Environmental Engineering Reference
In-Depth Information
steady state undrained strength, S
us
, or liquefied shear strength, S
u(LIQ)
, depending on how
the strength is assessed.
Saturated zones of the dam or foundation which are subject to limited strain softening
(temporary liquefaction), or even strain hardening, will experience pore pressure build up
in the earthquake which should be modelled in this “post earthquake stability analysis”.
If the factor of safety is greater than 1.0 in these conditions, the slope will have experi-
enced deformations during the earthquake, but no further deformation should occur. If
however the factor of safety is less than 1.0, a flow slide will occur, with large, rapid
deformations.
In 1997, a National Science Foundation Workshop was held on “Shear Strength of
Liquefied Soils” (NSFW, 1998). The main objectives of the workshop included seeking
consensus on practice related issues concerning the shear strength of liquefied soils.
Unfortunately, unlike the NCEER workshop the year before (reported in Youd et al.,
2001), the participants did not reach consensus on important issues. The proceedings are
however a valuable reference. For the purposes of this chapter, the authors have relied
upon those proceedings, a later paper by Olsen and Stark (2002), Fell et al. (2000) and
references on the topic referred to in that paper.
As discussed in Fell et al. (2000), the steady state undrained strength, S
us
, of a soil is
dependent on the void ratio, confining stresses, stress path of loading and soil fabric. The
use of results from triaxial compression tests alone may be unconservative, because S
us
in
triaxial extension and direct simple shear may be lower.
Poulos et al. (1985) developed a laboratory procedure using triaxial compression tests
on reconstituted specimens. However later studies, summarized in NSFW (1998), ques-
tion the validity of this. NSFW (1998) recommend for high risk or high consequence of
failure projects, laboratory testing using undisturbed samples obtained by ground freez-
ing, but could not reach a consensus on test procedures.
There was a consensus that empirical corrections relating S
u(LIQ)
to back-analysed field
failures represented an economical and reasonable means for estimating the liquefied
shear strength. This has been the authors' experience and hence these procedures are
given in some detail below.
If the outcomes of such analyses are marginal, and the consequences of failure are suf-
ficiently high, it is recommended the services of an expert in the area be obtained to advise
further, particularly if laboratory testing is being contemplated. It would seem that sim-
plified laboratory procedures e.g. triaxial compression tests on reconstituted samples,
have little merit.
12.5.2
Method for estimating S
u(LIQ)
The authors have for some time used the method of Stark and Mesri (1992) to assess the
steady state undrained strength, S
us
, for use in post earthquake analysis. This method uses the
Standard Penetration Test. It requires adjustment of the SPT (N
1
)
60
values for fines content
(% passing 0.075 mm) to give (N
1
)
60cs
, or (N
1
)
60
clean sand equivalent. Stark and Mesri
(1992) back analysed failures, but also gathered a lot of laboratory test data and seem to
have relied on the latter rather than the back analysis to reach their recommended S
us
.
NSFW (1998) question the use of fines content corrections and recommend momentum
effects be accounted for in the analysis of case studies.
Olsen and Stark (2003) have followed this procedure in the analysis of 33 case studies
of flow liquefaction (excluding lateral spreading as recommended by NSFW, 1998). They
demonstrate that it appears that the Stark and Mesri (1992) recommendation was gener-
ally conservative.
Given this, it is recommended that the Olsen and Stark (2003) method be adopted. This
uses (N
1
)
60
and q
c1
values (for SPT and CPT), corrected to 100 kPa overburden stress, but
