You are on page 1of 23

Danziger, B. R., Costa, A. M., Lopes, F. R. & Pacheco, M. P. (1999). Geotechnique 49, No.

6, 777799

Back analysis of offshore pile driving with an improved soil model


B. R . DA N Z I G E R ,  A . M . C O S TA , { F. R . L O P E S { a n d M . P. PAC H E C O }
Nous avons pratique des retro analyses poussees
sur les comportements sur le terrain des pilotis
offshore fonces dans des sables calcaires en
utilisant un modele de sol ameliore, developpe
par Simons. Les dossiers ont ete obtenus pendant l'installation des pilotis a tuyaux fermes
pour les plate-forme offshore du site Northeast
Pole dans le bassin de Campos, au Bresil. Ce
modele est particulierement adequat pour des
conditions de battage difciles, quand les piles
sont foncees presque jusqu'au refus. Dans ces
cas, un amortissement du a l'inertie (amortissement par rayonnement) prevaut alors que
l'amortissement visqueux est negligeable. Les
retro analyses utilisant ce modele ont permis de
calculer le deplacement relatif entre le pilotis et
le sol en divers points le long du pilotis grace a
la methode d'elements nis en une dimension.
La rigidite du sol ks et le coefcient d'amortissement cs ont ete retro calcules pour differentes
profondeurs de pilotis. Nous avons egalement
retro calcule les valeurs du rapport entre le
module de cisaillement et la resistance de l'arbre (G/) et nous offrons quelques suggestions
sur les valeurs supposees du rapport a choisir
dans les etudes des caracteristiques de battage
dans les sables calcaires.

Extensive back analyses of eld records of offshore piles driven in calcareous sands were
performed using an improved soil model developed by Simons. The records were obtained
during the installation of closed-end pipe piles
for offshore platforms at the Northeast Pole site
of the Campos Basin in Brazil. This model is
especially adequate for hard driving conditions
when piles are driven almost to refusal. In these
cases, inertial (radiation) damping prevails,
whereas viscous damping may be neglected.
Back analyses with this model allowed the relative displacements between pile and soil to be
calculated at various points along the pile in a
one-dimensional nite-element method solution.
The soil stiffness ks and the damping coefcient
cs were back-calculated for different pile depths.
Values of the ratio of the shear modulus to the
shaft resistance (G=) were also back-calculated
and suggestions are given concerning tentative
values of this ratio to be selected in drivability
studies in calcareous sands.

KEYWORDS: calcareous soils; model tests; numerical


modelling and analysis; piles.

number of theoretical and practical limitations of


Smith's (1960) model. Simons (1985) also emphasized that for an applied load giving rise to purely
elastic displacements, the soil would provide two
distinct rate-dependent components of soil reaction
to the motion of an embedded pile: one due to
material viscosity, and the other due to its inertia,
known as radiation damping. Simons (1985) pointed
out that inertial (radiation) damping is more important when nearly elastic displacement occurs. On the
other hand, for plastic displacements, the material
viscosity becomes the more important effect. Therefore, when piles are driven almost to refusal, the
rst condition prevails and radiation damping appears to be the most important form of energy loss.
For the closed-end pipe piles used in the Campos Basin, offshore from Brazil, hard driving conditions occurred at the end of pile installation and
therefore the radiation damping of Simons' (1985)
model can be considered the sole rate-dependent
component of soil reaction.

INTRODUCTION

When back-analysing eld records of pile driving,


one of the main problems is related to the adequacy
of the model adopted to represent the soilpile
interaction. The most commonly used model in
computer programs (for back analysis) was developed by Smith (1960). Limitations of this model
have been pointed out by many authors. Forehand &
Reese (1964), for instance, observed that many
combinations of quake values and damping coefcients in Smith's model could be selected to t eld
records reasonably well. Simons (1985) described a
Manuscript received 8 Oct. 1997; revised manuscript
accepted 24 June 1999.
Discussion on this paper closes 30 June 2000; for further
details see p. ii.
 Fluminense Federal University, Brazil.
{ PETROBRAS, Rio de Janeiro, Brazil.
{ Federal University of Rio de Janeiro.
} Rio de Janeiro State University.

777

778

DANZIGER, COSTA, LOPES AND PACHECO

The geotechnical conditions in the Northeast


Pole site of the Campos Basin consist mainly of
calcareous sand through which very low shaft
resistance is usually developed. A conical steel
point was added to each pile in order to increase
its toe resistance. Pile driving was monitored by
the usual method of recordings of accelerometers
and force transducers at the pile head.
Danziger et al. (1992) presented some back
analyses of the same piles using Smith's (1960)
model. In the present paper, the records examined
previously are reinterpreted, using Simons's (1985)
model. Comparisons of the results presented in this
paper with those reported before show interesting
aspects concerning the damping effects derived
from the two models.
GEOTECHNICAL CONDITIONS AND PILE
CHARACTERISTICS

A total of seven platforms were installed at the


Campos Northeast Pole site (Fig. 1), Pargo 1A

being the largest and consequently the most extensively monitored. Results from four borings with
soil sampling and one cone penetration test indicated a granular stratum throughout the whole area,
with variable cementation. The soil prole consists
of a supercial layer of silty sand (ne to medium,
with quartz grains, to a depth of about 13 m) and
layers of calcareous sand (with localized shell and
coral fragments in addition to some clay lenses) to
a depth of 120 m (McClelland, 1985). The geotechnical characteristics are summarized in Fig. 2. In
Fig. 2, the results of cone penetration tests (CPTs)
as presented by McClelland (1985) are presented as
thin lines, whereas the simplied prole given by
Geomecanica (1986) is indicated by the thick line.
Laboratory direct shear tests were performed
with soil samples prepared at both their minimum
and maximum densities. The friction angle varied
from 268 to 348 for the minimum density, and from
338 to 408 for the maximum density. The greater
values were found for the shallower depths, down
to 13 m, where the sand layer had a high quartz

At
O lant
ce ic
an

Pacific Ocean

Brazil

50

m
100
200 m

Campos

228

Northeast
Pole
Carapeba
Field

Vermelho
Field

Pargo
Field

22830

North
Pole
00

00
10

South
Pole

10

00

20

41830

418

Fig. 1. Location of construction site

40830

238
408

779

BACK ANALYSIS OF OFFSHORE PILE DRIVING

Depth:
m
0.00
2.25

Soil description

Fine to medium loose sand with quartz


and calcareous grains

Mean point resistance


q c: MN/m2
0

10

20

30

40

50

Fine to medium silty sand, dense to very


dense with clay lenses

12.50
Fine to medium silty sand, dense to very
dense with calcareous grains and clay
lenses

33.00

Interrupted test

.70
Fine to medium calcareous silty sand,
very dense

.50

.60

72.75
Fine to medium silty sand, dense to very
dense with quartz and calcareous grains

Interrupted test

.60
.60
.60

97.50
Fine to medium silty sand, dense to
very dense

.50

116.00

Fig. 2. Geotechnical characteristics at Pargo 1A site (McClelland, 1985;


Geomecanica, 1986)

content, and the lower friction angles were found


for the greater depths, where the sand layer had a
high content of calcium carbonate.
Figures 3 and 4 show the pile toe characteristics.
The choice of closed-end pipe piles resulted in a
considerable reduction in the foundation cost
(Costa et al., 1988; Mello et al., 1989).
The main piles A1, A2, B1 and B2 were driven
with an MRBS 4600 and an MRBS 5000 hammer.
The nal penetration of the piles varied from 42 m
to 45 m below the sea-bed. Part of the record
number in this paper indicates the pile penetration
below the sea-bed (e.g. record A24425 refers to
pile A2 with pile toe 4425 m below sea-bed). The
nal blow counts were close to 250 blows=25 cm
(regarded as refusal).
SIMONS'S MODEL

In Simons's (1985) model (also described in


Randolph & Simons, 1986) the theory of dynamic
elasticity is used to determine the soil stiffness and
damping coefcients. In addition, a yielding me-

chanism consistent with the physical process involved is employed.


The dynamic shear stress at the pilesoil
interface resulting from a harmonic motion u(t)
u exp(it) was given by Novak (1977) as

G
[Sw1 (a0 ) iSw2 (a0 )]u
2r0

(1)

where a0 r0 =vs is the dimensionless frequency


ratio, G is the soil shear modulus, r0 is the pile
radius, vs is the soil shear wave velocity, Sw1 and
Sw2 are functions of the frequency ratio a0 and
is the frequency.
In fact, during installation the pile is submitted
not to a harmonic motion but to a transient one,
resulting in the propagation of stress and strain
waves along its length. Clough & Penzien (1975)
obtained a solution for the wave propagation from
an analysis of modal superposition, where a nite
number of harmonic responses can be superimposed to obtain a reasonable approximation of the
wave propagation. In this approach, however, the
higher frequencies of the wave spectrum are not

780

DANZIGER, COSTA, LOPES AND PACHECO

Dia. 5 1.67

quency-independent form of the solution (equation


(1)) is then required. For that purpose, Simons
(1985) proposed that the function Sw1 and the ratio
Sw2 =a0 could be approximated by frequency-independent constants such as Sw1 2:9 and Sw2 =a0
2 (Fig. 5). Equation (1) then becomes

0.15

0.038

0.355

1.21
0.29
1.17

0.352

0.352

u(t) u exp(it)

(3)

and
@u
iu(t)
@t
equation (2) becomes


G
2r0 @u

2:9u
2r0
vs @ t

(4)

(5)

Equation (5) represents the equation of motion of a


spring and dashpot system with spring stiffness k s
and damping coefcient cs , where

0.384

0.384

0.
40

458

(2)

Since

0.355

0.355

0.075

G
(2:9 i2a0 )u
2r0

20

S w1, S w2

15

Fig. 3. Characteristics of the conical steel point (Costa


et al., 1988) (dimensions in metres)

S w2
10

S w1

incorporated into the analysis and a reasonable


match between measured and back-calculated results is generally difcult. In addition, in order to
incorporate the pile slip, the pile response must be
obtained using a direct time integration scheme,
instead of a frequency domain approach. A fre-

Fig. 4. View of the conical steel point

0.5

1.0

1.5

2.0

2.5

3.0

a0

Fig. 5. Dynamic stiffness coefcients for soils (Simons,


1985)

BACK ANALYSIS OF OFFSHORE PILE DRIVING

ks

2:9G
2r0

(6)

G
:
(Gr)0 5
vs

(7)

and
cs

Therefore, Simons (1985) assumed that the soil


resistance could be modelled by a series of such
spring and dashpot systems (Fig. 6).
When the shear stress at a point at the pilesoil
interface (equation (5)) reaches the soil yielding
stress s , the frictional bond between the pile
and the soil is broken. Then the spring and dashpot
system is disconnected from the pile and the soil
and pile displacements are calculated separately.
The soil continues to resist with its yielding stress
until the stress level reduces below the yielding
limit and the bond is restored. When the spring
dashpot system is disconnected, the pile displacement is calculated in the usual manner assuming
that the soil continues to resist at its limiting
strength. The soil displacement, on the other hand,
may be updated independently by solving the equation of motion of the springdashpot system with
the yielding stress at each time step (Simons,
1985). From equation (5) it follows that
2r0 ls K s u Cs

@u
@t

(8)

where K s and Cs represent the values of k s and cs


multiplied by 2r0 l (l being the length of the pile
element).
Equation (8) is a partial differential equation
which allows the determination of the change in
soil displacement us under the stress s over the
time interval t. The solution of equation (8) as
obtained by Simons (1985) is




s
t
t
us 2r0 l
us 1 exp K s
Ks
Cs
(9)
uts being the soil displacement at the beginning of
the interval.
Pile

mp
l

kp
Soil

ks

Cs

Fig. 6. Pile and soil displacement at the interface


(Simons, 1985)

781

As yielding proceeds, the stress which the soil


adjacent to the pile would sustain if the pile and
soil were to rejoin is continuously calculated according to Simons (1985) as


@up
1
Ks us Cs
(10)

2r0 l
@t
where us is the soil displacement and @up =@ t is
the pile velocity. When the stress given by equation
(10) falls below the yielding stress s , the pile and
soil are assumed to rejoin and the subsequent
increments in soil displacements are equal to the
increments in pile displacement.
The soil reaction to the pile motion at the toe
due to radiation damping can be modelled by the
well-known Lysmer (1965) analogue. Similarly to
the case of shaft reaction, an analytical solution is
rst determined and then adjusted to an equivalent
frequency-independent spring and dashpot system.
In the present analysis, however, the Lysmer analogue was not used, owing to the more complex
representation of the conical point. Instead, this
sharp variation of pile impedance was modelled by
masses consistently concentrated at closely spaced
nodes at the pile toe. Similarly, the toe resistance
was replaced by equivalent shaft resistances concentrated at the same nodes (by application of
equations (5)(10)). Danziger et al. (1992) showed
that this procedure allowed a good match between
measured and back-calculated velocities in the
region around the pile toe using Smith's (1960)
model.
Appendix 1 presents the equations used for the
calculation of the element area and the consistent
mass added to each pile node, for both cylindrical
and truncated conical elements. Cylindrical elements were adopted for the pile shaft and truncated
conical elements for the toe elements. The derivation of the equations was presented by Danziger
(1991) and was based on Mello (1990).
It is worth mentioning that before analysing the
data presented in this paper some analyses were
performed in order to check the assumptions made
regarding the validity of the adopted length of the
elements and of the mass distribution near the pile
toe. Well-known boundary conditions with responses calculated by the analytical formulation of
the differential wave equation were compared with
the results of the one-dimensional nite-element
method (1-D FEM) approach. The responses obtained from the FEM agreed with those obtained
from the analytical formulation, adding condence
to the assumptions made.
Simons's (1985) model has also been analysed
by other authors, e.g. Chow et al. (1988a,b), Lee
et al. (1988), Nguyen et al. (1988), Matsumoto &
Takei (1991) and Randolph & Deeks (1992). Lee
et al. (1988) included the effect of viscous damp-

782

DANZIGER, COSTA, LOPES AND PACHECO

ing, which had not been considered explicitly in


the original model. As stated before, viscous damping is not considered in this paper as radiation
damping is the most signicant portion of energy
loss for hard driving conditions.
BACK ANALYSES

Several back analyses of eld records were performed using a 1-D FEM program originally developed by Costa (1988), with the soilpile interaction
following Simons's (1985) model as implemented
by Danziger (1991). The application of Simons's
(1985) model in a 1-D FEM solution to match eld
records was carried out in a way similar to the
CAPWAP analysis (Goble et al., 1980). This implementation had the following main objectives:
(a) to implement a more realistic soil model,
applicable to hard driving conditions, where
radiation damping is believed to prevail
(b) to investigate the relative displacements between the pile shaft and the soil in the 1-D
FEM solution
(c) to back-calculate the soil stiffness k s and the
damping coefcient cs at different pile depths
(d ) to compare the mobilized soil resistance and
the resistance distribution along the pile shaft
obtained from Smith's and Simons's models for
the same eld records
(e) to compare the penetration for one blow and
the relative displacement between pile and soil
at the pile toe (rst node).
The soil parameters obtained from the back
analyses are presented in Table 1. A typical plot
showing the measured and simulated velocities at
the pile head versus time is indicated in Fig. 7 for
pile A2, with 4425 m of penetration below the
sea-bed. In Fig. 7 ll refers to the part of pile length
above the sea-bed whereas lent refers to the embedded portion of pile length. C is the stress wave
velocity. The occurrence of successive disconnections and reconnections between the pile and soil
is clearly depicted in Figs 8 and 9.
Figure 8(a) shows measured and simulated pile
head displacements for pile A2, while Figs 8(b)
and (c) indicate simulated velocities and displacements, respectively, at the pile toe. It can be seen
that the maximum toe displacement corresponds to
null velocity, when soil unloading begins. For this
eld record, the pile and soil disconnected at the
instant of maximum velocity. Up to this point
(before yielding) the pile and soil displacements
followed the same curve. When slip began, the
soil displacements became smaller than the pile
displacements, the difference being the plastic displacements at the given depth. The plastic displacements also increased up to the point of null
velocity. It can also be seen that when the pile and

soil rejoined, the subsequent increments of soil and


pile displacements were equal and both curves
followed a parallel pattern.
The plots of static and total soil reaction were
interpreted at some specic pile nodes, including
those simulating shaft and toe resistance (Fig. 9).
At the nodes simulating shaft resistance, such as
node 25 in Fig. 9(a), a large contribution of radiation damping can be inferred from the back analysis. In this case, the dominant effect of radiation
damping at the earlier stages of shaft loading
during the short time interval between points A
and B in Fig. 9(a)causes the limiting soil resistance s to be reached when the static resistance is
still lower than s . This means that the disconnection between the pile and soil was initiated at point
A, remaining effective between points A and B,
until the spring component of increasing static soil
reaction reached s at point B. At this point both
particle velocity and radiation damping are zero,
and therefore reconnection resumes. Beyond point
B the velocity changes its sign and therefore the
static and dynamic components have opposite
signs. It may happen in some nodes that reconnection starts before the particle velocity becomes
zero, i.e. before the limiting resistance s related to
those nodes is fully mobilized. This indicates that
the assigned values of s are probably too high and
should be revised.
At the conical steel point, the sudden sharpening
of the pile section causes the modelled radiation
damping to decrease linearly to zero at the toe.
Therefore, the total and static responses are coincident at node 1 in Fig. 9(b), with no contribution
from radiation damping. At node 2 (r0 0:22 m),
in spite of the very small contribution from radiation damping the total and static responses are no
longer coincident, although the difference is small
(Fig. 9(c)).
PENETRATION FOR ONE BLOW

As mentioned before, for the present case of


closed-end pipe piles driven in the Campos Basin,
the toe resistance was modelled by means of shaft
resistances concentrated at closely spaced nodes at
the pile toe (Figs 3 and 4). Since the soil behaviour
is assumed to be linear elastic in Simons's model,
the permanent displacement under hard driving
conditions can be consistently determined from the
difference between the pile displacement and the
soil displacement.
Actually, the penetration for any particular blow
at the nal stage of pile driving does not necessarily
correspond to predominantly plastic displacements;
it also incorporates a component of elastic displacement that is not fully recoverable, associated with
residual stresses. In fact, the ultimate soil displacement inferred from Fig. 8(c) indicates that the static

Table 1. Back analyses of eld records from piles A1, A2 and B2 by Simons's (1985) model
Record
quality

A1
4300
(A14300)

Good

A2
4325
(A24325)

Excellent

A2
4425
(A24425)

Excellent

B2
4200
(B24200)

Good

Bearing
capacity:
kN

Toe
resistance:
%

23 500

54

25 160

28 000

38 800

75

75

70

Damping coefcient cs : kN s=m3

Toe: 1000 (1013)


Shaft:
From 0 to 34 m:
From 34 to 415 m:
Toe: 750 (1235)
Shaft:
From 0 to 34 m:
From 34 to 417 m:
Toe: 750 (1348)
Shaft:
From 0 to 35 m:
From 35 to 427 m:
Toe: 800 (1370)
Shaft:
From 0 to 33 m:
From 33 to 405 m

Soil stiffness k s : MN=m3

10 (83)
70 (211)

Toe: 667
Shaft:
From 0 to 34 m:
From 34 to 415 m:

170 (55)
350 (197)

Toe: 991
Shaft:
From 0 to 34 m:
From 34 to 417 m:

210 (64)
350 (194)

Toe: 1182
Shaft:
From 0 to 35 m:
From 35 to 427 m:

290 (88)
350 (188)

Toe: 1220
Shaft:
From 0 to 33 m:
From 33 to 405 m

Maximum pile
displacements: mm
Top

Toe (soil)

Shaft
(elastic)

185

81
(52)

104

210

67
(52)

143

212

62
(47)

150

318

104
(61)

214

23
145

10
125

13
122

25
114

BACK ANALYSIS OF OFFSHORE PILE DRIVING

Pile depth: m
(Record)

783

784

DANZIGER, COSTA, LOPES AND PACHECO

A24425
0.7
Mobilized resistance:
(kN/m2)
Sea floor

200 400 600 800 1000

2kN/m2
20.3
Velocity: m/s

2l /c
2ll /c

2lent /c
14kN/m2

21.3
Measured velocity
Simulated velocity

3kN/m2
20kN/m2
246kN/m2 4689kN/m2

22.3
0

20

40

60

80

100

120

140

Time: ms

Fig. 7. Typical measured and simulated velocities at the pile head

component of soil resistance was not totally unloaded, indicating the presence of residual stresses.
The values of penetration for the blows backanalysed are presented in Table 2. The penetration
for one blow in the eld can be estimated from the
inverse of the number of blows for a penetration of
025 m. Although such a criterion is somewhat
arbitrary and represents only a rough approximation, the authors' experience with many instrumented high-capacity offshore piles in the Campos
Basin indicates that this simple criterion may be
useful to check for consistency of the back-calculated parameters.
The back analysed penetration in Table 2 was
determined by subtracting the maximum soil displacement from the maximum pile displacement at
the pile toe, both listed in Table 1. Except for pile
A2 at 4425 m penetration, Table 2 shows that the
penetration obtained from the eld records is greater than the back-analysed value. The penetration
obtained from the eld records is greater because
it incorporates the elastic component of soil displacement that is not fully recoverable, associated
with residual stresses.
The unrecoverable soil displacement is also depicted by the position of the dashed line at the end
of the soil displacement plot in Fig. 8(c).
RESIDUAL STRESSES

It can be inferred from the plot of axial stresses


versus time for the lowest element (Fig. 10) that
hard driving gives rise to important residual stres-

ses. In fact, a nal axial stress different from zero


at the end of the plot, after the dynamic resistance
is reducted to zero, indicates the presence of
residual stresses.
In Simons's model, the presence of residual
stresses can also be visualized by plotting the soil
displacement at the pile toe. A nal stabilized soil
displacement different from zero indicates the occurrence of residual stresses. For pile A2 (4425 m
of embedment) a plot of axial stress is presented in
Fig. 10, which indicates a nal stress at the pile
toe of about 10% of the maximum stress.
A common consequence of hard driving is the
occurrence of residual stresses superimposed from
previous blows. As a result, the mobilized resistance near the pile toe during the blow may be
lower than the ultimate soil resistance. In those
cases, according to Holloway et al. (1978), the
resistance distribution along the pile shaft inferred
from a back analysis may be different from the
real distribution.
Goble & Hery (1984) presented a new version
of the WEAP program, called CUWEAP, that
follows the procedure suggested by Holloway et al.
(1978). In this program successive blows are analysed without zeroing displacements and stresses in
the meantime. This procedure is repeated until a
convergence criterion is accomplished.
In the work described in the present paper
successive force records were not available
and, therefore, it was not possible to perform a
multiple-blow analysis similar to that suggested by
Holloway et al. (1978). In spite of this, an impor-

785

Displacement: m

BACK ANALYSIS OF OFFSHORE PILE DRIVING

20.03

Simulated displacement
Measured displacement
20.023
0

50

100

150

Time: ms
(a)

Velocity: m/s

0.00

Simulated velocity

22.00
0

50

100

150

Time: ms
(b)

Displacement: m

0.002

20.003
Disconnection
Reconnection
Soil displacement
Pile displacement
20.008
0

50

100

150

Time: ms
(c)

Fig. 8. (a) Displacement at the pile head; (b) velocity at the pile toe; (c)
pile and soil displacement at the pile toe

tant effect of residual stress in hard driving conditions was observed and was evidenced by Simons's
(1985) model, as mentioned previously.
SOIL STIFFNESS

The values of the back-calculated soil stiffness


k s related to discrete pile segments, selected approximately according to the soil stratication, are

shown in Table 1. From these results, the corresponding values of the soil shear modulus G were
determined from equation (6) as
G

k s 2r0
2:9

(11)

Table 3 and Fig. 11 show simplied proles of


back-calculated values of the shear modulus of the
soil and the corresponding mobilized shaft resis-

786

Force: kN

DANZIGER, COSTA, LOPES AND PACHECO

Radiation damping

20

Static resistance (node 25)


Total resistance (node 25)

s
A

2180

0
(Disconnection)

B (Reconnection)
50

100

Time: ms

150

Force: kN

(a)

2800

21800

Static resistance (node 1)


Total resistance (node 1)

(Disconnection) A
0

B (Reconnection)
50

100

150

Time: ms

Force: kN

(b)

2800

22800

Static resistance (node 2)


Total resistance (node 2)

(Disconnection) A
0

B (Reconnection)
50

100

150

Time: ms
(c)

Fig. 9. Static and total resistances for pile A1 at 43 m of penetration: (a)


at node 25; (b) at node 1; (c) at node 2

tance s . The values of s representing the toe


resistance were assumed to be the point ultimate
resistance divided by the surface area of the conical point.
Down to 34 m, the back-calculated values of
shaft friction were very low, different from the
CPT results. This was attributed to the reduced
resistance offered by the soil during continuous
driving. Similar behaviour is obtained in CAPWAP

analyses of piles in sandy soils, where the skin


friction mobilized during continuous driving can be
very low. Visser et al. (1985), for instance, obtained a very low soil resistance during continuous
pile driving in a dense to very dense sand layer in
the North Sea. In this case, redriving of an instrumented pile after a 24 h interval indicated a set-up
factor of 26 for the frictional resistance.
It must also be remembered that even in the

787

BACK ANALYSIS OF OFFSHORE PILE DRIVING

Table 2. Comparison of penetration for a blow inferred from back analyses with that obtained from
eld records
Record

Penetration (backanalysed): mm

Penetration (eld
records): mm

A14300
A24325
A24425
B24200

29
15
15
43

35
23
08
12

CPT there is a considerable decrease in the lateral


resistance when it is measured along the whole
length of the rod. This was the reason why
Begemann (1963) proposed measurement of the
lateral friction close to the cone with the so-called
cone sleeve.
It can be concluded from Table 3 and Figs 2
and 11 that s increases signicantly as the soil
resistance increases, whereas the ratio G=s shows
a relatively small variation, with a mean value of
166 along the shaft and 209 at the toe. Therefore,
for the purpose of preliminary drivability studies in
calcareous granular soils using Simons's (1985)
model, the shear modulus can be assumed as
G  175s

(12)

Equations (11) and (12) can be very useful to


estimate preliminary values of k s in the rst trials
of a back analysis or for drivability studies.
RADIATION DAMPING

It should be emphasized that in Simons's (1985)


model both soil stiffness and damping coefcient
depend on the shear modulus in such a way that
they are dependent variables. However, in order to
make the computed velocity at the pile head match
the measured velocity properly, it was necessary to
vary independently the soil stiffness and the damping coefcient in the back analyses. Although these
two parameters were treated as independent vari-

Table 3. Soil shear modulus G and mobilized shaft


resistance s obtained from back analyses
Record

Depth from sea


oor: m

G:
MN=m2

s :
kN=m2

G=s

A14300

00 to 340
340 to 415
415 to 430

41
262
6035

24
149
2833

171
176
213

A24325

00 to 340
340 to 417
417 to 432

18
228
8966

10
124
4213

175
184
213

A24425

00 to 350
350 to 427
427 to 442

24
221
10690

13
126
4689

185
175
228

B24200

00 to 330
330 to 405
405 to 420

46
207
11035

32
179
6064

144
116
182

 Depth corresponding to the pile toe.

ables, similar trends could be veried in their


behaviour.
The values of damping coefcient presented in
Table 1 are back-calculated ones. The values in
parentheses are related to cs calculated from equation (7) on the basis of the G values shown in
Table 3, adopting a soil density of 1:7 Mg=m3 .
Apart from pile A1 at 43 m penetration, the
three remaining records showed similar trend. The
:
cs values calculated as (rG)0 5 , on the basis of G
values from Table 3, can now be compared to the
back-calculated damping coefcients shown in
Table 1. The two values differ typically by a factor
of 030060 for the pile shaft. The differences in
behaviour between shaft and toe can be explained
by the residual stress located at the pile toe,
equilibrated by shear stresses acting downwards
along the pile shaft. The mobilized toe resistance
during a blow is underestimated and the mobilized
shaft resistance is overestimated in cases where
residual stresses are present at the pile toe. In fact,
underestimated s values at the pile toe may have

Axial stress: MN/m2

Record A24425

280

Simulated axial stress

2180
0

50

100
Time: ms

Fig. 10. Axial stress at the pile toe

150

788

DANZIGER, COSTA, LOPES AND PACHECO

Mobilized shaft resistance s: kN/m2

43.0 m
41.5 m
34.0 m

200 400 600 800 1000

10
20
30
40
Soil shear modulus G : MN/m2

50

60

(a)

43.2 m
41.7 m
34.0 m

Mobilized shaft resistance s: kN/m2


200 400 600 800 1000

10
20
30
40
Soil shear modulus G : MN/m2

50

60

(b)

Fig. 11. Simplied proles of shaft friction and shear


modulus: (a) pile A1 at 430 m penetration; (b) pile A2 at
432 m penetration; (c) pile A2 at 442 m penetration; (d) pile
B2 at 420 m penetration

789

BACK ANALYSIS OF OFFSHORE PILE DRIVING


2

44.25 m
42.7 m
35.0 m

Mobilized shaft resistance s: kN/m


200 400 600 800 1000

10
20
30
40
Soil shear modulus G : MN/m2

50

60

(c)

Mobilized shaft resistance s: kN/m2

42.0 m
40.5 m
33.0 m

200 400 600 800 1000

10

20

30

40

50
2

Soil shear modulus G : MN/m


(d)

Fig. 11. (Cont.)

60

790

DANZIGER, COSTA, LOPES AND PACHECO

been responsible for the higher G=s ratio at the


pile toe, and overestimated s values at the pile
shaft may have caused the lower G=s ratios near
the pile shaft. As a consequence, the damping
:
coefcient cs , calculated as (rG)0 5 , resulted in
high values at the pile toe when compared to the
back-calculated parameters shown in Table 1. The
opposite has been observed near the pile shaft.
Randolph & Deeks (1992) presented a eld case
in which a pipe pile behaved as unplugged during
driving. In that case, residual stresses are not supposed to occur. Furthermore, Randolph & Deeks
(1992) discuss a redriving record in which the soil
response is usually distinct from that in continuous
driving. Although a direct comparison between the
Randolph & Deeks (1992) results and those in the
present paper cannot be made, mainly because they
deal with different situations, Randolph & Deeks
(1992) found it necessary to halve (at the pile shaft)
:
the dashpot values obtained from (rG)0 5 in order to
achieve a good match between the measured and
computed velocity curves.
Randolph & Deeks (1992) mentioned that the
trend found is consistent with the comment by
Mitwally & Novak (1988) that a zone of soft soil
around a driven pile would produce a much lower
damping than a homogeneous soil. In fact there are
not enough data to enable a comprehensive evaluation of the radiation damping during continuous

driving. More data are needed from different sites


before conclusions on the overall behaviour of
radiation damping can be arrived at.
In order to examine the inuence of the choice
of the damping coefcient and soil stiffness on the
signal matches, the following analyses were carried
out (a) a sensitivity analysis for the blow on pile
A1 at 43 m and (b) a comparison of the simulated
velocities (i) using the parameters in Table 1 and
(ii) using the best consistent set of stiffness and
damping coefcient, based on equations (6) and
(7).
The sensitivity analysis
A sensitivity analysis was performed for the
blow on pile A1 at 43 m penetration. This blow
was chosen in an effort to clarify its behaviour,
which was distinct from that of the other records.
The results are summarized in Table 4 and Figs
1217. The gure numbers listed in Table 4 are
those of the gues showing the matches based on
the radiation damping and soil stiffness values
shown in the same row of the table. Fig. 12
illustrates the original match, with the parameters
also shown in Table 1.
The match in Fig. 13 indicates a lower toe
damping than that indicated in Table 1 (Fig. 12).
On the other hand, the bearing capacity and the

Table 4. Sensitivity analysis of A14300 record (pile A1 at 43 m penetration)


Figure
number

Bearing capacity:
kN

12
(original t as
in Table 1)

23 500

13

25 000

14

15

16

17

Toe resistance: Damping coefcient cs : kN s=m3


%

10
70

Toe: 667
Shaft
(83) From 0 to 34 m:
23
(211) From 34 to 415 m: 145

60

Toe: 950 (893)


Shaft:
From 0 to 34 m:
100
From 34 to 415 m: 150

Toe: 518
Shaft:
(75) From 0 to 34 m:
18
(188) From 34 to 415 m: 116

25 000

60

Toe: 950 (893)


Shaft:
From 0 to 34 m:
70
From 34 to 415 m: 100

Toe: 518
Shaft:
(75) From 0 to 34 m:
18
(188) From 34 to 415 m: 116

25 000

60

Toe: 950 (893)


Shaft:
From 0 to 34 m:
From 34 to 415 m:

50
70

Toe: 518
Shaft:
(75) From 0 to 34 m:
18
(188) From 34 to 415 m: 116

50
70

Toe 470
Shaft:
(68) From 0 to 34 m
(168) From 34 to 415 m

15
93

30
60

Toe: 470
Shaft:
(68) From 0 to 34 m
(168) From 34 to 415 m:

15
93

25 000

25 000

54

60

60

Toe: 1000 (1013)


Shaft:
From 0 to 34 m:
From 34 to 415 m:

Soil stiffness k s : MN=m3

Toe: 950 (850)


Shaft:
From 0 to 34 m:
From 34 to 415 m:
Toe: 950 (850)
Shaft:
From 0 to 34 m:
From 34 to 415 m:

791

BACK ANALYSIS OF OFFSHORE PILE DRIVING

1.0

Velocity: m/s

0.5
0.0
20.5
21.0

Measured velocity
Simulated velocity

21.5
22.0
0

10

20

30

40

50

60

70

80

90

100

110

120

Time: ms

Fig. 12. Measured and simulated velocities at the pile head for pile A1 at
43 m penetration. Match corresponding to data in Table 1, also shown in
Table 5
1.0

Velocity: m/s

0.5
0.0
20.5
21.0
Measured velocity
Simulated velocity

21.5
22.0
0

10

20

30

40

50

60

70

80

90

100

110

120

Time: ms

Fig. 13. Measured and simulated velocities at the pile head for pile A1 at
43 m penetration. Match corresponding to data shown in Table 5

percentage of toe resistance are somewhat higher.


The match is quite good up to 2l=c; after this time
the simulated velocity needs to be reduced in order
to t the measured velocity. Fig. 14 shows the
same blow with a reduced damping coefcient
along the pile shaft. A better agreement was found
after 2l=c but the need to reduce the damping
coefcient is still apparent, to a lesser extent. Fig.
15 indicates a fairly good t after wave reection
at the pile toe but the simulated velocity now
seems to be higher than the measured velocity near
the pile toe. A reduction in stiffnes can have an
effect in this direction, as shown in Fig. 16. In fact
this last measure helped to t the record in the
vicinity of 2l=c but also produced an effect after
that time. An additional reduction in the damping

coefcient along the pile shaft produced a better t


as shown in Fig. 17.
The data shown in Table 4 suggest that the t
presented in Table 1 at 43 m penetration could be
replaced by the one related to Fig. 17. In this case,
the soil response would not be so different from
that of the remaining records.
The comparative analysis
A comparison between the simulated velocities
(a) using the parameters in Table 1 and (b) using
the best consistent set of stiffness and damping,
based on equations (6) and (7), was made. The
comparison was performed for piles A2 at 4325 m
penetration and B2 at 42 m penetration.

792

DANZIGER, COSTA, LOPES AND PACHECO

1.0

Velocity: m/s

0.5
0.0
20.5
21.0

Measured velocity
Simulated velocity

21.5
22.0
0

10

20

30

40

50

60

70

80

90

100 110

120

Time: ms

Fig. 14. Measured and simulated velocities at the pile head for pile A1 at
43 m penetration. Match corresponding to data shown in Table 5

1.0

Velocity: m/s

0.5
0.0
20.5
21.0

Measured velocity
Simulated velocity

21.5
22.0
0

10

20

30

40

50

60

70

80

90

100

110

120

Time: ms

Fig. 15. Measured and simulated velocities at the pile head for pile A1 at
43 m penetration. Match corresponding to data shown in Table 5

The consistent set of stiffness and damping coefcients was based on equations (6) and (7),
adopting a soil shear modulus G taken as an
average of those indicated in Table 3, and the
actual soil density of 1:7 Mg=m3 . The resulting
parameters are indicated in Table 5.
Figures 18 and 19 show the measured and
simulated velocities at the pile head for pile A2 at
4325 m penetration and for pile B2 at 42 m penetration, respectively. Figs 18(a) and 19(a) indicate
the simulated velocities related to k s and cs from
Table 1, whereas Figs 18(b) and 19(b) indicate the
simulated velocities related to the consistent set of
stiffness and damping coefcients based on the soil
shear modulus G and equations (6) and (7).

From Figs 18 and 19 it can be concluded that


although a better t between the measured and
simulated velocities is obtained when both the
stiffness and the damping coefcient are varied
independently, as indicated in Figs 18(a) and 19(a),
a reasonable agreement can also be found when
adopting a consistent set of stiffness and damping
coefcients, as indicated in Figs 18(b) and 19(b).
Figs 20 and 21 compare on a single plot for each
pile the simulated velocity based on k s and cs from
Table 1 and the simulated velocity based on the
consistent set of stiffness and damping coefcients.
Figs 20 and 21 indicate that the differences are not
signicant and are mainly concentrated near the
pile toe, where signicant residual stresses are

793

BACK ANALYSIS OF OFFSHORE PILE DRIVING

1.0

Velocity: m/s

0.5
0.0
20.5
21.0
Measured velocity
Simulated velocity

21.5
22.0
0

10

20

30

40

50

60

70

80

90

100

110

120

Time: ms

Fig. 16. Measured and simulated velocities at the pile head for pile A1 at
43 m penetration. Match corresponding to data shown in Table 5

1.0

Velocity: m/s

0.5
0.0
20.5
21.0

Measured velocity
Simulated velocity

21.5
22.0
0

10

20

30

40

50

60

70

80

90

100 110

120

Time: ms

Fig. 17. Measured and simulated velocities at the pile head for pile A1 at
43 m penetration. Match corresponding to data shown in Table 5

Table 5. Consistent set of stiffness and damping coefcients based on soil shear
modulus G
Depth from sea
oor

Soil shear modulus:


MN=m2

Soil stiffness
k s : MN=m3

Damping coefcient
cs : kN s=m3

00 to 340
340 to 415
Pile toe

32
230
9182

18
128
10149

738
1977
12494

supposed to occur. It should also be emphasized


that both analyses were performed using the same
resistance distribution along the pile shaft and toe,
as indicated in Table 1. Small variations in resistance distribution were found to be much more
signicant in signal matching than the differences

resulting from using the consistent set of stiffness


and damping coefcients or the stiffness and
damping coefcients from Table 1.
Danziger et al. (1996) presented a discussion on
the uniqueness of the CAPWAP-type analysis
based on Smith's model. The authors suggested

794

DANZIGER, COSTA, LOPES AND PACHECO

1.0

Velocity: m/s

0.5
0.0
20.5
21.0
21.5

Measured velocity
Simulated velocity

22.0
0

10

20

30

40

50

60

70

80

90

100

110

120

Time: ms
(a)

1.0

Velocity: m/s

0.5
0.0
20.5
21.0
21.5

Measured velocity
Simulated velocity

22.0
0

10

20

30

40

50

60
70
Time: ms

80

90

100

110

120

(b)

Fig. 18. Measured and simulated velocities at the pile head for pile A2 at
4325 m penetration: (a) simulated velocity related to ks and cs from Table
1; (b) simulated velocity related to consistent set of ks and cs

some useful procedures to arrive at a unique solution. The authors' suggestions were based on a
considerable number of back analyses using
Smith's model. Unfortunately, Simons's (1985)
model has not been extensively applied so far and
a similar set of procedures could not be proposed
for using this model. Extensive analyses need to be
performed before any conclusions can be drawn
concerning the quality of the match obtained using
Simons' model.
The need to treat the soil stiffness and damping
coefcient as independent variables in order to
produce better signal matches can only be claried
when other documented cases of large closedended piles become available. Gathering more data
from different sites will enable verication of
whether the trend found in the present paper is a

particular feature of closed-end pipe piles at the


end of continuous driving in calcareous sands or
can be found in other situations.
BEARING CAPACITY

Table 6 was prepared from the results of soil


resistance distribution obtained from back analyses.
This table presents, for each eld record, both the
toe and the shaft resistance during a single blow;
the mean, maximum and minimum shaft resistance
mobilized along the pile length; the mean shaft
resistance mobilized in both layers of calcareous
sand; and the mobilized toe resistance. The mobilized toe resistance was estimated as the toe ultimate resistance divided by the lateral area of the
conical point. In the same column, in parentheses,

795

BACK ANALYSIS OF OFFSHORE PILE DRIVING

1.0
0.5

Velocity: m/s

0.0
20.5
21.0
21.5
22.0
Measured velocity

22.5

Simulated velocity

23.0
0

10

20

30

40

50

60

70

80

90

100

110 120

Time: ms
(a)

1.0
0.5

Velocity: m/s

0.0
20.5
21.0
21.5
22.0
Measured velocity

22.5

Simulated velocity

23.0
0

10

20

30

40

50

60

70

80

90

100

110

120

Time: ms
(b)

Fig. 19. Measured and simulated velocities at the pile head for pile B2 at
42 m penetration: (a) simulation velocity related to ks and cs from Table 1;
(b) simulated velocity related to consistent set of ks and cs

the unit toe resistance, calculated as the toe ultimate resistance divided by the cross-sectional area
at the top of the cone, is also indicated.
From this table the following observations can
be made.
(a) The mean shaft resistance is around 40 kN=m2 .
(b) The larger values of shaft resistance are concentrated near the pile toe, as expected. In the
remainder of the shaft and, mainly, close to the
sea oor, the shaft resistance shows low values.
(c) For the rst layer of calcareous sand, from 13
to 33 m depth, the mean shaft friction mobilized according to Simons's model is about
21 kN=m2 .
(d ) For the second layer of calcareous sand, from
33 m to the pile toe, the mean shaft friction

mobilized according to Simons's model is


about 127 kN=m2 .
(e) The end-bearing stress values, qp , are very low
in relation to the cone resistance. This may be
due to the following reasons: (i) scale effect, a
signicant factor for a 167 m dia. pile; (ii) the
lack of accounting for residual stresses; (iii)
the fact that the end-bearing stress qp actually
mobilized during pile monitoring may be lower
than the end-bearing capacity, mainly in hard
driving conditions when small penetrations
occur. In fact, redriving some piles with a
heavier hammer indicated a higher toe resistance than that back-analysed at the end of
driving (Danziger et al., 1992).
A comparison between the results in Table 6

796

DANZIGER, COSTA, LOPES AND PACHECO

1.0

Velocity: m/s

0.5
0.0
20.5
21.0
21.5

Parameters from Table 1


Consistent set of parameters

22.0
0

10

20

30

40

50

60 70
Time: ms

80

90

100 110 120

Fig. 20. Simulated velocities at the pile head for pile A2 at 4325 m
penetration

1.0
0.5

Velocity: m/s

0.0
20.5
21.0
21.5
22.0
Parameters from Table 1

22.5

Consistent set of parameters

23.0
0

10

20

30

40

50

60

70

80

90

100 110 120

Time: ms

Fig. 21. Simulated velocities at the pile head for pile B2 at 42 m


penetration

Table 6. Mobilized resistances during continuous driving


Pile

Depth:
m

Toe
resistance:
kN

Skin
resistance:
kN

s mean :
kN=m2

s max :
kN=m2

s min :
kN=m2

s
13 to 33 m:
kN=m2

s
33 m to pile
toe: kN=m3

qp :
kN=m2 

A1

4300

12 690

10 810

47

175

23

141

A2

4325

18 870

6 290

28

225

10

113

A2

4425

21 000

7 000

30

246

13

105

B2

4200

27 000

11 800

53

352

20

38

150

2 833
(5 795)
4 213
(8 616)
4 689
(9 589)
6 064
(12 402)

 Unit toe resistance estimated as the toe ultimate resistance divided by the lateral area of the conical point (values in
parentheses: unit toe resistance estimated as the toe ultimate resistance divided by the cross-sectional area at the top of
the cone).

BACK ANALYSIS OF OFFSHORE PILE DRIVING

and similar results obtained with Smith's model


(described by Danziger et al., 1992) leads to the
following conclusions.
(a) The ultimate resistance back-calculated with
Simons's model exceeded that obtained with
Smith's model by some 15%.
(b) The mean shaft friction mobilized with Simons's model was nearly the same as that
obtained with Smith's model.
(c) As far as the toe resistance is concerned, the
differences between the two models reached
35%, Simons's ultimate resistance exceeding
Smith's.
A reason why Simons's model gives different
results from Smith's near the pile toe may be the
effect of viscous damping, which is not considered
explicitly by Simons (1985). Although the most
signicant portion of the energy loss for hard
driving conditions is due to radiation damping, the
viscous damping still plays a minor role. Its effect,
which is usually greater at the pile toe, may have
been incorporated into the static resistance, leading to a Simons's ultimate resistance exceeding
Smith's.

(d )

(e)

( f)

CONCLUSIONS

The paper presents results of back analyses of


eld records of offshore closed-end pipe piles
driven in calcareous sands, performed with an improved soil model proposed by Simons (1985). The
reasons for implementing this model were the
consistency in its formulation, its adequacy for
hard driving conditions and the possibility of following the relative displacement between the pile
and the soil at the pile shaft. At the same time the
model preserves the simplicity of Smith's (1960)
approach.
The main conclusions drawn in the paper are
the following.
(a) The process of matching the predicted behaviour with eld records presents no additional
difculties when compared with the use of
Smith's model.
(b) Although in Simons's model the soil stiffness
and the damping coefcient depend on the
shear modulus in such a way that they are
dependent variables, a better agreement between simulated and measured velocities was
found in the present analysis when both
parameters varied independently. However,
when adopting the best consistent set of
stiffness and damping coefcients as dependent variables, reasonable agreement between
simulated and measured velocities was also
obtained.
(c) Back calculated values of the ratio of the shear

( g)

797

modulus to the shaft resistance (G=s ) indicated similar values for depths along the shaft
and near the pile toe, around (G=s ) 175.
For drivability studies involving calcareous
sands, in the absence of a soil shear modulus
prole, the authors suggest the following
procedure: rst estimate s for each foundation
depth; the value of the shear modulus is then
obtained by taking G 175s . On the basis of
the shear modulus, the value of the soil
stiffness and a rst indication of the damping
coefcient can be determined.
A comparison of the pile capacities obtained
from Smith's (1960) model and Simons's
(1985) model showed that the ultimate resistance back-calculated with Simons's model
exceeded that obtained with Smith's model
by some 15%; the mean mobilized shaft
friction was nearly the same and the difference
in toe resistance was about 35%, Simons's
value exceeding Smith's.
For hard driving conditions, where elastic
displacements are believed to prevail, radiation
damping appears to be the most appropriate
form of energy loss to consider. In these cases,
Simons's model should be preferred. On the
other hand, for easy driving, where plastic
displacements are dominant, viscous damping
is perhaps the most appropriate form of
dynamic soil reaction.
More data need to be gathered from different
sites to verify whether the trend found in the
present paper is a particular feature of closedend pipe piles at the end of continuous driving
in calcareous sands or can be found in other
situations.

ACKNOWLEDGEMENTS

The authors are grateful to PETROBRAS (the


Brazilian State Oil Company) for allowing the
publication of the data, and to CNPq (Brazilian
Research Council) for nancial support to the
senior author.

APPENDIX 1. CYLINDRICAL AND TRUNCATED


CONICAL ELEMENTS

Cylindrical element

The element area A is


A e( e)
where e and are dened in Fig. 22.
The consistent mass matrix is
2
3
1
AL 6 1 2 7
M r
4
5
3 1 1
2
The consistent mass added to the nodes is

(13)

(14)

798

DANZIGER, COSTA, LOPES AND PACHECO

The consistent mass added to the node corresponding to


the diamter 1 is
rel
[4(1 e) (2 1 )]
(18)
12
The consistent mass added to the node corresponding to
the diameter 2 is
m1

m2

Fig. 22. Denition of and e

mr

AL
3

(15)

Truncated cone element


The element area A is
A

e(2 1 )
ln[(2 e)=(1 e)]

(16)

where 1 and 2 are dened in Fig. 23.


The consistent mass matrix is
rel
M
12

4(1 e) (2 1 )
2(1 e) (2 1 )

2(1 e) (2 1 )
4(1 e) 3(2 1 )


(17)

2
1

Fig. 23. Denition of 1 and 2

rel
[4(1 e) 3(2 1 )]
12

(19)

REFERENCES
Begemann, H. K. S. (1963). The friction jacket cone as
an aid in determining the soil prole. Proc. 6th Int.
Conf. Soil Mech. Found. Engng, Montreal 1, 1720.
Chow, Y. K., Karunaratne, G. P., Wong, K. Y. & Lee,
S. L. (1988a). Prediction of load-carrying capacity of
driven piles. Can. Geotech. J. 25, 1323.
Chow, Y. K., Wong, K. Y., Karunaratne, G. P. & Lee,
S. L. (1988b). Wave equation analysis of pilesa
rational theoretical approach. Proc. 3rd Int. Conf.
Applicat. Stress Wave Theory to Piles, Ottawa,
208218.
Clough, R. W. & Penzien, J. (1975). Dynamics of structures. McGraw-Hill, New York.
Costa, A. M. (1988). DINEXP-1D Program. Rio de
Janeiro: Cenpes (Research Centre), PETROBAS.
Costa, A. M., Moreira, L. F. R., Ebecken, N. F. F.,
Coutinho, A. L. G. A., Landau, L. & Alves, J. L. D.
(1988). Recent application of computer methods for
drivability analysis of offshore piles in Brazil. Proceedings of the International Conference on Computer
Modelling in Ocean Engineering, Venice, pp.
665672.
Danziger, B. R. (1991). Dynamic analysis of pile driving.
DSc thesis, COPPE/Federal University of Rio de
Janeiro (in Portuguese).
Danziger, B. R., Pacheco, M. P., Costa, A. M. & Lopes,
F. R. (1992). Back-analyses of closed-end pipe piles
for an offshore platform. Proc. 4th Int. Conf. Applicat. Stress Wave Theory to Piles, The Hague,
557562.
Danziger, B. R., Lopes, F. R., Costa, A. M. & Pacheco,
M. P. (1996). A discussion on the uniqueness of
CAPWAP analyses. Proc. 5th Int. Conf. Applicat.
Stress Wave Theory to Piles, Orlando, 394408:
Forehand, P. W. & Reese, J. L. (1964). Prediction of pile
capacity by the wave-equation. J. Soil Mech. Found.
Div., ASCE 90, 125.
Geomecanica (1986). Evaluation of soil properties at the
Pargo SiteCampos Basin, Report RJ-3274/007. Geomecanica (in Portuguese). Rio de Janeiro, Brazil.
Goble, G. G. & Hery, P. (1984). Inuence of residual forces
on pile driveability. Proc. 1st Int. Conf. Applicat. Stress
Wave Theory on Piles, Stockholm, 131161.
Goble, G. G., Raushe, F. & Likins, G. E. (1980). The
analysis of pile drivinga state-of-the-art report.
Proc. 2nd Int. Conf. Applicat. Stress Wave Theory on
Piles, Stockholm, 154161.
Holloway, D. M., Clough, G. M. & Vesic, A. S. (1978).
The effects of residual driving stresses on pile performance under axial loads. Proceedings of the Offshore Technology Conference, OTC 3306, Houston,
22252236.

BACK ANALYSIS OF OFFSHORE PILE DRIVING

Lee, S. L., Chow, Y. R. & Karunaratne, G. P. (1988).


Rational wave equation model for pile driving analysis. J. Geotech. Engng Div., ASCE 114, No. 3,
306325.
Lysmer, J. (1965). Vertical motion of rigid footings,
Report 3-115. Vicksburg, MS: US Army Corps of
Engineers, Waterways Experiment Station.
Matsumoto, T. & Takei, M. (1991). Effects of soil plug
on behaviour of driven piles. Soils Found. 31, No. 2,
1434.
McClelland, B. (1985). Geotechnical site investigation,
preliminary eld report, Pargo Site, Campos Basin,
Job No. 0184-1237.
Mello, J. R. C., Amaral, C., Costa, A. M., Rosas, M. M..
Coelho, P. S. D. & Porto, E. C. (1989). Closed-ended
pipe piles: testing and piling in calcareous sand.
Proceedings of the Offshore Technology Conference,
OTC 6000, Houston, 95106.
Mello, L. F. N. (1990). Internal report on the conical pile
base model. Rio de Janeiro: Cenpes (Research Centre), PETROBRAS (in Portuguese).
Mitwally, H. & Novak, M. (1988). Pile driving analysis
using shaft models and FEM. Proc. 3rd Int. Conf.
Applicat. Stress Wave Theory to Piles, Ottawa, 455466.

799

Nguyen, T. T., Berggren, B. & Hansbo, S. (1988). A new


soil model for pile driving and driveability analysis.
Proc. 3rd Int. Conf. Applicat. Stress Wave Theory to
Piles, Ottawa, 353367.
Novak, M. (1977). Vertical vibration of oating piles.
J. Engng Mech. Div., ASCE EM1, 153168.
Randolph, M. F. & Deeks, A. J. (1992). Keynote lecture:
dynamic and static soil models for axial pile response.
Proc. 4th Conf. Applicat. Stress Wave Theory to Piles,
The Hague, 314.
Randolph, M. F. & Simons, H. A. (1986). An improved
soil model for one dimensional pile driving analysis.
Proc. 3rd Int. Conf. Numer. Meth. Offshore Piling,
Nantes, 117.
Simons, H. A. (1985). A theoretical study of pile driving.
PhD thesis, University of Cambridge.
Smith, E. A. L. (1960). Pile driving analysis by the waveequation. J. Soil Mech. Found. Div., ASCE 86, No. 4,
3561.
Visser, M., Van der Zwaag, G. L. & Pluimgraaff, D. J. H.
M. (1985). Application of monitoring systems during
pile driving in North Sea sands. Proc. 4th Int. Conf.
Behaviour of Offshore Structures, Amsterdam,
623631.

You might also like