Geoscience Reference
In-Depth Information
loose sand layer to liquefy and the embankment crests to displace longitudinally toward
thechannelbyabout0.75and1.7mfortheembankmentswithandwithoutpiles,respec-
tively. Thus, the pile pinning effect reduced longitudinal embankment displacements by
50% to 60% for this model. The embankments also developed substantial transverse
spreading and surface settlements, which are important considerations for evaluating the
post-earthquake accessibility orserviceability of a bridge.
Thedesignproceduresforestimatingpilepinningeffectsonembankmentsaredescribed
in the next section, after which they are illustrated by application to the dynamic cen-
trifuge model test shown in Figures12.14 and 12.15.
3.2. PROCEDURES FOR ESTIMATING PILE PINNING EFFECTS
Proceduresforestimatingpilepinningeffectsonlongitudinalembankmentdisplacements
have been used in practice and incorporated in recommended specifications for seismic
design of bridges (e.g., Martin et al., 2002). The first evaluation of these types of proce-
dures against physical data was based on centrifuge data (Boulanger et al., 2006), from
whichanumberofmodificationstothedetailsoftheproceduresdescribedinMartinetal.
(2002) were recommended. These procedures can be represented by threeprimary parts:
•
Estimate the longitudinal displacement of the embankment soil mass for a range of
restraining forces fromthepiles and bridge superstructure.
•
Estimate the longitudinal restraining force exerted on the embankment mass by the
piles and bridge superstructure for arange of imposed embankment displacements.
•
Determine the compatible displacement and interaction force between the embank-
ment mass and the piles and bridge superstructure.
Each of thesethree parts is discussed in more detail inthe following sections.
3.2.1. Estimating embankment displacements for a range of restraining forces
The estimation of embankment displacements due to liquefaction is complicated by the
needtoaccountforarangeofpossiblerestrainingforcesfromthepilesandbridgesuper-
structure. Of the four methods that were previously described for estimating free-field
lateral spreading displacements, only two provide a means to account for the effect of
restrainingforces:(1)Newmarkslidingblockmethodsand(2)nonlineardynamicnumer-
ical analyses. The method of integrating potential liquefaction-induced shear strains can
onlybeusedforthecaseofzerorestrainingforce,andtheavailableempiricalmodelsfor
lateral spreading are not applicable to bridge abutments. For this reason, designers have
turned to Newmark (1965) sliding block methods as the first step in the design process,
after which the potential benefits of more complicated dynamic numerical analyses can
be considered.
Thefirststepistoperformslopestabilityanalysesoftheembankment,suchasillustrated
inFigure12.16forthecentrifugemodelfromFigure12.14.Residualshearstrengths
S
r
)
of the liquefied layers may be estimated using case history based correlations between
(
