Proceedings of the 18
th
International Conference on Soil Mechanics and Geotechnical Engineering, Paris 2013, volume 6, 2016
The history of cyclic loadings applied to foundations, whether
calculated or measured, comprises a succession of variable loads
with an irregular amplitude and somewhat random distribution.
However, the cyclic tests that are feasible to run in the laboratory
on material samples have normally been designed as series of
cycles with a regular amplitude and constant period. A software
application called "Cascade", built to transform a single random
series of cyclic loads into a succession of ordered, constant-
amplitude series, was developed within the scope of this project.
This software employs cycle counting methods, of the "rain
flow" type (ASTM E 1049-85, NF A03-406, 1993). The concept
of damage, as intended by Miner, is applied herein to estimate
material damage on the basis of S-N type curves (also so-called
"Woelher curves"), as experimentally obtained by bringing to
failure the samples subjected to series of cycles with a constant
stress amplitude.
8.2.2
The SOLCYP project's experimental base
Pile tests were conducted at two experimental sites in northern
France. The first campaign took place at MERVILLE, where the
stiff and very highly overconsolidated Flanders clay was present
as of a depth of 3 m. Ten test piles were installed, featuring four
closed ended driven tubular metal piles, four CFA-type bored
piles, and two screwed piles. All piles were 13-m long with
diameters of either 406 mm (for the driven piles) or 420 mm
(bored piles). They underwent standard incremental static
loading tests, rapid monotonic loading tests and series of cyclic
tests including high-amplitude tests leading to failure at a small
number of cycles, and lastly low-amplitude tests run until 10,000
cycles. All loading modes were applied (tension, compression,
one-way, two-way). The main results were published in Benzaria
et al.
(2012, 2013a).
The second site, at LOON-PLAGE near Dunkirk, contained
dense sands. Two driven piles were set up along with five bored
CFA piles, as both types featured the same characteristics as
those at MERVILLE but with different lengths (10.5 m for the
driven piles and 8 m for the bored piles). The loading program
was also similar (Benzaria
et al.
, 2013b).
Many test series on instrumented model piles were completed
in both Fontainebleau sand and Speswhite clay. These tests were
conducted in the centrifuge at the Nantes-based IFSTTAR lab
(formerly LCPC). The objective targeted was to establish cyclic
stability diagrams in both types of reference soils (i.e. sands and
clays) and for both types of piles under study (driven and bored)
by considering a wide array of initial conditions (density,
consistency, consolidation) and loading modes. The first results
were the subject of publications (Guefresh
et al.
, 2012; Puech
et
al.
, 2013); this type of approach led to confirming the
representativeness of data acquired during
in situ
tests and then
extending their range of validity.
A third experimental approach consisted of performing tests
on very heavily instrumented model piles in the large calibration
chamber at the 3S-R Laboratory in Grenoble. These tests in
Fontainebleau sand, carried out in collaboration with Imperial
College London, yielded some outstanding information on the
mobilization of friction at the soil-pile interface and how it
evolves with changes in intensity and number of cycles (e.g. see
Tsuha
et al.
, 2012; Silva
et al.
, 2013).
8.2.3
Responses of piles to cyclic loadings
Figure 17 illustrates the type of behavior observed at the
MERVILLE site with highly overconsolidated clay. This figure
depicts the load-displacement relationship at the head of the
bored F2 pile loaded in compression. The ultimate load in
compression Q
uc
, as measured via a standard static test on pile
F1 (which is identical to F2), equaled 900 kN. The F2 pile,
which did not undergo any preliminary loading, was initially
subjected to a series of three cyclic loadings exceeding 3,000
cycles. The first two series (CC1 and CC2) did not cause any
significant permanent pile displacement. The third one (CC3),
characterized by a maximum loading Q
max
on the order of
800 kN, however generated fairly sizable permanent
displacements (nearly 20 mm). The test was stopped after 3,000
cycles and followed by a rapid static loading (CR1), which
indicated a post-cyclic capacity of 900 kN. Next, seven series of
cycles were applied. Series CC4 through CC7 did not bring
about any permanent pile head displacement upon completing
1,000 cycles per series. (Let's note that these tests were
arbitrarily separated to allow for visualization.) Tests CC8
through CC10, which once again reached a maximum force of
800 kN, generated permanent displacements that quickly began
to accumulate (each series was run for fewer than 100 cycles).
Post-cyclic capacity remained on the order of 900 kN (tests CR2
through CR4).
Fig. 17: Force-displacement relations at the head during
one-way compression tests on the MERVILLE F2 bored pile
(according to Benzaria
et al.
, 2013a)
This type of behavior was found during all tests carried out at
Merville, regardless of either pile type (bored, screwed, driven)
or loading mode (pure compression, pure tension, alternating
tension/compression). In sum:
-
A critical operating threshold is in place;
-
For a maximum load Q
max
below the given threshold, the pile
is stable (no significant permanent displacement, constant
cyclic stiffness) even with a high number of cycles (N>1000);
-
Once this threshold is reached, permanent displacements are
generated, and cyclic failure quickly ensues, typically in fewer
than 100 cycles;
-
The threshold is high in the regime of one-way loadings
(80%-90% of Q
us
) but decreases in the regime of two-way
loadings (see Fig. 19);
-
The post-cyclic capacity is not affected by prior cyclic
loadings.
The behavior observed at the LOON-PLAGE dense sand site
is altogether different. Figure 18 shows the response of two
identical bored F4 and F5 piles. The F4 pile underwent a
standard static loading test, yielding an ultimate reference load
Q
uc
=1,100 kN. The F5 pile was subjected to a cyclic loading
characterized by a Q
max
value of ~0.62 Q
uc
. The pile very quickly
accumulated permanent displacements (3% relative displacement
after just 14 cycles). The test was halted and the cyclic amplitude
substantially reduced (Q
max
~0.35 Q
uc
). The pile continued to
accumulate displacements (14 mm over 5,000 cycles).
From a general perspective, these observations were noted:
-
Bored piles are highly sensitive to cyclic loadings;
-
The post-cyclic capacity was significantly affected by the
cyclic loadings;
-
The cyclic failure criteria in compression need to be defined
in terms of tolerable displacements. An analysis of all tests
performed has led to defining the cyclic failure for a 3%
relative displacement.
Volume 6 - Page 90