Liquefied Strength Ratio From Liquefaction Flow PDF

Color profile: Generic CMYK printer profile
Composite Default screen
629
Liquefied strength ratio from liquefaction flow

failure case histories
Scott M. Olson and Timothy D. Stark
Abstract: The shear strength of liquefied soil, su(LIQ), mobilized during a liquefaction flow failure is normalized with re-
′ ) prior to failure to evaluate the liquefied strength ratio, su(LIQ)/σ vo
spect to the vertical effective stress (σ vo ′ . Liquefied
strength ratios mobilized during 33 cases of liquefaction flow failure are estimated using a procedure developed to
directly back-analyze the liquefied strength ratio. In ten cases, sufficient data regarding the flow slide are available to
incorporate the kinetics, i.e., momentum, of failure in the back-analysis. Using liquefied strength ratios back-calculated
from case histories, relationships between liquefied strength ratio and normalized standard penetration test blowcount
and cone penetration test tip resistance are proposed. These relationships indicate approximately linear correlations
between liquefied strength ratio and penetration resistance up to values of qc1 and (N1)60 of 6.5 MPa and 12 blows/ft
(i.e., blows/0.3 m), respectively.
Key words: liquefaction, flow failure, liquefied shear strength, stability analysis, kinetics, penetration resistance.
Résumé : La résistance au cisaillement du sol liquéfié, su(LIQ), mobilisée lors d’un écoulement par liquéfaction est nor-
′ ) antérieure à la rupture pour évaluer le rapport de résistance
malisée par rapport à la contrainte effective verticale (σ vo
′ . Les rapports de résistance liquéfiée mobilisés au cours de 33 cas de rupture par liquéfaction sont
liquéfiée, su(LIQ)/σ vo
évalués au moyen d’une procédure développée pour rétro-analyser directement le rapport de résistance liquéfiée. Dans
dix cas, il y avait suffisamment de données disponibles pour incorporer dans la rétro-analyse la cinétique de la rupture,
c’est-à-dire la quantité de mouvement. Utilisant les rapports de résistance liquéfiée rétro-calculés pour les histoires de
cas, on propose des relations entre les rapports de résistance liquéfiée, le nombre de coups de pénétration standard nor-
malisés, et la résistance en pointe de l’essai de pénétration au cône. Ces relations indiquent approximativement des cor-
rélations linéaires entre le rapport de résistance liquéfiée et la résistance à la pénétration jusqu’à des valeurs de qcl et
(Nl)60 de 6,5 MPa et de 12 coups/pied (coups/0.3 m) respectivement.
Mots clés : liquéfaction, rupture par écoulement, résistance au cisaillement, analyse de stabilité, cinétique, résistance à
la pénétration.
[Traduit par la Rédaction] Olson and Stark 647
Introduction cause it is generic and does not imply correspondence to any

laboratory test condition. Regardless of terminology, the
The liquefied shear strength, su(LIQ), is the shear strength concepts for the liquefied shear strength stem from
mobilized at large deformation after liquefaction is triggered Casagrande’s (1940) seminal work on the critical void ratio.
in saturated, contractive, sandy soils. This strength differs
from the shear strength available to a soil at the triggering of Several investigators have proposed procedures to esti-
liquefaction, which is referred to as the yield shear strength mate the shear strength of liquefied soils (Poulos et al. 1985;
(Terzaghi et al. 1996). The liquefied shear strength has been Seed 1987; Seed and Harder 1990; Stark and Mesri 1992;
referred to as the undrained steady-state shear strength, sus Ishihara 1993; Konrad and Watts 1995; Fear and Robertson
(Poulos et al. 1985), the undrained residual shear strength, sr 1995; among others). The most widely used of these proce-
(Seed 1987), and the undrained critical shear strength, dures are those developed by Poulos et al. (1985), Seed
su(critical) (Stark and Mesri 1992). Based on a National Sci- (1987), Seed and Harder (1990), and Stark and Mesri
ence Foundation (NSF) international workshop (Stark et al. (1992). These methods are discussed subsequently. Other
1998), the term liquefied shear strength is used herein be- methods to estimate the liquefied shear strength are critiqued
by Olson (2001).
Poulos et al. (1985) presented a laboratory-based proce-
Received 26 July 2001. Accepted 6 November 2001.
Published on the NRC Research Press Web site at dure to estimate the in situ liquefied shear strength from un-
http://cjg.nrc.ca on 9 May 2002. drained triaxial compression tests on reconstituted and high
quality piston samples. Slight errors in estimating the in situ
S.M. Olson.1 URS Corporation, 2218 Millpark Drive, void ratio from the piston samples can cause large differ-
St. Louis, MO 63043, U.S.A. ences in the laboratory steady-state strength because of the
T.D. Stark. University of Illinois at Urbana-Champaign,
small slope of the steady-state line (Kramer 1989). Poulos et
Department of Civil and Environmental Engineering, Urbana,
IL 61801, U.S.A. al. (1985) indicated that the steady-state shear strength is de-
pendent solely on the void ratio after consolidation (and
1
Corresponding author (e-mail: scott_olson@urscorp.com). prior to shear). However, more recent studies (e.g., Vaid and
Can. Geotech. J. 39: 629–647 (2002) DOI: 10.1139/T02-001 © 2002 NRC Canada
I:\cgj\Cgj39\Cgj-03\T02-001.vp
Monday, May 06, 2002 11:46:09 AM
630 Can. Geotech. J. Vol. 39, 2002
Thomas 1995) indicate that the laboratory steady-state shear pressure) should result in an increase in liquefied shear
strength also may be influenced by the mode of shear, effec- strength due to consolidation under an increased confining
tive confining pressure, and sample preparation method. stress. Accordingly, Stark and Mesri (1992) developed an
As an alternative, Seed (1987) presented a relationship approach to estimate the liquefied shear strength from lique-
(later updated by Seed and Harder 1990) to estimate the liq- faction case histories as a function of the prefailure vertical
uefied shear strength from the equivalent clean sand stan- effective stress. Utilizing the cases from Seed and Harder
dard penetration test (SPT) blowcount, (N1)60-cs, where (1990) and three additional cases, Stark and Mesri (1992)
presented a relationship between the liquefied strength ratio,
 ER  ′ and (N1)60-cs, where σ vo ′ was a “representative”
[1] (N1) 60-cs =  NC N + ∆(N1) 60 su(LIQ)/σ vo
 60  prefailure vertical effective stress in the zone of liquefaction.
Their re-analysis did not reduce the scatter of the case histo-
where N is the field SPT blowcount, ER is the energy ratio ries in comparison to the relationship proposed by Seed and
of the SPT hammer system used (in percent), and ∆(N1)60 is Harder (1990) as a result of many of the difficulties dis-
a fines content adjustment to generate an “equivalent clean cussed above for the Seed and Harder (1990) approach.
sand” blowcount. The use of a fines content adjustment is Stark and Mesri (1992) also suggested that many of the liq-
discussed in a later section of this paper. CN is the overbur- uefaction failures experienced drainage during flow, result-
den correction factor (slightly modified from Liao and Whit- ing in back-calculated shear strengths that did not represent
man 1986) undrained conditions. Stark and Mesri (1992) discerned this
0.5 because some of the liquefied strength ratios exceeded the
P  level ground yield strength ratios for the same SPT blow-
[2] CN =  a 
 σ v′  count. (See Stark and Mesri 1992 for further details and a
description of “yield strength ratio.”)
where Pa is atmospheric pressure and σ vo ′ is the vertical ef- In summary, there are a number of practical difficulties
fective stress (same units as Pa). and uncertainties in using existing methods to estimate the
The Seed (1987) approach is based on the back-analysis liquefied shear strength and the strength ratio. In addition,
of 17 case histories of liquefaction flow failures and lateral laboratory testing is an expensive and difficult means to esti-
spreads. The Seed and Harder (1990) relationship is the mate liquefied shear strength. This paper presents improved
state-of-practice to estimate the liquefied shear strength, de- relationships to estimate the liquefied strength ratio from
spite numerous uncertainties implicit in back-calculating the CPT or SPT penetration resistance. A back-analysis proce-
liquefied shear strength or in determining the “representa- dure is presented to directly evaluate the liquefied strength
tive” SPT blowcount. For example, 6 of the 17 cases involve ratio rather than using values of su(LIQ) and σ vo ′ estimated
liquefaction-induced lateral spreading, not liquefaction flow separately. This procedure incorporates the entire range of
failure. Recently, NSF workshop participants (Stark et al. prefailure vertical effective stress acting throughout the zone
1998) concluded that the shear strength back-calculated of liquefaction. Thirty-three case histories of liquefaction
from cases of lateral spreading may not correspond to the flow failure involving loose clean sands, silty sands, sandy
shear strength mobilized during a liquefaction flow failure silts, and tailings sands were back-analyzed to develop the
and should be considered separately. For some cases, Seed improved relationships. Lateral spreading case histories are
(1987) used the prefailure geometry to back-calculate an up- not considered in this study.
per bound liquefied shear strength, while for most cases the Sufficient information is available in ten cases to incorpo-
postfailure geometry was analyzed. Seed and Harder (1990) rate the kinetics, i.e., momentum, of failure in the stability
considered the kinetics of the failure movements to back- analysis. These analyses suggest that kinetics only influ-
calculate liquefied shear strengths for an unknown number ences the back-analysis of liquefied shear strength in
of the 17 case histories, but did not present their kinetics- embankments/slopes larger than 10 m in height. For clarity,
based methodology, and in many cases their results differ the term “kinetics” describes the forces, accelerations, and
from the results presented herein, which explicitly incorpo- displacements associated with flow of the failure mass. The
rate kinetics. In 7 of the 17 cases, SPT blowcounts are not term “dynamics” is not used because it typically is associ-
available and had to be estimated from an appraisal of rela- ated with seismic forces.
tive density, and numerous cases had only a limited number
of penetration tests from which to select a representative Discussion of concepts and terminology
value of (N1)60. Finally, Seed and Harder (1990) recom-
mended that “the lower-bound, or near lower-bound relation- Liquefied shear strength
ship between [liquefied shear strength] and (N1)60-cs be The liquefied shear strength is the shear strength mobi-
used… at this time owing to scatter and uncertainty, and the lized at large deformation by a saturated, contractive soil fol-
limited number of case studies back-analyzed to date.” Nu- lowing the triggering of strain-softening response. In the
merous studies have suggested that the use of the lower laboratory, where drainage conditions are controlled, the term
bound relationship to estimate liquefied shear strength re- “undrained” applies. However, in the field, as evidenced by
sults in conservative factors of safety, particularly for large observation and analysis of flow failures, drainage may oc-
structures with large static shear stresses (e.g., Pillai and cur (Stark and Mesri 1992; Fiegel and Kutter 1994). There-
Salgado 1994; Finn 1998; Koester 1998). fore, the shear strength mobilized in the field may not be
Stark and Mesri (1992) concluded that an increase in pre- undrained. The term “liquefied shear strength” is used to
failure vertical effective stress (i.e., vertical consolidation describe the shear strength actually mobilized during a
© 2002 NRC Canada
Monday, May 06, 2002 11:46:09 AM
Olson and Stark 631
liquefaction flow failure in the field, including any potential Fig. 1. Relationship between liquefied shear strength and initial
effects of drainage, pore-water pressure redistribution, soil major principal effective stress for remolded layered specimens
mixing, etc. of silty sand, Batch 7, Lower San Fernando Dam (Baziar and
Laboratory shear tests on loose to medium dense labora- Dobry 1995). Kc, ratio of major principal effective stress to
tory specimens or loose specimens under low effective con- minor principal effective stress at the end of consolidation.
solidation pressure indicate that dilation often occurs at large
strain. These soils may exhibit a “quasi-steady state” (Ishi-
hara 1993), or minimum strength prior to strain hardening.
However, as noted by Yoshimine et al. (1999), once lique-
faction is triggered and deformation begins in the field, the
“…behavior may become dynamic and turbulent due to iner-
tia effects [i.e., kinetics]…” and the dilation observed in the
laboratory may not occur in the field. Despite some of these
difficulties in interpreting the liquefied shear strength mobi-
lized in the field, one purpose of this study was to determine
if the liquefied shear strength can be interpreted in terms of
the critical void ratio concept proposed by Casagrande (1940).
Liquefied strength ratio

The liquefied strength ratio is defined as the liquefied
shear strength normalized by the prefailure vertical effective
′ . Recent laboratory testing (e.g., Ishihara
stress, su(LIQ)/σ vo
1993; Baziar and Dobry 1995; Vaid and Sivathalayan 1996)
has shown that the liquefied shear strength of many loose
(i.e., contractive) cohesionless soils is linearly proportional
to the major principal effective stress after consolidation.
Figure 1 presents an example of this behavior for remolded
layered specimens of silty sand, Batch 7, from Lower San parallel to the steady-state line, for the range of effective
Fernando Dam (LSFD) (Baziar and Dobry 1995). stress of engineering interest. While this concept is not ex-
Following critical state soil mechanics theory, for a given pected to hold for dense sands, the authors suggest that if a
method of deposition and mode of shearing of a co- material is loose enough to be susceptible to liquefaction
′
hesionless soil, there is a unique value of su(LIQ)/σ mean for a flow failure, then this concept applies. Figure 2 presents an
given value of state parameter, ψ, as follows: example of this concept for specimens of silty sand, Batch 7,
from LSFD. Olson (2001) presents examples of this behav-
su (LIQ ) 1
[3] = M10− ψ / λ ss ior for other sands.
σ mean
′ 2 This study applies the premise that liquefied shear
strength is proportional to effective consolidation stress to
where σ mean
′ is the mean effective stress, M is the slope of the interpretation of liquefaction flow failures. Specifically,
the failure surface in stress path space (q′, σ mean
′ ), q′ is the as the prefailure vertical effective stress increases, void ratio
effective deviator stress, and λ ss is the slope of the steady- decreases, and therefore, liquefied shear strength increases.
state line (SSL) when mean effective stress is plotted on a The prefailure vertical effective stress is used for normaliza-
logarithmic scale. Been and Jefferies (1985) defined the state tion because alternate prefailure effective confining stresses
parameter as such as the mean effective stress (σ mean ′ ), effective stress
[4] ψ = eo – ess normal to the failure surface (σ n′ ), and major or minor prin-
cipal effective stress (σ lc′ or σ 3c
′ ), are difficult to determine
where eo is the in situ void ratio at a given mean effective accurately. Furthermore, because Ko is not expected to vary
confining stress prior to shearing, and ess is the void ratio at greatly in loose sandy soils, σ vo ′ (rather than σ mean
′ ) can be
the steady-state line for the same effective confining stress. used to determine the liquefied shear strength without intro-
Because the value of Ko (the ratio of horizontal to vertical ducing significant uncertainty. Lastly, the failure surface
effective stress) does not vary significantly in loose sands, within the zone of liquefaction for the majority of the flow
′ , is also uniquely re-
the liquefied shear strength, su(LIQ)/σ vo failures studied approaches direct shear conditions; thus, the
lated to state parameter. prefailure σ vo
′ is nearly equal to σ n′ .
For many loose compressible sandy soils, the consolida-
tion line and steady-state line are approximately parallel Relation between liquefied strength ratio and
(Olson 2001; Cunning 1994). Therefore, for a given increase penetration resistance
in effective confining stress, the value of state parameter re- Several researchers (e.g., Jefferies et al. 1990; Stark and
mains nearly constant and the liquefied shear strength in- Mesri 1992; and Ishihara 1993) have related liquefied strength
creases in proportion to the effective confining stress. Even ratio to penetration resistance. A correlation between lique-
if the consolidation behavior and steady-state line are not ex- fied strength ratio and penetration resistance is reasonable
actly parallel, it may be reasonable to assume that in many because both liquefied strength ratio and penetration resis-
field conditions the consolidation behavior is approximately tance are functions of soil density and effective confining
© 2002 NRC Canada
Monday, May 06, 2002 11:46:10 AM
Fig. 2. A comparison of the consolidation behavior and the steady- su (LIQ)

[8] = 0.003Q
state line for remolded layered specimens of silty sand, Batch 7, σ vo
′
Lower San Fernando Dam (after Baziar and Dobry 1995). σ ′3us ,
minor principal effective stress at the steady-state condition. Stark and Mesri (1992) recommended the following relation-
ship between liquefied strength ratio and normalized pene-
tration resistance
su (LIQ)
[9] = 0.0055(N1) 60-cs
σ vo
′
to describe the constant volume conditions present at the

start of flow slides. Ishihara (1993) presented theoretical
′ – N1 relationships for individual sands but also
su(LIQ)/σ vo
found that many available case histories fell below these the-
oretical relationships.
Back-analysis of liquefaction flow failures

Three levels of stability analysis were used to back-
calculate the liquefied strength ratio for the 33 flow failures.
The appropriate level of analysis depended on the detail and
quality of information available for each case history. For
cases with minimal available information, a simplified anal-
stress. For example, Been et al. (1987b) related normalized ysis was conducted to estimate the liquefied strength ratio.
CPT tip resistance to state parameter, as follows: For the majority of cases (21 of 33), sufficient information
was available to allow a rigorous slope stability analysis and
q c − σ mean
[5] = k exp− mψ the direct back-calculation of the liquefied strength ratio. For
σ mean
′ cases with appropriate documentation and failure conditions,
an additional kinetics analysis was conducted to back-
where k and m are soil constants that can be related to the calculate the liquefied strength ratio. The following sections
slope of the steady-state line. Equation [5] indicates that describe the simplified, rigorous, and kinetics analyses.
higher values of normalized penetration resistance corre-
spond to lower values of state parameter. Sladen (1989) dis- Simplified stability analysis of postfailure geometry
cussed difficulties with the exact form of this relationship, Several investigators (e.g., Meyerhof 1971; Lucia 1981;
suggesting that the relationship is influenced by stress level. and Ishihara et al. 1990a) have developed simplified meth-
However, Shuttle and Jefferies (1998) showed that the rela- ods to estimate the shear strength mobilized during flow
tionship is influenced by shear modulus, not stress level. slides. The Ishihara et al. (1990a) method was adopted for
It is generally accepted that the shear modulus of sand in- the simplified analysis and is briefly reviewed.
creases with the square root of effective confining stress and Ishihara et al. (1990a) used the following assumptions to
Ko does not vary greatly in loose sands. Therefore, (qc – estimate the liquefied shear strength mobilized during sev-
σ mean )/σ mean
′ can be reasonably replaced by another form of eral flow failures: (i) the ground surface and the surface of
normalized CPT penetration resistance, qc1, to account for the flowed material are approximately parallel when the mass
the increase in shear modulus with depth. This form of nor- comes to rest; (ii) side forces are equal, opposite, and co-
malized CPT tip resistance is used throughout the remainder linear; and (iii) the shear strength mobilized at the moment
of this paper and is defined as the failed mass comes to rest is equal to the liquefied shear
strength. If the flowed material has an average thickness of
1.8
[6] q c1 = q cC q = q c ≈ q cC N H and a unit weight of γ t , then force equilibrium in the di-
0.8 + (σ v′ / Pa ) rection of flow (similar to an infinite slope analysis) indi-
cates that
where qc is CPT tip resistance and the form of Cq in eq. [6]
was defined by Kayen et al. (1992). [10] su (LIQ) = γ t H sin α cos α
On the basis of this discussion, Jefferies et al. (1990) sug-
gested the following theoretical relationship (assuming rea- where α is the angle of inclination of the sliding surface and
sonable values of k, m, Ko, and λ ss): surface of flowed material to the horizontal.
Equation [10] is used to estimate values of liquefied shear
su (LIQ) strength for cases with insufficient information to conduct
[7] = 4.4 × 10–7(Q) 3.83
σ vo
′ the rigorous stability analysis and where the above assump-
tions are satisfied. Numerous flow failure cases in the litera-
where Q was defined as qc/σ v′ . However, Jefferies et al. (1990) ture did not meet these minimum assumptions and therefore
found that liquefied strength ratios back-calculated from are not included in this study.
seven flow failures fell below the theoretical relationship and The simplified procedure only provides a value of lique-
followed the trend fied shear strength. Therefore, a representative vertical
© 2002 NRC Canada
Monday, May 06, 2002 11:46:11 AM
Olson and Stark 633
effective stress is estimated from the prefailure geometry to above the phreatic surface, or otherwise known not to have
calculate a liquefied strength ratio. Because the simplified liquefied (e.g., segments 10–14 in Fig. 3), are assigned ap-
analysis does not explicitly consider kinetics, values of propriate drained or undrained shear strengths (see Olson
′ estimated using this method may be smaller
su(LIQ)/σ vo 2001 for details of shear strength assignment for each of the
than the actual su(LIQ)/σ vo ′ of the liquefied soil (Olson case histories). The liquefied strength ratio is then varied
2001). (which in turn varies the liquefied shear strength mobilized
along each segment of the postfailure geometry) until a fac-
Rigorous stability analysis of postfailure geometry tor of safety of unity is achieved.
Olson (2001) developed a rigorous stability analysis pro- This analysis considers the entire range of prefailure verti-
cedure to back-calculate the liquefied strength ratio directly. cal effective stress within the zone of liquefaction rather than
This analysis was conducted for the majority (21 of 33) of a single “representative” value of prefailure σ vo′ . Therefore,
cases. The rigorous stability analysis uses Spencer’s (1967) liquefied strength ratios back-calculated using this technique
slope stability method as coded in the microcomputer pro- are considered more appropriate than those reported else-
gram UTEXAS3 (Wright 1992). Because this analysis also where, e.g., Stark and Mesri (1992) and Ishihara (1993). For
does not explicitly consider kinetics, values of su(LIQ)/σ vo ′ use in the kinetics analyses, the weighted average prefailure
estimated using this method may be smaller than the actual vertical effective stress was calculated as follows:
su(LIQ)/σ vo′ of the liquefied soil.
∑ i=1 σ v′ ,i Li
n
As explained subsequently, this analysis requires an esti-
mate of the range of σ vo′ for the liquefiable material prior to [11] σ vo
′ (ave) =
∑ i=1 Li
n
failure. To accurately assess the range of prefailure σ vo ′ for
the liquefiable material, either the geometric extent of the
liquefied soil must be known, or the initial failure surface where σ v,i
′ is the prefailure vertical effective stress for seg-
must be assumed to pass approximately through the center ment i and Li is the length of segment i.
of the zone of liquefaction. Olson (2001) examined a limited
number of well-documented flow failures and found that ini- Stability analysis considering kinetics of failure mass
tial failure surfaces do pass approximately through the center movements
of the zones of liquefaction. Therefore, this assumption appears To obtain the best estimate of liquefied shear strength mo-
valid and is used for the analysis of the other flow failures. bilized during failure, the back-analysis should consider the
In this procedure, the postfailure sliding surface is divided kinetics of failure. The reason for this is illustrated in Fig. 4
into a number of segments. Based on the lengths of the using the calculations made for the liquefaction flow failure
postfailure segments, corresponding lengths of liquefied soil of the North Dike of Wachusett Dam (Olson et al. 2000). At
are defined within the prefailure geometry, i.e., within the the onset of a liquefaction flow failure, only small strains are
zone of liquefaction or near the initial failure surface. Fig- required to reduce the shear strength from the yield (or peak)
ure 3 illustrates this procedure for the flow failure of LSFD. shear strength to the liquefied shear strength (Davis et al.
The postfailure sliding surface (Fig. 3b) is broken into 14 1988). These strains occur while the driving shear stress re-
segments of various lengths. Based on the postfailure geom- mains relatively unchanged. For simplification, the liquefied
etry segment lengths, corresponding lengths of liquefied soil soil is assumed to be in a post-peak condition and the mobi-
are defined in the prefailure zone of liquefaction (Fig. 3a). lized strength at the beginning of failure (at time t = 0) is
The distance between the prefailure segments (i.e., thickness equal to the liquefied shear strength (as indicated in Fig. 4a).
t in Fig. 3a) is approximately equal to the average final The initial driving shear stress in the zone of liquefaction is
thickness of the liquefied soil estimated from the postfailure determined from a static slope stability analysis assuming a
geometry. Segments 1–9 of the postfailure sliding surface factor of safety of unity (Castro et al. 1989; Seed et al.
consist of liquefied soil, while segments 10–14 consist of 1989). Because the initial driving shear stress is larger than
nonliquefied soil (namely, segments 10–13 correspond to the liquefied shear strength (this is a prerequisite for a lique-
core material, and segment 14 corresponds to ground shale faction flow failure), the mass begins to accelerate
hydraulic fill and rolled fill). Therefore, only segments 1–9 downslope (Fig. 4b). Therefore, the velocity of the failure
are located in the liquefied zone in the prefailure geometry mass increases from zero (Fig. 4c), and downslope displace-
in Fig. 3b. The prefailure σ vo
′ is determined for each segment ment occurs (Fig. 4d). The downslope displacement of the
in the liquefied soil (segments 1–9) and is assigned to the failure mass, in turn, decreases the driving shear stress of the
corresponding segment in its postfailure position. Olson failure mass because of the curvature of the failure path.
(2001) found that rearranging the positions of the segments When the driving shear stress is reduced to the liquefied
has little effect on the back-calculated liquefied strength ra- shear strength, the failure mass has an acceleration of zero
tio, as long as the segments are equally spaced in the zone of and has attained its maximum velocity (Figs. 4b and 4c). (In
liquefaction or centered around the initial failure surface. Fig. 4a, the mobilized shear resistance is lower than the liq-
Using the individual σ vo′ values for each segment and a uefied shear strength as a result of hydroplaning, as dis-
single value of su(LIQ)/σ vo ′ , individual values of liquefied cussed subsequently.) Because the failure mass has a finite
shear strength are assigned to each segment of the post- velocity, it continues to displace and deform, decreasing the
failure geometry for the stability analysis (i.e., segments 1–9 driving shear stress to a value less than the liquefied shear
in Fig. 3). This allows the variation in prefailure σ vo
′ within strength, thereby decelerating the failure mass (i.e., upslope
the zone of liquefaction to be reflected in variable liquefied acceleration; Figs. 4a and 4b). When the failure mass
shear strengths along the final sliding surface. Soils initially reaches a velocity of zero and comes to rest, the driving
© 2002 NRC Canada
Monday, May 06, 2002 11:46:12 AM
Fig. 3. (a) Simplified prefailure geometry of the Lower San Fernando Dam for determination of prefailure vertical effective stresses
used in liquefied strength ratio stability analysis. (b) Simplified postfailure geometry and assumed final positions of the liquefied soil
segments (slices 10–14 did not liquefy).
shear stress may be considerably less than the liquefied the movement of the sliding mass center of gravity (see
shear strength (Fig. 4a). At the instant the failure mass co- Fig. 4), su is the mobilized shear resistance, and L is the
mes to rest, the mobilized shear resistance decreases to that length of the failure surface. At the start of sliding, the
required for static stability, i.e., the driving shear stress ob- weight term is larger than the shear strength term, and accel-
tained from the postfailure geometry (Fig. 4a). eration is downslope. Near the end of sliding, the weight
The kinetics analysis used for this study was adapted from term is smaller than the shear strength term, and acceleration
the procedure outlined by Davis et al. (1988), and is re- is upslope (thereby decelerating the mass; Figs. 4a and 4b).
viewed briefly. This analysis is based on Newton’s second Olson (2001) examined a limited number of well-
law of motion, as follows: documented flow failure case histories and found that most
initial and final failure surfaces could be approximated using
[12] ΣF = ma third-order polynomials. Therefore, the movement of the
where F are the forces acting on the moving mass (in vector center of gravity of the sliding mass for the ten cases ana-
form), m is the mass of the failed material (weight divided lyzed using kinetics was assumed to follow a third-order
by acceleration due to gravity, g), and a is the acceleration polynomial. A third-order polynomial has the form
of the center of gravity of the failed material. Referring to
[14] y = ax3 + bx2 + cx + d
Fig. 4, the net force, ΣF, acting on the failure mass in the di-
rection of the movement of the center of gravity is given by where a, b, c, and d are constants that can be calculated based
the driving weight of the failure mass minus the mobilized on (i) the x and y coordinates of the initial and final positions of
shear resistance of the soil, as follows: the center of gravity of the failure mass; and (ii) the curvature of
[13] ΣF = [(W sin θ) − ( su L)] = ma the travel path of the center of gravity. This curvature was as-
sumed to parallel the curvature of the final sliding surface. Using
where W is the weight of the failure mass, θ is the angle be- the slope (dy/dx) of the tangent to the curve described in eq. [14]
tween the horizontal and the tangent to the curve describing at any point, the sine of the angle θ is given by
© 2002 NRC Canada
Monday, May 06, 2002 11:46:14 AM
Olson and Stark 635
Fig. 4. Freebody diagram used for kinetics analysis (top) and kinetics analysis for the North Dike of the Wachusett Dam (a–d).
© 2002 NRC Canada
Monday, May 06, 2002 11:46:18 AM
dx L 
su −  d sd 
dy  100 
[15] sin θ = [18] su (LIQ) =
 Ld 
2
 dx  1.0 − 
1+  
 dy   100 
where dy and dx are the vertical and horizontal displace- where su is determined by the solution of eq. [17], Ld is the
ments, respectively, of the center of gravity of the failure percentage of the total length of the postfailure sliding sur-
mass along the curve defined by eq. [14]. face that incorporates soils that did not liquefy, and sd is the
The acceleration of the failure mass center of gravity is average shear strength of the soils that did not liquefy.
estimated using the second derivative of the displacement, ∆, Values of sd are provided subsequently and Olson (2001) de-
with respect to time, t, as tails the assignment of shear strength to nonliquefied soils
for each case history.
The kinetics analysis only provides the “best estimate” of
d2 ∆
[16] a = su(LIQ). This value of su(LIQ) is divided by the weighted
dt 2 average prefailure σ vo
′ (eq. [11]) to obtain the “best estimate”
of liquefied strength ratio. These values of liquefied shear
Substituting into eq. [12] yields the following: strength and strength ratio are “best estimates” because they
incorporate the kinetics of failure, potential hydroplaning and
W d2 ∆ mixing effects, and the shear strength of nonliquefied soils.
[17] [(W sin θ) − ( su L)] =
g dt 2
Case histories of liquefaction flow failure

Equation [17] can be solved for displacement numerically or
by direct integration. Olson (2001) employed a time-step nu- Olson (2001) collected 33 liquefaction flow failure case
merical solution using a spreadsheet program. An initial histories for which CPT or SPT results were available or
value of liquefied shear strength is assumed, and eq. [17] is could be reasonably estimated. Olson (2001) also described
solved to estimate the total displacement and duration of the available information, the analyses conducted, the evalua-
movement. The assumed value of liquefied shear strength tion of penetration resistance, and the uncertainties involved
then is revised to obtain reasonable agreement with the ob- in each case history. The case histories and corresponding
served displacement of the center of gravity of the failure references are summarized in Table 1. The liquefied strength
mass. The approximate duration of sliding is only available ratios back-calculated using the simplified or rigorous stabil-
for two cases, therefore agreement is based solely on center ity analysis and weighted average prefailure σ vo ′ are pre-
of gravity movement. sented in Table 2. Table 2 also includes values of liquefied
The kinetics analysis also should account for potential shear strength back-calculated independently for each case.
hydroplaning (slide material “riding” on a layer of water as The back-calculation of these values will be discussed sub-
flow occurs), mixing with water, and the increase in void ra- sequently. For the ten cases with sufficient documentation to
tio of the liquefied material if the failure mass slid into a perform a kinetics analysis, the “best estimate” liquefied
body of water (Castro et al. 1992). To account for these po- shear strength and strength ratios considering the kinetics of
tential effects, the shear strength mobilized along the failure failure are presented in Table 3. As expected, the “best esti-
surface in the body of water (beyond the original limits of mate” liquefied shear strengths and strength ratio values esti-
the prefailure geometry) is assumed to be equal to 50% of mated from the kinetics analysis are greater than the values
the value of shear strength mobilized within the limits of the estimated from the rigorous stability analysis (Table 2). Ta-
prefailure slope geometry. For each case, reduction factors ble 3 also includes strengths of nonliquefied soils (sd) and
of 0% and 100% also are used to ascertain the sensitivity of the percentage of the postfailure sliding surface that incorpo-
the liquefied shear strength to the effect of hydroplaning and rates nonliquefied soils. Measured or estimated penetration
to obtain a range of possible values of liquefied shear resistances and selected soil properties available for the case
strength. The same reduction factors were used by Castro et histories are presented in Table 4. For cases where either
al. (1992) to back-calculate the possible range of liquefied CPT or SPT penetration alone was measured, the corre-
shear strength mobilized during the flow failure of LSFD. sponding value of the other penetration resistance was esti-
The kinetics analysis also considers the change in weight mated using the qc /N60 relationship presented by Stark and
of the failure mass if the mass slid into a body of water and Olson (1995) and the median grain size of the liquefied soil.
the change in the length of the failure surface during flow. Other details and uncertainties regarding the penetration
These changes during failure are incorporated into the solu- tests are discussed in the following section. Figure 5 pres-
tion of eq. [17] as a function of the distance traveled by the ents the “best estimate” liquefied strength ratios and mean
center of gravity of the failure mass with respect to its total qc1 values for each of the cases. Figure 6 presents “best esti-
distance of travel. Lastly, appropriate drained or undrained mate” liquefied strength ratios and mean (N1)60 values. The
shear strengths of nonliquefied soils are incorporated after numbers adjacent to each of the data points are the average
the solution of eq. [17] using the following adjustment to fines content of the liquefied soil. The role of fines content
su(LIQ): will be discussed subsequently.
© 2002 NRC Canada
Monday, May 06, 2002 11:46:20 AM
Table 1. Case histories of liquefaction flow failure and their corresponding references.
Olson and Stark
Case History Structure Apparent Cause of Sliding References
Monday, May 06, 2002 11:46:20 AM

1 Zeeland – Vlietepolder 1889 High tide Koppejan et al. (1948)
2 Wachusett Dam – North Dike 1907 Reservoir filling Olson et al. (2000)
3 Calaveras Dam 1918 Construction Hazen (1918, 1920)
4 Sheffield Dam 1925 Santa Barbara eq. (ML = 6.3) Engineering News-Record (1925); U.S. Army Corp of Engineers (1949); Seed et al. (1969)
5 Helsinki Harbor 1936 Construction Andresen and Bjerrum (1968)
6 Fort Peck Dam 1938 Construction U.S. Army Corps of Engineers (1939); Middlebrooks (1942); Casagrande (1965)
7 Solfatara Canal Dike 1940 Imperial Valley eq. (ML = 7.1) Ross (1968)
8 Lake Merced bank 1957 San Francisco eq. (ML = 5.3) Ross (1968)
9 Kawagishi-Cho building 1964 Niigata eq. (MW = 7.5) Yamada (1966); Ishihara et al. (1978); Ishihara and Koga (1981); Seed (1987)
10 Uetsu Railway embankment 1964 Niigata eq. (MW = 7.5) Yamada (1966)
11 El Cobre Tailings Dam 1965 Chilean eq. (ML = 7–7.25) Dobry and Alvarez (1967)
12 Koda Numa highway embankment 1968 Tokachi-Oki eq. (M = 7.9) Mishima and Kimura (1970)
13 Metoki Road embankment 1968 Tokachi-Oki eq. (M = 7.9) Ishihara et al. (1990a)
14 Hokkaido Tailings Dam 1968 Tokachi-Oki eq. (M = 7.9) Ishihara et al. (1990a); Ishihara (1993)
15 Lower San Fernando Dam 1971 San Fernando eq. (MW = 6.6) Seed et al. (1973); Castro et al. (1989); Seed et al. (1989); Vasquez-Herrera and
Dobry (1989); Castro et al. (1992)
16 Tar Island Dyke 1974 Construction Mittal and Hardy (1977); Plewes et al. (1989); Konrad and Watts (1995)
Mochi-Koshi Tailings Dam 1978 Izu-Oshima-Kinkai eq. (ML = 7.0) Marcuson et al. (1979); Okusa and Anma (1980); Ishihara et al. (1990a)
17 Dike 1
18 Dike 2
Nerlerk Berm 1983 Construction Sladen et al. (1985a, 1985b, 1987); Been et al. (1987a); Sladen (1989); Rogers et al. (1990);
19 Slide 1 Konrad (1991); Hicks and Boughraruo (1998)
20 Slide 2
21 Slide 3
22 Hachiro-Gata Road embankment 1983 Nihon-Kai-Chubu eq. (M = 7.7) Ohya et al. (1985)
23 Asele Road embankment 1983 Pavement repairs Ekstrom and Olofsson (1985); Konrad and Watts (1995)
La Marquesa Dam 1985 Chilean eq. (Ms = 7.8) de Alba et al. (1987)
24 U/S slope
25 D/S slope
26 La Palma Dam 1985 Chilean eq. (Ms = 7.8) de Alba et al. (1987)
27 Fraser River Delta 1985 Gas desaturation and low tide Chillarige et al. (1997a, 1997b); Christian et al. (1997)
28 Lake Ackerman highway embankment 1987 Seismic reflection survey Hryciw et al. (1990)
29 Chonan Middle School 1987 Chiba-Toho-Oki eq. (M = 6.7) Ishihara et al. (1990a); Ishihara (1993)
30 Nalband Railway embankment 1988 Armenian eq. (Ms = 6.8) Yegian et al. (1994)
31 Soviet Tajik – May 1 slide 1989 Tajik, Soviet Union eq. (ML = 5.5) Ishihara et al. (1990b)
32 Shibecha-Cho embankment 1993 Kushiro-Oki eq. (ML = 7.8) Miura et al. (1998)
33 Route 272 at Higashiarekinai 1993 Kushiro-Oki eq. (ML = 7.8) Sasaki et al. (1994)
Note: ML, local or Richter magnitude; MW, moment magnitude; Ms, surface-wave magnitude; M, magnitude scale not available; eq., earthquake; U/S, D/S, upstream and downstream, respectively.
© 2002 NRC Canada

637
Table 2. Back-calculated liquefied strength ratios and liquefied shear strengths from liquefaction flow failure case histories.
Postfailure geometry strength ratio Postfailure geometry shear strength
Weighted average
Case Calculation Best Best estimate Lowerbound Upperbound prefailure vertical
history methoda estimate Lowerbound Upperbound (kPa) (kPa) (kPa) effective stress (kPa)
1 1 0.048 0.032 0.096 5.5 4.5 6.5 114.7
2 2 and 3 0.026 — — 3.8 — — 151.2
3 2 and 3 0.012 0.007 0.033 3.6 2.2 10.5 307.5
4 1 0.053 0.035 0.07 3.6 2.4 4.8 68.4
5 1 0.06 0.037 0.098 1.55 1.1 2.0 25.0
6 2 and 3 0.011 0.002 0.041 3.8 0.7 15.1 351.5
7 1 0.08 0.04 0.12 2.4 1.2 3.6 29.9
8 2 0.108 0.073 0.12 6.9 4.8 7.4 65.7
9 2 0.075 0.057 0.093 5.3 4.5 5.7 70.6
10 2 and 3 0.009 0.005 0.031 0.6 0.3 1.9 61.3
11 1 0.02 0.017 0.024 1.9 1.8 2.0 93.2
12 2 and 3 0.04 0.036 0.082 1.0 0.8 1.9 23.2
13 1 0.043 0.034 0.051 1.8 1.4 2.2 41.9
14 1 0.073 0.062 0.116 4.8 4.1 6.6 65.9
15 2 and 3 0.029 0.026 0.076 4.8 4.3 12.2 166.7
16 1 0.058 0.037 0.105 12.0 7.7 21.6 205.9
17 1 0.06 0.033 0.12 3.6 2.4 7.2 59.9
18 1 0.104 0.071 0.136 5.4 4.8 6.0 52.2
19 2 0.086 — — 2.5 — — 29.5
20 2 0.06 0.025 0.094 1.7 1.0 2.4 31.1
21 2 0.034 0.029 0.041 1.5 1.2 1.7 44.3
22 2 and 3 0.042 0.033 0.05 1.4 1.1 1.6 32.1
23 1 0.104 0.083 0.125 6.3 5.0 7.5 59.9
24 2 0.07 0.04 0.10 3.1 1.9 4.3 43.6
25 2 0.11 0.04 0.18 5.3 2.2 9.8 47.9
26 2 0.12 0.08 0.20 4.8 2.4 7.9 37.8
27 4 0.10 — — — — — —
28 2 and 3 0.066 0.054 0.097 3.4 2.9 4.8 51.5
29 2 0.091 0.066 0.116 4.8 3.8 6.0 53.6
30 2 0.109 0.10 0.12 5.7 5.3 6.2 52.7
31 2 0.082 0.026 0.16 8.4 2.9 15.6 103.9
32 2 and 3 0.078 0.066 0.095 5.0 4.1 6.2 64.7
33 2 and 3 0.059 0.059 0.061 2.9 2.9 3.0 49.3
Note: —, data not available.
a
Method 1, simplified analysis; method 2, rigorous stability analysis; method 3, stability analysis considering kinetics (see Table 3); method 4,
laboratory steady-state testing.
Table 3. Back-calculated liquefied shear strength and strength ratios for the 10 cases that consider the kinetics of failure.
Shear strength and strength ratio considering kinetics Nonliquefied soils
Case Best estimate shear Lowerbound shear Upperbound shear Best estimate Lengtha Shear strength
history strength (kPa) strength (kPa) strength (kPa) strength ratio (%) (kPa)
2 16.0 10.4 19.1 0.106 12 52.6
3 34.5 28.7 37.8 0.112 7 104
6 27.3 16.8 34.0 0.078 25 4.8
10 1.7 — — 0.027 0 —
12 1.2 — — 0.052 0 —
15 18.7 15.8 21.8 0.112 33 38.1
22 2.0 1.0 3.2 0.062 18 8.3
28 3.9 3.4 4.7 0.076 14 7.3
32 5.6 3.9 8.3 0.086 9 10.5
33 4.8 3.0 5.7 0.097 16 21.5
a
Percentage of final sliding surface that incorporates soils that did not liquefy.
Sources of uncertainty in the analyses and their sured penetration resistance (see Tables 2, 3, and 4). The un-
importance certainty in back-calculated strength ratios resulted from
For a given case history, there often was a considerable several factors, including (i) limits of the zone of liquefac-
range of back-calculated liquefied strength ratios and mea- tion; (ii) shear strength of the nonliquefied soils; (iii) loca-
© 2002 NRC Canada
Monday, May 06, 2002 11:46:20 AM
Table 4. Measured and estimated penetration resistances for the liquefaction flow failure case histories.
Normalized Penetration Resistance Soil grain properties
Lowerbound Upperbound Stark and
Case Available Lowerbound Upperbound Reported Approximate Olson (1995)
Mean qc1 Mean (N1)60 (N1)60 (N1)60

history dataa (MPa) qc1 (MPa) qc1 (MPa) (blows/0.3 m) (blows/0.3 m) (blows/0.3 m) DR (%) Approximate D50 (mm) FCb (%)
Olson and Stark
qc/N60
1 CPT 3.0 1.7 4.4 7.5 4.2 10.9 — 0.12 3–11 0.4
Monday, May 06, 2002 11:46:21 AM

2 SPT 4.6 2.6 6.5 7 4 10 — 0.42 5–10 0.65
3 DR 5.5 1 6 8 2 12 20–50 — 10 to >60 ~0.5
4 DR 2.2 1.8 2.6 5 4 6 20–40 0.10 33–48 0.35
5 Est. 4.0 — — 6 — — ~40 — — ~0.5
6 SPT 3.4 1.6 5.6 8.5 4 14 40–50 0.06–0.2 ~55 0.3–0.5

7 DR 2.5 — — 4 — — ~32 0.17 6–8 0.47
8 DR; SPT 3.2 3 6.2 7.5 6.5 12.3 ~40 0.21 (0.18–0.25) 1–4 0.5 (0.48–0.55)
9 CPT; SPT 3.1 1.7 3.8 4.4 3.7 5.6 ~40–50 0.35 <5 0.59
10 Est. 1.8 — — 3 — — — 0.3–0.4 0–2 0.6 (0.57–0.62)
11 SPT 0 — — 0 — — — 0.08 to ?? 55–93 0.3 to ??
(desiccated to NC tailings)
12 Est. 1.35 — — 3 — — — 0.15 to 0.20 ~13 ~0.45
13 Est.c 1.05 0.9 1.2 2.6 2.3 3 — silty sand ~0.4
14 CPT 0.36 0.35 0.38 1.1 1 1.2 — ~0.074 ~50 ~0.32
sandy silt – silty sand
15 CPT; SPT 4.7 2.1 6.2 11.5 5 15 ~48 (d/s) ~0.074 (0.02–0.3) ~50 (5–90) ~0.32
16 CPT; SPT 3.0 2 4 7 4 15 ~30–40 ~0.15 ~10–15 ~0.45
17 CPT; SPT 0.5 0.25 1 2.7 0 6 — 0.04 85 0.28
18 CPT; SPT 0.5 0.25 1 2.7 0 6 — 0.04 85 0.28
19 CPT 4.5 2.6 7.8 8.7 5 15 ~30–50 0.22 2–12 0.52
20 CPT 3.8 1.9 8.0 7.2 3.6 15.3 ~30–50 0.22 2–12 0.52
21 CPT 3.8 1.9 8.0 7.2 3.6 15.3 ~30–50 0.22 2–12 0.52
22 CPT; SPT 3.0 1.1 4.9 4.4 3.1 5.8 — 0.2 10–20 0.5
23 SPT 4.0 3.4 4.6 7 6 8 — 0.3 (0.15–0.55) 32 (23–38) 0.57
24 SPT 2.0 1.8 2.3 4.5 4 5 — ~0.15 ~30 ~0.45
25 SPT 4.1 3.2 5 9 7 11 — ~0.15 ~20 ~0.45
26 SPT 1.8 1.0 2.5 3.5 2 5 — ~0.2 ~15 ~0.5
27 CPT 2.9 1.3 4.5 5.3 2.4 8.2 –25 to 5 0.25 0–5 0.55
28 SPT 1.9 0.6 4.4 3 1 7 ~0 0.4 0 0.63
29 SPT 2.6 1.8 4.4 5.2 2.6 9 — ~0.2 18 ~0.5
30 SPT 6.0 2.3 8.1 9.2d 3.6d 12.4d — ~1.5 ~20 0.65e
31 CPT 1.9 1.1 2.4 7.6 4.4 9.6 — 0.012 100 0.25
32 Est.c 2.8 1.5 5.4 5.6 2.9 10.7 — 0.2 (0.12–0.4) 20 (12–35) 0.5
33 SPT 3.2 1.2 5.0 6.3 2.4 10 — ~0.2 f 20 f ~0.5
a
CPT, measured cone penetration resistance; SPT, measured standard penetration resistance; DR, relative density; Est., estimated.
b
FC, fines content.
c
Values of SPT and CPT penetration resistance were estimated from measured Swedish cone penetration test results.
d
Values of N60 were corrected for gravel content as described in Terzaghi et al. (1996).
e
qc /N60 determined from data presented by Yegian et al. (1994); D50 is outside range reported by Stark and Olson (1995).
f
Values of D50 and FC were estimated from same parent soil deposit described in Miura et al. (1998).
© 2002 NRC Canada

639
Fig. 5. A comparison of liquefied strength ratio relationships based on normalized CPT tip resistance.
tion of the initial and final surfaces of sliding; (iv) location etration resistance was estimated from relative density and
of the phreatic surface within the slope in a few cases; vertical effective stress, additional uncertainty is introduced
(v) potential of drainage or pore-water pressure redistribu- in the estimate of representative penetration resistance.
tion occurring during flow (i.e., undrained condition is not Other uncertainties in interpreting penetration resistance in-
maintained); and (vi) location of the postfailure slope toe in clude (i) effects of flow, reconsolidation, and aging when
a few cases. Olson (2001) describes the uncertainties in- the penetration tests were conducted some time after failure;
volved in each case history. Uncertainty due to the potential (ii) position of the phreatic surface at the time of testing;
of drainage or pore-water pressure redistribution occurring (iii) differences in penetration resistance when penetration
during flow is inherent in all studies of liquefaction case his- tests were conducted near (or opposite to) the location of
tories, and simplified methods to account for this potential failure (e.g., LSFD); and (iv) upper limit of the overburden
effect have not been developed. correction for conditions of low vertical effective stress (a
There is also considerable uncertainty in defining a “rep- maximum correction of two was used for this study). It
resentative” penetration resistance due to the inherent vari- should be noted that the majority of available penetration
ability of natural deposits and the typical segregation or tests were conducted following liquefaction; however, pene-
layering encountered in some man-made deposits (Popescu tration tests for a number of cases were conducted prior to
et al. 1997). This uncertainty is apparent for some large val- liquefaction. Olson (2001) indicates the timing of penetra-
ues of upper bound penetration resistance (see Table 4). In tion tests with respect to the occurrence of liquefaction; de-
some cases, sufficient penetration resistance results are tails the interpretation of representative, lower bound, and
available to interpret reasonable upper and lower bounds to upper bound penetration resistances; and describes the un-
the data. Unfortunately in many cases, insufficient data are certainties applicable to each case history.
available to make a reasonable judgment. Therefore, the Examining Tables 2, 3, and 4, the ranges of reported values
upper bound value for these cases is the highest value of appear large, particularly for penetration resistance. How-
penetration resistance measured near or in the zone of lique- ever, these are the same ranges of upper and lower bound
faction, despite the fact that the highest value is unlikely to strength and penetration resistance implicit in the relation-
be representative of the material that liquefied. The selection ships developed by Seed (1987), Seed and Harder (1990),
of a “representative” penetration resistance is discussed fur- Stark and Mesri (1992), and Ishihara (1993), although these
ther in a subsequent section. investigators do not describe the magnitude or sources of un-
In cases where penetration resistance was converted using certainty involved in the case histories. For example, for
the qc /N60 relationship (Stark and Olson 1995) or where pen- LSFD, the “representative” value of (N1)60 within the zone
© 2002 NRC Canada
Monday, May 06, 2002 11:46:23 AM
Olson and Stark 641
Fig. 6. A comparison of liquefied strength ratio relationships based on normalized SPT blowcount.
of hydraulic fill that likely liquefied has been reported as 15 taken as the mean values. So that the actual ranges can be
by Seed (1987), 5.5 by Davis et al. (1988), 11.5 by Seed et examined, Table 4 includes upper and lower bound values of
al. (1989) and Seed and Harder (1990), 12 (with a range of penetration resistance. As previously mentioned, sufficient
9–15) by Jefferies et al. (1990), 7 (with a minimum repre- penetration resistance results are available for some cases to
sentative value of 4) by McRoberts and Sladen (1992), and interpret reasonable upper and lower bounds to the data.
8.5 by Poulos (1988) and Castro (1995). These “representa- However, in many cases, insufficient data are available to
tive” values vary from 4 to 15; a considerable range in itself. make a reasonable judgment. For these cases, the upper
In the downstream shell of the dam, the actual measured val- bound is the maximum value of penetration resistance mea-
ues of (N1)60 ranged from 6 to over 40. Correcting these val- sured near or in the zone of liquefaction, despite the fact that
ues to correspond to the upstream slope (Seed et al. 1989), the highest value is very unlikely to be representative of the
the (N1)60 values are approximately 3 to 37. material that liquefied.
The true range of (N1)60 values (from 3 to 37) for LSFD is
not shown in existing relationships between liquefied shear
strength or strength ratio and (N1)60 because this range Interpretation and discussion
would plot off the chart. Individual investigators determined
mean and (or) median values of (N1)60 within the zone of Despite the uncertainties for each case, a reasonable trend
liquefaction and used engineering judgment to evaluate if in the data is apparent, particularly for the cases where the
these values were “representative” of the hydraulic fill that most information is available (cases plotted with a solid,
liquefied and led to the observed failure. Large (N1)60 values half-solid, or open circle in Figs. 5 and 6). Upper bound,
(probably above 15–20) are likely too dense to be contrac- lower bound, and average trendlines are proposed in Figs. 5
tive under the effective stresses present in the upstream slope and 6. The average trendlines are linear regressions of the
of LSFD, and are therefore too large to be representative. data excluding the cases where only the simplified analysis
Small (N1)60 values (probably less than about 6) are some- was conducted (cases plotted as triangles in Figs. 5 and 6).
what anomalous and probably do not represent the overall The average trendlines are described as
density of the hydraulic fill. Therefore, mean, median, or
values of (N1)60 based on judgment are reported in the litera- su (LIQ)
[19a] = 0.03 + 0.0143(q cl ) ± 0.03
ture and used in existing relationships between liquefied shear σ vo
′
strength or strength ratio and penetration resistance.
In this study, “representative” values of qc1 and (N1)60 are for q c1 ⱗ 6.5 MPa
© 2002 NRC Canada
Monday, May 06, 2002 11:46:26 AM
su (LIQ) Fig. 7. An evaluation of the strength ratio concept using lique-

[19b] = 0.03 + 0.0075[(N l ) 60 ] ± 0.03
σ vo
′ faction flow failure case histories.
for (N1) ⱗ 12
The upper and lower trendlines in Figs. 5 and 6 approxi-
mately correspond to plus and minus one standard deviation
(the standard deviation for both trendlines was ±0.025).
Included in Fig. 5 is the design line presented by Olson
(1998). The Olson (1998) design line is conservative for all
values of qc1. Included in Fig. 6 are the boundaries for lique-
fied strength ratio proposed by Stark and Mesri (1992) and
the design lines proposed by Stark and Mesri (1992) in
eq. [9] and Davies and Campanella (1994). The data in Fig. 6
show considerably less scatter compared to the bounds pre-
sented by Stark and Mesri (1992) as a result of the improved
analyses conducted in this study. The design line proposed
by Stark and Mesri (1992) in eq. [9] is conservative for all
values of (N1)60, while that proposed by Davies and
Campanella (1994) is unconservative for (N1)60 values
greater than 8. It should be noted that part of the conserva-
tism of the Olson (1998) and Stark and Mesri (1992) design
lines results from incorporating a fines content adjustment,
while data in Figs. 5 and 6 are plotted without any adjust-
ment for fines content.
Back-calculation of liquefied shear strength

To evaluate the correlation between liquefied shear strength
and prefailure vertical effective stress, “single values” of liq-
uefied shear strength also were back-calculated from the
case histories. As suggested by Seed (1987) and Seed and
Harder (1990), “single values” of su(LIQ) were evaluated
from a static slope stability analysis of the postfailure geom- the improved analysis procedures of this study yielded
etry using Spencer’s (1967) method as coded in the com- different values of su(LIQ) and prefailure σ vo ′ than those re-
puter program UTEXAS3 (Wright 1992). The “single value” ported by Stark and Mesri (1992). Figure 7 illustrates that
of su(LIQ) was varied until a factor of safety of unity was this re-evaluation of case histories results in considerably
achieved. Appropriate drained or undrained shear strengths less scatter between the upper and lower bound relationships
were assigned to nonliquefied soils. As previously men- of su(LIQ) and prefailure σ vo ′ . In summary, the case history
tioned, ten cases had sufficient documentation to perform a data presented in Fig. 7 confirm laboratory data that indicate
kinetics analysis of the failure. This analysis also estimated a linear relationship between liquefied shear strength and
su(LIQ). Values of su(LIQ) back-calculated using the simpli- initial vertical effective stress (e.g., Fig. 1 from Baziar and
fied and static stability analyses are presented in Table 2, Dobry 1995; Ishihara 1993).
while values incorporating kinetics are found in Table 3.
Olson (2001) details the individual case history analyses. Effect of fines content on liquefied shear strength and
Figure 7 presents the “single values” of su(LIQ) (from Ta- strength ratio
bles 2 and 3) and weighted average prefailure σ vo ′ (eq. [3] and Previous case history studies (Seed and Harder 1990; Stark
Table 2). Despite differences in density, mode of deposition, and Mesri 1992) incorporated fines content adjustments to
grain size distribution, grain shape, state parameter, modes of generate an “equivalent clean sand” blowcount and to evalu-
shear, and steady-state friction angle of the liquefied soils, the ate the liquefied shear strength or strength ratio. The purpose
data in Fig. 7 illustrate that an approximately linear relationship of the adjustment is to increase the penetration resistance of
exists between su(LIQ) and weighted average prefailure σ vo ′ for silty sands to that exhibited by clean sands with identical rel-
liquefaction flow failures. The relationship ranges from approx- ative densities. The reason for this difference in penetration
imately su(LIQ) = 0.05 to 0.12σ vo ′ with an average value (from resistance is, in part, related to differences in soil compress-
linear regression) of 0.09σ vo ′ . As illustrated in Figs. 5 and 6, ibility (Been et al. 1987b).
the variation in su(LIQ)/σ vo′ from 0.05 to 0.12 is explained in No fines content adjustment was adopted in this study. In
terms of increasing normalized penetration resistance. Figs. 5 and 6, the fines content of the liquefied soil is pro-
Baziar and Dobry (1995) also presented bounding rela- vided next to each data point. The data reveal no trend in
tionships between liquefied shear strength and prefailure liquefied strength ratio with respect to fines content. The au-
vertical effective stress as shown in Fig. 7. With the excep- thors anticipate that although soils with higher fines contents
tion of LSFD, Baziar and Dobry (1995) used the values of should exhibit lower values of penetration resistance (as a
su(LIQ) and prefailure σ vo ′ presented by Stark and Mesri result of greater soil compressibility and smaller hydraulic
(1992) to develop these boundaries. As discussed previously, conductivity), these soils are more likely to maintain an un-
© 2002 NRC Canada
Monday, May 06, 2002 11:46:27 AM
Olson and Stark 643
drained condition during flow. The combination of these sistance. Therefore, mean values of penetration resistance
factors may, in effect, offset each other, resulting in no ap- were used in this study.
parent difference in values of liquefied strength ratio for When assessing liquefaction triggering and post-triggering
cases of clean sands and sands with higher fines contents. stability in practice, minimum values of penetration resis-
Therefore, this study recommends no fines content adjust- tance often are used with empirical relationships. If a mini-
ment for estimating liquefied strength ratio from the proposed mum value of penetration resistance is used in conjunction
relationships. with the relationships proposed in Figs. 5 and 6, an engineer
may consider selecting a liquefied strength ratio greater than
Effect of kinetics on liquefied shear strength and the value corresponding to the average relationship. In a
strength ratio small parametric study of three existing (unfailed) dams, the
The ten cases that explicitly consider the kinetics of failure authors found that using the mean penetration resistance
(Table 3) provide the “best estimates” of liquefied strength ra- with the average relationships in Figs. 5 and 6 provided
tio because the kinetics analysis accounts for the momentum nearly the same liquefied strength ratios as using the mini-
of the failure mass in the back-calculation of su(LIQ). How- mum penetration resistance with the upperbound relation-
ever, in cases where the center of gravity of the failure mass ships in Figs. 5 and 6. In addition, because the upper- and
did not move a considerable vertical distance, the effect of lower-bounds of the relationships proposed in Figs. 5 and 6
kinetics was unclear and thus investigated. correspond approximately to plus and minus one standard
The effect of kinetics on the liquefied shear strength was deviation, the desired level of conservatism can be used to
examined with respect to (i) the loss of potential energy of the estimate the liquefied strength ratio.
failure mass as a result of sliding; and (ii) the prefailure
height of the embankment or slope. The loss of potential en- Applications of liquefied strength ratio
ergy was calculated as the average weight of the failure mass
(from the pre- and post-failure geometry) multiplied by the The liquefied strength ratio allows the variation in lique-
change in vertical position of the centroid of the failure mass fied shear strength throughout a zone of liquefied soil to be
as a result of sliding. The effect of kinetics on the back- incorporated in a post-triggering stability analysis. Increases
calculated su(LIQ) was examined in terms of the difference in in su(LIQ) can be the result of increases in prefailure vertical
liquefied shear strength considering kinetics [su(LIQ, Kinet- effective stress, increases in normalized penetration resis-
ics)] minus the liquefied shear strength not considering kinet- tance, or both. To incorporate a strength ratio in a stability
ics [su(LIQ)]. As illustrated in Fig. 8a, the effect of kinetics analysis, a liquefied soil layer can be separated into a num-
on the back-calculation of liquefied shear strength is not sig- ber of sublayers of equal σ vo ′ (stress contours) and (or) equal
nificant unless the loss of potential energy of the failure mass penetration resistance (penetration contours). For example,
is greater than approximately 103 to 104 kJ/m. Considering each vertical effective stress contour would have an equal
this issue in a simpler manner, Fig. 8b illustrates the effect of value of su(LIQ), and su(LIQ) would increase as the σ vo ′ con-
the prefailure height of the embankment/slope on the back- tours increased.
calculated liquefied shear strength. As shown in Fig. 8b, ki- Additionally, liquefied strength ratios can be used to facil-
netics has a minor effect on the liquefied shear strength for itate remediation studies. Two common remediation tech-
embankments/slopes less than about 10 m in height. Only 1 niques for seismic dam stability are the use of stabilizing
of the 23 case histories where a kinetics analysis was not con- berms and soil densification. If a stabilizing berm is added,
ducted involves a slope with a height greater than 10 m. new σ vo′ contours can be developed to estimate the liquefied
Therefore, liquefied strength ratios back-calculated for the shear strength for various berm heights. The increase in ver-
other 22 cases using the simplified or rigorous stability analytical effective stress caused by the weight of the stabilizing
ses also represent “best estimates.” Further, for design and berm decreases the void ratio of the liquefiable material and
remediation, kinetics does not appear to play a significant role results in an increase in su(LIQ). If soil densification is used,
in embankments/slopes that are less than 10 m in height. penetration tests typically are conducted to verify the suc-
cess of the densification effort. These additional penetration
Effect of penetration resistance on liquefied strength ratio tests can be used to revise the liquefied strength ratio, and
As noted previously, mean values of penetration resistance thus revise values of su(LIQ). Densification often increases
are plotted in Figs. 5 and 6. However, most failures occur the horizontal effective stress, σ ho ′ (and thus penetration
through the weakest zones of soil, not through the mean resistance), without significantly increasing the vertical ef-
value zones. Popescu et al. (1997, 1998) showed that pore- fective stress. However, the increase in σ ho ′ caused by
water pressure buildup during seismic shaking is bracketed densification should decrease the void ratio of the treated
when soil properties are estimated from penetration resis- material and result in an increase in su(LIQ).
tance values between the median (50th percentile) and 20th On high-risk projects, the compressibility of the liquefiable
percentile. (In their nomenclature, this is the 80th percen- soil should be compared to the slope of the steady-state line to
tile.) Therefore, it may be more appropriate to use the mini- confirm the applicability of the strength ratio concept. If the
mum or 20th percentile values of normalized penetration compressibility of the soil is not found to be reasonably paral-
resistance (Popescu et al. 1997, 1998; Yoshimine et al. 1999) lel to the slope of the SSL, at least over the range of effective
to develop the relationships proposed in Figs. 5 and 6. Un- stresses of interest, the strength ratio concept may not be appli-
fortunately, in most flow failure case histories there are in- cable for the particular soil. If the compressibility is signifi-
sufficient penetration test results available to reasonably cantly smaller than the slope of the SSL, the strength ratio will
estimate a 20th (or other) percentile value of penetration re- lead to unconservative estimates of su(LIQ). Engineers should
© 2002 NRC Canada
Monday, May 06, 2002 11:46:28 AM
Fig. 8. The difference in back-calculated shear strength considering and not considering kinetics compared to (a) loss of potential
energy resulting from flow failure; and (b) prefailure height of the embankment.
be particularly wary of this potential for remediation projects be susceptible to flow failure, the assumption of parallel con-
involving large berms. However, as discussed above and solidation and steady-state lines may be reasonable, particularly
shown by Olson (2001), generally, if a soil is loose enough to for soils with greater than 12% fines content.
© 2002 NRC Canada
Monday, May 06, 2002 11:46:33 AM
Olson and Stark 645
Conclusions Kramer, and David R. Gillette for their discussions on this

topic and their review of this manuscript.
This paper evaluates 33 liquefaction flow failure case histo-
ries using a stability analysis specifically developed to back-
calculate the liquefied strength ratio. This improved ap- References
proach allows the entire range of vertical effective stress act- Andresen, A., and Bjerrum, L. 1968. Slides in subaqueous slopes
ing on the liquefied material prior to failure to be considered in loose sand and sil. Norwegian Geotechnical Institute Publica-
in the back-calculation of the liquefied strength ratio, rather tion No. 81, pp. 1–9.
than using a single “representative” value of prefailure verti- Baziar, M.H., and Dobry, R. 1995. Residual strength and large-
cal effective stress. In addition, analyses that incorporate the deformation potential of loose silty sands. Journal of Geotechnical
kinetics of failure are conducted to obtain the “best esti- Engineering, ASCE, 121(12): 896–906.
mate” of liquefied strength ratio. These analyses show that Been, K., and Jefferies, M.G. 1985. A state parameter for sands.
the effect of kinetics on the back-calculation of liquefied Géotechnique, 35(2): 99–112.
strength ratio is important for embankments/slopes greater Been, K., Conlin, B.H., Crooks, J.H.A., Fitzpatrick, S.W., Jefferies,
than 10 m in height. M.G., Rogers, B.T., and Shinde, S. 1987a. Back analysis of the
The factors contributing to the uncertainty of both the Nerlerk berm liquefaction slides: Discussion. Canadian Geo-
back-calculation of liquefied strength ratio and normalized technical Journal, 24: 170–179.
penetration resistance are discussed. Despite these uncertain- Been, K., Jefferies, M.G., Crooks, J.H.A., and Rothenburg, L. 1987b.
ties, there are clear trends of increasing liquefied strength ra- The cone penetration test in sands: part II, general inference of
tio with increasing normalized standard and cone penetration state. Géotechnique, 37(3): 285–299.
resistance. Casagrande, A. 1940. Characteristics of cohesionless soils affect-
The resulting relationships between liquefied strength ra- ing the stability of slopes and earth fills. In Contributions to Soil
tio and penetration resistance exhibit considerably less scat- Mechanics, 1925–1940, Boston Society of Civil Engineers, Oc-
ter than relationships previously proposed (e.g., Stark and tober, (Originally published in the Journal of the Boston Society
of Civil Engineers, January, 1936), pp. 257–276.
Mesri 1992). The average trendlines presented in Figs. 5 and
Casagrande, A. 1965. Second Terzaghi lecture: role of the “calcu-
6 (or eqs. [19a] and [19b]) can be used to estimate the lique-
lated risk” in earthwork and foundation engineering. Journal of
fied shear strength ratio from CPT or SPT normalized pene- the Soil Mechanics and Foundations Division, ASCE, 91(SM4):
tration resistance, respectively. The average trendlines are 1–40.
described as Castro, G. 1995. Empirical methods in liquefaction evaluation. In
su (LIQ) Proceedings of the 1st Annual Leonardo Zeevaert International
= 0.03 + 0.0143(q cl ) ± 0.03 Conference. Vol. 1, pp. 1–41.
σ vo
′ Castro, G., Keller, T.O., and Boynton, S.S. 1989. Re-evaluation of
for q c1 ⱗ 6.5 MPa the Lower San Fernando Dam: Report 1, an investigation of the
February 9, 1971 slide. U.S. Army Corps of Engineers Contract
and Report GL-89–2, Volumes 1 and 2. U.S. Army Corps of Engi-
neers Waterways Experiment Station, Vicksburg, Mississippi.
su (LIQ)
= 0.03 + 0.0075[(N l ) 60 ] ± 0.03 Castro, G., Seed, R.B., Keller, T.O., and Seed, H.B. 1992. Steady-
σ vo
′ state strength analysis of Lower San Fernando Dam slide. Jour-
nal of Geotechnical Engineering, ASCE, 118(3): 406–427.
for (N1) 60 ⱗ 12 Chillarige, A.V., Robertson, P.K., Morgenstern, N.R., Christian,
The CPT-based relationship is preferred for design because H.A. 1997a. Evaluation of the in situ state of Fraser River sand.
of the fundamental advantages of the CPT over the SPT in Canadian Geotechnical Journal, 34: 510–519.
most liquefaction problems (see Stark and Olson 1995). Chillarige, A.V., Morgenstern, N.R., Robertson, P.K., and Christian,
These relationships provide values of liquefied strength ra- H.A. 1997b. Seabed instability due to flow liquefaction in the
Fraser River delta. Canadian Geotechnical Journal, 34: 520–533.
tios that are greater than previously proposed and can be
Christian, H.A., Woeller, D.J., Robertson, P.K., and Courtney, R.C.
used in post-triggering stability analyses to assess the poten-
1997. Site investigation to evaluate flow liquefaction slides at
tial for flow failure. By using a liquefied strength ratio,
Sand Heads, Fraser River delta. Canadian Geotechnical Journal,
rather than a representative value of liquefied shear strength, 34: 384–397.
the increase in liquefied shear strength with vertical effective Cunning, J.C. 1994. Shear wave velocity measurements of co-
stress can be incorporated in a post-triggering stability anal- hesionless soils for evaluation of in situ state. M.Sc. thesis, De-
ysis and evaluation of remedial measures. Applications for partment of Civil Engineering, University of Alberta, Edmonton.
utilizing a liquefied strength ratio are presented. Davies, M.P., and Campanella, R.G. 1994. Selecting design values
of undrained strength for cohesionless soils. In Proceedings of
Acknowledgments the 47th Canadian Geotechnical Conference, Halifax, Nova Sco-
tia, September 1994, BiTech Publishers. Vol. 1, pp. 176–186.
This study was funded by the National Science Founda- Davis, A.P., Jr., Poulos, S.J., and Castro, G. 1988. Strengths back-
tion (NSF), Grant Number 97–01785, as part of the Mid- figured from liquefaction case histories. In Proceedings of the
America Earthquake (MAE) Center headquartered at the 2nd International Conference on Case Histories in Geotechnical
University of Illinois at Urbana-Champaign. This support is Engineering, 1–5 June, St. Louis, MO, pp. 1693–1701.
gratefully acknowledged. The authors would also like to de Alba, P., Seed, H.B., Retamal, E., and Seed, R.B. 1987. Resid-
thank Gholemreza Mesri, Youssef Hashash, Steven L. ual strength of sand from dam failures in the Chilean earthquake
© 2002 NRC Canada
Monday, May 06, 2002 11:46:33 AM
of March 3, 1985. Earthquake Engineering Research Center Re- Konrad, J.M. 1991. The Nerlerk berm case history: some consider-
port No. UCB/EERC-87–11, University of California, Berkeley, ation for the design of hydraulic sand fills. Canadian Geotechnical
CA. Journal, 28: 601–612.
Dobry, R., and Alvarez, L. 1967. Seismic failures of Chilean tail- Konrad, J.-M., and Watts, B.D. 1995. Undrained shear strength for
ings dams. Journal of the Soil Mechanics and Foundations Divi- liquefaction flow failure analysis. Canadian Geotechnical Jour-
sion, ASCE, 93(SM6): 237–260. nal, 32: 783–794.
Ekstrom, A., and Olofsson, T. 1985. Water and frost — stability Koppejan, A.W., van Wamelen, B.M., and Weinberg, L.J.H. 1948.
risks for embankments of fine-grained soils. In Proceedings of Coastal flow slides in the Dutch province of Zeeland. In Proceed-
the Symposium on Failures in Earthworks, Institution of Civil ings of the 2nd International Conference Of Soil Mechanics and
Engineers, London, 6–7 March, Vol. 1, pp. 155–166. Foundation Engineering, Rotterdam, Netherlands, 21–30 June,
Engineering News-Record. 1925. What happened to municipal utili- pp. 89–96.
ties at Santa Barbara. Engineering News-Record, 95(4): 146–149. Kramer, S.L. 1989. Uncertainty in steady-state liquefaction evalua-
Fear, C.E., and Robertson, P.K. 1995. Estimating the undrained tion procedures. Journal of Geotechnical Engineering Division,
strength of sand: a theoretical framework. Canadian Geotechnical ASCE, 115(10): 1402–1419.
Journal, 32(4): 859–870. Lucia, P.C. 1981. Review of experiences with flow failures of tail-
Fiegel, G.F., and Kutter, B.L. 1994. Liquefaction induced lateral ings dams and waste impoundments. Ph.D. thesis, University of
spreading of mildly sloping ground. Journal of Geotechnical En- California, Berkeley, Calif.
gineering, ASCE, 120(12): 2236–2243. Marcuson, W.F., III, Ballard, R.F., Jr., and Ledbetter, R.H. 1979.
Finn, W.D.L. 1998. Seismic safety of embankment dams: develop- Liquefaction failure of tailings dams resulting from the Near Izu
ments in research and practice 1988–1998. In Specialty Confer- Oshima earthquake, 14 and 15 January 1978. In Proceedings of
ence on Geotechnical Earthquake Engineering and Soil Dynamics the 6th Pan-American Conference on Soil Mechanics and Foun-
III, Seattle. American Society of Civil Engineers, Geotechnical dation Engineering, Lima Peru, Vol. 2, pp. 69–80.
Special Publication No. 75, Vol. 2, pp. 812–853. McRoberts, E.C., and Sladen, J.A. 1992. Observations on static and
Hazen, A. 1918. A study of the slip in the Calaveras Dam. Engi- cyclic sand-liquefaction methodologies. Canadian Geotechnical
neering News-Record, 81(26): 1158–1164. Journal, 29: 650–665.
Hazen, A. 1920. Hydraulic-fill dams. Transactions of the American Meyerhof, G.G. 1971. The mechanism of flow slides in cohesive
Society of Civil Engineers, Paper No. 1458, pp. 1713–1821 (in- soils. Géotechnique, 1: 41–49.
cluding discussions). Middlebrooks, T.A. 1942. Fort Peck slide. Transactions of the
Hicks, M.A., and Boughrarou, R. 1998. Finite element analysis of the American Society of Civil Engineers, 107: 723–764.
Nerlerk underwater berm failures. Géotechnique, 48(2): 169–185. Mishima, S., and Kimura, H. 1970. Characteristics of landslides and
Hryciw, R.D., Vitton, S., and Thomann, T.G. 1990. Liquefaction and embankment failures during the Tokachioki earthquake. Soils and
flow failure during seismic exploration. Journal of Geotechnical Foundations, 10(2): 39–51.
Engineering, ASCE, 116(12): 1881–1899. Mittal, H.K., and Hardy, R.M. 1977. Geotechnical aspects of a tar
Ishihara, K. 1984. Post-earthquake failure of a tailings dam due to sand tailings dyke. In Proceedings of the Conference on Geo-
liquefaction of the pond deposit. In Proceedings of the Interna- technical Practice for Disposal of Solid Waste Materials, ASCE
tional Conference on Case Histories in Geotechnical Engineering, Specialty Conference of the Geotechnical Engineering Division,
Rolla, Missouri, 6–11 May, Vol. 3, pp. 1129–1143. Vol. 1, pp. 327–347.
Ishihara, K. 1993. Liquefaction and flow failure during earthquakes. Miura, K., Yoshida, N., Nishimura, M., and Wakamatsu, K. 1998.
Géotechnique, 43(3): 351–415. Stability analysis of the fill embankment damaged by recent two
Ishihara, K., and Koga, Y. 1981. Case studies of liquefaction in the major earthquakes in Hokkaido, Japan. In Proceedings of
1964 Niigata earthquake. Soils and Foundations, 21(3): 35–52. the1998 Geotechnical Earthquake Engineering and Soil Dynam-
Ishihara, K., Yasuda, S., and Yoshida, Y. 1990a. Liquefaction-induced ics Specialty Conference, Seattle, Washington, 3–6 August.
flow failure of embankments and residual strength of silty sands. ASCE Geo-Institute Geotechnical Special Publication No. 75,
Soils and Foundations, 30(3): 69–80. Vol. 2, pp. 926–937.
Ishihara, K., Okusa, S., Oyagi, N., and Ischuk, A. 1990b. Liquefac- Ohya, S., Iwasaki, T., and Wakamatsu, M. 1985. Comparative study
tion-induced flow slide in the collapsible loess deposit in Soviet of various penetration tests in ground that underwent liquefaction
Tajik. Soils and Foundations, 30(4): 73–89. during the 1983 Nihon-Kai-Chubu and 1964 Niigata earthquakes.
Jefferies, M.G., Been, K., and Hachey, J.E. 1990. Influence of In Proceedings of the Workshop on In-Situ Testing Methods for
scale on the constitutive behavior of sand. In Proceedings of the Evaluation of Soil Liquefaction Susceptibility, San Francisco,
43rd Canadian Geotechnical Engineering Conference, Laval California, Vol. 1, pp. 56–88.
University, Quebec, Vol. 1, pp. 263–273. Okusa, S., and Anma, S. 1980. Slope failures and tailings dam
Kayen, R.E., Mitchell, J.K., Seed, R.B., Lodge, A., Nishio, S., and damage in the 1978 Izu-Ohshima-Kinkai earthquake. Engi-
Coutinho, R. 1992. Evaluation of SPT-, CPT-, and shear wave- neering Geology, 16: 195–224.
based methods for liquefaction potential assessments using Loma Olson, S.M. 1998. Post-liquefaction shear strength from laboratory
Prieta data. In Proceedings of the 4th Japan–U.S. Workshop on and field tests: field tests. In Proceedings, Shear Strength of Liq-
Earthquake Resistant Design of Lifeline Facilities and Counter- uefied Soils. Edited by T.D. Stark, S.M. Olson, S.L. Kramer, and
measures for Soil Liquefaction, Honolulu, Hawaii. Vol. 1, T.L. Youd. University of Illinois at Urbana-Champaign, Illinois,
pp. 177–192. 130–152. (Available only on the World Wide Web at
Koester, J.P. 1998. Discussion group: theoretical/conceptual issues: http://mae.ce.uiuc.edu/info/resources.html.)
soil mechanics perspective. In Proceedings, Shear Strength of Olson, S.M. 2001. Liquefaction analysis of level and sloping ground
Liquefied Soils. Edited by T.D. Stark, S.M. Olson, S.L. Kramer, using field case histories and penetration resistance. Ph.D. thesis,
and T.L. Youd. University of Illinois at Urbana-Champaign, Illi- University of Illinois at Urbana-Champaign, Urbana, Illinois. (Avail-
nois. (Available only on the World Wide Web at able at http://pgi-tp.ce.uiuc.edu/olsonwebfiles/olsonweb/index.htm
http://mae.ce.uiuc.edu/info/resources.html.) or from the author on CD-ROM.)
© 2002 NRC Canada
Monday, May 06, 2002 11:46:34 AM
Olson and Stark 647
Olson, S.M., Stark, T.D., Walton, W.H., and Castro, G. 2000. 1907 Sladen, J.A., D’Hollander, R.D., Krahn, J., and Mitchell, D.E. 1987.
Static liquefaction flow failure of the North Dike of Wachusett Back analysis of the Nerlerk berm liquefaction slides: Reply. Ca-
Dam. Journal of Geotechnical and Geoenvironmental Engi- nadian Geotechnical Journal, 24: 179–185.
neering, ASCE, 126(12): 1184–1193. Spencer, E. 1967. A method of analysis of the stability of embank-
Pillai, V.S., and Salgado, F.M. 1994. Post-liquefaction stability and ments assuming parallel inter-slice forces. Géotechnique, 17(1):
deformation analysis of Duncan Dam. Canadian Geotechnical 11–26.
Journal, 31: 967–978. Sasaki, Y., Oshiki, H., and Nishikawa, J. 1994. Embankment fail-
Plewes, H.D., O’Neil, G.D., McRoberts, E.C., and Chan, W.K. ure cased by the Kushiro-Oki earthquake of January 15, 1993.
1989. Liquefaction considerations for Suncor tailings pond. In In Proceedings of the 13th International Conference on Soil Me-
Proceedings of the Dam Safety Seminar, Edmonton, Alberta, chanics and Foundation Engineering, New Delhi, India, Vol. 1,
Sept., Vol. 1, pp. 61–89. pp. 61–68.
Popescu, R., Prevost, J.H., and Deodatis, G. 1997. Effects of spa- Stark, T.D., and Mesri, G. 1992. Undrained shear strength of lique-
tial variability on soil liquefaction: some design recommenda- fied sands for stability analysis. Journal of Geotechnical Engi-
tions. Géotechnique, 47(5): 1019–1036. neering, ASCE, 118(11): 1727–1747.
Popescu, R., Prevost, J.H., and Deodatis, G. 1998. Characteristic Stark, T.D., and Olson, S.M. 1995. Liquefaction resistance using CPT
percentile of soil strength for dynamic analysis. In Specialty and field case histories. Journal of Geotechnical Engineering,
Conference on Geotechnical Earthquake Engineering and Soil ASCE, 121(12): 856–869.
Dynamics III, Seattle. American Society of Civil Engineers, Stark, T.D., Olson, S.M., Kramer, S.L., and Youd, T.L. 1998. Shear
Geotechnical Special Publication No. 75, Vol. 2, 1461–1471. strength of liquefied soils. In Proceedings of the Workshop on
Poulos, S.J. 1988. Liquefaction and related phenomena. In Ad- Post-Liquefaction Shear Strength of Granular Soils, University
vanced dam engineering for design, construction, and rehabilita- of Illinois at Urbana-Champaign, Urbana, Illinois, 17–18 April
tion, Van Nostrand Reinhold, New York. 1997, 288 p. (Available only on the World Wide Web at
Poulos, S.J., Castro, G., and France, W. 1985. Liquefaction evalua- http://mae.ce.uiuc.edu/info/resources.html.)
tion procedure. Journal of Geotechnical Engineering Division, Terzaghi, K., Peck, R.B., and Mesri, G. 1996. Soil mechanics in
ASCE, 111(6): 772–792. engineering practice, 3rd ed. John Wiley & Sons, Inc., New
Rogers, B.T., Been, K., Hardy, M.D., Johnson, G.J., and Hachey, York, 549 pp.
J.E. 1990. Re-analysis of Nerlerk B-67 berm failures. In Proceed- U.S. Army Corps of Engineers. 1939. Report on the slide of a por-
ings of the 43rd Canadian Geotechnical Conference – Prediction tion of the upstream face of the Fort Peck Dam, Fort Peck,
of Performance in Géotechnique, Quebec, Canada, Vol. 1, Montana. United States Government Printing Office, Washing-
pp. 227–237. ton, D.C.
Ross, G.A. 1968. Case studies of soil stability problems resulting U.S. Army Corps of Engineers. 1949. Report on investigation of
from earthquakes. Ph.D. thesis, University of California, Berke- failure of Sheffield Dam, Santa Barbara, California. Office of
ley, Calif. the District Engineer, Los Angeles, California, June.
Seed, H.B. 1987. Design problems in soil liquefaction. Journal of
Vaid, Y.P., and Sivathayalan, S. 1996. Static and cyclic liquefaction
Geotechnical Engineering Division, ASCE, 113(8): 827–845.
potential of Fraser Delta sand in simple shear and triaxial tests.
Seed, H.B., Lee, K.L., and Idriss, I.M. 1969. Analysis of Sheffield
Canadian Geotechnical Journal, 33(2): 281–289.
Dam failure. Journal of the Soil Mechanics and Foundations Di-
Vaid, Y.P., and Thomas, J. 1995. Liquefaction and post-liquefaction
vision, ASCE, 95(SM6): 1453–1490.
behavior of sand. Journal of Geotechnical Engineering, ASCE,
Seed, H.B., Lee, K.L., Idriss, I.M., and Makdisi, F. 1973. Analysis
121(2): 163–173.
of the slides in the San Fernando Dams during the earthquake of
Feb. 9, 1971. Earthquake Engineering Research Center Report Vasquez-Herrera, A., and Dobry, R. 1989. Re-evaluation of the
No. EERC 73–2, University of California, Berkeley, Calif. Lower San Fernando Dam: Report 3, the behavior of undrained
Seed, H.B., Seed, R.B., Harder, L.F., and Jong, H.-L. 1989. Re- contractive sand and its effect on seismic liquefaction flow fail-
evaluation of the Lower San Fernando Dam: Report 2, examina- ures of earth structures. U.S. Army Corps of Engineers Contract
tion of the post-earthquake slide of February 9, 1971. U.S. Report GL-89–2, U.S. Army Corps of Engineers Waterways Ex-
Army Corps of Engineers Contract Report GL-89–2, U.S. Army periment Station, Vicksburg, Mississippi.
Corps of Engineers Waterways Experiment Station, Vicksburg, Wright, S.G. 1992. UTEXAS3: A computer program for slope sta-
Mississippi. bility calculations. Geotechnical Engineering Software GS86–1,
Seed, R.B., and Harder, L.F., Jr. 1990. SPT-based analysis of cyclic Department of Civil Engineering, University of Texas, Austin.
pore pressure generation and undrained residual strength. In Yamada, G. 1966. Damage to earth structures and foundations by
Proceedings of the H.B. Seed Memorial Symposium, Bi-Tech the Niigata earthquake June 16, 1964, in JNR. Soils and Foun-
Publishing Ltd., Vol. 2, pp. 351–376. dations, 6(1): 1–13.
Shuttle, D., and Jefferies, M. 1998. Dimensionless and unbiased Yegian, M.K., Ghahraman, V.G., and Harutinunyan, R.N. 1994.
CPT interpretation in sand. International Journal for Numerical Liquefaction and embankment failure case histories, 1988 Ar-
and Analytical Methods in Geomechanics, 22: 351–391. menia earthquake. Journal of Geotechnical Engineering, ASCE,
Sladen, J.A. 1989. Problems with interpretation of sand state from 120(3): 581–596.
cone penetration test. Géotechnique, 39(2): 323–332. Yoshimine, M., Robertson, P.K., and Wride (Fear), C.E. 1999. Un-
Sladen, J.A., D’Hollander, R.D., and Krahn, J. 1985a. The liquefac- drained shear strength of clean sands to trigger flow liquefac-
tion of sands, a collapse surface approach. Canadian Geotechnical tion. Canadian Geotechnical Journal, 36(5): 891–906.
Journal, 22: 564–578.
Sladen, J.A., D’Hollander, R.D., Krahn, J., and Mitchell, D.E. 1985b.
Back analysis of the Nerlerk berm liquefaction slides. Canadian
Geotechnical Journal, 22: 579–588.
© 2002 NRC Canada
Monday, May 06, 2002 11:46:34 AM

Liquefied Strength Ratio From Liquefaction Flow PDF

Uploaded by

Document Informationclick to expand document information

Document Informationclick to expand document information

Copyright:

Available Formats

Liquefied Strength Ratio From Liquefaction Flow PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Liquefied Strength Ratio From Liquefaction Flow PDF

Uploaded by

Copyright:

Available Formats

Color profile: Generic CMYK printer profile

Composite Default screen

Liquefied strength ratio from liquefaction flow

Introduction cause it is generic and does not imply correspondence to any

630 Can. Geotech. J. Vol. 39, 2002

© 2002 NRC Canada

Olson and Stark 631

Liquefied strength ratio

© 2002 NRC Canada

632 Can. Geotech. J. Vol. 39, 2002

Fig. 2. A comparison of the consolidation behavior and the steady- su (LIQ)

to describe the constant volume conditions present at the

Back-analysis of liquefaction flow failures

© 2002 NRC Canada

Olson and Stark 633

© 2002 NRC Canada

634 Can. Geotech. J. Vol. 39, 2002

© 2002 NRC Canada

Olson and Stark 635

© 2002 NRC Canada

636 Can. Geotech. J. Vol. 39, 2002

Case histories of liquefaction flow failure

© 2002 NRC Canada

Monday, May 06, 2002 11:46:20 AM

© 2002 NRC Canada

638 Can. Geotech. J. Vol. 39, 2002

Mean qc1 Mean (N1)60 (N1)60 (N1)60

Monday, May 06, 2002 11:46:21 AM

6 SPT 3.4 1.6 5.6 8.5 4 14 40–50 0.06–0.2 ~55 0.3–0.5

© 2002 NRC Canada

640 Can. Geotech. J. Vol. 39, 2002

Olson and Stark 641

642 Can. Geotech. J. Vol. 39, 2002

su (LIQ) Fig. 7. An evaluation of the strength ratio concept using lique-

Back-calculation of liquefied shear strength

© 2002 NRC Canada

Olson and Stark 643

© 2002 NRC Canada

644 Can. Geotech. J. Vol. 39, 2002

© 2002 NRC Canada

Olson and Stark 645

Conclusions Kramer, and David R. Gillette for their discussions on this

© 2002 NRC Canada

646 Can. Geotech. J. Vol. 39, 2002

© 2002 NRC Canada

Olson and Stark 647

© 2002 NRC Canada

You might also like

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.