Professional Documents
Culture Documents
Assessment of The Undrained Response of Sands Under Limited and Complete Liquefaction
Assessment of The Undrained Response of Sands Under Limited and Complete Liquefaction
Assessment of The Undrained Response of Sands Under Limited and Complete Liquefaction
Abstract: The technique presented deals with the assessment, based on drained test behavior and formulation, of the undrained postcy-
clic stress-strain behavior of sands under limited or complete 共full兲 liquefaction and its associated strength. At present, there is no
Downloaded from ascelibrary.org by Tokyo Univ Seisan Gijutsu on 06/13/15. Copyright ASCE. For personal use only; all rights reserved.
particular procedure that allows assessment of such undrained postcyclic behavior that could develop full 共pore-water pressure ratio, ru
= 1兲 or limited 共ru ⬍ 1兲 liquefaction. The prediction of the undrained postliquefaction 共full or limited liquefaction兲 response presented here
is based on basic properties of sand such as its relative density 共Drc兲 关or 共N1兲60 blowcount兴, the effective angle of internal friction 共兲, the
roundness of the sand grains 共兲, and the drained axial strain at 50% stress level 共50兲. The technique presented accounts for the excess
pore-water pressure induced by cyclic loading 共⌬uc兲 and the postcyclic excess pore-water pressure generated under undrained monotonic
loading 共⌬ud兲.
DOI: 10.1061/共ASCE兲GT.1943-5606.0000114
CE Database subject headings: Soil-liquefaction; Drainage; Sand, soil type; Stress strain relations; Assessment.
Postliquefaction Phenomenon in Saturated Sands the completely liquefied soil. It should be noted that the charac-
teristics of the postliquefaction response of sand do not rely on
The postliquefaction stress-strain behavior and strength of a com- the method of liquefying the sand, i.e., whether it is from static or
pletely liquefied soil 共pore-water pressure ratio, ru = 1兲, or a soil cyclic loading 共Vaid and Thomas 1995兲. It depends on the sand’s
under limited liquefaction 共ru ⬍ 1兲, are still subjects of consider- relative density and effective stress. The technique presented al-
able research interest. Seed 共1979兲 addressed this issue experi- lows assessment of the postliquefaction liquefaction stress-strain
mentally showing the strain generated resistance of saturated sand behavior of sand with limited or complete liquefaction based on
under undrained monotonic loading after first being completely drained test behavior. Such assessment requires only basic prop-
liquefied by cyclic loading and losing all of its static strength erties of the sand such as its relative density 关or 共N1兲60兴, the ef-
共Fig. 1兲. Upon undrained compressive monotonic loading follow- fective angle of internal friction 共兲, the roundness of the sand
ing complete or limited liquefaction with its associated drop in grains 共兲, the drained axial strain at 50% stress level 共50兲, and
effective confining pressure 共 ¯ 3兲, sand eventually responds in a the residual confining pressure 共¯ 3兲 if greater than zero.
共suppressed兲 dilative fashion 共Fig. 2兲. However, sand under lim-
ited liquefaction with a drop in confining pressure short of com-
plete liquefaction 共 ¯ 3 ⬎ 0兲 may experience initial 共restrained兲 Method of Analysis
contractive behavior that is then followed by dilative behavior
共Fig. 2兲. The undrained postliquefaction response of sand after Norris et al. 共1997兲 developed a technique that was formulated
complete liquefaction reflects a concave upward undrained devia- later by Ashour and Norris 共1999兲 to assess the undrained behav-
toric stress stiffening versus undrained axial strain response re- ior of sand under compressive monotonic loading from its drained
gardless of its initial conditions 共density or confining pressure兲 behavior. That technique employs a series of drained tests, with
before cyclic loading/liquefaction. The same is true for sand with volume change measurements, on samples isotropically consoli-
limited liquefaction after initial restrained contractive behavior dated to the same confining pressure, ¯3c, and void ratio, ec, to
whereby its effective stress path reaches the failure line and there- which the undrained test is to be subjected. However, the drained
after marches up the failure line due to restrained dilative behav- tests are rebounded to different lower values of effective confin-
ior. The difference is that it does not start from ¯3 = 0, as does ing pressure, ¯3, before being sheared. Such a technique allows
the assessment of undrained behavior of sand isotropically con-
1
Associate Professor, Dept. of Civil and Environmental Engineering, solidated to ¯3c that is subjected to compressive monotonic load-
Univ. of Alabama, Huntsville, AL 35806 共corresponding author兲. E-mail ing 共no cyclic loading兲. During an isotopically consolidated
ashour@eng.uah.edu. undrained 共ICU兲 test, the application of a deviatoric stress, d, in
2
Professor, Dept. of Civil Engineering, Univ. of Nevada, Reno, compressive monotonic loading causes an additional pore-water
NV 89557. pressure, ⌬ud, that results in a lower effective confining pressure,
3
Project Engineer, California Dept. of Transportation 共CALTRANS兲. ¯3, i.e.
Note. This manuscript was submitted on October 29, 2007; approved
on March 6, 2009; published online on October 15, 2009. Discussion ¯3 = ¯3c − ⌬ud 共No cyclic loading兲 共1兲
period open until April 1, 2010; separate discussions must be submitted
for individual papers. This technical note is part of the Journal of Geo- and an associated isotropic expansive volumetric strain, v,iso, the
technical and Geoenvironmental Engineering, Vol. 135, No. 11, same as recorded in an isotropically rebounded drained triaxial
November 1, 2009. ©ASCE, ISSN 1090-0241/2009/11-1772–1776/ test 共prior to shear loading兲. However, in the undrained test, the
$25.00. volumetric change or volumetric strain must be zero. Therefore,
0
0 10 20 30 40 50 Liquefaction
共
¯ 3cc ⬎ 0 or ⌬uc ⬍ ¯3c and ru ⬍ 1兲. In order to establish undrained
-1
behavior from drained response, it is necessary to characterize the
Initial Static
Loading drained volume change due to shear of Fig. 3共c兲 that must be
-2 equal and opposite to the isotropic volume change due to the
0 10 20 30 40 50 change in effective confining pressure 共 ¯ 3兲 indicated in Fig. 3共a兲.
Axial Strain, ε1, % This implies formulating equations for the v − 1 curve共s兲 in
shear loading at a constant confining pressure, ¯3. On the v − 1
Fig. 1. Undrained behavior of completely 共fully兲 liquefied Sacra- curve of Fig. 3共c兲 associated with the point r and path r̄-s̄
mento River sand 共Seed 1979兲 of Figs. 3共a and b兲 共for ¯3 艌 ¯3cc兲, the initial slope 共SA兲, 共1兲B,
and 共v,shear兲max at point B, and 共1兲C and 共v兲C at point C 共Fig. 5兲
are assessed based on Eqs. 共5兲–共10兲 initially developed by Ashour
and Norris 共1999兲 based on ¯3c and then modified in this section
there must be a compressive volumetric strain component, v,shear,
to account for the postcyclic effective stress ¯3cc and the associ-
due to the deviatoric stress, d. This shear induced volumetric
ated 50. It should be noted that the following equations were
strain, v,shear, must be equal and opposite to v,iso, so that the total
established based on experimental data from different types of
volumetric strain, v = v,iso + v,shear, in undrained response is zero,
sands 共Ashour and Norris 1999兲 that were constituted by using
or
different sample preparations. However, the sand relative den-
sity and grain shape still play a significant role in the soil struc-
v,shear = − v,iso 共2兲 ture 共fabric兲 formation. 50 of sand 共from drained stress-strain
In the isotropically rebounded drained shear test, v,iso and then curve兲 in Eq. 共6兲 still reflects the sand fabric effect 共i.e., sample
v,shear 共to match v,iso兲 are obtained separately and sequentially in preparation兲
the undrained test, they occur simultaneously.
1
The current technique is extended to incorporate the excess 共SA兲¯3cc = 共5兲
pore-water pressure 共uxs兲 induced by cyclic loading 共⌬uc兲 and its exp共 + Drc兲
2
冋 册
influence on the undrained behavior of sands under the compres-
sive monotonic loading 共⌬ud兲 whether the sand is subjected to 250
共v,shear兲B,¯ 3cc = 共v,shear兲max,¯ 3cc = 2 共6兲
limited or complete liquefaction. It should be noted that cyclic exp共Drc兲 ¯3cc
loading 共e.g., free-field seismic excitation兲 could induce either
limited 共ru ⬍ 1兲 or complete liquefaction in the sand 共ru = 1兲 as
6共v,shear兲max,¯ 3cc
seen in Fig. 2. However, the sand can be described as nearly 共1兲B,¯ 3cc = 共7兲
liquefied soil when ru is less than but close to 1 exp共Drc兲
q = σd /2
Post-liquefaction stress-
e
strain behavior of partially lop
liquefied sand (∆uc < σ3c Env e
Deviator Stress, σd
e
lu r
and. ru <1) Fai
Full Liquefaction
Post-liquefaction
Limited liquefaction
stress-strain behavior
of completely liquefied
σd = 2 Sr sand (∆uc = σ3c and ru =1) σ3 = 0 Sr p = σ3 + σd/2
xo σ3 > 0
Axial Strain, ε Contractive response
Fig. 2. Subsequent undrained stress-strain behavior of sand that has experienced limited or complete liquefaction
Iso
OCR
.c
on
q = σd / 2
so
Is
s o. R where OCR= ¯3cc / ¯3 for ¯3 艋 ¯3cc and OCR= ¯3 / ¯3cc for ¯3
li
εv
da
ne
tio
e
艌 ¯3cc
Li
bo
nc
un
re
urv
d
εv,iso
lu
e
冋 册
i
r
Fa
(εv)c
Cyclic Loading
r εv,iso
εv, iso = - εv, shear s
s Effective Stress Path
σ3c 共v,shear兲max,¯ 3 0.5
σ3 r
p 共1兲B,¯ 3 = 共1兲B,3cc 共14兲
∆ud ∆uc σ3c 共v,shear兲max,¯ 3cc
σ3 σ3 σ3cc = σ3c ∆uc
σ3cc
(c) 共v,shear兲C,3c
共v,shear兲C,¯ 3 = 共v,shear兲max,¯ 3 共15兲
Downloaded from ascelibrary.org by Tokyo Univ Seisan Gijutsu on 06/13/15. Copyright ASCE. For personal use only; all rights reserved.
冋 册
Steady State
σ3
Increasing
σ3 < σ 3cc 共1兲B,¯ 3 0.25
共1兲C,¯ 3 = 共1兲C,¯ 3cc 共16兲
r (εv,shear)net dilative 共1兲B,¯ 3cc
(εv,shear)contractive Axial Strain,ε1
共18兲
共1兲C,¯ 3cc = 6共v,shear兲max,¯ 3ccexp共tan2 兲 共9兲 共v兲c and are calculated based on ¯3c as presented by Ashour
and Norris共1999兲.
The above procedure can be applied as long the excess pore-
共S f 兲¯3cc = − 0.5Drc tan2 共10兲 water pressure ratio 共ru兲 induced by cyclic loading is less than 1
and the residual confining pressure 共 ¯ 3兲 is greater than zero at
Such formulation for shear loading is the same as that for shear point r 共soil with limited liquefaction兲. Under monotonic loading,
loading to match isotropic compression beyond 共v兲c on the back- sand with limited liquefaction may experience a contractive re-
bone isotropic consolidation curve of Fig. 3共a兲. In effect, cyclic sponse associated with a reduction in ¯3 关to point s in Figs. 3共a
loading decreases ¯3 but without changing the sand’s subsequent and b兲兴 to reach the lowest value of ¯3, and then rebound 共dilate兲
response, as if point r were at 共v兲c and ¯3c兲. Note that Drc 共the with increasing ¯3 until ¯3 = ¯3cc again 关point r̄ in Figs. 3共a and
relative density of consolidation兲 is a decimal value correspond- b兲兴. Sand continues to dilate beyond ¯3cc 关Figs. 3共a and b兲兴 with
ing to the point 共v兲c at ¯3c. increasing ¯3 and net negative monotonically induced pore-water
The slopes and coordinates at points A, B, and C on the v pressure 共⌬ud兲. When ¯3 ⬍ ¯3cc, v,iso rebounds to point s and then
− 1 drained shear curve 关Fig. 5兴, which are empirically calculated recompresses. This is associated with an equal net compressive
by Ashour and Norris 共1999兲 for ¯3c, are modified to deal with v,shear. However, when ¯3 ⬎ ¯3cc, v,iso moves from r̄ to s̄ and an
¯3cc for ¯3 ⬍ ¯3cc 共OCR= ¯3cc / ¯3兲 as given in Eqs. 共11兲–共17兲. The equal dilative v,shear develops simultaneously
following equations are associated with the path r-s-r̄ as seen in
Fig. 3共a兲:
冋 冉 冊 册
d = SL ¯3 tan2 45 +
−1 共19兲
冋 册
2
共SA兲¯3cc 共v,shear兲max,¯ 3 0.25
共SA兲¯3 = 共11兲 The varying stress level 共SL兲 is a function of the drained 1, 50,
OCR 0.5
共v,shear兲max,¯ 3cc
and current ¯3 are calculated as presented in Eq. 共19兲. It should be
noted that the presented work focuses on the undrained behavior
of liquefiable sands pre- and postpeak where contractive and di-
共v,shear兲max,¯ 3cc lative behavior continues without reaching the steady state defor-
共v,shear兲max,¯ 3 = 共12兲
OCR mation 共dv / d1 = 0兲 at very large normal strain 共Figs. 3–5兲.
q = σd / 2
on s
so 600 Ion Sand
lid ine 200 Nevada Sand
a ti
on r eL 400 Dr = 15%
εv
r, r, s (εv)c σ3c
r, r, s p 0 0
εv, iso = - εv, shear
0 10 20 30 0 10 20 30
σ3cc σ ∆uc
Undrained Axial Strain, ε1, % Undrained Axial Strain, ε1, %
3c
Contraction Zone s
Fig. 6. Postliquefaction undrained stress-strain behavior of com-
σ3
0.0
∆uc = σ3c σ3c pletely liquefied Ione and Nevada sands
σ3
Downloaded from ascelibrary.org by Tokyo Univ Seisan Gijutsu on 06/13/15. Copyright ASCE. For personal use only; all rights reserved.
σ3cc = 0
becomes negative, ¯3 will be larger than ¯3c and the undrained
(c) dε v / dε 1 = 0 resistance will be greater than the drained strength.
Volumetric Strain, v,shear
Steady State Based on its Drc, the completely liquefied sand may ex-
σ3 perience a virtually zero-strength transition zone with soil strain
共1 艋 xo兲 and ru = 1 before it starts to show some resistance, con-
ε
Increasing
(εv,shear)net dilative fining pressure 共 ¯ 3兲 and dilative response 共Fig. 1兲. This value of
x0 decreases with the increase of sand relative density 共Dr兲 and
Axial Strain,ε1
becomes approximately zero for dense sand 共Ashour 2009兲.
As a result of the development of complete liquefaction by
Dilation
Contraction
cyclic loading and the subsequent dilative response under an
ICU loading, two equal and opposite components of volume
For the same σ3 (εv,shear)dilative = (εv,iso)contractive changes 共strains兲 develop in sand. In the undrained test, the total
volumetric change or volumetric strain must be zero. As stated
Fig. 4. Interrelationships among 共a兲 isotropic consolidation followed before, in the isotropically rebounded drained shear test, v,iso and
by cyclic loading; 共b兲 undrained effective stress path; and 共c兲 drained then v,shear 共to match v,iso兲 are obtained separately and sequen-
stress-strain behavior under different values of ¯3 for completely tially; in the undrained test, they occur simultaneously 共Ashour
共i.e., fully兲 liquefied sand 2009兲.
Fig. 4 shows the drained dilative response of sand when
v,shear is expansive and v,iso is compressive starting with ¯3 = 0.
Postliquefaction Behavior of Completely Liquefied As a result of the complete liquefaction under cyclic loading,
Sands ¯3 = ¯3cc ⬵ 0 关point r in Figs. 4共a and b兲兴 and the associated v,iso
at the start of undrained monotonic loading 关points r, r̄, and s in
¯ 3cc = 0 or ⌬uc = ¯3c and ru = 1兲. Once the soil is completely liq-
共
uefied 共i.e., ru = 1, ¯3 and d are equal to zero兲 due to cyclic
loading, the above procedure must be modified in order to handle 400 400
Predicted
Predicted
a different type of behavior. As seen in Fig. 1, the completely
Deviator Stress, σd, kPa
Measured
Deviator Stress, σd, kPa
Measured
300 (Vaid and Thomas 1995) 300 (Vaid and Thomas 1995)
liquefied soil loses its strength when the excess pore-water pres-
sure due to cyclic loading is equal to the effective confining pres- 200
Dr = 19%
200 Dr = 40%
σ3c = 100 kPa
sure 共uxs = ⌬uc = ¯3c兲 and the pore-water pressure ratio 共ru兲 = 1. By σ3cc = 0 kPa
σ3c = 100 kPa
σ3cc = 0 kPa
applying monotonic loading thereafter, uxs decreases and causes a 100 ε50 = 0.005 100 ε50 = 0.0035
φ = 32ο φ = 36ο
corresponding growth in confining pressure 共effective stress兲.
This is accompanied by a growth in sand resistance 共d兲. As 0
0 5 10 15 20
0
0 5 10
seen in Fig. 1, uxs continues to decrease to zero and then to nega- Undrained Axial Strain, ε1, % Undrained Axial Strain, ε1, %
dε v / dε 1 = 0 600 600
Volumetric Strain, v,shear
Predicted
Measured Dr = 40%
Deviator Stress, σd, kPa
ε50 = 0.0035
after point C φ = 36ο
Sf Dr = 40%
A σ3c = 400 kPa
200 σ3cc = 105 kPa 200
C Axial Strain,ε 1 ε50 = 0.0035 Predicted
B ε 1,C , ε v, C φ = 36ο Measured
(Vaid and Thomas 1995)
ε 1,Β , ε v,max 0 0
Dilation 0 2 4 6 8 0 2 4 6 8
SA
Undrained Axial Strain, ε1, % Undrained Axial Strain, ε1, %
Contraction
Fig. 8. Postliquefaction undrained stress-strain behavior of Fraser
Fig. 5. Volumetric strain curve and its major points sand under limited liquefaction
Figs. 4共a and b兲兴 are one and the same. The change in the volu- Summary
metric strain v,iso due the increase in ¯3 is represented by the
variation in v,iso 关Fig. 4共a兲兴 associated with 共v,shear兲net dilative in The technique presented has been shown to successfully predict
Fig. 4共c兲. Eq. 共18兲 for v,iso is modified as follows: the postliquefaction stress-strain behavior of sands subjected to
limited or complete liquefaction by cyclic loading. Such capabil-
共v兲c ity is vital to such applications as evaluating the lateral response
v,iso = − 共v,iso兲a 共20兲
OCR of laterally loaded piles/shafts in liquefiable sands with/without
lateral soil spreading. The postcyclic excess water pressure and
and
the associated residual effective confining pressure govern the
共v兲c ¯3c postliquefaction undrained behavior 共stress-strain relationship兲 of
共v,iso兲a = where OCR = sand. Basic properties and conditions of the sand, which are usu-
共OCR 兲at Point r̄ ¯3 ally available to the designer, represent the foundation for the
It should be noted that ¯3 at point 共r̄兲 is approximately equal to prediction of the sand’s postliquefaction behavior.
zero. As observed experimentally and based on its relative den-
sity, the liquefied sand may experience a zero-resistance zone References
共
¯ 3 = 0 and d = 0 as seen in Fig. 1兲 with a progressive axial strain
共up to 1 = x0兲 under the compressive monotonic loading. x0 is Ashour, M. 共2009兲. “P-Y curve and lateral response of piles in fully
determined from the drained rebounded relationship 1 − v,shear at liquefied soils.” J. Geotech. Geoenviron. Eng., submitted.
very small values of ¯3 ⬵ 0 关Fig. 4共c兲兴. x0 defines the end of Ashour, M., and Norris, G. 共1999兲. “Liquefaction and undrained response
evaluation of sands from drained formulation.” J. Geotech. Eng.,
complete liquefaction zone 共⌬uc = ¯3c兲 and indicates the subse-
125共8兲, 649–658.
quent growth in ¯3 and d, the degradation in the excess pore- Fukushima, S., and Tatsuoka, F. 共1984兲. “Strength and deformation char-
water pressure 关Fig. 4共a兲兴, and the development of dilative acteristics of saturated sand under extremely low pressures.” Soil
response 关Fig. 4共c兲. It should be noted that v,shear for the dilative Found., 24共4兲, 30–48.
sand represents the suppressed volume increase beyond the origi- Nguyen, T. Q. 共2002兲. “Sand behavior from anisotropic and isotropic
nal volume of sand. static and dynamic triaxial tests.” Ph.D. thesis, Civil Engineering
As seen in Fig. 4共b兲, the resistance of completely liquefied Dept., Univ. of Nevada, Reno, Nev.
sand under compressive monotonic loading lies on the failure Norris, G., Siddharthan, R., Zafir, Z., and Madhu, R. 共1997兲. “Liquefac-
line with SL equal to 1 关Eq. 共19兲兴. The computer program 共S- tion and residual strength of sands from drained triaxial tests.”
Liquefaction兲 which incorporates the technique presented allows J. Geotech. Eng., 123共3兲, 220–228.
Norris, G., Zafir, Z., and Siddharthan, R. 共1998兲. “An effective stress
the assessment of the whole postliquefaction undrained stress-
understanding of liquefaction behavior.” Environ. Eng. Geosci., 4共1兲,
strain behavior of sands with limited and complete liquefaction. 93–101.
Figs. 6–8 show comparisons between the measured and predicted Seed, H. B. 共1979兲. “Considerations in the earthquake-resistance design
resistance of different sands 共Table 1兲 under full or limited lique- of earth and rockfill dams.” Geotechnique, 29共3兲, 215–263.
faction induced by cyclic loading 共ru兲 and then subjected to com- Vaid, Y. P., and Thomas, J. 共1995兲. “Liquefaction and post-liquefaction of
pressive monotonic loading. behavior of sand.” J. Geotech. Eng., 121共2兲, 163–173.