qc
Liquefied Monaco et al. (2005)
Non-liqufied
Marginal liquefied
Grasso and Maugeri (2006)-Eq. (10)
Tsai et al. (2009)
and Dr and between KD and Dr were used to develop
0.5 the CRR curve. Note that there were three CRR
0.4 models proposed by Grasso and Maugeri (2006), of
which, only the one developed based on the rela-
CSR7.5
The DMT-based methods for evaluating lique- CRR 7.5 0.0308e0.6054 KD . (3)
faction resistance of soils in terms of CRR have been
presented by Marchetti (1982), Robertson and Cam- Note that Eq. (3) corresponds to Eq. (10) in
panella (1986), Reyna and Chameau (1991), Monaco Grasso and Maugeri (2006). Tsai et al. (2009) em-
et al. (2005), Grasso and Maugeri (2006), Monaco ployed the results of in-situ tests to establish the cor-
and Marchetti (2007), and Tsai et al. (2009). The relation between CPT-qc and DMT-KD rather than the
recent development of CRR curves by Monaco et al. conventional qc-Dr-KD and N-Dr-KD relationships
Kung et al. / J Zhejiang Univ-Sci A (Appl Phys & Eng) 2011 12(11):807-817 809
used by Monaco et al. (2005) and Grasso and Maugeri cally, qc and fs are directly measured by the CPT.
(2006) to develop the DMT-based method for evalu- Then, qc1N,cs and Ic can be calculated based on
ating the CRR. A total of six sites were selected to Robertson and Wride (1998). In Figs. 3–8, SM
conduct the in-situ side-by-side CPT and DMT tests. represents silty sand, CL denotes low-plasticity silty
The regression analysis was then performed to di- clay, and ML represents low-plasticity sandy silt. For
rectly establish the relationship between the corrected each of six sites, DMT and CPT tests were conducted
cone resistance (qc1N,cs) and horizontal stress index at the same depth, which means KD and qc1N,cs of a soil
(KD), which can be expressed as at the same depth are available. In this regard, using
those data to examine the existing methods becomes
qc1N,cs 0.4 K D3 7.7 K D2 56 K D 20. (4) feasible. Therefore, the test results were collectively
employed to examine the difference between the ex-
isting DMT-KD-based liquefaction evaluation meth-
Note that qc1N,cs is so-called the clean-sand
ods of Monaco et al. (2005), Grasso and Maugeri
equivalence of the corrected cone tip resistance ac-
(2006), and Tsai et al. (2009), as shown in Fig. 1.
cording to Robertson and Wride (1998). Once the
qc1N,cs-KD relationship is available, the KD-based CRR
curve can be easily established through the existing N
er
Note that the above KD-based CRR curve was Site 1 hu i Riv
Yens Hsinhua Fault
established based on the widely accepted SPT- and Main sand boiling area in
1946 Hsinhua earthquake
Site 6
CPT-based CRR curves (Robertson and Wride, 1998;
Youd et al., 2001; Idriss and Boulanger, 2006) as well
as the correlations between qc and KD. The CRR curve Taiwan
Site 4
proposed by Tsai et al. (2009) has been additionally
Tainan
validated by Kung et al. (2009). More detailed in- area Site 5
Erjen
formation can be found in their paper. Rive
r
5 0 5 (km)
3 Side-by-side CPT and DMT tests Fig. 2 Layout of six study sites in Tainan
sand
10
CL: 5
gray and pale
brown silty clay
ML:
Depth (m)
SM:
gray silty fine sand
with clay
15
SM:
gray silty sand
with thin clay 10
sand
10
20
methods can be achieved by examining the correla- KD is only equal to a half of the measurements at
tion between CPT-qc and DMT-KD using the results depth of around 8 m. Similar results can be obtained
of side-by-side DMT and CPT tests (Figs. 3–8). in a deeper sandy layer at depths of 15.5 to 16.5 m.
Figs. 9–14 compare KD values measured by Fig. 10 displays the comparison of KD on site 2.
DMT with those computed based on CPT-qc values The KD measured at shallow depths near ground sur-
and qc-KD correlations. Note that only the data of KD face rapidly increases with the decrease of depth in
measurements at depths of 0–20 m are compared this site. Both methods from Tsai et al. (2009) and
since the liquefaction potential of soil at larger depth Grasso and Maugeri (2006) cannot capture this be-
is considered relatively low. As shown in Fig. 1, the havior. For the sandy layer at depths of 10.2 to
method by Monaco et al. (2005) is not included in the 17.8 m, the method by Tsai et al. (2009) would
comparison because the qc-KD correlation is not overestimate KD at depths of 10 to12 m, but the es-
clearly given in their paper. Therefore, only the timations at depths of 12 to 17.8 m are satisfactory.
methods by Tsai et al. (2009) and Grasso and Maugeri The method by Grasso and Maugeri (2006) generally
(2006) are selected to further study the intended issue. underestimates KD in this sandy layer at depths of
As shown in Figs. 9–14, the soil behavior type index 10.2 to 17.8 m.
Ic from CPT is applied in the present study to filter the
test data. Only the data points with Ic2.6, which can KD
0 2 4 6 8 10 12 14 16 18 Stratigraphy
be identified as the sandy soil according to Youd et al. 0
(2001), are adopted in the comparison (Figs. 9–14). ML: pale brown
2 sandy silt with thin,
Fig. 9 displays the comparison of KD at depths fine sand
4
within 0–20 m on site 1. The black points represent
CL: gray and pale
the DMT-KD measurements at various depths on this 6 brown silty clay
sites 2 and 3, the trend that KD increases with the The comparison of KD on site 5 is shown in
decrease of depth at shallow depths cannot be simu- Fig. 13. The method by Grasso and Maugeri (2006)
lated by the approaches of Tsai et al. (2009) and always underestimates KD in this case, irrespective of
Grasso and Maugeri (2006). The measured KD at depth. As to the performance of the method by Tsai et
depths of 7 to 9 m and 14 to 18 m can be accurately al. (2009), KD is underestimated at shallow depths (2
estimated by Tsai et al. (2009). For the depths of 9 to to 10 m), but can be adequately estimated at greater
12 m, both methods obviously underestimate KD. The depths (10 to 20 m).
KD estimated by Tsai et al. (2009) is generally greater
KD
than that estimated by Grasso and Maugeri (2006). 0 2 4 6 8 10 12 Stratigraphy
0
Backfill
KD Stratigraphy 2
CL: pale brown silty
0 2 4 6 8 10 12 Backfill clay with fine sand
0 4
SM: pale brown
silty sand with clay SM: pale brown
2 6 silty sand
SM: gray silty
4 sand with clay
8 CL: gray silty clay with
Depth (m)
fine sand
6 CL: gray silty clay
10
SM: gray silty sand
Depth (m)
8 12
CL: gray silty clay
10
SM: gray silty sand
14
12
CL: gray silty clay 16 SM: gray silty sand
6
with thin sand method by Tsai et al. (2009) on estimating KD through
CPT-qc is satisfactory. Specifically, KD can be rea-
Depth (m)
8
SM: gray silty sand sonably estimated by Tsai et al. (2009) at various
with thin clay
10 depths. Nevertheless, KD of sandy layers is signifi-
12 cantly underestimated by Grasso and Maugeri (2006)
CL: gray silty clay with
14 thin sand at depths of 0–20 m although the variation of KD
profiles estimated by Tsai et al. (2009) and Grasso
16
SM: gray silty
sand with thin
and Maugeri (2006) with depth is similar.
18 clay
Overall, the results reveal that the method by
20 Tsai et al. (2009) can reasonably estimate the KD
DMT measurements measurements, while the KD estimated by Grasso and
Estimated using Tsai et al. (2009)
Estimated using Grasso and Maugeri (2006)-Eq. (10) Maugeri (2006) are significantly smaller than the
Fig. 12 Comparison of KD measured by DMT and esti- measured values. It is not surprising that the per-
mated from qc-KD correlations on site 4 formance of the method by Tsai et al. (2009) is more
814 Kung et al. / J Zhejiang Univ-Sci A (Appl Phys & Eng) 2011 12(11):807-817
satisfactory, because their method was developed 0.92, while the slope for the Grasso and Maugeri
directly through regression analysis using the side-by- (2006) is equal to 1.67, which is far away from the 1:1
side CPT and DMT data of site 1 to site 5. perfect line. This result could be used to interpret the
trend of CRR curves shown in Fig. 1. Based on the
KD
Stratigraphy
preliminary investigation of this study, adopting
0 2 4 6 8 10 12 14 16 18
0 KD-Dr-qc relationship to correlate DMT-KD with
Backfill
10 R2=0.83
Site 1 Site 2 Site 3
Site 4 Site 5 Site 6
12 SM: gray sandy silt or
sand with clay
14
SM: gray silty
15
16 sand with thin clay
KD measured by DMT 0.92
18 1
5
Fig. 14 Comparison of KD measured by DMT and esti-
mated from qc-KD correlations on site 6
(a)
0
Although the CPT and flat-plat DMT have been 20
used for over 30 years, relatively little has been pub- Site 1 Site 2 Site 3 R2=0.89
Site 4 Site 5 Site 6
lished regarding comprehensive correlations between
the two in-situ tests (Robertson, 2009). Only a few 1
15
DMT-based liquefaction evaluation models have 1.67
KD measured by DMT
KD measured by DMT
underestimated by Grasso and Maugeri (2006) at
depths of 0–4 m and 4–20 m simultaneously. In this
10
regard, the capability of the method by Tsai et al.
(2009) in estimating KD is more satisfactory. How-
ever, it can be concluded that both methods by Tsai et
al. (2009) and Grasso and Maugeri (2006) are inca- 5
1
20 15
Site 1 Site 2 Site 3 2 1.59
KD measured by DMT
R =0.73
Site 4 Site 5 Site 6
1.37
KD measured by DMT
15 1 10
10 5
(b)
5 0
0 5 10 15 20
KD estimated using CPT-qc or qc1N,cs
(a)
Fig. 17 Performance of the method by Grasso and Mau-
0
20 geri (2006)-Eq. (10) with various depths. (a) 0–4 m; (b)
Site 1 Site 2 Site 3 R2=0.89 4–20 m
Site 4 Site 5 Site 6
0.88
type. Fig. 18 shows the fines content measured by the
1
soil samples taken from each of the six sites and ma-
10 terial index measured by the DMT tests. In practice, a
soil can be classified as sandy soil with ID>1.8, silty
soil with 0.6<ID<1.6, and clayey soil with ID<0.6.
5 Based on the test results conducted in the present
study, the upper bound may be slightly adjusted to be
(b) 1.6 for more adequately identifying the sandy soil
0
(Fig. 18). However, additional test results are required
0 5 10 15 20 to further verify this founding.
KD estimated using CPT-qc or qc1N,cs
Finally, this study collects liquefaction case
Fig. 16 Performance of the method by Tsai et al. (2009) histories, in which the DMT data are available, to
with various depths. (a) 0–4 m; (b) 4–20 m examine the accuracy of CRR curves proposed by
816 Kung et al. / J Zhejiang Univ-Sci A (Appl Phys & Eng) 2011 12(11):807-817
Tsai et al. (2009) and Grasso and Maugeri (2006). on SPT, CPT, and shear wave velocity, use of DMT
According to the comparison in the previous section for liquefaction resistance evaluation has received
of this study, KD based on CPT data would be sig- greater attention in recent years. The DMT is capable
nificantly underestimated by Grasso and Maugeri of measuring horizontal stresses and has an excellent
(2006), which results in the overestimated CRR operational repeatability. Thus, any improvement to
curve. This can be effectively verified by a number of the existing DMT-based methods for liquefaction
data points from actual liquefaction case histories resistance evaluation should be of interest to geo-
(Fig. 19). technical engineers. This study collected and con-
ducted the side-by-side DMT and CPT data and used
3.0 these data to examine the existing DMT-based
CL
ML
Classified by Unified Soil
Classification System (USCS)
methods for evaluating liquefaction resistance of
SM
soils. Specifically, this study employed the CPT-qc
2.4 Sandy soil
data and the correlation between CPT-qc and
Material index, ID
Pasqualini, E., 1985b. Penetration Resistance and Liq- Proceedings of the 16th International Conference on Soil
uefaction of Sands. Proceedings of the 11th International Mechanics and Geotechnical Engineering, p.2693-2698.
Conference on Soil Mechanics and Foundation Engi- Reyna, F., Chameau, J.L., 1991. Dilatometer Based Liquefac-
neering, 4:1891-1896. tion Potential of Sites in the Imperial Valley. Proceedings
Kung, G.T.C., Lee, D.H., Tsai, P.H., 2009. Performance of of the 2nd International Conference on Recent Advances
DMT-based liquefaction evaluation methods on case in Geotechnical Earthquake Engineering and Soil Dy-
histories of Chi-Chi earthquake. Journal of Marine Sci- namics, p.385-392.
ence and Technology, 17(4):283-292. Robertson, P.K., 2009. CPT-DMT correlations. Journal of
Marchetti, S., 1982. Detection of Liquefiable Sand Layers by Geotechnical and Geoenvironmental Engineering,
Means of Quasi-Static Penetration Tests. Proceedings of 135(11):1762-1771. [doi:10.1061/(ASCE)GT.1943-5606.
the 2nd European Symposium on Penetration Testing, 0000119]
p.689-695. Robertson, P.K., Campanella, R.G., 1986. Estimating lique-
Marchetti, S., Monaco, P., Totani, G., Calabrese, M., 2001. faction potential of sands using the flat plate dilatometer.
The Flat Dilatometer Test (DMT) in Soil Investigations– Geotechnical Testing Journal, 9(1):38-40. [doi:10.1520/
GTJ10610J]
A Report by the ISSMGE Committee TC16. Proceedings
Robertson, P.K., Wride, C.E., 1998. Evaluating cyclic lique-
of the 2nd International Flat Dilatometer Conference,
faction potential using the cone penetration test. Cana-
p.7-48.
dian Geotechnical Journal, 35(3):442-459. [doi:10.1139/
Mitchell, J.K., Lodge, A.L., Coutinho, R.Q., Kayen, R.E., t98-017]
Seed, R.B., Nishio, S., Stokoe, K.H., 1994. Insitu Test Tsai, P.H., Lee, D.H., Kung, G.T.C., Juang, C.H., 2009. Sim-
Results from Four Loma Prieta Earthquake Liquefaction plified DMT-based methods for evaluating liquefaction
Sites: SPT, CPT, DMT and Shear Wave Velocity. Report resistance of soils. Engineering Geology, 103(1-2):13-22.
No. UCB/EERC-94/04, Earthquake Engineering Re- [doi:10.1016/j.enggeo.2008.07.008]
search Center, University of California, Berkeley. Youd, T.L., Idriss, I.M., Andrus, R.D., Arango, I., Castro, G.,
Monaco, P., Marchetti, S., 2007. Evaluating Liquefaction Christian, J.T., Dobry, R., Finn, W.D.L., Harder, L.F.,
Potential by Seismic Dilatometer (SDMT) Accounting Hynes, M.E., et al., 2001. Liquefaction resistance of soils:
for Aging/Stress History. Proceedings of the 4th Interna- summary report from the 1996 NCEER and 1998
tional Conference on Earthquake Geotechnical Engi- NCEER/NSF workshops on evaluation of liquefaction
neering, paper No. 1626. resistance of soils. Journal of Geotechnical and Geoen-
Monaco, P., Marchetti, S., Totani, G., Calabrese, M., 2005. vironmental Engineering, 127(10):817-833. [doi:10.
Sand Liquefiability Assessment by Flat Dilatometer Test. 1061/(ASCE)1090-0241(2001)127:10(817)]