SEISMIC DESIGN AND PERFORMANCE OF SURFACE FOUNDATIONS - Earthquake Geotechnical Engineering - page 218

it is known that the undrained shear strength of clay is increased at high strain rates
(e.g. Ishihara, 1996; Ahmed-Zeki et al., 1999) and that, for insensitive clays, the post-
cyclic loading undrained shear strength is not greatly diminished (Andersen et al., 1980;
Ishihara, 1996). This observation indicates that the assessment of the bearing strength of
shallow foundations on saturated clay can often ignore the effect of earthquake accelera-
tion, as indicated in EC8 Part 5, and that traditional bearing strength approaches, which
have greater scope for shape and embedment effects, can be used.

H owever, there is one important modifying factor that needs to emphasised. The
F undrained dimensionless parameter indicates that there is a size effect on the undrained
bearing strength of a sh all ow foundation associated with the inertia loading of the soil
beneath the foundation. F undrained is a function of the product of a g B , so that an earth-
quake acceleration that is of negligible consequence for a foundation, say 2m wide,
mightbeofsignificanceforone20mwide.ThiswasalsonotedbyPender(1995)onthe
basis of avery simpleanalysis usingatwoblock failure mechanism.

The undrained shear strength, s u ,in F undrained suggests a clay soil. However, saturated
sand also has an undrained shear strength (although it is rather more difficult to estimate
than s u forclay).Thusthecommentsinthissectionapply,inprinciple,toshallowfounda-
tions on saturated sand. Saturated sand presents problems as the behaviour is very much
affected by cyclic loading. Leaving aside the question of liquefaction (as one would not
consider a shallow foundation on a liquefiable deposit of sand) and considering dense
sandsonestillobservesthatcyclicloadingleadstoasignificantsofteningofthematerial
(Ishihara, 1985) although not necessarily a loss in strength. Eurocode 8 suggests a way
to handle this phenomenon. A model factor is introduced into the top line of the defini-
tion of the dimensionless parameters in Equation (10.3). This is intended to make some
allowance for the uncertainty in the material and in particular to accommodate any dif-
ferences between the actual material behaviour and that assumed in deriving the bearing
strength surface specified in Equation (10.3). Values given in EC8 for this model para-
meter are given in Table 10.1. Clearly the main “target” in this table is loose saturated
sand. It is well known that shallow foundations in these materials are likely to be subject
to liquefaction, sothat consideration of this case is unlikely.

2.3. DRAINED RESPONSE

Depositsofdrycohesionlesssoilwillrespondinadrainedmannertoearthquakeloading.
Figure 10 .6 has plotted sections of the drained bearin g s tre ng th surface for a r an ge of
values of F drained . The top part of the figure gives the V

− H boundary in the M = 0

Table 10.1. Bearing strengthmodelfactors given inEC8

Medium-dense to Loose dry Loose saturated Nonsensitive clay Sensitive clay

dense sand sand sand

1.00 1.15 1.50 1.00 1.15