Marine Structures 32 (2013) 113–135

Contents lists available at SciVerse ScienceDirect

Marine Structures
journal homepage: www.elsevier.com/locate/
marstruc

Structural mechanics for design of grouted

connections in monopile wind turbine structures
Inge Lotsberg*
DNV, Veritasveien 1, 1322 Oslo, Norway

a r t i c l e i n f o a b s t r a c t

Article history: Over the last 30 years it has become standard practice to connect
Received 19 August 2012 offshore oil and gas structures to their foundation piles using cy-
Received in revised form 1 March 2013 lindrical shaped grouted connections with shear keys or weld
Accepted 4 March 2013
beads. Circumferential shear keys, or weld beads, are provided
around the outside of the piles and the inside of the pile sleeves in
Keywords:
jacket structures to transfer forces through the grouted connec-
Grouted connections
tion. The same methodology is also now being used by the wind
Shear keys
Axial capacity energy industry to connect wind turbine support structures to
Bending moment their foundation piles. These structures are subjected to rather
Dynamic loading severe dynamic loading, it is therefore important to document the
Wind turbine structures fatigue capacity of these grouted connections. As a direct result of
Monopiles this need, a joint industry project focusing on the capacity of cy-
Transition pieces and offshore structures lindrical shaped grouted connections with shear keys was initiated
by DNV in January 2011 and continued through until completion in
May 2012. This project has involved fatigue testing of grouted test
specimens in the laboratory, finite element analyses and assess-
ment and development of a recommended design methodology.
The design methodology includes the Ultimate Limit State and the
Fatigue Limit State. Fatigue testing of full-scale specimens would
require very large test setup and hydraulic actuators. Therefore
special box specimens were designed with a representative radial
stiffness similar to that of large diameter connections, with both
full size grout thickness and geometry of the shear keys. An
analytical approach for design of these specimens and for the
design of grouted connections in monopiles is presented in this
paper.
Ó 2013 Elsevier Ltd. All rights reserved.

* Tel.: þ47 93429833.

E-mail address: [email protected].

0951-8339/$ – see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.marstruc.2013.03.001
114 I. Lotsberg / Marine Structures 32 (2013) 113–135

Definition of symbols

D shell stiffness
Dp shell stiffness of pile
DTP shell stiffness of transition piece
E modulus of elasticity for steel
Fh force per unit length of circumference in radial direction
Fv force per unit length of circumference acting in vertical direction
L total length of grouted section
Lg effective length of grouted section ¼ L 2tg
Mtot global bending moment on grouted connection
M0 bending moment over wall thickness per unit length of circumference
Rp outer radius of pile
RTP outer radius of transition piece
Q0 shear force over wall thickness per unit length of circumference
h height of shear keys
k radial stiffness parameter
kv stiffness in vertical direction for one shear key connection per unit circumferential
length
kvn stiffness in vertical direction for n shear key connections per unit circumferential length
kTP stiffness per unit circumferential length of the transition piece from top of pile to centre
of reaction force from the shear keys
le elastic length
lep elastic length of the monopile
leTP elastic length of the transition piece
m ratio elasticity module steel to grout
n number of effective shear keys on each side of the grouted connection
p nominal contact pressure in radial direction between grout and steel
r radius of shell
s distance between centres of shear keys
seff effective distance between shear keys
t thickness of shell
tg thickness of grout
tp thickness of pile
tTP thickness of transition piece
w width of shear keys
dh displacement in the horizontal direction
dv displacement in the vertical direction at top of pile
dv0 maximum relative vertical displacement around circumference of transition piece at
centre region with shear keys
4 angle
n Poisson’s ratio for steel
m grout to steel interface coefficient of friction
 I. Lotsberg / Marine Structures 32 (2013) 113–135 115

1. Introduction

Cylindrical shaped grouted connections have a positive track record stretching back over many
years in jacket structures in the oil and gas industry. On older jacket structures the piles were driven
through the jacket legs and then welded to the top of the legs. Grouting was performed to fill the space
between the outer surface of the pile and the inner surface of the leg. Shear keys were not required in
these structures to achieve sufficient structural capacity. This methodology was used for the first
platforms installed in the North Sea in the seventies. In older design recommendations from the
American Petroleum Institute [1] and UK [6]; there was no explicit guidance on how to determine the
capacity when mechanical shear keys were used. However, both these recommendations did open for
the use of shear keys.
At the end of the seventies a number of tests on the capacity of grouted connections were per-
formed at Wimpey laboratories in London [4,5]. A design equation accounting for the capacity with
shear keys was developed. This design methodology was used in the design of the pile to sleeve
connections in the Magnus platform and it was concluded that a significant amount of steel was saved
[15]. This design methodology became more formalised by the issue of Amendment No. 4 to the UK
Department of Energy’s guideline in 1982 where the design equation from Billington and Tebett was
included. An assessment of the capacity of grouted connections was also performed within API [14].
This resulted in new design recommendations in [2]. Also some joint industry projects were performed
by DNV at end of the seventies and beginning of eighties investigating the capacity of grouted con-
nections, e.g. [16] and [12]. The second reference is a summary paper of work performed in the
aforementioned joint industry projects.
During the 1990s the content of ISO 19902 was developed [10]. This standard was issued in
2007. Even if the safety format in this standard looks different from that of UK Department of
Energy [7] and HSE UK [11] it provides similar or somewhat larger interface shear capacity for
axial load depending on grout strength and radial stiffness. It is noted that the allowed geometry
of the shear keys in terms of shear key height divided by distance between the shear keys is
increased from 0.04 in [11] to 0.10 in [13]. The latest version of API RP2A [3] includes the same
formulation as in API RP2A [2] with respect to capacity of cylindrical shaped connections with
shear keys.
A grouted connection is used to connect the transition piece to the monopile in wind turbine
structures as indicated in Fig. 1. The transition piece is installed on top of the monopile resting on
temporary supports. The transition piece is then jacked up to the correct verticality before the
grouting is carried out. After curing, the jacks are removed leaving a gap between the supports
and the monopile. The influence from dynamic moment seen in large diameter monopile con-
nections in the wind industry is significantly larger than the axial force. Reference is made to
[17,21,23].
Some test results from plain cylindrical shaped grouted connections performed at the university
of Aalborg around 2000 indicated that sufficient capacity could be achieved without the use of
shear keys in these connections. In addition, due to significant dynamic moments in these struc-
tures the use of shear keys was questioned from a fatigue point of view. During 2009 the axial
capacity of large diameter plain grouted connections became questioned. Based on uncertainty
related to capacity of these connections, a joint industry project was initiated regarding the capacity
of large diameter grouted connections in offshore wind turbine structures (for plain connections
without shear keys). This project was started in November 2009 and finished in January 2011. A
design methodology based on grouted conical connections was established in this project [18].
During early 2010 it was confirmed by inspection that a number of wind farms with plain grouted
connections had settled such that the temporary supports were again resting on the top of the
monopiles (Fig. 1) and the force flow through the structures was different from that assumed in
design.
In January 2011 another joint industry project with an objective to provide a sound data basis and to
establish a reliable design methodology for cylindrical shaped grouted connections with shear keys
subjected to alternating loading was started [19].
116 I. Lotsberg / Marine Structures 32 (2013) 113–135

Fig. 1. Principle of grouted connection in monopile structure.

2. Plain cylindrical shaped grouted connections (without shear keys)

2.1. Nominal contact pressure

The axial stresses in the transition piece and the monopile due to bending moment loading are
usually much larger in a wind turbine structure than that from the axial permanent load from the
weight of the structure above the connection. The moment action will lead to a tension load in the
circumferential direction of the grouted connection that may exceed the tensile capacity of the grout.
This may lead to cracking of the grout in radial and vertical sections as indicated in Fig. 2. Due to the
relatively high local slenderness (diameter to thickness ratio) of the pile and the transition piece,
ovalisation of the cylinders will also occur and a gap will open up between the grout and the steel in
the case of large moments. This will lead to a relative sliding between the steel and grout. One can
thus argue that the main purpose of the grout is to transfer pressure from the transition piece to
the pile.
The moment is transferred from the transition piece to the monopile through horizontal contact
forces as indicated in the sketch in Fig. 3. There will also be vertical friction forces due, to the contact

Fig. 2. Behaviour of grouted connection subjected to a large bending moment.

 I. Lotsberg / Marine Structures 32 (2013) 113–135 117

Fig. 3. Illustration of reaction forces in the grouted connection due to moment loading.

pressure, which contribute to the moment capacity of the grouted connection (red arrows in Fig. 3).
Based on performed tests it is found that a characteristic friction coefficient value of 0.40 between steel
and grout can also be applied to grouted connections subjected to long-term sliding. As long as there is
friction force between the steel and grout due to contact pressure, there will also be vertical friction
forces due to the surface irregularity (or fabrication tolerances) in the connection (black arrows in
Fig. 3). This effect is not recommended for use in design; however, it must be kept in mind when
assessing laboratory test data. If shear keys are installed around the circumference of the monopile and
the transition piece, these shear keys will also transfer vertical shear forces, contributing to the
moment capacity of the grouted connection; however, this first phase of the project did not consider or
test connections with shear keys.
In the following, connections without shear keys are considered. The contact pressure shown in Fig. 3
will act around most parts of the circumference. This contact pressure will provide some horizontal shear
resistance due to the friction between the steel and grout. The horizontal shear forces shown in Fig. 3 will
also contribute to the moment capacity of the grouted connection (green arrows in Fig. 3).
The actual behaviour of the grouted connections subjected to a bending moment may be simulated
by a finite element analysis that accounts for compressive contact between the steel and the grout, but
without transferring tensile stresses, and with a proper friction coefficient where contact pressure is
present.
118 I. Lotsberg / Marine Structures 32 (2013) 113–135

An analytical expression of the relationship between the contact pressure and bending moment
acting on the grouted connection can be derived based on certain assumptions concerning pressure
distribution as indicated in Fig. 3. The moment resistance due to contact pressure (blue arrows in Fig. 3)
can be derived by integrating the contact pressure and area times its moment arm around the
circumference as

Rp L2g
Mp ¼ p (1)
3

where p is the maximum nominal pressure at the top and bottom of the grouted section as shown in
Fig. 3; Rp, outer radius of pile; Lg, effective height of grouted section.
Reference is made to Appendix A for a more detailed derivation of this equation.
The moment resistance due to the horizontal friction force (green arrows in Fig. 3) can be derived by
integrating the contact pressure and area times the moment arm within the stipulated green line from
a to c in Fig. 3 with pressure 0.75p at positions b and d as

Rp L2g
Mmh ¼ mp (2)
p
Similarly the moment resistance due to the vertical friction force (red arrows in Fig. 3) can be
derived by integrating the contact pressure and area times the moment arm outside the stipulated line
from a to c in Fig. 3 with pressure 0.5p at positions b and d as

Mmv ¼ mpR2p Lg (3)

The pressure used to derive the equations for moments is based on an assessment that a full friction
cannot be obtained in two orthogonal directions. This assessment was also based on comparison with
the results of a finite element analysis. Reference is made to Appendix A for further details.
Moment equilibrium with the external bending moment gives

Mtot ¼ Mp þ Mmh þ Mmv (4)

In addition to these moments, there is a friction moment due to surface irregularity which is
neglected for large diameter connections. From Eqs. (1)–(4) an estimate of the maximum nominal
contact pressure is derived as

3pMtot
p ¼
 (5)
Rp L2g p þ 3m þ 3pmR2p Lg

The contribution to the nominal contact pressure from the global shear force at the grouted
connection in a monopile is considered to be small and is neglected in the following.

2.2. Opening between the steel and the grout at the top and bottom of the grouted connection

An estimate of the maximum opening between the steel and the grout at the top and bottom of the
grouted connection is derived from radial flexibility of the pile and the transition piece in addition to
geometric considerations as

 
dh ¼ 3 dTP þ dp (6)

where dp is the radial decrease in pile diameter for contact pressure p; dTP is the radial increase in
diameter of the transition piece for contact pressure p.
Here, the compression of the grout is neglected as the contribution from this to the total defor-
mation is small for typical wind turbine connections. The change in radius due to contact pressure for
the pile and the transition piece is derived as
 I. Lotsberg / Marine Structures 32 (2013) 113–135 119

pR2TP
dTP ¼
EtTP
(7)
pR2p
dp ¼
Etp
The total radial displacement dh is derived as the sum of displacements from the transition piece
and the pile as illustrated in Fig. 4a. For moment loading the transition piece will move on the pile until
contact around the circumference is achieved as illustrated through steps shown in Fig. 4b and c. The
last step until position c implies some ovalisation of the transition piece.
Furthermore based on a geometrical consideration the vertical relative displacement between the
transition pieces and the pile is derived as

2Rp
dv ¼ dh (8)
Lg
The basis for this equation is that a small rotation of the transition piece about a centre of the
grouted connection gives a horizontal displacement at top of pile equal to dh and a vertical displace-
ment equal to dv at a distance Rp from the rotation centre.
The derived analytical approach has been compared with numerical analysis results using the
ABAQUS program for a full size model with pile diameter equal 5000 mm.
A significant sliding between the steel and grout is expected to occur in the case of a combination of
a flexible transition piece (large diameter to thickness ratio) and large bending moments even with a
high friction coefficient. Thus, it is not realistic to try to improve the structural behaviour of the large
diameter connections subjected to dynamic bending moments by increasing the roughness of the steel
surfaces.

3. Cylindrical shaped grouted connections with shear keys

3.1. Load on shear keys

In the following a cylindrical shaped grouted connection with shear keys as illustrated in Figs. 5
and 7 is considered. A resistance moment from the shear keys now adds to the resistance moments
considered earlier in Section 2.1. Thus the total resistance moment can be presented as

Mtot ¼ Mp þ Mmh þ Mmv þ Mshear keys (9)

Fig. 4. Illustration of derivation of equation for opening.

120 I. Lotsberg / Marine Structures 32 (2013) 113–135

where the notations were explained in Section 2.1 except for Mshear keys which is the resistance moment
due to the shear keys.
The opening between the pile and the transition piece at top and bottom of the grouted connection
can be expressed by the same equation as for a condition without shear keys (however, the contact
pressure and the ovalisation are reduced due to force action from the shear keys). The relative vertical
displacement between the pile and the monopile is derived as earlier from Eq. (8). This relative vertical
displacement can be further expressed by use of Eqs. (6) and (7) as

R2p
!
6p Rp R2
dv ¼ þ TP (10)
E Lg tp tTP

where E is the Young’s modulus for steel; p, maximum nominal contact pressure at top and bottom of
the grout section as indicated in Fig. 3.
The moment resistance due to the shear keys can be calculated based on Fig. 6 as

Z p=2
Mshear key ¼ 4 kv Rp d4Rp sin 4ðdv sin 4Þ (11)
0

where kv is the spring stiffness of a shear key per unit length of that shear key around the
circumference acting in the vertical direction; dv, maximum relative vertical displacement
calculated at the positions of the shear keys. The following capacity equation is derived by
integration

Fig. 5. Example of grouted connection with shear keys in monopile.

 I. Lotsberg / Marine Structures 32 (2013) 113–135 121

MShear key

dφ
φ
RP
RP dφ
RP sin φ

δv = max vertical displacement

δ = δv sin φ

Fig. 6. Model for analysis of moment resistance from shear keys.

Mshear key ¼ pdv kv R2p (12)

It is assumed that the resistance from several shear keys can be added together such that

kvn ¼ kv n (13)
It is assumed that the shear keys are placed in the centre region of the grouted connection of the
monopile; see also Fig. 5. The reason for this is that one would prefer to place the shear keys in a region
without significant opening during a moment loading. The vertical stiffness of the transition piece
representing half the effective height in the grouted connection reads

2tTP E
kTP ¼ (14)
Lg
The inverse of stiffness is flexibility (in general terms)

1
f ¼ (15)
k
It is assumed that flexibilities in a series system can be added according to the following relation

1 1 1
¼ þ (16)
keff kvn kTP
From this equation keff can be derived. Then Eq. (12) is modified to include the stiffness of n number
of effective shear keys and the axial stiffness of the transition piece. The resulting resistance moment
due to shear keys can then be expressed as

Mshear key ¼ pdv keff R2p (17)

N
keff has unit
mm2
From Eqs. (1)–(3), (9), (10) and (17) the following expression for the nominal contact pressure is
derived
122 I. Lotsberg / Marine Structures 32 (2013) 113–135

Fig. 7. Analysis model for derivation of expression for stiffness for vertical spring between shear keys.

3pMtot ELg
p ¼ (18)
R2p R2TP
!
n o
ELg Rp L2g ðp þ 3mÞ þ 3pmR2p Lg þ 18p2 keff R3p þ
tp tTP
 I. Lotsberg / Marine Structures 32 (2013) 113–135 123

where m is the friction coefficient; m ¼ 0.4 is used as characteristic value for design; m ¼ 0.7 is used for
assessment of laboratory test data.
From finite element analysis of grouted connections it is observed that compressive struts develop
between shear keys on the transition piece and the shear keys on the monopile when the transition
piece is moved vertically relative to that of the monopile. It is further observed that tensile stresses
develop in the grout normal to the compressive strut which will lead to cracking of the grout in a
direction parallel with the compression strut as loading from the transition piece is increased.
The spring stiffness representing the stiffness of the grouted connection with shear keys is derived
from an analysis model as shown in Fig. 7. It is assumed that a compression strut develops between a
shear key at the inner side of the transition piece and a shear key at the outer side of the monopile.
The Young’s modulus of high strength grout used in monopile structures is large, typically in the
order of 50,000 MPa. Thus the grout becomes stiff with respect to radial compression and shear
deformation. The main flexibility of the connection is due to radial deformation of the pile and the
transition piece at the ends of the compression strut, Fig. 7a. These flexibilities can be represented by
springs as illustrated in this figure.
The relative sliding between the transition piece and the monopile is in the vertical direction.
Therefore it is most convenient to define a spring that acts in the same vertical direction; Fig. 7c.
Reference is made to Fig. 7b for assessment of reaction force between the compression strut and
transition piece. From this figure the following relation applies (force per unit length of circumference)

Fh
Q0 ¼ (19)
2
For the moment in the transition piece the following relation applies

Q0 leTP
M0 ¼ (20)
2
where leTP is elastic length of the transition piece which is defined as (see also Appendix B for classical
shell theory)
pffiffiffiffiffiffiffiffiffiffiffiffiffiffi
RRP tTP
leTP ¼ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ffi (21)
4
3 1 n2


The relation between the force and the radial displacement of the transition piece can be developed
as

8DTP
Fh ¼ dhTP (22)
l3eTP

where dhTP is the radial deflection of the transition piece for horizontal force Fh; DTP, shell stiffness of
the transition piece is defined as

Et 3
DTP ¼  TP  (23)
12 1 n2

Similarly the relation between force and radial displacement for the pile can be developed as

8Dp
Fh ¼ dhp (24)
l3ep

where dhp is the radial deflection of the pile for horizontal force Fh; Dp, shell stiffness of the pile defined as

Et 3
Dp ¼  p  (25)
12 1 n2

The resulting deflection from the transition piece and the pile is derived as
124 I. Lotsberg / Marine Structures 32 (2013) 113–135

dh ¼ dhTP þ dhp (26)

The following relation is obtained from moment equilibrium of the compression strut in Fig. 7b
s
F ¼ tg Fv (27)
2 h
where s is distance between shear keys (or s/2 as shown in Fig. 7b as it is assumed that the shear keys
on the monopile side is between those on the side of the transition piece).
From Fig. 7c the following relation with respect to vertical and horizontal displacements are derived
(assuming the displacements leads to a small rotation of the compression strut)

dh dv
¼ (28)
s=2 tg
where dv is relative vertical displacement of the transition piece. From Eqs. (22)–(28) the following
vertical spring stiffness representing a unit length of a shear key is derived as

s2 E
kv ¼ 3=2 3=2
(29)
Rp RTP
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ffi  
2 4 3 1 n2 tg2
 
þ
tp tTP
where now

Fv
kv ¼ (30)
dv
There is also some flexibility in the transition piece from the top of the pile to the centre of location
of the shear keys as represented by Eq. (14).
By combining stiffnesses as in Eq. (16) the following expression is now derived for the effective
spring stiffness for n shear keys

2tTP s2eff nE
keff ¼ 3=2 3=2
(31)
Rp RTP
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi  
ffi
4 4 3 1 n2 tg2 tTP þ ns2eff Lg

þ
tp tTP
This spring stiffness enters Eq. (18) for calculation of nominal contact pressure.
Then the relative vertical displacement between the transition piece and the pile can be calculated
from Eq. (10).
The loading on the shear keys can then be calculated as

FvShk ¼ keff dv (32)

The action force (per unit length around the circumference) from bending moment and vertical
force transferred to the shear keys is derived as

R2p
!
6pkeff Rp R2 P
FvShk ¼ þ TP þ (33)
E Lg tp tTP 2pRp

where P is the self-weight of structure above the pile including full weight of the transition piece for the
ultimate limit state; P ¼ 0 for the fatigue limit state.
The loading on one shear key is obtained as

Fv Shk
Fv1 Shk ¼ (34)
n
It is assessed that the effective distance between the shear keys is somewhat shorter than the
distances measured from centre to centre. The following expression for effective distance between
shear keys is derived
 I. Lotsberg / Marine Structures 32 (2013) 113–135 125

seff ¼ s w (35)

where w is the width of one shear key. More laboratory testing is proposed to investigate this further
when the distance between the shear keys is increased.
For the purpose of design it is necessary to assess the above average action force against the capacity
from [13] for the ultimate limit state. For the fatigue limit state the capacity should be assessed against
an S–N curve for grouted connections with shear keys [20].

3.2. Design of box test specimens

Most of the laboratory fatigue tests on grouted connections with shear keys reported in the
literature were performed with specimens with a diameter less than 400 mm [12]. It has been
questioned whether these specimens were representative for the structural behaviour of large
diameter grouted monopile connections which are subjected to a static axial load and a significant
dynamic bending moment. Therefore an effort was made to design test specimens with a represen-
tative radial stiffness similar to that of grouted connections of large diameter piles in the order of 5 m
or even larger.
From laboratory testing of grouted connections it has been observed that compression struts are
developed in the grout between shear keys on the transition piece and the pile, as indicated in Fig. 7a.
As described earlier the struts are rather stiff compared with the radial stiffness of the pile and the
transition piece. Therefore the radial deflection is mainly governed by the thickness of the steel, the
geometry of the connections in terms of radius, grout thickness, shear key geometry and Young’s
modulus for the steel. From this information it was possible to design box specimens that could
simulate the structural behaviour of large diameter connections. Reference is made to Fig. 8 showing a
photo of a box specimen. This figure also shows a vertical and a horizontal section through the grouted
section. The box specimens were designed with a full size grout thickness of 100 mm, full size of shear
keys (height ¼ 12 mm) and a distance between the shear keys similar to that used in the full size
structures. Because of the symmetry two volumes of grouted connections were tested as shown in
Fig. 8. To date four box specimens have been designed, fabricated and tested under reversed cyclic
loading. The stiffness of the box tests corresponds to that of cylindrical diameters of approximately
800 mm, 2200 mm and 5000 mm. In addition to the four box tests; a cylindrical shaped grouted
connection, of diameter 800 mm, with shear keys (as shown in Fig. 9) was tested under reversed
bending moment with a constant axial force.
The smallest box, box specimen no. 1, was made with a similar radial stiffness for comparison
with the cylindrical test specimen. Box specimen no. 2 was made to simulate the radial stiffness of
pile to sleeve grouted connections in a typical jacket structure. Box specimens no. 3 and 4 were
designed with a similar radial stiffness as that of large diameter monopile connections (D ¼ 5 m).
The representative radial stiffness of the box specimens was achieved through the choice of gov-
erning dimensions (plate thickness and span length) of the plate segments between the grouted
section and the bolt arrangements. This is shown in the horizontal section through box specimen in
Fig. 8. The radial stiffness of the boxes was verified by measurements with hydraulic actuators
inside the boxes before grouting was performed. When the box specimens were loaded in the axial
direction under reversed cyclic loading, the side plates of the boxes were subjected to “radial”
contact pressure from the grout, and a similar behaviour as found in a large diameter grouted
connection was simulated, as illustrated in the lower part of Fig. 8. It is noted that there is no
confinement in the circumferential direction in the box test specimens. This is considered to be of
less significance due to a rather low tensile capacity of the grout such that radial cracks in the
meridian direction are observed to occur in cylindrical tests specimen when they are subjected to
large bending moments as indicated in Fig. 2.
The laboratory testing using these box test specimens has been shown to be successful. Therefore a
further test programme is being developed for better assessment of different parameters such as ge-
ometry of shear key arrangement and type of grout material. By use of these test boxes it is also
possible to open up the sides after grouting (sides in radial direction) such that degradation of the
126 I. Lotsberg / Marine Structures 32 (2013) 113–135

Fig. 8. Box test specimen for simulation of large diameter connection.

connections can be followed during testing and a better physical understanding of the structural
behaviour of grouted connections with shear keys can be achieved.

3.3. Comparison of design procedure with test data

The derived analytical procedure has been assessed against measured data from testing of the
box test specimens in Section 3.2 and a cylindrical test specimen subjected to large bending mo-
ments as shown in Fig. 9. The calculated and the measured relative displacement in the axial
 I. Lotsberg / Marine Structures 32 (2013) 113–135 127

Fig. 9. Cylindrical test specimen.

Fig. 10. Measured relative sliding between transition piece and pile compared with that calculated based on the proposed approach.
128 I. Lotsberg / Marine Structures 32 (2013) 113–135

direction between the transition piece (outer sleeve) and the pile for the cylindrical test specimen
are shown in Fig. 10. The measured relative axial displacement was derived from the LVDT placed
horizontally as shown in Fig. 9c. The measured amplitudes were derived from one of the first load
cycles the specimen was subjected to under constant amplitude loading. The plus and minus
bending moments were approximately equal during the test. The measurements are referred to the
top and bottom of the specimen as placed in a horizontal position in the test laboratory. These
regions were subjected to the largest bending moments during the test. It is noted that there is less
good correspondence between calculated and measured at the bottom than at the top. This graph
presents amplitude displacement that involves zero external loading. This load level may involve
some internal residual stresses due to friction after unloading. This load level is removed from the
graph if one considers a full load cycle by adding the resulting graphs from maximum and negative
moments together. Then a better correspondence between measured and calculated displacements
also for the bottom LVDT reading is achieved.
A similar comparison has also been made for the tests performed on cylindrical test specimens at
university of Leibniz [23]. This comparison is shown in Fig. 11 for relative axial displacement and
Fig. 12 for opening displacement. It is observed that the derived stiffness provides displacements
that in general are in good agreement with those measured for different grout strengths, different
grout lengths and different number of shear keys as shown in Table 1. The largest differences be-
tween calculated and measured displacement is observed for test no. 6 in Fig. 11. Test no. 6 in Fig. 11
is similar in geometry to that of test nos. 2 and 4. One may note that the main difference between
these tests is only the grout compressive strength which is 90 MPa for test no. 6, 70 MPa for test no.
4 and 130 MPa for test no. 2. The grout strength is not a parameter in the analytical derived sliding
(due to large Young’s modulus for all these grout qualities). As 90 MPa fall in between that of test
nos. 2 and 4 a similar measured value would also be expected for test no. 6 as for test nos. 2 and 4.
For the opening displacement in Fig. 12 there is acceptable correspondence between calculated and
measured values also for test no. 6. Based on this a higher measured value for test no. 6 would be
expected in Fig. 11.

Fig. 11. Measured relative sliding between transition piece and pile compared with that calculated based on proposed approach with
stiffness keff.
 I. Lotsberg / Marine Structures 32 (2013) 113–135 129

Fig. 12. Measured opening between transition piece and pile compared with that calculated based on proposed approach with
stiffness keff.

We have to realize that the laboratory testing is made on grouted connections where some un-
certainty in structural behaviour and measurements is to be expected also due to other parameters
than that accounted for in analysis such as surface irregularity.
A similar comparison is also made between measured and calculated displacement for the box test
specimens described in Section 3.2. For box specimen no. 3, which simulates a typical monopile ge-
ometry, a somewhat non-linear displacement was measured. This gives less force on the shear keys
than predicted by the derived approach. This is thus to the safe side for design of shear keys. However, it
is not conservative for assessment of contact pressure. Therefore, it is proposed to reduce the stiffness
by a factor equal 2 for design with respect to contact pressure. The resulting displacement for this
stiffness is also shown in Fig. 13 for box specimen no. 2 tested in this project. This specimen was
subjected to reversed dynamic loading. The loading was gradually increased. The load levels in the
figure correspond to the same value in the reversed direction as that presented in the notation.

4. Closure

Two joint industry projects on grouted connections have been performed since autumn 2009
[19,20]. In the first project it was shown that cylindrical shaped grouted connections without

Table 1
Parameters of the test specimens from Leibniz [23].

Test no Uniaxial compressive strength, fc (MPa) Relative grout length, L/(2Rp) Number of shear keys, n
1 130 1.3 0
2 130 1.3 7
3 130 1.0 5
4 70 1.3 7
5 70 1.0 5
6 90 1.3 7
130 I. Lotsberg / Marine Structures 32 (2013) 113–135

Fig. 13. Load displacement curve for box specimen no. 2.

shear keys can no longer be recommended due to the low long-term axial capacity of such
connections. A design methodology using conical shaped connections was developed to account
for large dynamic bending moments on monopile structures. It was further shown that the contact
pressure between the steel structure and the grout has to be limited due to the potential for
cracking of the grout at the grout ends and abrasive wear. The resulting design methodology has
been included in [8].
A second joint industry project on the capacity of cylindrical shaped grouted connections with
shear keys subjected to alternating dynamic loading has also been performed. It was assessed that it
is difficult to achieve reliable fatigue test results from scaled tubular test specimens, i.e. using small
diameter cylinders, subjected to reversed dynamic loading when attempting to simulate the
behaviour of large diameter monopile connections. Large actuators would be required to perform
tests with specimens of a similar size to that built offshore. Therefore a methodology with test box
specimens simulating structural properties similar to that of large diameter connections was
developed. Four different boxes have been tested in the laboratory so far and more testing is being
planned.
A design methodology to account for the dynamic bending moment and the axial force in con-
nections with shear keys has been developed for the ultimate limit state and the fatigue limit state. The
results have been compared with laboratory test data from simulated capacity of large diameter
connections. It was found that the design procedure provides design data that are in good agreement
with the measured laboratory data. The basis for the analytical part of this design procedure has been
presented in this paper.
The new design methodology for cylindrical shaped grouted connections with shear keys will be
included in a revised version of DNV-OS-J101 during 2013.

Acknowledgement

Acknowledgement is made to the JIP Partners for their support and contribution to this work:
Ballast Nedam Engineering, BASF Construction Chemicals, Centrica, Densit A/S, DNV, DONG Energy,
Germanisher Lloyd Industrial Services GmbH, Gottfried Wilhelm Leibniz Universität Hannover, Joint
Venture East Anglia Offshore Wind, Keppel Offshore and Marine Technology Centre Pte Ltd,
Kvaerner Jacket Technology AS, MT Højgaard A/S, Nippon Steel Corporation, Per Aarsleff A/S, RWE
 I. Lotsberg / Marine Structures 32 (2013) 113–135 131

Innogy GmbH, SSE Renewables Developments (UK) Limited, Statoil Petroleum AS, WTM Engineers
GmbH.

Appendix A. Derivation of moment resistances from contact pressure

This appendix presents some more detailed derivation of moment resistances from contact pres-
sure between steel and grout in the grouted connection than that shown in Section 2 of this paper.

Fig. A.1. Pressure distribution used for calculation of moment resistance.

The moment resistance due to contact pressure is derived by integration around half the circum-
ference b–c–d in Fig. A.1 as
Z p=2
Lg 1 Lg 2

Mp ¼ 2 sin 4Rp d4 p 2 (A.1)
0 2 2 2 3
where the term in the bracket is moment arm about the centre of the grouted connection. The left part
of this equation is force and there is one reaction force at the top and one at the bottom of the grouted
connection in Fig. 3. Then the moment resistance due to contact pressure is derived by integration as

Rp L2g
Mp ¼ p (A.2)
3
The pressure distribution in Fig. A.1 within the stipulated red line (for interpretation of the refer-
ences to colour in this figure, the reader is referred to the web version of this article) can be expressed
as

3 4
pð4Þ ¼ p (A.3)
2 p
for 0  4  p/2. This equation provides a pressure corresponding to 0.75p at positions b and d in
Fig. A.1. Then an expression for the moment due to friction forces in the horizontal direction can be
presented as
132 I. Lotsberg / Marine Structures 32 (2013) 113–135

Z p=2
  Lg 1 Lg 2

Mmh ¼ 4 sin 4Rg d4 p 4 m 2 (A.4)
0 2 2 2 3

where the term in the bracket is moment arm.

Z p=2
p=2
4sin 4 d4 ¼ ½sin 4 4cos 4Š0 ¼ 1:0 (A.5)
0

Then the moment due to horizontal friction force is derived as

1
Mmh ¼ pmRp L2g (A.6)
p
The moment due to friction in the axial direction is derived as

Z p=2
Lg 1  
Mmv ¼ 2 Rp d4 pm Rp sin 4 2 (A.7)
0 2 2

where the term in the bracket is moment arm. The left part of the equation is force and there is one
reaction force at the top and one at the bottom of the grouted connection in Fig. 3. The integration is
performed around half the circumference b–c–d with a constant contact pressure p. This corresponds
to integration around the full circumference where the contact pressure is zero at position a, equal to p
at c and equal to 0.5p at b and d. Then by performing the integration the moment due to vertical friction
force is derived as

Mmv ¼ mpR2p Lg (A.8)

The calculated moment resistances have been compared with finite element analysis (FEA) as
shown in Fig. A.2. A rather large friction coefficient equal to 0.8 was used in the finite element analysis.
As the friction coefficient is being reduced, relatively more of the resistance is provided by pressure
only. It should be added that this comparison was performed before a final decision on pressure values
to be used at positions b and d in Fig. A.1 was made.
The selected pressures at b and d for derivation of the two resistance friction moments were also
based on a consideration that the force due to friction in any direction cannot be larger than pressure
time friction time area. Thus one cannot have a full friction force both in vertical and horizontal di-
rection around the circumference as action from the two components cannot be larger than that of a
vector representing full friction.
It might be added that the finite element analysis was performed for a connection with a rather
large diameter to thickness ratio. As this ratio is being reduced the steel structure will start to pick
up some more local bending moment over the thickness in the circumferential direction due to the
contact pressure. This will reduce the stresses at positions b and d in Fig. A.1 as compared with that
of position c. Thus, analysts may show a slightly different pressure distribution around the
circumference depending on diameter to thickness ratio used in analysis [21]. One does not want to
rely too much on resistance from friction also due to such uncertainties. Therefore, it is recom-
mended to use a characteristic friction coefficient for design purpose equal to 0.40 while 0.70 was
derived from testing as a mean friction coefficient. It is also recommended to reduce the spring
stiffness for the shear keys in order not to underestimate the pressure for control of the grout ca-
pacity. A design of a typical grouted connection in a monopile with a different number of shear keys
has been assessed based on the derived resistance equations in this paper. The results are shown in
Fig. A.3. It is observed that the contribution to moment resistance from friction is significantly less
than that from the shear keys and the contact pressure when a relevant number of shear keys is
used.
 I. Lotsberg / Marine Structures 32 (2013) 113–135 133

1.0
Mp
0.9
Mµh
0.8 Mµv

Contribution to total moment

Mp FEA
0.7 Mµh FEA
0.6 Mµv FEA

0.5

0.4

0.3

0.2

0.1

0.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Friction coefficient

Fig. A.2. Contribution to moment resistances from contact pressure as function of friction coefficient.

Fig. A.3. Example of contribution to resistance moment as function of number of effective shear keys.

Appendix B. Classical shell theory

The differential equation for the cylindrical shell shown in Fig. B.1 can be found in the literature [22]
and [9]

v4 w Et   n
D þ w ¼ p x N (B.1)
vx4 r2 r
where D is the flexural rigidity of the shell defined as
134 I. Lotsberg / Marine Structures 32 (2013) 113–135

Et 3
D ¼ (B.2)
12 1 n2
 

where E is the Young’s modulus; n, Poisson’s ratio; t, shell thickness; r, radius of the cylindrical shell
measured from the axis of the cylinder to the middle of the shell surface; p(x), internal radial loading;
N, axial force per unit circumferential length.

Fig. B.1. Circular cylindrical shell loaded symmetrically with respect to its axis.
The solution of the homogenous part of the differential equation (Eq. (B.1)) can be expressed as

l2e
wh ¼ ðM g ðxÞ þ Q0 le g1 ðxÞÞ (B.3)
2D 0 4

where an elastic length is defined as

pffiffiffiffi
rt
le ¼ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ffi (B.4)
4
3 1 n2


and with reference to Fig. B.1 the following definitions are made

x
x¼ ; g1 ðxÞ ¼ e x cos x; g2 ðxÞ ¼ e x sin x; g3 ðxÞ ¼ g1 ðxÞ þ g2 ðxÞ; g4 ðxÞ ¼ g1 ðxÞ g2 ðxÞ
le
(B.5)

The particular part can be expressed for p(x) as a polynomial in less or equal third degree as

r2   nr
wpart ¼ p x N (B.6)
Et Et

The total displacement, w, is the sum of the homogenous and the particular part. Then for constant p

vw le
¼ ð2M0 g1 ðxÞ þ Q0 le g3 ðxÞÞ (B.7)
vx 2D

The moment (per unit circumferential length) at a section x (Fig. B.1) is calculated as

v2 w
Mx ¼ D ¼ M0 g3 ðxÞ þ Q0 le g2 ðxÞ (B.8)
vx2

and the shear force (per unit circumferential length) is derived as

 I. Lotsberg / Marine Structures 32 (2013) 113–135 135

vMx 2M0 g2 ðxÞ

Qx ¼ ¼ þ Q 0 g4 ð x Þ (B.9)
vx le

References

[1] API RP2A. Recommended practice for planning, designing and constructing fixed offshore platforms. 9th ed. Dallas, Texas,
USA: The American Petroleum Institute; November 1977.
[2] API RP2A. Recommended practice for planning, designing and constructing fixed offshore platforms. 15th ed. Dallas, Texas,
USA: The American Petroleum Institute; October 1984.
[3] API RP2A-WSD. Recommended practice for planning, designing and constructing fixed offshore platforms – working stress
design. 21st ed. Dallas, Texas, USA: The American Petroleum Institute; 2000.
[4] Billington CJ, Lewis GHG. The strength of large diameter grouted connections. Paper presented at 10th offshore technology
conference, Houston, Texas, 8–11 May, OTC paper no. 3083; 1978. p. 291–301.
[5] Billington CJ, Tebbett IE. The basis for new design formulae for grouted jacket to pile connections. Paper presented at 12th
offshore technology conference, Houston, Texas, 5–8 May, OTC paper no. 3788; 1980. p. 449–58.
[6] Offshore installations: guidance on design and construction. 2nd ed.UK: Department of Energy, ISBN 0114106126; 1977.
[7] Amendment no. 4 to offshore installations: guidance on design and construction. Department of Energy, ISBN 0114107661;
1982. First published April 1982.
[8] DNV-OS-J101. Design of offshore wind turbine structures; October 2011.
[9] Flügge W. In: Stresses in shells. 2nd ed. Berlin, Heidelberg, New York: Springer-Verlag; 1973.
[10] Harwood RG, Billington CJ, Buitrago., J, Sele A, Sharp JV. Grouted pile to sleeve connections: design provisions for the new
ISO standard for offshore structures. ASME; 1996. OMAE 1996.
[11] Pile/sleeve connections. HSE, ISBN 0717623904; 2002.
[12] Ingebrigtsen T, Løset Ø, Nielsen SG. Fatigue design and overall safety of grouted pile sleeve connections. Paper presented at
22nd offshore technology conference, Houston, Texas, 7–10 May, OTC paper no, 6344; 1990. p. 615–28.
[13] ISO 19902. Fixed steel offshore structures; 2007.
[14] Karsan DI, Krahl NW. New API equation for grouted pile-to-structure. Paper presented at 16th offshore technology con-
ference, Houston, Texas, 7–9 May, OTC paper no. 4715; 1984. p. 49–54.
[15] Lewis GHG, Livett JG, McLaughlin RTP, Mead KC. A cost saving design for pile to structure connections as applied to BP
Magnus. Paper presented at 12th offshore technology conference, Houston, Texas, 5–8 May, OTC paper no. 3789; 1980. p.
459–66.
[16] Løseth Ø. Grouted connections in steel platforms – testing and design, institute of structural engineers informal study
group – model analysis as a design tool, joint I. Two day seminar on the use of physical models in the design of offshore
structures; November 15–16, 1979. Paper no. 8.
[17] Lotsberg I, Serednicki A, Cramer E, Bertnes H, Enggaard Haahr P. On structural capacity of grouted connections in offshore
structures. Presented at OMAE 2011-46169; 2011.
[18] Lotsberg I, Serednicki A, Bertnes H, Lervik A. Design of grouted connections for monopile offshore structures. Results from
two joint industry projects. Stahlbau 2012a;81(9):695–704.
[19] Lotsberg I, Serednicki A, Cramer E, Bertnes H. Behaviour of grouted connections of monopile structures at ultimate and
cyclic limit states. The Structural Engineer 2012b;February:51–7.
[20] Lotsberg I, Serednicki A, Oerleans R, Bertnes H, Lervik A. Capacity of cylindrical shaped grouted connections with shear
keys in offshore structures reported from a joint industry project. The Structural Engineer 2012c;91(1):42–8.
[21] Prakhya G, Zhang C, Harding N. Grouted connections for monopiles – limits for large wind turbines. The Structural En-
gineer 2012;March:30–45.
[22] Timoshenko SP, Woinowsky-Krieger S. In: Theory of plates and shells. 2nd ed. Mc Graw-Hill Book Company, Inc; 1959.
[23] Schaumann P, Lochte-Holtgreven S, Wilke F. Bending tests on grouted joints for monopile support structures. DEWEK2010
– 10th German wind energy conference, Bremen; 2010.