You are on page 1of 24

Results from long-term measurement in piles of

drag load and downdrag

Bengt H. Fellenius, Dr.Tech., P.Eng.
Bengt Fellenius Consultants Inc.,
Calgary, Alberta, T2G 4J3

Canadian Geotechnical Journal, April 2006

Vol. 43, No. 4, pp. 409 - 430.

The Bäckebol site in June 1968


Results from long-term measurement in piles of

drag load and downdrag
Bengt H. Fellenius, Dr.Tech., P.Eng.

SUMMARY Several full-scale, long-term tests on instrumented piles performed since

the 1960s through the 1990s are presented. The results of the tests show that large
drag load will develop in piles installed in soft and loose soils. The test cases are from
Norway, Sweden, Japan, Canada, Australia, Hawaii (USA), and Singapore and
involve driven steel pile piles and precast concrete piles. The test results show that
the transfer of load from the soil to the pile through negative skin friction, as well as
from the pile back to the soil through positive shaft resistance, is governed by
effective stress and that already very small movement will result in mobilization of
ultimate values of shaft shear. The pile toe resistance, on the other hand, is
determined by downdrag of the pile and the resulting pile toe penetration.
Reconsolidation after the pile installation with associated dissipation of pore pressures
will result in significant drag load. An equilibrium of force in the pile will develop,
where the sustained loads on the pile head and the drag load are equal to the positive
shaft resistance plus the pile toe resistance. The location of the force equilibrium —
the neutral plane — is also where the pile and the soil move equally. The drag load is
of importance mostly for very long piles, piles longer than 100 pile diameters, for
which the pile structural strength could be exceed. However, downdrag, i.e.,
settlement of the piled foundation, is a very important issue and particularly for low-
capacity short piles. Soil settlement at the neutral plane will result in a downdrag of
the pile. The magnitude of the downdrag will determine the magnitude of the pile toe
penetration into the soil, which will determine the pile toe resistance and affect the
location of the neutral plane. Nature's iteration of force and soil settlement will decide
the final location of the neutral plane.

The text and figures in the following are reformatted from the original journal presentation, but no other
changes have been made.

Received June 2, 2005. Accepted January 10, 2006.

Published on the NRC Research press web site [] on March 22, 2006

Bengt H. Fellenius, Bengt Fellenius Consultants Inc., 1905 Alexander Street SE, Calgary, Alberta, T2G 4J3.

The 1960s saw an upsurge of interest in performing full-scale 0 500 1,000 1,500 2,000
instrumented piles to study the magnitude and development of 0
negative skin friction due to soil settling around the piles. Water
Surfac e FILL
Tests were performed in Norway, Sweden, Japan, and Canada.
The author has revisited the pioneering papers publishing the
results and presents here a summary of a reanalysis of the 10
records. One of these tests was started in June 1968 by the
Swedish Geotechnical Institute and consisted of recording load
distribution in two instrumented, 300 mm diameter, single, 20
precast concrete piles driven through a 40 m thick clay deposit Marine
and 15 m into an underlying sand deposit. The results obtained

during the first 160 days and 1,300 days were published by
Fellenius and Broms (1969) and Fellenius (1971; 1972), 30
respectively. The study continued beyond 1972 with the
placing in September 1973 of an earth fill over a 41 m diameter
circular area around the piles followed by frequent monitoring 40
of forces in the piles and settlement until the site was closed-up
in 1983. The measurements taken until mid-1975 were written
up by the Swedish Geotechnical Institute in Swedish (Bjerin
1977). This paper reports the results until the end of the field 50
test in August 1983, 5,500 days after the start. Bedrock
'May 1963

April 1964

NGI Sörenga Site

In April 1962, the Norwegian Geotechnical Institute started a Fig. 1 Sörenga site. Site profile and measured distribu-
full-scale test on a telltale instrumented steel pile at Sörenga tion of settlement. Data from Bjerrum
near Oslo (Bjerrum and Johannessen 1965). The test consisted and Johannessen (1965).
of measurements on a 55 m long, 470 mm diameter pipe pile
driven in an area to be reclaimed from the sea by means of
a 10 thick fill. The natural soil at the site consisted of a 43 m PILE SHORTENING (mm)
thick deposit of marine clay on bedrock, a calcareous schist. 0 5 10 15 20
The natural water content of the clay was about 40 % at the sea 0
bed reducing to about 30 % at the bedrock. The liquid limit
was about 10 % larger than the water content. The undrained
shear strength ranged from about 20 KPa at the sea bed
increasing to about 70 KPa at the bedrock. 10

The placing of the fill initiated a consolidation process in the

clay deposit. Fig. 1 presents the measured distribution of 20
settlement, and Fig. 2 the distribution of excess pore pressures
measured in April 1964, two years after the pile driving and

start of placing the fill. Fig. 2 also shows the calculated

effective stress after full dissipation of the excess pore 30
pressures. The difference, as indicated, is the effective stress
existing in April 1964. The settlement resulted in negative skin
friction along the test pile, accumulating to drag load in the 40
pile. The drag load caused the pile to shorten, and Fig. 3 shows
the pile shortenings measured one year (May 1963) and two
years (April 1964) after the pile driving. Bjerrum and
Johannessen used the pile shortenings to calculate the 50
distribution of stress in the pile. The stress data have been May 1963
converted to load in the pile and are presented in Fig. 4. (The
'April 1964
load values are plotted, not at mid-point of each measuring 60
length, L, but at 0.58L down from the upper end of the
measuring length, L, as recommended by Fellenius (2002) for
plotting of loads evaluated from telltale measurements).
Fig. 3 Sörenga site. Measured pile shortening Data from
Bjerrum and Johannessen (1965).

As indicated in Fig. 4, a calculated drag load distribution for a

constant beta-coefficient of 0.25 matches reasonably the
0 200 400 600 measured drag load distribution of April 1964. Also included
0 is the drag load distribution after full dissipation of pore
pressures calculated using same beta-coefficient. A calculation
σ'z after full using the undrained shear strength resulted in a calculated drag
dissipation of load distribution very similar to that calculated for a beta-
excess pore coefficient of 0.25. However, in contrast to the effective stress
pressure calculations, the calculations using the undrained shear strength
cannot be used to reproduce the measured changes in drag load
20 for the actual distributions of excess pore pressure.

Bjerrum et al. (1969) reported that in 1967, after an additional

three years of measurements, the maximum load in the pile and
the settlement of the clay had increased to 4,000 KN and
2,000 mm, respectively, and that the movement of the pile toe
Δu into the calcareous schist, i.e., the downdrag, amounted to
40 100 mm. The observations mean that the neutral plane was
σ'z located close to the pile toe and the maximum load in the pile is
also the toe resistance of the pile in the calcareous schist. As
50 the maximum load of 4,000 KN also corresponds to a stress of
April about 250 MPa, the structural strength of the steel, it is possible
1964 that instead of the pile toe being forced into the bedrock, the
mentioned 100-mm toe movement is due to that the steel pile
yielded structurally at the pile toe.

NGI Heröya Site

Fig. 3 Sörenga site. Distribution of excess pore pressure
In 1969, the Norwegian Geotechnical Institute published results
in April 1964 and calculated effective stress.
of additional full-scale investigations (Bjerrum et al. 1969) now
Data from Bjerrum and Johannessen (1965)
performed at three sites in 1962 through 1963. One of the
additional test sites was located at Heröya in the Oslo Harbor,
LOAD (KN) where, to construct a dock, land was reclaimed from the sea by
0 1000 2000 3000 4000 placing a 6.5 m fill over the sea bed. The soil consists of an
0 approximately 25 m thick layer of marine clay, having a water
'May 1963 content of about 30 % very close to the liquid limit —
somewhat different to the clay at the Sörenga site. A gravel
'April 1964 layer deposited on bedrock existed below the clay at depths
10 ranging from 30 m to 35 m. Before the placing of the fill, the
ß=0.25 and σ' at undrained shear strength ranged from about 20 KPa at the top
April 1964 of the clay layer to about 40 KPa immediately above the gravel
20 ß=0.25 and σ' at layer. After two years of consolidation, the undrained shear
Full Dissipation strength had increased to about 40 KPa at the sea bed, but no
change in undrained shear strength was found to have occurred

From Undrained
Shear Strength. at the lower boundary of the clay layer.
In 1966, four 300 mm diameter, telltale-instrumented pipe piles
were driven at the Heröya site. The piles were driven closed-
40 toe with an enlarged size, 400 mm rock shoe, "Oslo-point".
The results from of two of the test piles are quoted in the
following. The surface of one of the two test piles was treated
with a 1 mm thick bitumen coat and the other was untreated.
50 The settlement of the sea bed amounted to 160 mm during the
first year after the driving of the test piles. Similar to that
shown in Fig. 2, Fig. 5 shows the measured excess pore
60 pressures remaining in May 1967 and the calculated effective
stress after full dissipation of the pore pressures. The
difference is the effective stress existing in May 1967. Fig. 6
Fig. 4 Sörenga site. Measured and calculated distribu-
shows the distribution of shortening measured in May 1967 for
tion of load. Data from Bjerrum and Johannessen
the bitumen-coated and the uncoated piles. That the bitumen is

0 100 200 300 400 0 5 10 15
0 0


5 5

σ'z after full

dissipation of
10 10
excess pore
coated uncoated


20 20
Δu σ'z

25 25

30 30
Start of Gravel
35 35

Fig. 5 Heröya site. Distribution of excess pore pressure Fig. 6 Heröya site. Distribution of pile shortening in
measured in May 1967 and calculated effective May 1967. Data from Bjerrum et al. (1969).
stress. Data from Bjerrum et al. (1969).

effective in reducing the shaft shear is due to its viscous The pile shortenings measured in May 1967 were converted
properties and inability to sustain shear forces and resist load in the pile, and the load distributions for the two test piles
movement. A small strain shear rate will mobilize small shear are shown in Fig. 8 (plotted at the 0.58L-point of each
resistance in the bitumen coat, sandwiched between the pile measuring length, L, indicated by the stepped curve). The load
and the soil, that is smaller than the soil shear resistance distribution of the uncoated pile shows how the negative skin
(Fellenius 1975; 1979). friction accumulates to a maximum drag load value of
about 1,100 KN at about 25 m depth. Below this depth, the
Figure 7 shows pile shortenings measured during a 400 day shaft shear turned to positive shaft resistance and the load in
period after the pile driving. The results demonstrate that the the pile reduced with depth to a mobilized toe resistance of
thin bitumen coat was highly effective in reducing — about 550 KN. The negative and positive direction loads are in
practically eliminating —the drag load causing the shortening. equilibrium. The location of the maximum value — the
In March 1967, about 300 days into the monitoring of the site, location of the force equilibrium — is called the neutral plane.
several piles, widely spaced, for the construction of the dock The paper by Bjerrum et al. (1969) is the first to show that the
were driven within a distance of about 10 m from the test piles. development of a force equilibrium is indeed the manifestation
The pile driving caused an increase of pore pressures in the soil of load transfer for piles in settling soil. For comparison to the
and a corresponding decrease of effective stress. Coinciding effective stress analysis, Fig. 8 also shows a distribution of drag
with the decrease of effective stress, the piles lengthened, as load calculated for a beta-coefficient of 0.3 applied to the
indicated in Fig. 7. Settlement or heave were not measured at effective stress of May 1967. The agreement is not as good as
near the test piles, and can be argued that soil heave due to the that shown for the Sörenga pile, but still remarkably good
driving would have contributed to the lengthening (unloading) considering the crude telltale instrumentation system employed
of the piles. However, in the author's opinion, this is not likely. in the test.
The reduced negative skin friction resulted in a reduction of the
drag load and, consequently, a lengthening of the piles. As the Bjerrum et al. (1969) reports that the ground surface settled
pore pressures introduced by the pile driving dissipated, the 200 mm and the pile head settled 33 mm. Combined with the
effective stress returned to its previous level, as did the load in about 4 mm pile shortening at the neutral plane, this means that
the pile — and the piles shortened to the length they had before the settlement at the neutral plane and, therefore, also the
the construction piles were driven. The observation confirms downdrag of the pile was 29 mm, and the pile toe penetration
the finding in the Sörenga test (Bjerrum and Johannessen 1965) into the gravel deposit above the bedrock was about 25 mm, or
that pile shaft shear in clay soil is a function of effective stress. about 6 % of the pile toe diameter (resulting in the 540 KN pile
toe resistance).

The results of the other full-scale tests in the Norwegian clay

DAYS AFTER START (Bjerrum et al. 1969) show similarly large drag loads, ranging
from 1,200 KN through 4,000 KN and that the negative skin
0 100 200 300 400 friction correlates well to an effective stress analysis applying a
0 beta-coefficient of 0.2 through 0.3. For five of the test cases,
the measured ground settlements ranged from 70 mm through
Bitumen 2,000 mm. For the sixth case, where the test piles were driven
Coated at a site where no new fill was placed, only very small
Reduced shortening settlement occurred, which was caused by reconsolidation after

due to pore pressure the driving and ongoing secondary compression. Yet, the
4 increase from driving measured drag load, 3,000 KN, on the piles driven at this site
of construction piles was as large as for where large settlement was measured.
near the test piles.
6 In summary, the tests in Norway showed that drag loads can be
very large, negative skin friction is a function of effective
stress, and a force equilibrium will develop somewhere down
8 the pile. It is notable that, although the observations were
Un- included in the 1969 paper, somehow the geotechnical practice
coated missed to draw the conclusion that the relative movement
Pile between the pile and the soil needed to fully develop the shaft
shear is insignificantly small. That is, ostensibly settling soils
are not a necessary requirement for large drag loads to develop
12 on long piles.

Fukagawa, Tokyo, Japan

14 The proceedings of the Mexico ISSMFE conference also
contained a paper by Endo et al (1969) reporting on a full-scale
test on three, strain-gage instrumented, 43 m embedment,
Fig. 7 Heröya site. Measured total pile shortening During single, 609 mm (24 inches) diameter steel pipe piles with a wall
May 1966 through June 1967. Data from Bjerrum et thickness of 9.5 mm (0.375 inch). Below a 2 m thick surface
al. (1969). fill, the soil profile consisted of silty sand followed by 37 m of
compressible sandy silt, clay, and silt, 4 m of silt, and loose
sand to large depth. The groundwater table was located at a
depth of 2 m. An ongoing pumping of water from the sand
0 300 600 900 1,200 1,500 below 43 m depth had created a steep downward gradient at the
0 site and a consolidation of the compressible soil layers. The
piles were instrumented with strain-gages placed at every 6 m
length of the piles, starting at a depth of 8 m, and seven
5 settlement gages and seven piezometers were installed at 6 m
intervals in the soil near the piles. Fig. 9 shows the soil profile
Distribution with basic soil parameters and Fig. 10 shows the pore pressure
10 calculated from distribution at the start of the monitoring in June 1964 and
ß=0.3 times σ'z for about two years later in April 1966 (672 days).
actual excess u
The pore pressures underwent only small changes during the

Measured monitoring period, reducing a further about 10 KPa to 20 KPa
distribution below 10 m depth.
A diesel hammer with a nominal energy of 115 KJ was used to
the drive the piles into the surface of the sand at 43 m depth
25 terminating the driving at a very light resistance of
25 mm/blow. One pile was driven open-toe and two closed-
Bitumen toe. One of the latter two was driven inclined by 8° to the
30 vertical (1H:7V).
The consolidation settlement caused the soil to hang on the
piles. For the vertical, closed-toe pile, Fig. 11 shows the loads
(converted from the strain-gage values) versus time over the
two years long monitoring period. Fig. 12 shows the load
Fig. 8 Heröya site. Measured and calculated distributions distributions in the same pile at five different times.
of load. Data from Bjerrum et al. (1969).


0 20 40 60 80 100 0 50 100 150 200
0 0

10 10

Clay April
20 20


30 30
wP wL Silt

40 40

Sand Distribution
50 50

Fig. 9 Fukagawa site. Soil profile with water content and Fig. 10 Fukagawa site. Measured distribution of pore
plastic and liquid limits. Data from Endo et al. (1969) pressure at start of monitoring and two years later.
Data from Endo et al. (1969).

3500 LOAD (KN)

0 500 1000 1500 2000 2500 3000 3500
#5 0
3000 #4 29 DAYS
#6 5 124 DAYS
2500 #6
275 DAYS
#7 10
#2 400 DAYS

2000 15 672 DAYS

#4 #3

#7 25 #4

1000 30
500 #6
#2 40
0 200 400 600 800 1000


Fig. 11 Fukagawa site. Load measured at different times and Fig. 12 Fukagawa site. Fukagawa site. Vertical distribution
depths in the vertical closed-toe test pile. Data from Of load in the vertical closed-toe pile. Data from
Endo et al. (1969). Endo et al. (1969).

As shown in Fig. 13A, while the open-toe pile mobilized about understanding that the neutral plane is determined by both
half as much toe resistance as the other two, the development these two aspects is important. While the location of the force
of drag load in the three piles was about the same. The closed- equilibrium can be determined by analysis of resistance
toe piles (one vertical and one inclined) show about the same distribution or directly by testing an instrumented pile, in
load distribution. Near the ground surface and to a depth of contrast, the location of no relative movement between the soil
about 25 m, the distribution of drag load in the open-toe pile and pile, is much more prone to be in error when determined
was about equal to that of the closed-toe piles, but the drag from settlement analysis. Determining the location of the
loads differed below this depth. However, over the about ten neutral plane from the location of the force equilibrium is a key
metre length above the pile toe, the slope of the distribution to the settlement analysis of a pile group, as indicated by
curve of the open-toe pile is similarly to that of the closed-toe Fellenius (1984; 1989; 2004).The location of the neutral plane
piles, indicating that the unit shaft shear, here in the positive is influenced by the pile toe resistance, which is determined by
direction, is the about the same for the three piles along this the magnitude of the toe penetration. Endo et al. (1969)
length. In contrast, in the in-between about ten metre long provides data on movement of the pile toe and the mobilized
zone, from depths of 25 m through 35 m, where the transition pile toe resistance for the start of the monitoring and at three
occurs from negative direction shear to positive direction, the times during the monitoring. Fig. 14 is prepared from
open-toe pile exhibits a longer transition zone, starting slightly combining these data and shows the mobilized toe resistance as
above the 25 m record, resulting in an appearance of a smaller a function of the net toe penetration. The four data points show
magnitude shaft shear immediately below 25 m. a surprisingly linear trend. A more gradual shape similar to the
dashed curve in the figure would perhaps have been expected.
The neutral plane is the term for the location of the force However, the 22 mm penetration after the start of the
equilibrium in the pile. It is also the term for the location of monitoring (the zero point is unknown) corresponds to only
where there is no relative movement between the pile and the about 4 % of the pile diameter and it is not unrealistic for the
soil, i.e., where the pile and the soil settle equally, as toe resistance measurements to show a linear trend with
demonstrated in Fig. 13B. The paper by Endo et al. is the first increasing penetration at such small penetration.
to present observations revealing and confirming this fact. The


0 500 1000 1500 2000 2500 3000 3500 0 50 100 150 200
0 0
5 5
10 10

C 15 15
a Soil
20 20

25 25

30 30
35 35
l Open-toe
40 40
S Closed-toe
a Piles
45 45
d Toe Closed-toe Pile
50 50 Penetration

Fig. 13A Fukagawa site. Distribution of load in Fig. 13B Fukagawa site. Distribution of soil and pile
full-length. piles. settlement 672 days after start of monitoring.
Data from Endo et al. (1969). Data from Endo et al. (1969).

0 2,000 4,000 6,000 8,000
1,500 0

5 ß = 0.40
500 15
ß = 0.35

0 10 20 30 load

Fig. 14 Fukagawa site. Pile toe penetration —Vertical ß = 0.30
closed-toe pile. Data from Endo et al. (1969). 35

The author has fitted an effective stress analysis to the load 40

distribution data points of the two-year measurements of ß = 0.25
April 1966 (Day 672) for the vertical closed-toe pile, applying 45
the then measured pore pressure distribution. The data points,
the calculated curve fitted to the data, and the so-determined 50
beta-coefficients are presented in Fig. 15. The beta-coefficients
are about 0.3, which is similar to the values back-calculated for Fig. 15 Fukagawa site. Load distribution measured
the Sörenga and Heröya cases. The measured toe resistance and calculated for actual pore pressures and
corresponds to a toe bearing coefficient, Nt, equal to 8, which is matched to measured loads for the beta-
a low value and signifies that the pile toe is in a very loose soil coefficients shown and the pore pressure
or that the net toe penetration into the sand is too small to distribution of April 1966.
generate a significant toe resistance. The thin dashed line
shows the load distribution when assuming that all shear forces The embankment settled significantly. At the time of the
act in the positive direction, representing the distribution in a driving the test pile, 550 days after the start of the placing of
static loading test to the same mobilized toe resistance. the embankment, the ground surface had settled about 1.7 m.
The vertical alignment of the road bed where it connected to
Berthierville, Quebec, Canada the bridge was continually corrected by adding pavement
asphalt. Fig. 17 shows that during five years of observations
Bozozuk and Labrecque (1969) and Bozozuk (1970, 1972;
after the driving of the test pile, the original ground surface
1981) described a long-term measurement case history from
next to the test pile settled an additional 540 mm. Settlement
Berthierville, Quebec, Canada. In October 1964, a highway
gages placed at depths of 6 m, 12 m, 30 m, and 44 m showed
embankment with a final 11 m height, a 300 m base, and 27 m
that most of the settlement occurred above 30 m depth.
road width was placed on an 80 m thick deposit of clay. In
Fig. 18A shows the vertical distribution of settlement and the
April 1966, a tell-tale instrumented, 324 mm diameter closed-
downdrag (i.e., settlement) of the test pile. The pile head
toe pipe pile was driven through the embankment to a depth
settled 470 mm, 70 mm less than the ground surface. The drag
of 40 m below the original ground surface. Fig. 16A shows the
load distribution in the pile is shown in Fig. 18B. The two
soil profile with basic soil parameters indicating that the upper
figures show the same location of the point of equal movement
about 18 m consists of stratified fine soils deposited on a
(Fig, 18A) and the point of equal negative and positive
Champlain clay deposit. Fig. 16B shows the calculated
direction forces (Fig. 18B), i.e., the neutral plane — at a depth
distributions of initial and final effective stresses (after full
of about 12 m to 14 m.
dissipation of excess pore pressures) and the distribution of
excess pore pressures measured five years after the pile driving,
The distribution of load in the pile was calculated from pile
i.e., April 1971. The excess pore pressure measured in the
shortenings measured by means of eight telltales. Fig. 18B
lowest piezometer, if believed, would indicate that the placing
indicates that the slope of the load distribution curve below the
of the embankment induced an excess pore pressure that would
neutral plane over an about 10 m length is about half of that
be larger than the embankment stress and remain so even five
above the neutral plane and that, in contrast, the distribution
years after the placing of the fill. The author thinks that a more
curve is almost vertical over the following about 20 m length.
plausible five-year pore pressure distribution would be more
The slopes are indicative of shaft resistance. If the two slopes
like the dashed line shown continuing to 50 m depth, which
are true, they would indicate that the positive shaft resistance
line is drawn assuming the excess pore pressure below 18 m
between the neutral plane and the next 10 m below would be
depth would have the same relative portion of the imposed
more than about twice as large as the negative skin friction
embankment stress as the one at 18m depth.

w n, w P, and w L (%); τu (KPa) STRESS (KPa)

0 20 40 60 80 100 0 200 400 600 800
Original Ground Surface 0
GW Measured
Silty CLAY Excess Pore
10 SAND, SILT 10 Pressure

Clayey SILT
20 20
τu ( )
30 30

wP wL Silty CLAY
40 40

wn ?
50 50
60 60 Stress

70 70 Initial
80 Stress

Fig. 16A Berthierville site. Soil profile with water contents, Fig. 16B Berthierville site. Distribution of initial and final
plastic and liquid limits, and vane shear strengths. Data from effective stress, and measured excess pore pressures. The
Bozozuk and Labrecque (1969). question mark indicates an envisaged pore pressure
distribution. The pressure data are from Bozozuk (1972).
0 500 1,000 1,500 2,000
0 1 2 3 4 5 Years very small over the 20 m length further below. The author
S-10 44 m
S- 1 30 m believes the three lower telltales are affected by friction in their
guide pipes, which has resulted in inaccurate load data below

200 the neutral plane.

Bozozuk (1981) reported the results of a static loading test
carried out ten years after the pile driving. Fig. 19 presents the
S-3 12 m
500 S-4 6 m measured load-movement curve. At the applied loads of
S-5 0 m 510 KN, 1,020 KN, 1.520 KN, 2,030 KN, and 2,820 KN, the
load levels were maintained for one day before continuing the
test. No significant movement developed during these load-
Fig. 17 Berthierville site. Settlement measured at gages placed holding occasions. At the applied load of 3,060 KN,
at ground surface, and at depths of 6 m, 12 m, 30 m, and 44 m movements became progressively large as the ultimate
during the first five years of monitoring after the driving (April resistance of the pile was reached. During the following two
1966 through April 1971). The notes "S-1 30 m identify the days, the load that the soil could sustain gradually reduced. In
gage number and installation depths. Data from Bozozuk the reloading cycle after unloading, the previous maximum
(1972). load could not again be reached.

immediately above the neutral plane. Moreover, they would The vertical distribution of load in the pile evaluated at the five
indicate that a next to zero positive shaft resistance would have levels of sustained load is shown Fig. 20. The figure also
been mobilized along the lowest 20 m length of the pile. shows the 10-year distribution determined immediately before
However, the soil above neutral plane is not that different from the static loading test and, for comparison, the distributions at
the soil below, and, definitely, the soil below the neutral plane 1 year, 3 years, and 5 years. The latter curves indicate how the
is quite homogeneous throughout the length of the pile and increasing values of applied load progressively made the shaft
beyond. Therefore, the positive shaft resistance mobilized shear to change from negative to positive direction and that the
cannot differ much along this length of the pile. Were the applied load did not add to the existing maximum drag load at
values true, the effective stress within the 10 m length below the neutral plane, demonstrating that live load does not add to
the neutral plane must be large, while at the same time being the drag load in a pile but replaces it.


0 100 200 300 400 500 600 0 1000 2000
-10 0


0 10



10 20
15 Settlement
20 30


30 40

35 Test
40 Pile 50
1 Year
45 3 Years
5 Years
50 60

Fig. 18A Berthierville site. Distribution of settlement of the Fig. 18B Berthierville site. Distribution of load in the test pile
soil and test pile with depth measured over 5 years after the at 1 year, 3 years, and 5 years after the driving of the test
driving of the test pile. Data from Bozozuk (1972). pile. Data from Bozozuk (1972).

2 days
3,000 forces slowly built up over a ten-year period and the forces
mobilized in a relatively rapidly performed loading test were

very similar in magnitude.

Fig. 21 shows a compilation of the data and analyses results.
1,500 Curves marked A and B show the load distributions at the 10-
year measurements (before the static loading test) and at the
maximum sustained load in the static loading test (2,820 KN),
500 respectively. Curves C and D show load distributions
calculated using effective stress in fitting the calculated
0 20 40 60 80 100 120
distribution to the 10-year measured load curve. Curve C
indicates the distribution for the shaft shear acting in the
negative direction along the full length of the pile, and Curve D
Fig. 19 Berthierville site. Load-movement curves from static similarly for the shaft shear acting in the positive direction.
loading test at ten years after the pile driving. Data from The intersection of Curves C and D is the location of the
Bozozuk (1981). neutral plane. Point E indicates ultimate resistance (plunging
load) in the static loading test. Finally, Curve F shows the load
distribution calculated from the assumption that the shaft shear
No measurements of load distribution could be obtained for the was equal to the undrained shear strength determined in the
3,060 KN plunging load. The drag load of about 1,500 KN vane shear test shown in Fig. 16A. There is little or no relation
was supported by an equal magnitude shaft resistance present between the evaluated shaft shear (negative skin friction) and
in the portion of the pile below the neutral plane—the point of the undrained in-situ shear strength values. However, the vane
force equilibrium. This means that half of the about 3,000 KN tests were performed before the placing of the fill. New vane
applied load served to eliminate the drag load and the soundings after the more than ten years of consolidation would
remaining half served to keep mobilizing the positive resistance have shown larger strength values. It is outside the scope of
below. In other words, the test results proved that the negative this text to discuss whether or not the new vane soundings
skin friction and the positive shaft resistance are fully would have produced a load distribution more in agreement
mobilized shear forces. Moreover, the results show that the with the measured load distribution, Curve C.

LOAD (KN) The evaluated load distribution in the embankment fill

corresponds to beta-coefficients of 3.0 in the upper 4 m and 0.6
0 1000 2000 3000
in the lower about 6 m. A beta-coefficient of 3 is a very high
0 value. As Bozozuk (1972) points out, the values reflect a
combination of the large horizontal stress created by the
compaction of the engineered fill and the effect of that the
10 embankment settled more in the center, as opposed toward the
sides. The latter caused the embankment to bow, which
Static increased the horizontal stress (the earth pressure) in the upper
portion of the embankment. The beta-coefficient value of 0.2
20 in the clay is low but within the range of values found in

1 3 5 10
normally consolidated marine clays. However, because
approximations associated with the use of telltale
30 instrumentation and necessary approximations in the author's
evaluation, such as the assumption made with regard to the
distribution of pore pressures over the lower half of the pile, the
indicated beta-coefficients are only approximate values. A
40 repeat of the test using modern methods of instrumentation
would probably have resulted in coefficients slightly different
to those shown.
That the data below the neutral plane and the lowest piezometer
value appear to be suspect should not be taken as rejection of
Fig. 20 Berthierville site. Berthierville site. Load the test results. The case history confirms the results of the
distributions at passive monitoring of the test pile at 1 year, tests in Norway and Japan with regard to that the load transfer
3 years, 5 years, and 10 years after the pile driving, and for five is governed by effective stress, that the shaft shear in negative
sustained loads from the static test. Data from Bozozuk (1972; and positive direction are equal, that shaft resistance acting
1981). over long term is equal to that mobilized in a short term static
loading test, and that the beta-coefficient is a function of not
just the overburden effective stress, but also of the horizontal
LOAD (KN) stress in the soil. These are first time findings and the papers
0 1,000 2,000 3,000 4,000 are enlightening major contributions to the advancement of the
0 state-of-the-art.

Melbourne, Australia
Walker et al. (1973) installed a 760 mm diameter, closed-toe,
strain-gage instrumented, pipe pile to an embedment depth of
F 27 m by driving the piles through 6 m silty sand deposited
on 15.5 m stiff silty clay, followed by 3 m of sandy silt on
ß = 3.0 0 m - 4 m
20 ß = 0.6 4 m - 10 m
dense sand and gravel. The pile reached 2.5 m into the sand
ß = 0.2 10 m - 50 m and gravel. An identical pile was coated with 1.5 mm of
bitumen though its entire length and driven to an embedment
of 24.5 m. The silty sand layer was compact with SPT N-
indices smaller than 20 bl/0.3m. The water content of the stiff
silty clay layer ranged from 60 % through 100 % and the liquid
and plastic limits ranged from 90 through 120 and 30 through
D C 50, respectively. After completion of the pile driving and site
instrumentation (piezometers and settlement gages), a 3.0 m
thick fill was placed over an area of 200 m by 100 m around
the piles. The silty clay is overconsolidated having a
preconsolidation margin of about 80 KPa, much larger than the
stress imposed by the 3 m fill.
Fig. 21 Berthierville site. Load distributions.
A. at the 10-year measurements (before the static test) The surcharge resulted in an increase of pore pressures of about
B. at the maximum sustained load in the test, 2,820 KN 20 KPa that did not change during the reported monitoring
C. calculated using effective stress, matching to the drag period. Settlement measured at ground surface was small.
loads measured above the neutral plane Over a monitoring period of 238 days after the placing of the
D. calculated using the same parameters as in "C". surcharge, only about 25 mm of settlement was recorded.
E. ultimate resistance in the static loading test. Walker et al. (1973) indicated that the measured settlement was
F. calculated from the vane shear strength due to creep rather than to consolidation. The strain gages in
Test data A, B, and E from Bozozuk (1981). the test pile registered increase of load. Fig. 22 shows also the

loads evaluated from the strain gages at 20 m embedment depth Keehi Interchange, Hawaii, USA
in the uncoated pile. The 238-day load distribution in the piles
Clemente (1981) reports a full-scale test involving three
is shown in Fig. 23. As seen, the bitumen all but eliminated the
instrumented, 419 mm diameter, octagonal, precast prestressed
shear forces along the pile. Over the 238 days, the loads in the
concrete piles in the island of Oahu, Hawaii. As indicated in
uncoated pile had built up to a maximum of about 1,800 KN at
Fig. 24, the soils at the site consisted of a 4 m thick sand fill
the neutral plane (the 20 m gage). Because of the small relative
placed a long time earlier over a 36 m thick deposit of soft clay
movements between the pile and the soil, the transition zone is
followed by 16 m of stiff silty clay. Hereunder, starting at the
long and covers essentially the full thickness of silty clay layer.
depth of 56 m, the soil consisted of a mixture of medium dense
No pile toe load was mobilized, which means that the
to dense coral sand alternating with firm to stiff silty coral clay.
settlement at the neutral plane was so small that no toe
The groundwater table was located at the interface between the
penetration was created. The dashed lines in Fig. 23 indicate
fill and the soft clay. In July 1977, three test piles,
the load distribution calculated in an effective stress analysis
numbered 6, 7, and 8, were driven and instrumented with
matched to the data. The beta-coefficient fitted to the load
multi-rod extensometers measuring pile shortening. Piles 6
from the fill layer is large, about 2.0. In the fine sand, the fitted
was driven to 40.0 m depth and Piles 7 and 8 were driven to
beta-coefficient reduces from 1.0 at the upper boundary to 0.8
49.7 m depth. Before the driving, Pile 7 was painted with a
at the lower boundary, still large for a silty sand. In the silty
thin asphalt primer and then coated with A85 bitumen over an
clay and sandy silt, it is 0.40 and 0.45, respectively. In the sand
upper 35 m length, leaving the rest of the pile, 15 m, uncoated.
and gravel, calculations using a beta-coefficient of 0.6 produces
Piles 6 and 8 were not coated. The thickness of the coat was
a load distribution that matches the measured load distribution.
aimed to be 1.5 mm, but because of the warm climate, much of
the bitumen flowed off the upper surface and sides of the pile.
Fill completed DAYS AFTER START
0 50 100 150 200 250 300
After the coating, the pile was driven to an embedment depth of
0 2000 50 m. Piles 6 and 8 were driven to 40 m and 50 m depth,
5 respectively. In August 1977, a 6 m high embankment with a

10 Load in pile 1500 crest area of 18 m by 27 m was placed around the piles and soil
at 20m depth settlements and pile shortenings were monitored for 180 days.


20 1000 The shortenings were converted to strain and loads in the piles.

30 500
35 27 m
40 0

Fig. 22 Melbourne site. Settlement of ground surface and 18 m PLAN

development in load at 20 m depth in the pile. Data VIEW
from Walker et al. (1973). PILE #6 PILE #8

0 500 1000 1500 2000 2500 6m EMBANKMENT

w = 70 %
10 LL = 80
30 PL = 45
Transition Zone

15 Stiff
Bitumen Silty 40
coated Clay

25 Sandy 60
30 Sand 70 CORAL SAND

Fig. 23 Melbourne site. Measured load distribution in

uncoated and coated piles, and calculated Fig. 24 Keehi Interchange site. Plan view and profile of fill,
distribution. Data from Walker et al. (1973). soil, and piles. Data from Clemente (1981)

Figure 25 shows the load distributions in the clay below the fill Bangkok, Thailand
and the settlement distribution after the 180 days. The symbols
Indraratna et al. (1992) presented a full-scale study of two
connected with dashed line show the evaluated loads. The
400 mm diameter, instrumented precast concrete piles driven
solid lines show the load distributions calculated by effective
into soft Bangkok clay. One of the two piles was coated with a
stress method. The author derived a beta-coefficient equal
layer bitumen along its upper 20 m length. The bitumen
to 0.25 by matching the calculated loads to the evaluated loads
amount was 200 cm3/m2, that is, the coat was theoretically
for all three piles. The paper does not report any piezometer
0.2 mm thick. The soil profile consisted of an upper 2 m thick
observations and the author has assumed that no excess pore
layer of weathered clay followed by 13 m of soft clay with a
pressures remained in the clay at the end of the observation
water content of about 80 %, deposited on 14 m of firm clay
period. The 0.25 value is a lower bound value, because any
with a water content of about 60 % above sand found at 29 m
remaining pore pressures, if known and considered in the
calculations, would result in higher beta-coefficients. All three
piles developed a force-equilibrium neutral plane. Because of
The piles were driven in five stages with the pile driving
the different conditions of lengths and coating, the neutral
interrupted at 8 m, 12 m, 16 m, and 20 m depths for a period of
planes are at different depths, but the evaluated effective stress
nine days and an uplift, quick method, static loading test
parameters are the same for all three piles.
performed before the driving was resumed. The final pile
depth was 25 m. After the driving, a 2 m high embankment
was placed around the piles over an area of 14 m x 24 m, as
indicated in Fig. 26. Pile loads and settlement were recorded
LOAD (KN) over a period of 265 days after the placing of the embankment.
0 500 1000 1500 2000 2500 Fig. 27 shows the load-movements for the pull tests. The
450 mm figure for the uncoated pile is supplemented with the plot of the
0 distribution of the ultimate load for each pull test versus depth.
Despite the quick static loading test and the very thin bitumen
SETTLEMENT coat, the bitumen reduced the maximum shaft resistance to
10 50 % of the shaft resistance of the uncoated pile.

Distance (m)
20 -15 -10 -5 0 5 10 15

ß = 0.25
10 mm
Distance (m)

Coated Uncoated

ß = 0.25
0 mm
50 10 embankment Fill
2 m high
ß = 0.25 -5-15
60 5


10 Soft
15 Clay

20 Load Cells
Piezometers Firm
Fig. 25 Keehi Interchange site. Load and settlement 25 Clay

distributions after 180 days. Data from 30

Clemente (1981).
Fig. 26 Bangkok site. Plan view and section of site.
The embankment fill caused the ground to settle. At the
ground surface, the settlement was 450 mm. Most of the
Figure 28A shows the distribution of load measured 3 days,
settlement developed in the upper ten metre depth. The
25 days, 92 days, 156 days, and 265 days after the placing of
measurements show that further down in the soil, the relative
the embankment. The figure includes also the ultimate
movement between the pile surface and the soil was no more
resistances from the four uplift static loading tests. The drag
than a few millimetre. Yet both negative skin friction and
load distribution measured at 265 days and the uplift resistance
positive shaft resistance appear to have been fully mobilized.
distribution are quite similar to each other. The figure is
The drag load along the bitumen coated length of Pile 7
supplemented with a curve showing the fitted load distribution
amounted to 375 KN, while the drag load along Pile 8 to the
for a beta-coefficient of 0.25 above the neutral plane and 0.30
same length was 1,770 KN. This means that the less than about
below. The measured toe resistance corresponds to a toe
1 mm thick coat of bitumen reduced the negative skin friction
bearing coefficient of 3, which is a small value even in marine
to 20 % of that found for the uncoated pile.
clay. The paper reports vane shear undrained shear strength

500 500
Uncoated Pile 8 m Length Bitumen Coated Pile 8 m Length

12 m Length 12 m Length
400 400
16 m Length 16 m Length

20 m Length 20 m Length


0 200 400 600 800

200 200


20 100

0 10 20 30 40 50
0 10 20 30 40 50

Fig. 27 Bangkok site. Pull test results. Inserted graph shows maximum load versus depth.
Data from Indraratna et al. (1992).

Axial Load (KN) SETTLEMENT (mm)

0 100 200 300 400 500 600 700 800
Ground Surface 0 100 200 300

Loads from
pull tests 0 5 10 15 20
10 10


3 days
15 15
β = 0.25 3 Days
25 days NEUTRAL PLANE 25 Days
20 20 20 53 Days
92 days 156 days
81 Days
265 days
25 25 25 141 Days
β = 0.30 β = 0.30 ENLARGEMENT 209 Days
265 Days
30 30

Fig. 28B Bangkok site. Settlement distribution. Data from

Fig. 28A Bangkok site. Load distributions. Data from Indraratna et al. (1992).
Indraratna et al. (1992).
Singapore Port
Leung et at. (1991) presented measurements in a strain-gage
values down to a depth of 16 m. A distribution calculated instrumented, square 280 mm precast concrete pile driven to
using the vane shear strength values as unit shaft resistance 24 m embedment through a 9.5 m silt and clay fill, 8.5 m of
results in a distribution of load that is close to the distribution marine clay, and 3 m of sandy clay on a shale bedrock. The
determined in the static uplifts tests and the drag-load piles were installed to support a container deck and the
distribution to 16m depth. (The distribution is not shown in allowable load on the piles was 800 KN. Fig. 29 shows the
Fig, 26A). Despite the thin bitumen coat, the maximum drag measured load distribution once the deck was built (57 days
load for the coated pile was reduced to 40% of the maximum after the driving) and two years later (745 days). The amount
drag load in the uncoated pile. of soil settlement (due probably to the fill) is not reported by
Leung et al. (1991), nor do they report settlement of the piled
The container loads constituted a variable, or live load on the
piles, which amounted to only about 200 KN/pile during the

two-year measurement period. The container loads affected the The Bäckebol site lies at an open area along the Göta River
load distribution in the upper 9 m of the pile. Below this depth, outside the city of Göteborg, in Southwestern Sweden. The
the load distribution was governed by the soil settlement that area is virgin, untouched by construction since it rose from the
was large enough to generate negative skin friction along the sea after the end of the Ice Age. As shown in Fig. 30A, the
pile. A neutral plane developed at the lower boundary of the main soil body consists of a 40 m thick layer of sensitive,
marine clay. The measured load distribution corresponds to a marine post-glacial clay, followed by silty sand to large depth.
beta-coefficient of 0.5 in the marine clay. Note the similarity To a depth of about 17 m, the natural water content is
between the distribution of load above the neutral plane to that about 90 % and larger than the liquid limit. Hereunder, the
shown in Fig. 20 for the second and third load levels of the water content reduces with depth, but stays close to the liquid
Berthierville test. Fig. 28B shows the measured distribution of limit. The percentage of clay size particles is about 80 % to
settlement and the settlement of the pile with an enlargement is 20 m depth. From here to 30 m depth, the percentage is about
small as it is governed by the soil settlement at the location of 55 %. Details about the site are available in Fellenius and
the neutral plane, which is smaller than 10 mm. The relative Broms (1969) and Fellenius (1972).
movement between the pile and the soil that mobilized the
positive shaft resistance below the neutral plane, the last about
As indicated by the vane shear strength distribution, shown in
7 m length of the pile, was about 2 mm and the toe penetration
Fig. 30B, the clay is soft to a depth of about 10 m and firm
was about 2 mm, which is commensurate with the very small
below this depth. Eleven consolidometer tests performed on
measured toe resistance.
samples from depths of 4 m through 32.5 m show that the
preconsolidation margin of the clay is about 20 KPa in the
Variable load
LOAD (KN) upper portion of the clay and increases to about 30 KPa in the
0 200 400 600
lower portion. Fig. 30C presents the compressibility of the clay
expressed in Janbu modulus numbers (CFEM 1992 and Janbu
Two Two years
1963; 1998) obtained from the tests. An example of the stress-
months later strain curves representative for the consolidation tests is
after start (745 days )
(57 days )
& presented in Fig. 31. The virgin Janbu modulus number, m,
5 Clay ranged from about 4 through 7 with an average of about 5,
ß = 0.5 Fill which indicates a very compressible clay. The reloading
modulus number, mr, ranged from 60 through 80 with an
10 average of 75. (The void ratios, e0, for the tested samples
ranged from 1.3 through 2.4, the virgin consolidation

Marine coefficients, Cc, ranged from 0.06 through 0.16, and the
15 Clay reloading consolidation coefficients, Cr, ranged from 0.07
through 0.11). The depth to the groundwater table varies
slightly seasonally from close to the ground surface and at
20 Clay
about 1.0 m depth. The pore water pressure is hydrostatically
Shale Before the pile driving, three pneumatic piezometers
25 Bedrock
(Kallstenius and Wallgren 1956) were installed at depths of
Residual 9 m, 23 m, and 30 m at a distance of about 0.7 m outside the
Soil intended location of each pile. In addition, an axially-flexible-
30 hose multipoint-settlement gage was placed 0.1 m outside each
pile location and one 5 m away from the piles, enabling
measurements of vertical movement at every two metre to
Fig. 29 Singapore Port site. Load distribution. Data from
depths to about 33 m. Instrumentation details are described by
Leung et al. (1991).
Fellenius (1972). The distances to the points in the gages was
measured with a tape inside the hose and related to elevation by
means of surveyor's leveling of the uppermost point. The
accuracy of a single tape measurements is about 1 mm and in
Background information
comparing two or more measurements the accuracy is at
In June 1968, a full-scale test was started in Bäckebol, Sweden, best 2 mm.
consisting of driving two single, 55 m embedment, 300 mm
diameter, hexagonal cross section, load-cell instrumented, A load cell was specially developed for the two test piles
spliced, precast concrete piles. The observations continued into designed to measure the load in the pile directly, eliminating
August 1983, i.e., for 5,500 days. The objective was to the need for using the pile modulus to determine the load in the
measure the forces developing in a pile due to reconsolidation pile. Details about the load cell are available in Fellenius and
of the soil after the pile driving, and as caused by soil Haagen (1969). The piles were made up of five 11.0 m long
settlement due to the placing of a fill around the piles. segments spliced in the field as each pile was driven. The load
cells were manufactured as 0.6 m long pile segments and
placed in between the first and second, the second and third,

PL, LL, and w n (%) STRENGTH and STRESS (KPa) MODULUS NUMBERS, m r and m

0 25 50 75 100 0 50 100 150 200 250 0 25 50 75 100

0 0 0

5 5 5

10 10 10

15 15 15



m mr
20 20 20

25 25 25

30 30 30
35 35 35

40 40 40

A. Distribution of Atterberg limits B. Distribution of vane shear strength, effective C. Distribution of modulus numbers.
and water content. stress and preconsolidation stress.

Fig. 30 Bäckebol site.

and the third and fourth pile segments in both test piles. Pile The testing activities consisted of the following steps:
PII was also supplied with a load cell at the pile toe. Pile PI
was driven to a depth of 53.1 m and Pile PII to 55.0 m. After 1. Taking last "zero" (initial) readings of site
the driving, a toe telltale was inserted into each pile to measure gages on June 24, 1968, before driving the
pile shortening. Details about the piles are available in piles and recording all gages at frequent
Fellenius (1972). intervals following the driving until
November 2, 1969 (Day 0 through Day 496)

2. Forming and casting a concrete slab on the
piles to provide a load of 440 KN (started
Sample Depth = 27.5 m
on Day 495 and the formwork was removed
25 on Day 517)
3. Recording all gages at frequent intervals until
20 October 31, 1970 (Day 518 through Day 859)
4. Placing a 360 KN of concrete blocks on to
Strain (%)

15 the concrete slab to increase pile load

to 800 KN (Day 860)
10 5. Recording all gages at frequent intervals until
September 29, 1973 (Day 860 through
Day 1,923)
mr = 80 6. Placing a 2 m high fill in a circle with 42 m
σ'0 = 147 KPa σ' c = 170 KPa diameter around the piles (Day 1,924
0 through Day 1,988)
10 100 1,000
Stress (KPa)
7. Recording all gages until end of test on
September 27, 1982 (Day 1,988 through
Day 5,206)
Fig. 31 Bäckebol site. Example of stress-strain curve from
consolidometer test; sample from 27.5 m depth 8. Last survey of multipoint settlement gage on
August 8, 1983 (Day 5,523)

The driving of the two piles created large excess pore pressures
-30 -25 -20 -15 -10 -5 0 5 10 15 20
in the clay. Fig. 32 shows pore pressures measured at a depth 0
of 23 m adjacent to the piles. At 23 m depth, the vertical
effective overburden stress before the pile driving was about
120 KPa. Obviously, for the first few days after the pile DAY 150
driving, the piles are standing in a highly remolded soil with EOD
zero effective stress—essentially a liquid. The pore pressures 10
required about five months to return to the before-pile-driving
level. 15

Fellenius and Broms (1969) present observations of soil
movements during and immediately following the driving. The
driving of the piles into the sand caused the sand to compact.
Precision leveling immediately after the end of the driving 25

(EOD) showed that the sand surface settled 9 mm at Pile I

and 7 mm at Pile II during the driving. The simultaneous 30
measurements showed that the clay layer expanded due to the
driving of the displacement piles. As shown in Fig. 33, the 35
upper 5 m portion of the ground next to the pile heaved
about 20 mm relative to the bottom of the clay layer, resulting
in a net heave of the ground surface. Below 2.5 m depth, the
upward movement of the clay was smaller than the settlement
of the boundary between the clay and the sand. Thus, but for
Fig. 33A Bäckebol site. Heave (negative numbers) near Pile
the uppermost about 2.5 m, the net effect during the driving of
PII during the first 150 days after driving.
the two single piles is a downward movement of the soil in
relation to the piles. During the first few days following the
driving, the concrete piles lengthened (swelled) about 1 mm SETTLEMENT OVER FIRST 150 DAYS (m m )
from absorbing water from the soil. During the first 150 days 0 2 4 6 8 10
following the driving, the clay compressed (settled) due to the 0


Clay Two piezometers
140 40 m at 23 m depth

80 15 m

60 25

20 30

0 30 60 90 120 150 180

Fig. 32 Bäckebol site. Pore pressure measured at 23 m

The survey method of measuring settlement at the site was not

Fig. 33B Bäckebol site. Settlement measured near Pile PII
accurate enough to record the settlement until the
during the first 150 days after driving.
measurements had continued for a few years. However, as the
monitoring proceed, it was realized that a small continued
The load-cell measurements showed no change during the very
settlement was occurring in the area that was equal in
first days when the measured pore pressures exceeded the
magnitude to the isostatic land rise, i.e., about a millimetre a
calculated effective overburden stress. That is, the small heave
year at the ground surface. The settlement of the ground
and lengthening of the pile due to absorption of water did not
surface away from the piles was measured to be about 10 mm
coincide with any change of load in the pile. Thereafter, the
to 15 mm over the first 500 days. Next to the piles it was
loads measured in the piles started to increase coinciding with
somewhat smaller (Fellenius 1971; 1972).

2,000 Second load 2 m thick fill Last of records

placed on piles placed over site from Pile PII
First load M1 & M5
placed on piles
M2 & M6
M3 & M7
1,000 M3 M7

M2 M6
M4 M1 M5

0 1,000 2,000 3,000 4,000 5,000

Fig. 34 Bäckebol site. Loads measured in the piles from end of driving.

the dissipation of the pore pressures. The forces are built up

from accumulation of negative skin friction due to the relative FORCE AT LOAD CELL (KN)
movement between the pile and the clay. Fig. 34 shows the 0 500 1,000 1,500 2,000
development of pile forces with time after the pile driving. 0

The shortening of the piles due to the measured forces during 496 859
the first 500 days before any load was placed on the pile head 10
amounted to about 5 mm, and the pile toe movement 518 860
(penetration) into the silty sand was about the same magnitude.
Load-cell M4 shows that the pile toe force increased from
80 KN to 200 KN.

At the end of driving, EOD, the forces in the pile were about
equal to the buoyant weight of the pile. As the drag load from 30
the clay increased, positive resistance was mobilized in the
lower portion of the pile. The development is best discerned in 0
2,650 &
Fig. 35 showing the vertical distributions of force in the piles. 40
When placing the first 440 KN on the pile head (Test Activity 1,923 1,988
No. 2), the drag load was essentially eliminated. The small soil
movements continued and, therefore, the drag load continued to
develop, as shown in Figs. 34 and 35. The rapid placement of
the additional 360 KN on the piles (Test Activity No. 4),
eliminated the drag load, temporarily, and actually created a
small positive shaft resistance along the piles. Each of the two 60
applications of load to the pile heads resulted in a total pile
shortening of about 2 mm to 3 mm over the length in the clay.
The shortening at the 20 m mid-depth is about 1 mm to 2 mm. Fig. 35 Bäckebol site. Distribution of load in the piles at
After the two loads had been placed on the pile heads, the total end of driving: Immediately before and after:
pile shortening since the pile driving was 9 mm, about equal or
adding the first load (Day 496 and Day 518),
marginally smaller than the settlement of the ground surface
close to the piles at that time. That is, the forces and force adding the second load (Day 859 and Day 860),
changes in the pile are associated with relative movements placing the fill (Day 1,923 and Day 1,988),
between the pile and the soil smaller than about 2 mm. and at final stabilized loads (Pile PI at Day 2,650 and
Pile PII at Day 3,128).

After adding the load to the pile heads and about a year of The starting point is the net toe load immediately after the end
continued monitoring, a 2 m high fill was placed around the of the pile driving. As shown, the load-movement curve is
piles in a circular area with a 41 m diameter (during approximately linear and does not show any sign of
Days 1,923 through 1,988). The fill accelerated the settlement approaching an ultimate value. The maximum toe penetration
and the development of drag load. Within about 3 months, the corresponds to about 5 % of the pile diameter.
forces in the piles reached their maximum values, which level
stayed constant for the duration of the following about 3,000 The measurements and analysis indicate that the total ultimate
days of monitoring the site. Pile PII was damaged above shaft resistance in the clay and underlying sand was about
ground breaking the cable connections to the load cells some 3,000 KN. The mobilized toe resistance was 300 KN
time after Day 2,650 and no further records from the load cells (including the pile weight) for a total toe penetration of
in PII were obtained. Monitoring of the site and of Pile I about 13 mm. The total pile shortening for the 800 KN applied
continued until Day 5,206, almost fifteen years after the end of load plus drag load was about 18 mm.
pile driving.
Measurements of settlement showed that most settlement
The maximum load measured in load cell M5 was 1,670 KN. occurred in the upper layers. Fig. 39 shows the settlement
Subtracting the load applied to the pile head, the maximum measured next to Pile PII at different days after the start. The
drag load recorded by load cell M5 is 870 KN. A calculation data show clearly that most settlement occurred above the ten
applying the vane shear strength as negative skin friction metre depth where the stress from the fill exceeded or was abut
results in a calculated drag load of 3,500 KN, about four times equal to the preconsolidation margin. The settlement data are
larger than the measured drag load. According to Swedish inconsistent if the survey information would be taken as
practice, described by Holtz and Wennerstrand (1972), a liquid deciding for the long-term settlement of the ground surface. If
limit of 85 necessitates reduction to 75 % of the measured so, all the three multipoint settlement hoses would indicate that
value, i.e., the calculated drag load would become about the settlement below the 32 m depth would be three times as
2,600 KN, still much larger than measured. large as the settlement of about 9 mm measured to have
occurred between the depths of 10 m and 32 m. Obviously, the
An effective stress analysis has been matched to the casing supporting the upper portion of the multipoint settlement
distribution of pile load measured on Day 2,650 in Pile PII and hose appears to have moved due, probably, to seasonal
Day 3,128 in Pile PI. The beta-coefficients in the clay for influences. The individual measurements presented in Fig. 39
fitting the calculations to the measured loads between the are plotted as if no settlement occurred below 32 m depth from
ground surface and first load cell level, between first and the time that the fill was placed. Clearly, a few millimetre of
second load cell levels, and between second and third load cell settlement must have occurred also below the lowest point.
levels, are 0.18, 0.15, and 0.14, which are low values. The
dashed line shown in Fig. 35 indicates the calculated In Fig. 40, the measured settlements below 10 m depth are
distribution. The measured forces suggest that the force shown as measured on Day 2,044, the day all of the fill had
equilibrium lies below load cells M1 and M5. The dashed line been placed and on Day 4,362 almost seven years later, when
is drawn, somewhat arbitrarily, to reflect this. The slope of the the last reading was taken. Below about 30 m depth, the
dashed line in the sand upward from the toe load cell is difference between the curves is only 2 mm. Probably,
calculated using a beta-coefficient of 0.6. assigning the settlements to be equal at 32 m depth has cut a
few millimetre off the true settlement. Still, it is clear that the
The averages of unit shear force along the pile are obtained settlement at depth was very small. Yet, it fully mobilized
from differentiating the loads between the load cells and negative skin friction between the load-cells M1 and M2 and
presented in Fig. 36. For clarity, the values are only plotted up M5 and M6 within a year after placing the fill, as shown in
to Day 3,000. The values are about constant beyond that date. Fig. 36.
The average shaft shear, showing a maximum value
corresponding to a beta-coefficient of 0.33 recorded in Pile PII A comparison between the measured settlement and that
between load cells M4 and M5 in the sand, includes a short calculated from the soil data is shown in Fig. 41. Fig. 41A
distance of negative skin friction above the neutral plane and shows the effective stress, the stress from the fill per
the positive shaft resistance below. Calculations using the Boussinesq distribution, and the distribution of
mentioned 0.6 beta-coefficient in the sand places the location preconsolidation stress. Fig. 41B shows the distribution of
of the neutral plane about 4 m below load cell M 5, at 46.5 m settlement calculated from the values plotted in A and the
depth. settlement distribution measured at Day 4,362 next to Pile PII.
Note the stress from the fill is so close to the preconsolidation
To measure the pile shortening, new telltales were installed in margin that a change of a 1 KPa or so, one way or the other,
Pile PI at the start of the placing of the fill. Together with the would significantly alter the calculated settlement distribution.
survey of the pile head, the telltales determined the pile toe However, the two curves show agreement and support the
movement, as shown in Fig. 37. Fig. 38 shows the pile toe conclusion that most of the settlement at the site from the
load-movement from the start of the test. The toe movement is placed fill occurred in the upper ten metre of soil and that only
accurately known only in Pile PI and the toe load only in Pile very small relative settlement occurred between the pile and the
PII (load cell M4). Fig. 38 shows a combination of the two clay below ten metre depth.
measurements: toe load versus toe movement. The buoyant
weight of the pile has been subtracted from all load values.

Load placed 2 m thick fill placed



on pile heads over site
Beta = 0.18
0 Ground to M3

-5 Ground to M7

0 500 1,000 1,500 2,000 2,500 3,000


20 Beta = 0.15
0 500 1,000 1,500 2,000 2,500 3,000

Load placed 2 m thick fill placed


on pile heads over site

25 Beta = 0.14
0 500 1,000 1,500 2,000 2,500 3,000


0 500 1,000 1,500 2,000 2,500 3,000



60 M5-M4

Beta > 0.33
in the sand

Fig. 36 Bäckebol site. Measured average shear forces between the load cells

Shortening and movement (mm) PIle PI
0 5 10 15 20

10 5
Days 2,044 & 122
PIle P I
5 10

0 15

1,900 2,400 2,900 3,400 3,900 4,400 4,900 5,400
Days after driving
Days 4,362 & 2,440
Fig. 37 Bäckebol site. Movement of pile heads and pile toe
of Pile PII since start of placing fill 25

300 30


Day 2,651
Start of 40
placing fill
Day 1,922

100 Fig. 40 Bäckebol site. Distributions of settlement measured

in multipoint hose below 10 m depth. Days after start (first
number) and days after placing of fill (second number) are
indicated. The dashed lines indicate the trend of the
0 5 10 15

Fig. 38 Bäckebol site. Toe load of Pile PII versus toe CONCLUSIONS
penetration of Pile PI
Several generally applicable conclusions of importance for the
SETTLEMENT (mm) design of piled foundations can be drawn from the reported
0 25 50 75 100 125 150 175 200 225 250 275 300 325 350 full-scale tests, as summarized below. However, presenting
design recommendations is outside the scope of the this paper.
5 For such, see Fellenius (2004).

15 Days after driving

• Piles are usually installed to transfer loads through soft or
loose soil layers to more competent soil. Under such

and days after

placing fill
conditions, negative skin friction will always develop
2,044 & 122
2,569 & 647
along the piles and accumulate to drag load.
2,953 & 1,031
• The case histories have shown that effective stress
Multipoint hose 3,641 & 1,719
35 4,362 & 2,440 governs the load transfer from soil to pile and from pile to
next to the piles surveyor based soil.
• Load distribution calculated from undrained shear
SETTLEMENT (mm) strength values may occasionally agree with actual load
0 25 50 75 100 125 150 175 200 225 250 275 300 325 350
0 distribution when no excess pore pressures exist, but this
appears more to be a coincidence as opposed to a general

15 Days after driving

• At all times, an equilibrium will exist between loads
sustained on the pile head and the drag load acting

and days after

placing fill
downward and the positive shaft resistance and toe
2,044 & 122
25 resistance acting upward. That is, a force equilibrium
2,652 &730
point, called neutral plane, will always develop. The
3,143 & 1,221
Multipoint hose 5m 5,523 & 3,601
location of the equilibrium will adjust to a change of
away from the piles Surveyor based
sustained load or change of pore pressure.
• If the soil settlement is large and the pile toe response is
Fig. 39 Bäckebol site. Distribution of settlement measured stiff, the neutral plane will be located very close to the pile
in multipoint hoses next to Pile II and 5 m away from the piles toe.


0 50 100 150 200 250 0 100 200 300
0 0
5 5

10 10
15 15


25 25

30 30

35 35

40 40

Fig. 41 A. Bäckebol site. Distribution of effective stress, preconsolidation stress, and stress from fill.
B. Distribution calculated settlement measured at Day 4,362 next to Pile PII.

• A few millimetre of movement between the pile surface • Two single piles of different length, or of same length but
and the soil can be sufficient to fully mobilize the shaft having different pile toe response, with the same dead load
shear in negative as well as positive directions. applied to the pile head, can have a different depths to the
neutral plane. They will therefore experience different
• The length of the zone of transition from negative downdrag (settlement), because settlement at the two
direction shear force to positive direction is a function of neutral plane locations is different. It follows that two
the magnitude of the movement between the pile surface such piles connected to a common pile cap will have the
and the soil, or more precisely, the relative settlement same depth to the neutral plane, if the cap is sufficiently
gradient. Small relative movement will result in a long rigid to enforce that the pile head movements are the same
transition zone and large relative movement will result in a for both piles. Therefore, the two piles cannot have the
short transition zone. same dead load, but must attract different magnitude dead
• The neutral plane location is the location of the force loads, as determined by the magnitudes necessary for the
equilibrium, which is where the transition from negative to piles to develop that same location of the neutral plane. It
positive shear direction is completed. It is also the location also follows that the pile toe penetration must be of similar
where there is no relative movement between the pile and magnitude, and, inasmuch the pile toe stiffnesses are
the soil. That is, at the neutral plane, the settlement of the different, the pile toe resistances will be different in
soil and the settlement of the pile are equal, from which magnitude.
follows that the settlement of the pile head is the soil • A very thin — about a millimetre or two — coat of
settlement at the neutral plane plus the pile shortening for bitumen of wide range of viscosity will significantly
the load. reduce the shear forces along a pile surface.
• The load in a pile at the neutral plane is the maximum • The effect of the bitumen is, of course, to reduce the
load in the pile and it is the sum of the sustained load (dead negative skin friction and the drag load. It will also move
load) and the drag load. A temporary load, such as a live the neutral plane deeper and, possibly to a depth where the
load, will not add to the load at the neutral plane. The settlement is smaller and, therefore, the downdrag will be
load at the neutral plane will not change (unless the live smaller. The drag load will have changed to a lesser
load is so large as to totally eliminate the drag load, that is, degree. Note that the bitumen coat will also reduce the
the live load has to be twice the drag load at the neutral pile capacity and the factor of safety of the pile in carrying
plane, a rather unlikely design case). the applied load.

Fellenius, B.H. and Haagen, T., 1969. A new pile-force gage for
The Bäckebol test was planned and started when the author was accurate measurements of pile behavior. Canadian
with the Swedish Geotechnical Institute headed by Bengt B. Geotechnical Journal, Vol. 6, No. 3, pp. 356 - 362.
Broms. The test was performed on the initiative of Sölve Fellenius, B.H., 1971. Negative skin friction on long piles
Severinsson, Göteborg, and J. Clemont Brodeur, Canada. driven in clay. Results of full scale investigation on
Financial support was provided by the Swedish Council for instrumented piles. Swedish Geotechnical Institute,
Building Research and the Axel Johnson Institute for Industrial Proc. 25, 38 p.
Research. The author is grateful to the Swedish Geotechnical Fellenius, B.H., 1972. Downdrag on piles in clay due to
Institute for the permission to publish the Bäckebol negative skin friction. Canadian Geotechnical Journal,
measurements of 1972 through 1982. Vol. 9, No. 4, pp. 323 - 337.
Fellenius, B.H., 1975. Reduction of negative skin friction with
REFERENCES bitumen slip layers. Discussion. American Society of Civil
Bjerin, L., 1977. Påhängslaster på långa betongpålar. Swedish Engineers, ASCE Journal of the Geotechnical Engineering
Geotechnical Institute, Linköping, Report, 62 p. Division, Vol. 101, GT4, pp. 412 - 414.
Bjerrum L. and Johannessen, I.J. 1965. Measurements of the Fellenius, B.H., 1979. Downdrag on bitumen coated piles.
compression of a steel pile to rock due to settlement of the American Society of Civil Engineers, ASCE Journal of
surrounding clay. Proceedings 6th International Conference Geotechnical Engineering, Vol. 105, GT10, pp. 1262 - 1265.
on Soil Mechanics and Foundation Engineering, Montreal, Fellenius, B.H., 1984. Negative skin friction and settlement of
September 8 - 15, Vol. 2, pp. 261 - 264. piles. Proceedings of the Second International Seminar, Pile
Bjerrum L., Johannessen, I.J., and Eide, O., 1969. Reduction of Foundations, Nanyang Technological Institute, Singapore,
negative skin friction on steel piles to rock. Proceedings 7th 18 p.
International Conference on Soil Mechanics and Foundation Fellenius, B.H., 1989. Unified design of piles and pile groups.
Engineering, Mexico City, August 25 - 29, Vol. 2, pp. Transportation Research Board, Washington, TRB Record
27 - 34. 1169, pp. 75 - 82.
Bjerrum L., 1972. Embankments on soft ground. Proc. Specialty Fellenius, B.H., 2002. Determining the resistance distribution in
Conference on Performance of Earth and Earth-supported piles. Part 1: Notes on shift of no-load reading and residual
Structures, Purdue Univ., Lafayette, September, Vol. 2, load. Part 2: Method for Determining the Residual Load.
pp. 1 - 54. Geotechnical News Magazine, Vol. 20, No. 2, pp. 35 - 38,
Bozozuk, M. and Labrecque, A., 1969. Downdrag measurement and No. 3 , pp. 25 - 29.
on 270-ft composite piles. Proceedings of Symposium on Fellenius, B.H., 2004. Unified design of piled foundations with
Performance of Deep Foundations, San Francisco, emphasis on settlement analysis. "Honoring George G.
June 23 - 28, 1968, American Society for Testing and Goble — Current Practice and Future Trends in Deep
Materials, ASTM, Special Technical Publication, STP 444, Foundations" Geo-Institute Geo TRANS Conference, Los
pp. 15 - 40. Angeles, July 27-30, 2004, Edited by J.A. DiMaggio and
Bozozuk, M., 1970. Field observations of negative skin friction M.H. Hussein. ASCE Geotechnical Special Publication,
loads on long piles in marine clay. Proceedings of GSP 125, pp. 253 - 275.
Conference on Design and installation of Cellular Structures, Holtz, R.D., and Wennerstrand, J. 1972. Discussion on
Lehigh University, Bethlehem, April 13 - 15, 1970, Embankments on soft ground, Proc. Specialty Conference on
pp. 273 - 279. Performance of Earth and Earth-supported Structures, Purdue
Bozozuk, M., 1972. Downdrag measurements on a 160-ft Univ., Lafayette, September, Vol. 3, pp. 59 - 60.
floating pipe pile in marine clay. Canadian Geotechnical Indraratna, B., Balusubramanimam, A.S., Phamvan P., and
Journal, Vol. 9, No. 4, pp. 127 - 136. Wong, Y. K., 1992. Development of negative skin friction
Bozozuk, M., 1981. Bearing capacity of a pile preloaded by on driven piles in soft Bangkok clay. Canadian Geotechnical
downdrag. Proc. 10th International Conference on Soil Journal, Vol. 29, No. 3, pp. 393 - 404.
Mechanics and Foundation Engineering, Stockholm, June Janbu, N., 1963. Soil compressibility as determined by
15 - 19, 1981, Vol. 2, pp. 631 - 636. oedometer and triaxial tests. European Conference on Soil
Canadian Foundation Engineering Manual, CFEM, 1992. Third Mechanics and Foundation Engineering, Wiesbaden, Vol. 1,
Edition. Canadian Geotechnical Society, BiTech Publishers, pp. 19 - 25, and Vol. 2, pp. 17 - 21.
Vancouver, 512 p. Janbu, N., 1998. Sediment deformations. University of
Clemente, F.M., 1981. Downdrag on bitumen coated piles in a Trondheim, Norwegian University of Science and
warm climate. Proceedings of the 10th International Technology, Geotechnical Institution, Bulletin 35, 86 p.
Conference on Soil Mechanics and Foundation Engineering, Kallstenius, T and Wallgren, A., 1956. Pore water pressure
Stockholm, June 15 - 19, Vol. 2, pp. 673 - 676. measurements in field investigation. Swedish Geotechnical
Endo M., Minou, A., Kawasaki T, and Shibata, T, 1969. Institute, Proceeding No. 13, 58 p.
Negative skin friction acting on steel piles in clay. Proc. 8th Leung, C.F., Radhakrishnan, R., and Tan, S.A., 1991.
International Conference on Soil Mechanics and Foundation Performance of precast driven piles in marine clay.
Engineering, Mexico City, August 25 - 29, Vol. 2, American Society of Civil Engineers, ASCE Journal of
pp. 85 - 92. Geotechnical Engineering, Vol. 117, No. 4, pp. 637 - 657.
Fellenius, B.H. and Broms, B.B., 1969. Negative skin friction Walker, L.K., Le, P., and Darvall, L., 1973. Dragdown on
for long piles driven in clay. Proc. 7th International coated and uncoated piles. Proc. 8th International Conference
Conference on Soil Mechanics and Foundation Engineering, on Soil Mechanics and Foundation Engineering, ICSMFE,
Mexico City, August 25 - 29, Vol. 2, pp. 93 - 97. Moscow, August 1973, Vol. 2, Paper 3/41, pp. 257 - 262.