DRAINED U P L I F T CAPACITY OF D R I L L E D SHAFTS

UNDER REPEATED AXIAL LOADING

By John P. Turner, 1 Associate Member, and Fred H. Kulhawy, 2
Fellow, ASCE
Downloaded from ascelibrary.org by University of Hong Kong on 01/08/22. Copyright ASCE. For personal use only; all rights reserved.

ABSTRACT: An experimental study of the effects of repeated axial loading on

drained uplift capacity of drilled shafts in granular soil is described. The mecha-
nisms that cause changes in drilled shaft side resistance are identified and the ef-
fects of initial soil density, shaft-depth-to-diameter ratio, and magnitude of re-
peated loading are evaluated. Uplift capacity changes are found to depend primarily
upon the magnitude of cyclic displacement, and a relationship between the mag-
nitude of cyclic displacement and uplift-capacity reduction is presented. Critical
levels of repeated loading (CLRL) are established, above which shafts fail in uplift
and below which failure under repeated loading does not occur. Implications for
the design of drilled shafts under repeated axial loading are presented.

INTRODUCTION
Drilled shafts often are used to support structures subjected to repeated
live loads that cause alternating uplift and compression forces to be trans-
mitted to the foundations. Examples include electrical transmission line and
microwave towers subjected to wind, ice, and maintenance loads. The re-
sulting axial forces transmitted to the foundations will reverse direction (uplift/
compression) repeatedly. The basic design criteria for drilled shafts sup-
porting such structures are that they must (1) be stable under the applied
loads; and (2) not move in such a way as to impair the function of the
structure. Under repeated uplift/compression loads, uplift capacity usually
is the controlling factor for foundation stability, and foundation movements
often are greater than for sustained uplift or compression loads of the same
magnitude.
Design methods for drilled shaft foundations under static axial loads are
well developed for a wide variety of subsurface conditions (O'Neill and Reese
1970; Reese and Wright 1977; Kulhawy et al. 1983) and have been sub-
stantiated by field performance data [e.g., Stas and Kulhawy (1984)]. Ad-
equate guidelines have not been developed, however, for evaluating the ef-
fects of repeated axial loading on drilled shaft uplift capacity. Existing data
on the behavior of drilled shafts under repeated axial loading are very limited
and inadequate to allow the formulation of design guidelines (Turner et al.
1987).
To address this lack of information, a study was conducted to examine
experimentally the behavior of drilled shafts during repeated uplift/compres-
sion loading under drained conditions. The purposes of this study were to
(1) determine the effects of repeated axial loading on the uplift capacity of
drilled shafts; (2) describe the mechanisms that cause failure during repeated
loading; and (3) systematically evaluate the effects of initial soil condition,
shaft geometry, and magnitude of repeated loading on shaft capacity. Static
'Asst. Prof., Dept. of Civ. Engrg., University Station Box 3295, Univ. of Wy-
oming,
2
Laramie, WY 82071.
Prof., School of Civ. and Envir. Engrg., Cornell Univ., Ithaca, NY 14853.
Note. Discussion open until August 1, 1990. To extend the closing date one month,
a written request must be filed with the ASCE Manager of Journals. The manuscript
for this paper was submitted for review and possible publication on December 29,
1988. This paper is part of the Journal of Geotechnical Engineering, Vol. 116,
No. 3, March, 1990. ©ASCE, ISSN 0733-9410/90/0003-0470/$1.00 + $.15 per
page. Paper No. 24461.
470

J. Geotech. Engrg., 1990, 116(3): 470-491

 uplift and repeated uplift/compression load tests were conducted in the lab-
oratory on model drilled shafts constructed in carefully prepared and instru-
mented deposits of dry sand. Detailed descriptions of the study and com-
prehensive results are given by Turner and Kulhawy (1987a).
UPLIFT CAPACITY OF DRILLED SHAFTS
Downloaded from ascelibrary.org by University of Hong Kong on 01/08/22. Copyright ASCE. For personal use only; all rights reserved.

Drilled shaft foundations develop their resistance to uplift loads from a

combination of side resistance, Qm, and tip resistance, Q,u, as illustrated in
Fig. 1. The resisting forces must be in vertical equilibrium with the applied
butt load and the weight of the foundation and enclosed soil within the vol-
ume A X D (in which A = shaft cross-sectional area; and D = shaft depth).
The uplift capacity then is given by
Qu = Qsu + Q,u + W (1)
in which the subscript u refers to uplift.
Tip resistance in uplift can be provided by either suction or tension. Tip
suction is a pore water stress effect that does not act during drained loading.
Tip tension usually is neglected because of the very low tensile strength of
most soils and because common construction practices usually result in a
thin smear zone of very low tensile strength at the tip (Kulhawy et al. 1983).
The side resistance of a deep foundation is expressed as

7(z)dz. (2)

in which T(Z) = shearing resistance with depth along the foundation side;
and the integral represents a summation of shear stress over the foundation
surface area. For a cylindrical shear mode of failure, the side resistance of
a drilled shaft under drained conditions is given by

tQ
I

FIG. 1. Equilibrium Forces Acting on Drilled Shaft Foundation in Uplift

471

J. Geotech. Engrg., 1990, 116(3): 470-491

 Qs = -nB ah(z) tan 8(z)rfz (3a)

Qs = TtB av(z)K(z) tan h(z)dz (3b)

Downloaded from ascelibrary.org by University of Hong Kong on 01/08/22. Copyright ASCE. For personal use only; all rights reserved.

Jo
in which oh = horizontal effective stress (which acts as a normal stress on
the soil-foundation interface); 5 = effective stress angle of friction for the
soil-foundation interface; z = depth; &„ = vertical effective stress (=yz); 7
= effective unit weight of soil; and K = coefficient of horizontal soil stress
(ah/av). The critical terms to evaluate in Eq. 3 are K and 8.
For concrete cast against soil, the interface friction angle,_ 8, is equal to
the soil friction angle, cj> (Kulhawy et al. 1983). The value of <>
j is determined
by appropriate shear testing. Repeated applications of load can change the
friction angle of soil by altering the void ratio, thereby changing the interface
shearing resistance of a drilled shaft.
The coefficient of horizontal soil stress, K, is the most difficult term to
evaluate in Eq. 3b. K is a function of the original in-situ horizontal soil
stresses (K0), the stress changes caused by construction, and the stress changes
that occur during loading. Stas and Kulhawy (1984) discuss the methods of
evaluating design values of K for deep foundations under static axial loads,
including in-situ measurement, geologic reconstruction, and empirical cor-
relations.
The interface angle of friction and the coefficient of horizontal soil stress
are often grouped as

Qs = TTB yz$(z)dz (4)

Jo
in which p = K tan 8. The experimental program described in this paper
was designed to determine the range of values to be expected for changes
in drilled shaft side resistance and uplift capacity caused by repeated axial
loading. Side resistance changes are evaluated quantitatively in terms of changes
in p, combining the effects of changes in 8 and K.
EXPERIMENTS
Laboratory uplift and two-way repeated load tests were conducted on model
drilled shafts at two scales. The first series, referred to as the small-scale
models, were nominally 76 mm diameter and either 305 mm (D/B = 4) or
610 mm (D/B = 8) in length and were constructed in deposits of dry sand
prepared in a steel testing chamber. The second series, referred to as the
large-scale shafts, were nominally 152 mm in diameter and 1,219 mm long
and were constructed and tested in a 2.1 -m diameter by 2.9-m deep testing
pit. The materials, equipment, and procedures are described briefly here.
The range of D/B ratios used in this study (D/B = 4 to 8) is typical of
drilled shafts used to support transmission line and microwave structures.
Soil Type
The soil used in the tests is a poorly graded sand, described locally as
filter sand. This sand is composed dominantly of subangular limestone and
quartz particles, with specific gravity of 2.74 and the following gradation:
472

J. Geotech. Engrg., 1990, 116(3): 470-491

 Relative Density, Dr
0 0.5 l.o
1
50 o-.(kN/m2)
/ 2 . 5 • 15 _
CD
a; CO
Downloaded from ascelibrary.org by University of Hong Kong on 01/08/22. Copyright ASCE. For personal use only; all rights reserved.

g1 45 OY /20 \~
10 <U

a.
-o- 4S/°
c
< tf' 4>cv=32.9'
c
o
35
-A*l
_ j
S.D.
o A/& -J. -

\_ —° n - ^ - 8 - * - A — AA"
u.
^
a
30
J
cu
0.
25
14 15 16 17 18 . 19

Density, y (kN/m3)

FIG. 2. Friction Angle versus Filter-Sand Density

D10 = 0.2 mm; Dm = 0.52 mm; Dm » 2.5-3 mm. The dry density ranges
from 15.5 kN/m 3 (minimum) to 18.3 kN/m 3 (maximum) in accordance with
ASTM D2049-69 (Stewart and Kulhawy 1981).
The soil-friction angle was measured in direct shear at normal stresses
ranging from 2.5 to 20 kN/m 2 by Trautmann et al. (1985). These normal
stress values are representative of those for the model tests. The variation
of friction angle with initial sand density for a normal stress of 5 kN/m 2 is
shown in Fig. 2. The peak values ((f>p) are given by circles and the softened
or critical state values (c|>ro) are given by triangles, showing plus or minus
one standard deviation. Using Bolton's (1986) general interpretation method,
the data are described well by the following:
«>, - 4>„ = 2[Dr(9.7 - In a„) - 1] (5)
in which Dr = relative density expressed as a ratio; o-„ = normal stress in
kN/m 2 ; and 9.7 = Bolton's Q parameter for the mixed limestone/quartz
mineralogy. The lines forCT„= 2.5, 5, and 20 kN/m 2 are given by Eq. 5
and describe the data well.
Facilities for Small-Scale Tests
The facilities for both test series each consisted of a testing chamber, soil
placement capabilities, loading system, instrumentation, and data acquisition
system. The small-scale facilities are described first. Further details are given
by Turner and Kulhawy (1987a).
Soil Placement
The small-scale tests were conducted within a. steel cylindrical dram, 565
mm in diameter and 867 mm deep. Filter sand was deposited in the test
drum by air pluviating from a suspended soil hopper. By adjusting the hop-
473

J. Geotech. Engrg., 1990, 116(3): 470-491

 102x102mm l.lm Reaction
aluminum sections frame

Load cell Loading

device
Downloaded from ascelibrary.org by University of Hong Kong on 01/08/22. Copyright ASCE. For personal use only; all rights reserved.

Model 2.0m
Stress cells shaft

-Wheel assembly

FIG. 3. Small-Scale Testing System

per height, the as-placed unit weight of the loose and medium-dense deposits
was controlled. Dense filter sand was obtained by air pluviating in 100 mm
lifts, followed by vibratory compaction of each lift. These placement meth-
ods resulted in the following relative density (Dr) ranges: loose (14 to 25%),
medium dense (39 to 50%), and dense (86 to 93%). As-placed densities for
loose sand were determined by weighing cylindrical pans placed on the sur-
face and filled with sand pluviated from the hopper. For medium and dense
sand, the Selig density scoop (Trautmann et al. 1985) was used. Both meth-
ods had a standard error of about 1%.
Loading Equipment and Instrumentation
Fig. 3 shows the small-scale test setup. For each test, instruments were
used to measure the soil total stresses, the axial load at the shaft butt (top),
and the shaft axial displacements. The in-situ soil horizontal stresses adjacent
to the shafts were monitored during construction and loading with deflecting-
diaphragm, resistance strain-gaged type stress cells, as described by Weiler
and Kulhawy (1978). Stress cells were placed approximately 20 mm from
the soil-shaft interfaces. For the 610-mm long shafts, stress cells were in-
stalled at depths of 127, 254, 381, 508, and 584 mm. For the 305-mm long
shafts, stress cells were installed at depths of 102, 203, and 208 mm. Axial
uplift and compression loads were applied with a displacement-control ac-
tuator and measured at the shaft butt with a load cell, and axial displace-
ments were measured with both a dial gage and a direct current differential
transducer (DCDT).
Data Acquisition
The analog outputs from all electrical transducers, including stress cells,
load cells, and DCDT, were monitored, digitized, and transferred to a desk-
top computer, where they were processed and/or stored.
Shaft Construction
Construction of the model shafts was designed to simulate, as closely as
possible, the casing method of drilled shaft construction. In the field, a tem-
porary steel casing is installed into the hole during or after excavation to
prevent the hole from caving. The casing is removed while concrete is placed
into the hole to form the shaft. In the laboratory tests, the casing was placed
474

J. Geotech. Engrg., 1990, 116(3): 470-491

 during preparation of the soil deposits. Therefore, the excavation phase of
construction was not simulated. All other phases of construction, including
concrete placement, casing extraction, and concrete curing, were simulated
as realistically as possible.
Shaft Loading
Downloaded from ascelibrary.org by University of Hong Kong on 01/08/22. Copyright ASCE. For personal use only; all rights reserved.

A modified version of the method of equilibrium (Mohan et al. 1967) was

used to conduct the static uplift and two-way repeated load tests described
herein. The axial load and displacement were monitored continuously as the
shaft displacement was increased until the desired axial load was achieved.
After achieving a state of equilibrium, which usually occurred within a few
seconds, all instruments were read, and then the loading was continued. The
direction of loading was reversed and the process was repeated to achieve
two-way repeated axial loads. Each complete cycle of loading took approx-
imately 30 s.
Test Variables
The testing program was designed to evaluate the effects of soil properties,
shaft geometry, and level of repeated loading on drilled shaft behavior. Table
1 summarizes the variables for the 24 small-scale tests conducted. Three soil
densities, two D/B ratios, and two loading modes were used, including static
uplift and two-way repeated loading. The magnitude of repeated load is the
maximum value of load applied on a repeated basis. For example, a mag-
nitude of 100 N consists of 100 N in uplift followed by 100 N in compres-
sion, and so on.
For each density and D/B ratio, at least one static uplift test was con-
ducted to determine the static uplift capacity. In addition, enough two-way
tests were conducted under the same conditions to determine the two-way
repeated load level above which a model shaft would pull out in 100 cycles
of load or less, and below which it would not. This threshold load level is
referred to as the critical level of repeated loading (CLRL) for 100 cycles
and was chosen for convenience to provide a reference number of cycles to
evaluate the effects of repeated loading. Shafts that did not fail within 100
cycles then were tested to failure in static uplift to determine the effects of
repeated loading on shaft capacity.
Facilities for Large-Scale Tests
The large-scale testing facilities consisted of an in-floor testing pit, sand
storage bin, materials handling equipment, dust control and ventilation sys-
tem, load testing equipment, and instrumentation. The same filter sand was
used in both scales of tests.
Soil Placement
The sand was deposited in the 2.1-m diameter, 2.9-m deep testing pit by
raining from a suspended soil hopper, a method similar to the one described
for the small-scale tests, to achieve the same density and stress history as
the small-scale, medium-dense deposits. The average measured relative den-
sities of the six large-scale deposits ranged from 44 to 49%.
Loading and Instrumentation
Data acquisition, shaft construction, and loading procedures for the large-
scale shafts were the same as those described for.the small-scale shafts. Ax-
ial loads were measured at the shaft butt using a load cell, and axial dis-
placements at the shaft butt were measured with both a dial gage and a
475

J. Geotech. Engrg., 1990, 116(3): 470-491

 TABLE 1. Summary of Small-Scale Load Tests
Magnitude of
Test Soil density" D/Bb Loading mode 0 repeated load (N)
(1) (2) (3) (4) (5)
SI M(39) 8 SU NA
Downloaded from ascelibrary.org by University of Hong Kong on 01/08/22. Copyright ASCE. For personal use only; all rights reserved.

S2 M(43) 8 SU NA
S3 M(50) 8 TR 267
S4 M(39) 8 TR 290
S5 M(50) 8 TR 470
S6 L(18) 8 SU NA
$7 L(18) 8 TR 20Q
S8 L(25) 8 TR 330
S9 D(89) 8 SU NA
S10 D(89) 8 TR 735
Sll D(86) 8 TR 1,450
S12 D (86) 8 TR 2,400
S13 M(50) 4 SU NA
S14 M(50) 4 TR 125
S15 M(50) 4 TR 145
S16 D(89) 4 SU NA
S17 D(89) 4 TR 240
S18 D(86) 4 TR 280
S19 L(25) 4 SU NA
S20 L(21) 4 SU NA
S21 L(14) 4 TR 65
S22 L(21) 4 TR 80
S23 L(21) 4 TR 100
S24 L(14) 4 TR 120
*L = loose; M = medium dense; D = dense; and number in parentheses is relative
density, in percent,
b
AU shafts are 76 mm diameter; for D/B = 4, depth = 305 mm; and for D/B = 8,
depth = 610 mm.
C
SU = static uplift; and TR = two-way repeated loading.

DCDT. In-situ soil stresses were measured using the same stress cells de-
scribed previously, placed approximately 20 mm from the soil-shaft interface
at depths of 305, 610, 914, and 1,067 mm.
Test Variables
Table 2 summarizes the test conditions for the six large-scale shafts, in-
cluding the loading mode, magnitude of repeated load, and the number of
loading cycles. All shafts were constructed in medium-dense filter sand, with
D/B equal to 8. Two of the shafts (LI and L2) were tested to determine
the uplift capacity with no repeated loading. Four large-scale shafts were
subjected to two-way repeated loading to determine the critical level of re-
peated loading and uplift capacity changes caused by repeated loading. Shaft
L3 was subjected to 1,000 cycles of repeated loading and then tested in
uplift. Shaft L4 underwent 200 cycles at ±2.00 kN followed by 111 cycles
at ±2.67 kN, at which point it failed in uplift. Shaft L5 underwent 216
cycles at +2.45 kN, at which point failure occurred, and shaft L6 was sub-
jected to a variable load history before being tested to failure in uplift.

476

J. Geotech. Engrg., 1990, 116(3): 470-491

 TABLE 2. Summary of Large-Scale Load Tests
Test" Loading mode" Magnitude of repeated load (kN) Number of cycles
d) (2) (3) (4)
LI (49) su — 1
L2 (49) su — 1
Downloaded from ascelibrary.org by University of Hong Kong on 01/08/22. Copyright ASCE. For personal use only; all rights reserved.

L3 (44) TR 1.43 1,000

L4(46) TR 2.00 200
L4(46) TR 2.67 111
L5 (46) TR 2.45 216
L6 (49) TR 2.00 100
L6 (49) TR 2.45 70
L6 (49) TR 2.23 100
L6 (49) TR 2.45 10
Note: All shafts constructed in medium dense sand, DIB = 8.
"Number in parentheses is relative density in percent.
b
SU = static uplift; and TR = two-way repeated loading.

SUMMARY OF TEST RESULTS

Static Uplift Capacities

Static uplift load tests were conducted for each test condition to provide
a reference state for evaluating the effects of repeated loading. The results
of these tests are given in Table 3. The displacement at maximum uplift load
ranged from 3 to 10 mm. These values are consistent with the values of
displacement required to mobilize the maximum side resistance in full-scale
drilled shafts, reported by Kulhawy et al. (1983) as ranging typically from
5 to 15 mm.
Table 3 includes values of (3 back-calculated from the tests. These values

TABLE 3. Summary of Static Uplift Load Test Results

Test Side resistance Displacement at uplift
number Soil density* D/B Qu-W (N) (3° capacity (mm)
d) (2) (3) (4) (5) (6)
SI M(39) 8 1,451 1.52 7.6
S2 M(43) 8 1,536 1.48 7.7
S6 L(18) 8 507 0.50 8.0
S9 D(89) 8 9,594 8.45 6.6
S13 M(50) 4 253 1.01 3.2
S16 D(89) 4 523 1.84 3.0
S19 L(25) 4 103 0.42 3.3
S20 L(21) 4 113 0.48 3.3
LI M(49) 8 5,380 0.94 6.2
L2 M(49) 8 5,710 0.97 10.0
"L = Loose; M = medium dense; D = dense; and number in parentheses is relative
density, Dr (percent).
b
2» = measured uplift load; and W = weight of foundation.
c
fS = K tan 8 = 2(g„ - Wj/'nByD^ (using actual as-built diameter plus a 7.5-mm soil
collar).

477

J. Geotech. Engrg., 1990, 116(3): 470-491

 2.0 n

1.5
z
Downloaded from ascelibrary.org by University of Hong Kong on 01/08/22. Copyright ASCE. For personal use only; all rights reserved.

o 1.0 -

X
0.5 -

0 5 10 15

(a) Shaft Displacement (mm)

Stress coll depth » 508 mm

5 10 15
0>) Shaft Displacement (mm)

FIG. 4. Load-Displacement and Stress Cell Measurements for Test S2: (a) Load
versus Displacement; and (ft) Stress Cell Measurements

of p are generally higher than would be expected by taking the product of

K and tan 8, where 8 for drilled shafts is taken as the soil friction angle, 4>,
as presented in Fig. 2. The primary cause is an increase in the horizontal
stress as the filter sand undergoes dilation when sheared during uplift load-
ing. For example, Fig. 4 shows the uplift load versus displacement for shaft
S2 and the corresponding horizontal stress measurements for the stress cells
at different depths. The measured horizontal stresses increased with uplift
478

J. Geotech. Engrg., 1990, 116(3): 470-491

 displacement as the sand dilated, increasing the horizontal normal stress at
the soil-shaft interface and therefore the back-calculated (3 value.
At low confining stresses, loose sand may undergo volume increases, as
does dense sand under moderate confining stress. In direct shear tests con-
ducted by Trautmann and O'Rourke (1983), filter sand initially prepared at
densities corresponding to the loose condition of this study exhibited dilative
Downloaded from ascelibrary.org by University of Hong Kong on 01/08/22. Copyright ASCE. For personal use only; all rights reserved.

behavior at normal stresses of 2.5 and 5.0 kN/m 2 . The horizontal stresses
in the loose test deposits ranged from zero at the surface to about 4 kN/m 2
at a depth of 600 mm. Therefore, increases in horizontal stress would be
expected in the model tests in loose sand.
The large value of side resistance measured for shaft S9 indicates that
boundary effects may be significant in the small-scale testing drum for dense
sand. The horizontal stress measured next to the inside wall of the testing
chamber (a distance of 203 mm from the shaft) showed an increase during
loading of 33 kN/m 2 , compard to 42 kN/m 2 measured at the same depth
but 20 mm from the shaft. These values indicate that significant increases
in horizontal stress from soil dilation are transmitted to the wall of the cham-
ber in dense sand. It is difficult to assess the influence of the boundary on
shaft capacity, but future testing of shafts in dense sand should be conducted
in a larger chamber. In the loose and medium dense deposits, horizontal
stress measurements next to the wall of the chamber showed no change dur-
ing loading.
Critical Levels of Repeated Loading (CLRL)
Figs. 5, 6, and 7 illustrate how the critical level of repeated load was
determined for the small-scale shafts tested in medium-dense sand at D/B
= 8. Two shafts (SI and S2) were constructed and tested in static uplift to
obtain an average static uplift capacity of 1,517 N. Two-way repeated load
tests then were conducted on three different shafts constructed identically to
SI and S2, using three magnitudes of repeated loading, for 100 cycles or
until failure occurred. Shaft S3 was loaded at ±267 N and withstood 100
cycles with very little accumulation of displacement. Shaft S3 then was tested
in uplift to failure, giving a slightly higher capacity than SI or S2. The load-
versus-displacement curve during repeated and uplift loading is shown in
Fig. 5, with the uplift curves for SI and S2. The displacements of shaft S3
during repeated loading were too small to be visible at the scale of Fig. 5.
Shaft S4 (Fig. 6) was loaded for 100 cycles at ±290 N, followed by uplift
to failure. S4 underwent maximum two-way cyclic displacements of ap-
proximately 2 mm after 100 cycles, and showed less uplift capacity than SI
or S2. Finally shaft S5 was loaded at ±470 N (Fig. 7). After 50 cycles, S5
was undergoing cyclic displacements of greater than 5 mm and, on the next
application of uplift load, the shaft failed. Tests on shafts S4 and S5 provide
a range for the CLRL for 100 cycles, with medium dense sand and D/B =
8. In this study, the level of repeated loading (LRL) is defined as the ratio
of the magnitude of repeated load minus the shaft weight, (Q - W)tc?eaKi,
to the (average) static uplift capacity minus the shaft weight (Qu — W), ex-
pressed as a percentage

(g
LRL = ~ W)repea'ed x 100% (6)
(G - W)
For the small-scale shafts, a range for the critical level of repeated load
was determined in the same way as previously described for the other soil
479

J. Geotech. Engrg., 1990, 116(3): 470-491

 2.0-r
Downloaded from ascelibrary.org by University of Hong Kong on 01/08/22. Copyright ASCE. For personal use only; all rights reserved.

x
<r

-0.5- 1 -

Shaft Displacement (mm)

FIG. 5. Load-Displacement Curves for Static Uplift (Shafts S1 and S2) and Two-
Way Repeated Loading (Test S3)

2.0

T3
(0
O

X
CE

-0.5

Shaft Displacement (mm)

FIG. 6. Load-Displacement Curves for Static Uplift (Shafts S1 and S2) and Two-
Way Repeated Loading (Test S4)

480

J. Geotech. Engrg., 1990, 116(3): 470-491

Downloaded from ascelibrary.org by University of Hong Kong on 01/08/22. Copyright ASCE. For personal use only; all rights reserved.

Shaft Displacement (mm)

FIG. 7. Load-Displacement Curves for Static Uplift (Shafts S1 and S2) and Two-
Way Repeated Loading (Test S5)

densities and D/B ratios. Table 4 summarizes the test results in which the
CLRL for 100 cycles is presented as a function of soil density and shaft
geometry (D/B). These results indicate that the CLRL decreases for increas-
ing density and increasing depth-to-diameter ratio.
Uplift Capacity Changes Caused by Repeated Loading
As described previously, model shafts that withstood 100 (small-scale tests)
or more (large-scale tests) cycles of two-way repeated loading then were
tested in uplift to determine the effects of repeated loading on uplift capacity
in terms of change in the parameter (3. Changes in |3 are expressed in terms
of a coefficient M p , defined as the ratio of (3 after repeated loading to 3 with
no repeated loading. Values of Mp are presented in Table 5 and indicate that
(3 (and therefore uplift capacity) may increase for shafts constructed in loose
sand, increase or decrease in medium-dense sand, depending on the level of
repeated loading, and decrease generally in dense sand.

TABLE 4. Critical Level of Repeated Load From Small-Scale Model Tests

CLRL (Percent)0
Soil condition D/B = 4 D/B = 8
(D (2) (3)
Loose >65 24-47
Medium dense 27-42 15-26
Dense 38-46 8-14
'Critical level of repeated load = level of repeated load above which failure occurs
within 100 cycles, below which failure does not occur.

481

J. Geotech. Engrg., 1990, 116(3): 470-491

 TABLE 5. Changes in the Parameter p Caused by Repeated Axial Loading
D/B = 4 D/B = 8
Soil condition Test LRL" Ma" Test LRLa Ma"
(1) (2) (3) (4) (5) (6) (7)
Downloaded from ascelibrary.org by University of Hong Kong on 01/08/22. Copyright ASCE. For personal use only; all rights reserved.

Loose S21 22 1.13 S7 24 1.05

Loose S22 29 1.68 S7 24 1.05
Loose S23 45 1.49 S7 24 1.05
Medium dense S14 27 1.18 S3 12 1.05
Medium dense S14 27 1.18 S4 15 0.65
Medium dense S14 27 1.18 L3 16 0.78
Medium dense S14 27 1.18 L6 Variable 0.68
Dense S17 38 1.00 S10 8 0.84
Dense S18 46 0.92 S10 8 0.84
Note: p = K tan 8 = 2(QU - Wj/rrByD1 (using actual as-built diameter, plus a 7.5-
mm soil collar).
"LRL = level of repeated loading defined by Eq. 4.
b
MQ = ratio of p after repeated loading to P determined for static uplift.

FAILURE MECHANISMS UNDER REPEATED LOADS

The effect of two-way repeated axial loading on the uplift capacity of the
model shafts falls into one of three categories: (1) no change; (2) increase;
or (3) decrease. Failure in less than 100 cycles is a case of capacity decrease.
The test results and general principles of soil behavior under repeated loading
are discussed herein to show that three mechanisms caused the observed
shaft capacity changes. These mechanisms are cyclic flow caused by sand
caving into the void beneath the tip during uplift, cyclic degradation of the
soil shearing resistance along the shaft-soil interface, and cyclic decreases
in the horizontal stress coefficient, K. All three mechanisms are dependent
on shaft displacements, and their combined effects are the cause of the ob-
served decreases in p.
Cyclic Flow
Stress cell measurements taken during static uplift and two-way repeated
loading suggest one of the mechanisms by which capacity decreases and
failures are triggered. As illustrated in Fig. 4 for test S2 (static uplift in
medium dense sand), most of the stress cells show an increase in the mea-
sured horizontal normal stress as the shaft is loaded, the soil is sheared, and
dilation of the soil is restrained. Measured initial stresses and stress increases
are greater with increasing stress cell depth. However, the horizontal stresses
close to the tip (bottom) of the model shafts exhibit a noticeable drop-off at
uplift displacements of 1 to 2 mm, which is smaller than the displacements
for shaft uplift failure, which lie in the range of 3 to 10 mm. This stress
relaxation near the tip may be caused initially by grains of sand "caving"
into the void created under the tip by uplift, leading to a decrease in the
radial effective stress around the tip. With further shaft displacement, the
zone of yield appears to progress up the shaft. This process also is illustrated
in Fig. 4(b), which shows that the stress cell closest to the bottom, at 584
mm, dropped off first, showing initial yield, followed by the stress cells at
depths of 508 mm and 381 mm, at which point the maximum side resistance
was mobilized and failure in uplift occurred. The stress cell closest to the

482

J. Geotech. Engrg., 1990, 116(3): 470-491

Downloaded from ascelibrary.org by University of Hong Kong on 01/08/22. Copyright ASCE. For personal use only; all rights reserved.

Shaft Uplift Displacement (mm)

FIG. 8. Peak Horizontal Stress versus Displacement for Test S11

surface (depth = 1 2 7 mm) shows little change because the low confining
stress does not restrict dilation.
Yielding because of soil caving has been observed in static uplift and re-
peated load tests on model anchors. Kalajian (1971) observed that sand grains
around the edge of a fluke anchor will cave into the cavity formed below
the anchor during uplift. During repeated loading, the soil grains that move
under the fluke prevent the anchor from returning to its initial position during
unloading, resulting in a net uplift displacement after each cycle. Bemben
et al. (1973) demonstrated that this phenomenon is the primary mechanism
that leads to failure of fluke anchors under repeated loading. For the model
shafts of this study, stress cell readings indicate that a similar process occurs
near the tips of the shafts. Decreases in the measured horizontal stresses near
the tip (as shown, for example, by the stress cell at a depth of 584 mm in
Fig. 8) are caused by decreases in confining stress as sand particles cave
into the void created by uplift. During repeated loading, if the uplift dis-
placement becomes large enough for soil caving to occur, which appears to
be in the range of 1 to 2 mm, then the uplift displacements become pro-
gressively larger with an increasing number of cycles, because the shaft is
prevented from returning to its initial position during unloading, as illus-
trated for shafts S4 and S5 in Figs. 6 and 7.

Cyclic Degradation of Side Resistance

The second mechanism is the degradation of shearing resistance along the
soil-shaft interface caused by two-way repeated loading. The magnitude of
shearing strain in repeated loading is a primary factor in determining the
response factors of granular soils, including soil strength (Leelanitkul and
Timmerman 1982), volume changes (Silver and Seed 1971a; Youd 1972),
and stiffness changes (Silver and Seed 1971b; Hardin and Drnevich 1972).
Once the mechanism of cyclic flow is triggered, the two-way cyclic dis-
placements become large enough to cause significant shearing strains in the
soil next to the shaft. The magnitude of strain is large enough to cause major
reorientation of sand grains and a decrease in the soil frictional resistance.
Without being able to measure changes in the interface friction angle di-
rectly, a reasonable lower bound approximation- can be made by assuming
that repeated loading reduces the soil (and interface) friction angle to its
minimum, or critical value, ()>„. According to Eq. 5, and as shown in Fig.

483

J. Geotech. Engrg., 1990, 116(3): 470-491

 TABLE 6. Expected Lower Bound Changes in interface Friction Angle Caused
by Repeated Loading
Range for Peak Friction
Angle (degrees)3 Range for Msb
Soil condition D/B = 4 D/B = 8 D/B = 4 D/B = 8
Downloaded from ascelibrary.org by University of Hong Kong on 01/08/22. Copyright ASCE. For personal use only; all rights reserved.

0) (2) (3) (4) (5)

Loose 33.6 to 34.9 34.1 to 35.4 0.93 to 0.97 0.91 to 0.96
Medium dense 40.7 37.9 to 40.0 0.75 0.77 to 0.83
Dense 43.4 to 43.9 41.9 to 43.2 0.67 to 0.68 0.69 to 0.72
"Peak friction angles computed by Eq. 5, assuming 4>„ = 32.9°.
b
M5 = tan 8„/tan hp.

2, the average value of $cv (and therefore 8C„) for the filter sand is 32.9°.
Decreases in (3 caused by changes in 8 to the minimum value can be ex-
pressed quantitatively by computing the ratio of tan 8„ to tan 8P. This ratio
is defined as the friction angle change factor, Ms. Table 6 lists the expected
values of Ms for each test condition listed in Table 5. The other component
of capacity change is assumed to be caused by average changes in the coef-
ficient of horizontal soil stress, as described next.
Cyclic Changes in K
Changes in the horizontal stress are illustrated by the response of the stress
cells measuring horizontal stresses in the soil-shaft interface. During re-
peated loading, each stress cell measures an increase in horizontal stress
during the uplift and compression portions of loading, just as for static uplift.
However, the amount of increase lessens with each cycle, indicating a cyclic
decrease of horizontal stress and, therefore, a decrease in side resistance
according to Eq. 3(a). The response of stress cell 18 versus shaft displace-
ment during test S l l is shown in Fig. 8 and illustrates this behavior. This
stress cell was located at a depth of 381 mm, so the process of caving prob-
ably was not causing the drop-off at small displacements. Shaft S l l failed
in 51 cycles, because of a yield zone moving up the shaft, decreasing side
resistance caused by degradation of the friction angle and decreasing hori-
zontal stress caused by cyclic degradation of the soil's ability to dilate. Ad-
ditional stress cell readings showing similar behavior were observed in tests
S5, S8, S l l , S12, S15, S18, and S24.
The change in side resistance caused by horizontal stress changes can be
evaluated quantitatively by computing the ratio of K calculated from tests
on shafts that failed under repeated loading to K from static tests with no
previous repeated loading. K is back-calculated from

K= (7)
tan 8
The ratio of Km/Kp then is computed from
Kcv Mp
(8)
K, = MK = M
—8
in which Mp = side resistance change factor for shafts that failed under
484

J. Geotech. Engrg., 1990, 116(3): 470-491

 TABLE 7. Changes in Horizontal Stress Coefficient Caused by Repeated Load-
ing

Test3 M„ M8 MK from Eq. 8 MK from cr4cycI1c/tT*,lM1c

(1) (2) (3) (4) (5)
S5 (50) 0.26 0.77 0.34 0.56
Downloaded from ascelibrary.org by University of Hong Kong on 01/08/22. Copyright ASCE. For personal use only; all rights reserved.

S8 (25) 0.47 0.91 0.52 0.66

S l l (86) 0.14 0.72 0.19 0.34
S12 (86) 0.25 0.72 0.35 0.38
S15 (50) 0.42 0.75 0.56 0.65
S24 (14) 0.63 0.97 0.65 0.81
L4(46) 0.38 0.81 0.47 0.51
b
L5 (46) 0.35 0.81 0.43
"Number in parenthesis is relative density, in percent.
b
No horizontal stress measurements for L5.

repeated loading; Mb = friction angle change factor; and MK is defined as

the horizontal stress change factor. This analysis is valid only for shafts that
reached the residual state (i.e., that failed under repeated loading), because
M8 is computed by assuming that the friction angle is reduced to its minimum
value of §m. Lower bound values of MK calculated according to Eq. 8 are
presented in Table 7. As a comparison, MK can be computed as the ratio of
peak horizontal stress measured at failure in repeated load tests, o"ftcyCHC, to
peak horizontal stresses measured in tests with no previous repeated loading,
octane- Values of MK computed from the stress cell measurements also are
presented in Table 7. The values of MK evaluated from stress cell measure-
ments are consistently higher than the values computed by Eq. 8. Both sets
of data exhibit the same trend of decreasing MK (more severe degradation
of horizontal stress) with increasing soil density. In the absence of direct
measurement of vertical and horizontal stresses at the soil/foundation inter-
face, these data provide a preliminary assessment of lower bound changes
in the parameter K caused by repeated loading.

Relationship between Capacity Changes and Shaft Displacements

For all shafts that withstood 100 or more cycles of repeated two-way load-
ing and showed either an increase or no change in uplift capacity, the uplift
displacements caused by repeated loading were less than 1 mm, and the
stress cell responses during the subsequent uplift test were similar to the
static uplift test without repeated loading (shaft S6, Fig. 7). Tests showing
this type of behavior were S3, S7, S10, S14, S17, and S21. Shafts that
withstood repeated loading, but underwent capacity decreases when tested
subsequently in uplift, all underwent uplift displacements greater than 1 to
2 mm, allowing a yield zone to form near the tip and for some degradation
of side resistance to occur. These test results, and research indicating that
shearing strain is a fundamental parameter influencing soil strength under
repeated loading (Leelanitkul and Timmerman 1982), suggest that capacity
reductions and failure may be related closely to the amount of displacement
caused by repeated loading.
Fig. 9 shows the coefficient Mp, which describes changes in (3, versus the
cyclic (peak to peak) displacement on the last load cycle before uplift to
failure. For shafts that withstood repeated loading without failing, the dis-
485

J. Geotech. Engrg., 1990, 116(3): 470-491

 u . . . | i i i i
-0 ...,....,...-
Small-Scale
D/B=8 D/B=4
9 Loose 0 Loose
D A Medium A Medium
m- Dense Dense
- a m Dense Q Dense
Downloaded from ascelibrary.org by University of Hong Kong on 01/08/22. Copyright ASCE. For personal use only; all rights reserved.

• \ o o
* \ •

• \ A •
»~ Pulloul in
A 100 cycles _

X
4> Lorge-Scale
. . i . , , ,
1 2 3 4
Cyclic Displacement Before Failure (mm)

FIG. 9. Capacity Change Coefficient versus Cyclic Displacement

placement is from the last cycle. For shafts loaded above the CLRL, the
displacements are from the last full load cycle before failure, and M? cor-
responds approximately to the level of repeated loading. As shown in Fig.
9, there is an apparent direct relationship between the cyclic displacement
and the change in (3 for all densities, stress histories, geometries, and model
scales.
Based on these observations, the behavior of model drilled shafts in sand
under two-way repeated axial loading can now be described. If the magni-
tude of repeated loading is high enough, uplift displacements will exceed
some threshold value, and soil caving will cause a softened zone near the
tip of the shaft. If loading continues, cyclic flow will occur, accompanied
by further yielding. The threshold displacement which will initiate caving
and cyclic flow appears to be in the range of 0.5 to 1.0 mm, which happens
to be close to the mean particle size of the sand used for this study. If uplift
displacements were less than about 0.5 mm, capacity increases were ob-
served in loose to medium-dense sand. If uplift displacements exceeded 0.5
mm, capacity reductions occurred, apparently because of a decrease in the
horizontal stress next to the tip. With increasing cyclic displacements, fur-
ther capacity reductions occurred, caused by: (1) Enlargement of the soft-
ened zone near the tip; (2) degradation of side resistance under two-way
repeated shearing; and (3) cyclic degradation of K. When the cyclic dis-
placement reaches 3-4 mm, major reorientation of grains occurs, accom-
panied by soil failure, and the shaft fails under repeated loading. This dis-
placement corresponds well to the maximum grain size of the sand used in
this study. Further research is necessary to determine the significance of soil
grain size.
IMPLICATIONS FOR ANALYSIS AND DESIGN

Foundation suitability ultimately is evaluated in terms of the factor of safety

with respect to the ultimate foundation capacity and limiting movements of
the soil-foundation-structure system. This study demonstrates that the portion
of ultimate drilled shaft capacity contributed by side resistance can be re-
duced by repeated loading. Factors of safety applied to the side resistance
must, therefore, account for these reductions.

486

J. Geotech. Engrg., 1990, 116(3): 470-491

 Design Based on CLRL
The critical level of repeated loading (CLRL) has been defined as the level
of repeated two-way axial loading, expressed as a percentage of ultimate
static side resistance, above which uplift failure will occur within 100 cycles
and below which failure will not occur. A rational approach to designing
drilled shafts for repeated loading is to keep the design level of repeated
Downloaded from ascelibrary.org by University of Hong Kong on 01/08/22. Copyright ASCE. For personal use only; all rights reserved.

loading below the critical level of repeated loading. This can be achieved by
designing a shaft so the ultimate static side resistance is greater than the
design magnitude of repeated load divided by the CLRL (in which the CLRL
is expressed as a decimal). As a guideline, the values of CLRL presented
in Table 4 are recommended for drilled shafts in cohesionless soils under
two-way repeated axial loading, for the ranges ofD/B and soil density used
in this study. Further research is required to establish values of the CLRL
for wider ranges of soil conditions, shaft geometry, number of loading cycles,
and methods of construction.
Many tower-type structures supported by drilled shafts are designed for
sustained dead loads and transient live loads (repeated loads). Typical factors
of safety applied to the drained side resistance for sustained loads range from
2 to 3. Higher or lower values may be warranted, depending on the vari-
ability of site conditions and the degree to which soil properties are known.
For low levels of repeated loading, the design side resistance based on ap-
plying a factor of safety of 2 to 3 to the side resistance required to support
the sustained dead loads is likely to be greater than that which is based on
repeated loading. The more stringent criterion from sustained dead loads or
repeated loading will control the foundation design.
Residual Uplift Capacity
As a lower bound on uplift capacity, it is useful to know the residual side
resistance in uplift of a drilled shaft after the applied load has exceeded the
available uplift resistance, either because of repeated loading or a single
extreme loading event. The results of this study and uplift model tests re-
ported by Stewart and Kulhawy (1981) and Kulhawy et al. (1979) can be
used to determine the residual uplift capacities for such cases. Fig. 10 shows
the ratio of residual side resistance to peak side resistance, expressed in terms
of the side resistance change coefficient Mp, versus initial sand density. As
shown, this ratio decreases with increasing density and, for a given density,
is greatest for monotonic loading and least for two-day repeated loading. For
comparison, the friction angle change coefficient, M8, also is shown in Fig.
10 versus initial sand density.
In the monotonic uplift tests, the residual capacity was determined by
complete unloading as soon as the shaft underwent initial yield (as soon as
the axial load began to decrease with increasing displacement) and then re-
loading the shaft to failure. This represents a single extreme loading event.
In the one-way repeated load tests, the magnitude of repeated load was in-
creased with each cycle until the shaft failed in uplift. All of the shafts were
failed in 5-7 cycles. Each shaft then was unloaded completely and reloaded
one time to failure to determine the residual uplift capacity. In the current
study, shafts that failed under two-way repeated loading reached their re-
sidual capacities after they underwent relatively large uplift displacements to
failure. These are the shafts shown in Fig. 9 that underwent failure within
100 loading cycles, after undergoing cyclic displacements of 3-4 mm.
The decrease in residual side resistance for repeated loading compared to
487

J. Geotech. Engrg., 1990, 116(3): 470-491

 A Friction angle reduction factor, Mg (computed)
® Monotonic (Stewart and Kulhawy, 1981)
I—I One-way repeated uplift (Kulhawy, et al., 1979)
X Two-way repeated, this study

1.0 - - •—
/////
Downloaded from ascelibrary.org by University of Hong Kong on 01/08/22. Copyright ASCE. For personal use only; all rights reserved.

Friction angle/
duction 7

0.8

0.6

M/3

0.4

0.2

0 _l I I L
0 20 40 60 80 100

Initial Sand Relative Density, %

FIG. 10. Reductions in Drilled Shaft Side Resistance versus Relative Density

monotonic loading, and for two-way repeated loading compared to one-way

repeated loading, can be explained by the behavior of soil under static and
repeated loading. The soil next to a shaft that fails under monotonic loading
does not undergo as much reorientation of grains as soil that is subjected to
repeated applications of load. If the shaft is unloaded before large displace-
ments occur, the geometric relationships between the individual grains are
not characteristic of the residual state.
Under repeated loading, the magnitude of shearing strain becomes large
and with each additional loading cycle there is further reorientation of sand
grains, bringing the soil shearing resistance closer to its residual (minimum)
value. Loading that causes two-way repeated shearing strains apparently is
a more efficient mechanism of bringing a soil to its residual state than one-
way loading, based on the lower residual capacities observed in this study.
Fig. 9 also illustrates that as the magnitude of cyclic displacement increases,
shaft side resistance decreases, again because of degradation of shearing re-
sistance as the soil grains undergo major reorientations. The larger the two-
way cyclic displacements, the larger are the grain reorientations and capacity
decreases.
Fig. 10 can be used as a guide for designing shafts with sufficient residual
capacity to sustain extreme loading events or repeated loading. The design
drained side resistance should be at least equal to the design uplift load di-
vided by the ratio of residual to peak side resistance, as given in Fig. 10.
Again, the more stringent criterion from sustained dead loads, two-way re-
488

J. Geotech. Engrg., 1990, 116(3): 470-491

 peated loading (based on CLRL), or residual uplift capacity will control the
foundation design.
CONCLUSIONS
A laboratory investigation has been described of the effects of repeated
axial loading on the uplift capacity of model drilled shafts under fully drained
Downloaded from ascelibrary.org by University of Hong Kong on 01/08/22. Copyright ASCE. For personal use only; all rights reserved.

conditions. The main results and conclusions are the following.

Repeated axial loading of drilled shaft foundations can change the uplift
capacity, and these changes depend primarily on the maximum cyclic dis-
placement. For the shafts in this study, a relationship has been established
between the magnitude of cyclic displacement and changes in side resis-
tance, expressed in terms of changes in the parameter (3. This relationship,
presented in Fig. 9, appears to be independent of soil density and stress
history, shaft geometry, and model scale.
A critical level of two-way repeated loading (CLRL) exists above which
a shaft will fail in uplift and below which it can withstand large numbers
of load applications with limited displacements and little or no change in
capacity. The CLRL was found to be a minimum for shafts in dense sand
having a depth to diameter (D/B) ratio of eight, and a maximum for shafts
in loose sand with D/B of four. In general, the tests showed increasing
CLRL for decreasing soil density and decreasing depth-to-diameter ratio.
The mechanisms that lead to capacity reductions and failure under two-
way repeated loading have been identified based on laboratory load tests on
instrumented model shafts. Soil caving beneath the shaft tip during uplift
forms a zone of reduced horizontal stress close to the tip and initiates cyclic
flow, a process of increasing cyclic displacements under repeated loading.
Continued loading subjects the soil at the shaft interface to large two-way
shearing strains, reducing both the shearing resistance and the horizontal
stress in the soil and decreasing the side resistance of the shaft. If the cyclic
displacements become greater than about 3 mm, the shaft will fail in uplift
under repeated loading. This displacement limit may correspond to the max-
imum soil particle size.
The design of drilled shafts in sand under two-way repeated axial loading
can be based on keeping the design level of repeated loading below the
CLRL. Table 4 is recommended for choosing values of the CLRL for the
ranges of soil density, shaft D/B ratios, and loading conditions evaluated in
this study. Further research is required to establish values of the CLRL for
other foundation and construction conditions. In addition to foundation ca-
pacity, the design of drilled shafts in sand under two-way repeated loading
apparently should be based on limiting cyclic displacements that will occur
for the design level of repeated loading. Although further data are needed
to define these limits precisely, an empirical method for predicting the cyclic
displacements has been suggested by Turner and Kulhawy (1987b).
For estimating the lower bound of uplift capacity for drilled shafts in cohe-
sionless soils, Fig. 10 presents a relationship between residual uplift capacity
and initial soil density for shafts under single loading events, one-way re-
peated uplift loading, and two-way repeated loading.
ACKNOWLEDGMENTS
This research was supported by the Electric Power Research Institute (EPRI)
under Project RP1493-4, for which V. J. Longo was the EPRI Project Man-
ager. K. J. McManus and C. N. Nicolaides assisted with laboratory testing.
489

J. Geotech. Engrg., 1990, 116(3): 470-491

 A. Avcisoy drafted the figures, and C. Froman and K. Menefee typed the
text. The writers thank these individuals for their assistance.

APPENDIX I. REFERENCES
Downloaded from ascelibrary.org by University of Hong Kong on 01/08/22. Copyright ASCE. For personal use only; all rights reserved.

Bemben, S. M., Kalajian, E. H., and Kupferman, M. (1973). "The vertical holding
capacity of marine anchors in sand and clay subjected to static and cyclic loading."
Proc, 5th Offshore Tech. Conf., 2, Houston, Tex., 871-880.
Bolton, M. D. (1986). "The strength and dilatancy of sands." Geotechnique, 36(1),
65-78.
Hardin, B. O., and Drnevich, V. P. (1972). "Shear modulus and damping in soils:
Measurement and parameter effect." J. Soil Mech. Found. Div., ASCE, 98(6),
603-624.
Kalajian, E. H. (1971). "The vertical holding capacity of marine anchors in sand
subjected to static and cyclic loading." Thesis presented to the University of Mas-
sachusetts, at Amherst, Mass., in partial fulfillment of the requirements for the
degree of Doctor of Philosophy.
Kulhawy, F. H., Kozera, D. W., and Withiam, J. L. (1979). "Uplift testing of model
drilled shafts in sand." J. Geotech. Engrg., ASCE, 105(1), 31-47.
Kulhawy, F. H., et al. (1983). "Transmission line structure foundations for uplift-
compression loading." Report EL-2870, Electric Power Res. Inst., Palo Alto, Calif.
Leelanitkul, S., and Timmerman, D. H. (1982). "Deformation of sand under cyclic
loading." Proc, Conf. on Soil Dynamics and Earthquake Engrg., Southampton,
U.K., 87-99.
Mohan, D., Mohan, G. S., and Jain, M. P. (1967). "A new approach to load tests."
Geotechnique, 17(3), 274-283.
O'Neill, M. W., and Reese, L. C. (1970). "Behavior of axially loaded drilled shafts
in Beaumont clay," Res. Report 89-8, Center for Highway Res., Univ. of Texas,
Austin, Tex.
Reese, L. C , and Wright, S. J. (1977). 'Drilled shaft design and construction guide-
lines manual." Implementation Package 77-21, U.S. Dept. of Transp., Washing-
ton, D.C.
Silver, M. L., and Seed, H. B. (1971a). "Deformation characteristics of sands under
cyclic loading." J. Soil Mech. Found. Div., ASCE, 97(8), 1081-1098.
Silver, M. L., and Seed, H. B. (1971b). "Volume changes in sands during cyclic
loading." J. Soil Mech. and Found. Div., ASCE, 97(9), 1171-1182.
Stas, C. V., and Kulhawy, F. H. (1984). "Critical evaluation of design methods for
foundations under axial uplift and compression loading." Report EL-3771, Electric
Power Res. Inst., Palo Alto, Calif.
Stewart, J. P., and Kulhawy, F. H. (1981). "Experimental investigation of the uplift
capacity of drilled shaft foundations in cohesionless soil." Contract Report B-49(6),
Niagara Mohawk Power Corp., Syracuse, N.Y.; also, Geotech. Engrg. Report 81-
2, Cornell Univ., Ithaca, N.Y.
Trautmann, C. H., Kulhawy, F. H., and O'Rourke, T. D. (1985). "Sand density
for laboratory studies." Geotech. Testing J., 8(Dec), 159-165.
Trautmann, C. H., and O'Rourke, T. D. (1983). "Behavior of pipe in dry sand under
lateral and uplift loading." Geotech. Engrg. Report 83-7, Cornell Univ., Ithaca,
N.Y.
Trautmann, C. H., O'Rourke, T. D., and Kulhawy, F. H. (1985). "Uplift force-
displacement response of buried pipe." J. Geotech. Engrg., ASCE, 111(9), 1061-
1076.
Turner, J. P., and Kulhawy, F. H. (1987a). "Experimental analysis of drilled shaft
foundations subjected to repeated axial loads under drained conditions." Report
EL-5325, Electric Power Res. Inst., Palo Alto, Calif.
Turner, J. P., and Kulhawy, F. H. (1987b). "Prediction of drilled shaft displacements
under repeated axial loads." Proc, Int. Symp. on Prediction and Performance in
Geotech. Engrg., Calgary, Alberta, Canada, 105-112.
Turner, J. P., Kulhawy, F. H., and Charlie, W. A. (1987). "Review of load tests

490

J. Geotech. Engrg., 1990, 116(3): 470-491

 on deep foundations under repeated loading." Report EL-5375, Electric Power Res.
Inst., Palo Alto, Calif.
Weiler, W. A., Jr., and Kulhawy, F. H. (1978). "Behavior of stress cells in soil."
Contract Report B-49(4), Niagara Mohawk Power Corp., Syracuse, N.Y.; also
Geotech. Engrg. Report 78-2, Cornell Univ., Ithaca, N.Y.
Youd, T. L. (1972). "Compaction of sands by repeated shear straining." J. Soil
Downloaded from ascelibrary.org by University of Hong Kong on 01/08/22. Copyright ASCE. For personal use only; all rights reserved.

Mech. and Found. Div., 98(7), 709-725.

APPENDIX II. NOTATION

The following symbols are used in this paper:

A = foundation cross-sectional area;

B = foundation diameter;
CLRL = critical level of repeated load;
D = foundation depth;
Dr = soil relative density;
K = coefficient of horizontal soil stress;
K0 = in-situ coefficient of horizontal soil stress;
LRL = level of repeated loading;
MK = horizontal stress change coefficient;
Mp = side resistance change coefficient;
Ms = interface friction angle change coefficient;
N = number of loading cycles;
Qsu = side resistance in uplift;
Q,u = tip resistance in uplift;
Qu = uplift capacity;
SU = static uplift loading;
TR = two-way repeated loading;
W = weight of foundation and enclosed soil;
z = depth;
7 = effective unit weight;
8 = angle of friction for soil-foundation interface;
6\ = horizontal effective stress;
ov = vertical effective stress;
T = shearing resistance;
$ = effective stress angle of friction;
<f>P = peak value of friction angle; and
4>cv = critical value of friction angle.

491

J. Geotech. Engrg., 1990, 116(3): 470-491