Thin–Walled Structures 154 (2020) 106880

Contents lists available at ScienceDirect

Thin-Walled Structures
journal homepage: http://www.elsevier.com/locate/tws

Numerical modeling and parametric analysis of grouted connections under

axial loading
Tao Chen a, b, *, Chengcheng Cao a, b, Chihai Zhang a, b, Xian Wang a, b, Ke Chen c, Guokai Yuan c
a
Key Laboratory of Performance Evolution and Control for Engineering Structures of Ministry of Education, Tongji University, Shanghai, 200092, China
b
Department of Structural Engineering, Tongji University, Shanghai, 200092, China
c
China Energy Engineering Group Guangdong Electric Power Design Institute Co., Ltd., Guangzhou, 510663, China

A R T I C L E I N F O A B S T R A C T

Keywords: With the rapid development of the offshore wind power industry, large-diameter grouted connections (GCs) with
GCs high-strength grout have become a developing trend. Previous studies on ordinary-strength and small-scale GCs
Axial mechanical properties may not meet current engineering needs. A numerical study was carried out to investigate how the axial me­
Radial stiffness
chanical properties of a large-diameter GC that has high-strength grout are influenced by geometric parameters,
Height to space ratio
Length to diameter ratio
such as radial stiffness, the height to space ratio (h/s) of shear keys, and the length to diameter ratio (Lg/Dp). The
finite element model was verified against the results of the experimental test. The mechanical properties of
ultimate bearing capacity, interface transfer strength, ductility, nominal average stress and grout contact pres­
sure distribution are discussed in detail. The results suggest that the increase in radial stiffness can strengthen the
axial ultimate bearing capacity and ductility of the GC. Besides, the h/s of the shear keys can affect the ultimate
bearing capacity and ductility of GC by influencing the number of diagonal compression struts. Although the Lg/
Dp can significantly improve the axial ultimate bearing capacity of GCs, conversely, the interface transfer
strength is continuously reduced. The research results can provide technical reference for the design of GCs in the
offshore wind turbine industry.

1. Introduction are much smaller than the diameters, which can be regard as thin-walled
structures. The load on the supporting structures of the offshore wind
As an important renewable source, wind energy has received turbine is predominantly transmitted to the foundation through the GC.
extensive attention from all over the world [1]. Due to the environ­ Therefore, the mechanical properties of the GC are extremely important
mental constraints on onshore wind power and the enormous reserves of for the safety of the entire structure.
offshore wind energy, offshore wind power has more development po­ To fully understand the mechanical performance of GCs, experi­
tential than onshore wind power [2,3]. For offshore wind turbines, the mental studies on the axial ultimate bearing capacity of GCs have been
foundation is critical to ensuring the integrity of wind turbines, which widely carried out in the past 40 years. Billington et al. [8]demonstrated
determines the success or failure of the offshore wind power industry. that the presence of shear keys could significantly improve the axial
Many foundation styles have been developed in recent years, such as the bearing capacity of GCs by testing 400 specimens. Subsequently, Karsan
gravity base, monopile, tripod, and jacket, to satisfy the needs of et al. [9]investigated how the length of GCs can be greatly shortened by
offshore wind power development [4]. Statistics show that monopiles using the shear key. Yamasak [10] and Lamport [11] compared the
and jackets have dominated the market [5], where the jacket accounted load-displacement curves of GCs with and without shear keys under
for 24.5% of all installed offshore wind turbines next to monopiles in axial load. They found that the failure mode of GCs without shear keys is
2018 [6]. The connection between the pile and substructure (Fig. 1) is a brittle failure, whereas the failure mode of GCs with shear keys is a
called a GC, which is the overlap of pile and sleeve tube, and the gap ductile failure. The above studies prove that the presence of shear keys
between them is filled with high-performance grout [7,8]. The tubes has a positive effect on the mechanical properties of GCs. Furthermore,
were rolled from steel plates. And the thicknesses of the pile and sleeve many scholars have carried out more detailed research to fully

* Corresponding author. Key Laboratory of Performance Evolution and Control for Engineering Structures of Ministry of Education, Tongji University, Shanghai,
200092, China.
E-mail address: [email protected] (T. Chen).

https://doi.org/10.1016/j.tws.2020.106880
Received 30 November 2019; Received in revised form 26 February 2020; Accepted 30 May 2020
Available online 13 June 2020
0263-8231/© 2020 Elsevier Ltd. All rights reserved.
T. Chen et al. Thin-Walled Structures 154 (2020) 106880

provide basic guidelines for the finite element analysis of GCs. However,
these two studies on finite element modeling did not consider the use of
shear keys. Subsequently, Schaumann et al. [24] analyzed the axial fa­
tigue of GCs with shear keys by the finite element method. Lo €hning et al.
[25] investigated the cause of slippage between the steel tube and grout
with the finite element method, and proved that the arrangement of
shear keys could reduce slip and improve axial bearing capacity. Lin and
Huang [26] discussed the location of the maximum stress distribution
and the most dangerous location of the GC under static force. Prakhya
[27] and Marion [28] considered the finite element model of the
large-sized GC. They found that the stress state of the large-size GC is
different from the previous model, and further research is needed on the
large-sized GC. Chen [29] carried out parameter studies on the GC
section under combined loads, and suggested that the thicknesses of pile
and sleeve tube have an inverse relationship with the maximum stress
and the maximum Tresca stress is insensitive to any geometry parame­
ters except for the grout thickness. Although there have been many
numerical studies on GCs, parameter analyses of GCs under axial loading
Fig. 1. GC details. are limited.
In recent years, the offshore wind power industry has developed
understand the failure mechanism under the axial force of GCs with rapidly. There is a tendency to locate wind farms further offshore with
shear keys. Aritenang [12] found that the failure of GCs is related to the larger wind turbines to obtain more power [30,31]. To meet the
circumferential buckling of the pile by studying the failure mechanism development demands of offshore wind power industry, the GC has
of GCs under axial load. At the same time, Aritenang derived his design shown new trends, which are the application of higher performance
formula by theoretical derivation [13]. Boswell et al. [14] proved that grout and larger diameter steel tubes [32] as well as the changes of shear
the strain of the steel tube at the shear keys is greater than that between keys arrangement. These trends may cause changes in the mechanical
the shear keys. Krahl et al. [15] proposed the theory of diagonal properties of GCs [33]. However, researches on the GCs currently used
compression struts based on experimental observations. They believe are limited which has led to insufficient understanding of the mechan­
that axial bearing capacity is mainly provided by two parts, one is the ical properties of the GCs currently used. Therefore, it is necessary to
friction and adhesion between the grout and the steel tube, and the other carry out further research on GCs. This study aims to investigate the
part is the axial force provided by diagonal compression struts. Sala et al. effect of different parameters on the mechanical properties of the GC
[16] also reached similar conclusions in 1989. However, earlier studies with a large diameter and high-strength grout under axial loading. This
above are based primarily on small diameter GCs using conventional paper first verifies the reliability of the analytical method by comparing
grouts. Nowadays, the diameter of the GC of jacket support structure and with the Jeong-Hwa’s tests [34]. The comparison mainly focus on three
monopile support structure used in offshore wind farm is generally aspects, including final failure modes, load-displacement curves and
2.0–2.5 m, 4.0–7.0 m, respectively. Larger diameter GCs will be devel­ ultimate bearing capacity, and the longitudinal strain at the shear key’s
oped in the future. The strength of grouting materials currently used in positions. Then, three important parameters affecting the mechanical
offshore wind power projects is above 100 Mpa [17]. To satisfy current properties of the GC are discussed, including radial stiffness, the h/s of
engineering demands, some scholars have carried out many experi­ the shear keys, and the Lg/Dp. The results can be an important reference
mental studies on the axial bearing capacity of large-diameter GCs using for the design of GCs.
high-performance grouts. Schaumann et al. [18] compared the axial
force behavior of GCs filled with high-performance grout and ordinary 2. Establishment of GCs models
Portland cement. The experimental results show that the ultimate
bearing capacity of GCs filled with ordinary Portland cement is signifi­ Pile, sleeve and grout are separately modeled and assembled as GCs
cantly lower than that of the high-performance grout. Meanwhile, [29]. The modeling parameters are carefully selected to ensure that the
Anders et al. [19] proved that the use of high-performance grout could model of GC is as consistent as possible with the actual situation, the
significantly increase the ultimate bearing capacity of GCs. However, it parameters include the shape of the shear key, the division of the
is very difficult to directly conduct large-scale experimental studies on element, the material properties, the boundary conditions, and the
large-diameter GCs. In order to simplify the test of the large-diameter interaction. The selection of modeling parameters will be described in
GCs, special box specimens with the same radial stiffness as the detail in the following sections.
large-diameter GCs were applied [17,20,21]. In general, full under­
standing of the stress distribution of GCs is needed in practical engi­ 2.1. Model geometry and meshing
neering design. However, it is hard to obtain the exact states of stress
and strain in GCs by experimental tests. Half of the model was used to improve computational efficiency, to
As an extensively used numerical method in the engineering field, take into account the symmetry of the shape and load conditions of GCs.
the finite element method can visually observe the condition of stress In engineering practice, the shear key is round bars welded to steel tube,
and strain inside the GC. At the same time, as long as the numerical which will cause mesh distortion and result in an inaccurate calculation.
research method is proven to be reliable, it can be used to study multiple The investigation carried out by Bosswell et al. shows that the cross-
parameters simultaneously. Besides, finite element analysis can save section of shear keys has little effect on the bond strength of GC [35].
time and resources compared to laboratory tests. Therefore, finite The curved edge shear key to the finite element model can be replaced
element analysis is promising as compensation for laboratory tests. A by the rectangular one of the same area [36]. So, in this paper, the
large number of studies on the numerical analysis of GCs have been irregularly shaped shear key is replaced by a straight-sided trapezoidal
reported. Based on large-scale physical tests, Andersen et al. [22] one to guarantee the quality of meshing.
initially attempted to employ a numerical simulation method to study The choice of element type is very important for finite element
the mechanical properties of GCs under a bending moment in 2004. analysis because it directly affects the accuracy and stability of the
Later, Nielsen et al. [23] discussed different modeling methods to calculation results. There are many element types available in the

2
T. Chen et al. Thin-Walled Structures 154 (2020) 106880

ABAQUS library, so special attention should be taken when selecting Table 1

suitable elements for modeling grout and the steel tube. The current Grout properties [7] and CDP parameters [42].
research results of finite element analysis of GCs show that simulation Symbol Value CDP Value
results using C3D8R elements are in good agreement with the experi­ 4
Eg 5.07 � 10 MPa ψ 38�
mental results [7,37,38]. Thus, both steel and grout are modeled by υg 0.2 ε 0.1
C3D8R elements in this study. In order to ensure the calculation results fcm 87.6 MPa fb0/fc0 1.16
convergence, the sleeve tube, pile tube and grout are divided into 50 fctm 4.63 MPa Kc 0.67
layers in the circumferential direction, the sleeve tube and pile tube in GF 0.1633 N/mm υ 0.001

the thickness direction are divided into 3 layers, and the grout is divided Note: Eg is young’s modulus of grout; υg is Poisson’s ratio of grout; fcm is the
into 6 layers. Shear keys are arranged at the central region of the GCs mean compressive strength of grout; fctm is the mean tensile strength of grout
and occupy half of the length of the GC [39]. The meshing at the position and GF is the fracture energy of grout, respectively. ψ is the dilation angle; ε is
of the shear key is property refined and divided into two layers in length the flow potential eccentricity; fb0/fc0 is the ratio of the biaxial/uniaxial
and width directions to ensure the accuracy of calculation results. The compressive strength ratio; Kc is the shape factor and υ is the viscosity param­
eter, respectively.
total number of elements of different models is between forty and eighty
thousand. The GC’s geometry modeling and meshing are shown in
Fig. 2. 2.3. Boundary conditions and interaction

2.2. Definition of material properties As mentioned above, symmetry was used in the modeling process, so
symmetric constraints in the z-direction on the symmetry plane were
For an accurate simulation of the mechanical properties of GCs, it is needed. The translational freedom of all nodes in the bottom section of
necessary to define the suitable material properties of grout and steel. At the pile was constrained, and a fixed constraint was imposed on the
present, grout material in the practical engineering field is a kind of bottom of the model. A reference point was introduced on the top sur­
cement-based mixture material, whose composition is similar to con­ face of the model, which was used to constrain the degrees of freedom of
crete. From the perspective of material properties, grout can be regarded all nodes on the top surface and to apply the load controlled by the
as high-performance or ultrahigh-performance concrete. Therefore, the displacement.
concrete damage plasticity model (CDP) in ABAQUS can be used to The surface-to-surface contact was set to describe the interaction
simulate the constitutive model of grout materials [25]. The CDP model between the grout and pile (sleeve) tube during analysis. The definition
was first proposed by Lubliner [40], and can demonstrate the plastic of contact between the grout and pile (sleeve) tube is divided into the
damage of concrete materials, including the development of cracks, the normal direction and tangential direction. In the normal direction, hard
closure of cracks, and the recovery of stiffness [41]. Therefore, the contact was used to simulate extrusion behavior between the pile
model can accurately simulate the mechanical properties of concrete (sleeve) tube and grout. The Coulomb friction model recommended by
members under tension or compression. In this paper, the value of CDP DNVGL-RP-0419 [46] was defined to simulate contact in the tangential
parameters refers to the research results of Nie et al. [42]. The uniaxial direction, and the friction coefficient of μ ¼ 0.7 is based on Lotsberg’s
compressive stress-strain curve of the grout is based on Guo’s research recommendation [20].
results [43]. The uniaxial tensile stress-strain curve of the grout is
assumed to be linear before the peak strain，as proposed by Ren and Li 3. Evaluating indices of GCs
[44]. When the strain exceeds the peak strain, the stress-strain curve of
the tensile softening stage of the grout material is described by the 3.1. Interface transfer strength
fracture energy [45]. Grouted properties and CDP parameters are shown
in Table 1. The “interface transfer strength” is proposed to simplify the calcu­
The constitutive model of the steel is assumed to be a bilinear elastic- lation of the ultimate bearing capacity of the GCs. The interface transfer
plastic model. The plastic behavior of the steel is simulated by the ki­ strength can be considered as the average of the bonding strength be­
nematic hardening rule and von Mises yield criterion. The Young’s tween the grout and the sleeve tube. This concept was first applied to
modulus of steel is 2.06 � 105 MPa, the yield strength is 360 MPa and offshore oil platforms to evaluate the ultimate bearing capacity of GCs
the Poisson’s ratio is 0.3. [47] and was adopted by API code [48]. The axial bearing capacity of
the GC can be calculated by the production of interface transfer strength
and the contact area of grout and sleeve tube. The interface transfer
strength can be calculated by equation (1).
pu
fcu ¼ (1)
π⋅Dp ⋅Lg

where Pu is the ultimate bearing capacity, Dp is the outer diameter of the

inner pile tube, and Lg is the effective length of GCs.

3.2. Ductility and energy dissipation

Ductility and energy dissipation are very important indicators for

steel-concrete composite structures. They evaluate the response of
structures under earthquakes and are an important reference for
assessing structural impact resistance.
In this paper, the ductility coefficient is used to evaluate the ductility
capacity of GCs. The expression of the ductility coefficient is in equation
(2)
�
μ ¼ Δu Δy (2)
Fig. 2. Modeling and meshing of GC.

3
T. Chen et al. Thin-Walled Structures 154 (2020) 106880

where △y is the yield displacement, and △u is the ultimate

displacement.
The yield displacement △y is determined by the method suggested
by Park [49]. This method uses an ideal elastoplastic polyline to replace
the actual load-displacement curve. The displacement corresponding to
the inflection point of the polyline is the yield displacement. According
to the yield ratio of the steel, the yield coefficient is taken as 0.75. The
product of the yield coefficient and the ultimate bearing capacity is
taken as a point on the vertical axis, and a horizontal line crossing the
load-displacement curve at point A is passed through this point. Con­
necting the origin of the coordinate to point A and extending upwards,
and intersect at point B with a horizontal line with the peak load point.
The transit point B is taken as the horizontal vertical line and intersects
the load-displacement curve at point C, which is the yield point. △u is
the displacement at 0.85Pmax after peak load as specified in Chinese
Specification for Seismic Test of Building [50] and suggested by Zhou
[51] and Jiang [52]. △y and △u are shown in Fig. 3.
Fig. 4. Schematic diagram of plastic and elastic energy dissipation zone.
The energy dissipation capacity of the GCs can be evaluated by the
area enclosed by the upward section and abscissa of the load-
displacement curve. The area enclosed by the rising section and the pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi σ y d
Rt ¼ 3ð1 v2 Þ (5)
abscissa before the yield point is the energy dissipation of the elastic E 2t
zone Eel. The area enclosed by the curve between the yield point and the
peak point is the energy dissipation of the plastic zone Epl. The total where v is Poisson’s ratio, σy is the yield stress, d is the diameter of the
energy dissipation Etot of the GCs is the sum of the energy dissipation of sleeve tube, and t is the sleeve tube thickness.
the elastic and plastic zones, as shown in Fig. 4. Considering the fact that local crushing of the grout and the failure of
The concept of the energy dissipation factor is introduced to quan­ the interface occur at ends of the GCs. Furthermore, the latest version of
titatively evaluate the energy dissipation capacity of GCs, and its the DNV code places shear keys in the central area of grout connections
expression is as shown in the following equation. [39]. Therefore, it is considered that the effective failure length le is
located in the central region of the GC. The definition of effective failure
Epl Etot Eel length and nominal average stress is shown in Fig. 5. As can be seen from
ηE ¼ ¼ (3)
Eel Eel Fig. 5, the black curve with straight lines indicates the actual stress
distribution within effective failure length of the sleeve tube. They were
3.3. Nominal average stress idealized as green shaded portion with same area. The value of hori­
zontal axis is the nominal average stress σN. With this procedure, the
When analyzing the stress distribution of the pile (sleeve) tube and actual stress in the sleeve tube is equivalently replaced by the average
grout in GCs through finite element analysis, the maximum stress is stress in the effective length le.
closely related to the type and size of the mesh used in the model. It A stress correlation factor is introduced based on nominal average
means that it is not appropriate to use local stress to evaluate the overall stress to simplify the design of GCs in the preliminary stage. The stress
stress state of GCs. In order to evaluate the overall stress state of GCs, correlation factor can be expressed as the ratio of the maximum stress
Chen et al. [29] proposed the concept of nominal average stress σ N. σ mises-max to the nominal average stress σN (see equation (6)). Therefore,
Nominal average stress is calculated by the effective length of damage in local stress can be predicted from a rough model without the details of
the thin-walled structures. The equation for effective failure length is as shear keys through the stress correlation factor.
follows [53].
σ mises max
1 ηcor ¼ (6)
le ¼ 1:2ð 0:08 1Þd (4) σN
Rt

When considering the effect of the radius-thickness ratio, Rt can be 3.4. Nominal contact pressure
expressed as
The contact pressure at the end of GCs is an important indicator for
the design of GCs. In engineering practice, the end grout is prevented
from cracking under repeated loading by controlling the contact stress at

Fig. 3. The definition of yield point and ultimate point. Fig. 5. Definition of nominal average stress and effective failure length.

4
T. Chen et al. Thin-Walled Structures 154 (2020) 106880

the end. In this study, the contact pressure at the end of GCs under the simulate the change of the stiffness of the specimen due to cracking and
axial load is extracted from the finite element calculation results to gain crushing of the grout. Other factors affecting the finite element calcu­
a deeper understanding of the interaction between the pile (sleeve) tube lation results include that the grout in the experiment is inhomogeneous,
and grout. the initial defects of the pile (sleeve) tube, and manufacturing errors,
among others.
4. Model verification
4.3. Comparison of longitudinal strain in steel tubes
The finite element method proposed by this paper is validated by
comparing it with Jeong-Hwa’s tests [34]. Jeong-Hwa et al. focused on Fig. 9 compares the longitudinal strain of the steel tube at the shear
the interfacial shear behavior of GCs with different shear key spacing keys. From Fig. 9, it is shown that finite element results have similar
under concentric and eccentric loading. The test overview is shown in strain distribution trends as the experimental results. It can be seen that
Fig. 6. The verification was performed on two specimens, CL-S65 and the maximum strain of the longitudinal strain is achieved at the fourth
CL-S85, which have the same geometry dimensions and loading pat­ shear key position. However, the differences are within the permissible
terns, and different shear key spacing. All material parameters were range and are considered acceptable. Therefore, we believe that the
taken from the original test results. The dimensions of the test specimens finite element model can provide a close prediction of the longitudinal
selected are shown in Table 2. The comparison between experimental strain distribution of a steel tube.
results and numerical simulation results mainly focuses on three aspects: In summary, the finite element analysis results are in good agree­
final failure modes, load-displacement curves and ultimate bearing ca­ ment with the experimental results. The finite element method proposed
pacity, and the longitudinal strain at the shear key’s positions. The in this paper can correctly reflect the mechanical behavior of the GC
following sections will describe the comparison of the results in detail. under axial loading, which can be used for the further study of the
mechanical behavior of GCs.
4.1. Comparison of failure modes
5. Parameters analysis
The failure mode comparison of GCs is shown in Fig. 7. From the
stress contour plot of the GCs, it can be seen that the grout between the To further study the influence of geometric parameters on the axial
shear keys positioned on each side of steel tube is subjected to greater force behavior of GCs, a parameter study is carried out. Three parame­
stress than the surrounding (as shown by the area surrounded by the red ters are considered, including radial stiffness, h/s, and Lg/Dp. The radial
line). This corresponds to the diagonal compression struts exposed by stiffness refers to the expression of the DNV specification [39], which
cutting the GC after the test (such as diagonal compression struts sur­ takes into account the effects of the geometry of the pile, grout, and
rounded by the red line in the test specimens). The ultimate failure mode sleeve on the radial stiffness. The equation for radial stiffness is defined
in the finite element calculation results is the buckling of the sleeve tube, as
which has a good agreement with the experimental results. At the same 1 1
Es Dg Dp DTp
time, it can be seen that the stress level at the position of shear keys is K ¼ð ⋅ Þ þð þ Þ (7)
Eg tg tp tTp
higher than other positions from stress contour plot, which corresponds
to the crushing of the grout at the position of the shear keys in the test
where Eg is young’s modulus of grout; Es is young’s modulus of pile and
specimens.
sleeve tube; Dp is outer diameter of grout layer; Dp is outer diameter of
pile tube; DTp is outer diameter of sleeve tube; tg is the thickness of grout
4.2. Comparison of load-displacement curves layer; tp is the thickness of pile tube; tTp is the thickness of sleeve tube;
The h/s is the ratio of the height of shear keys to the space of shear
Fig. 8 shows the comparison of the axial load-displacement curves. keys, which was introduced by DNV standard [39] to consider influence
As can be seen, the finite element calculation results show reasonable of the arrangement of shear keys on axial bearing capacity. Different
agreement with experimental results. The stiffness of finite element arrangement of shear keys will affect the failure states of grout layer.
calculation results is slightly lower than the experimental results, but The Lg/Dp is the length of the GC to the diameter of the GC, which is
still within a reasonable error range. The ultimate bearing capacity of another important parameter considered in DNV standard [39]. The
the finite element calculations for the CL-S65 and CL-S85 are 5645.6 kN Lg/Dp has a directly effect on the contact areas between steel tube and
and 5390.2 kN, corresponding experimental results are 5395.5 kN and grout, which will change the ultimate bearing capacity of GC. The di­
5829.9 kN, respectively. The main cause of the error is that the con­ mensions of GCs are shown in Fig. 10, and numerical simulation pa­
tinuum model is adopted in the finite element model, which cannot rameters are designed and shown in Table 3.

Fig. 6. Overview of Jeong-Hwa’s test [34].

5
T. Chen et al. Thin-Walled Structures 154 (2020) 106880

Table 2
Dimensions of Jeong-Hwa’s test specimens.
Specimens No Pile Sleeve Shear key tg (mm) Lg (mm)

Dp (mm) tp (mm) DTp (mm) tTp (mm) h (mm) S (mm)

CLS-65 508 12 609 12 4 65 38.5 520

CLS-85 508 12 609 12 5 85 38.5 680

Note: h and S respectively represent the height and spacing of the shear key, as shown in Fig. 6.

Fig. 7. Comparison of failure modes.

displacement curves of four different radial stiffnesses have obvious

ascending and descending branches. The four specimens reached the
ultimate bearing capacity when the vertical displacement was about 4
mm. When the ultimate bearing capacity was achieved, buckling
occurred in the bottom sleeve tube of all four specimens. With the in­
crease in radial stiffness, the ultimate bearing capacity of the GCs
increased continuously, and the axial ultimate bearing capacity of the
GCs increased by about 14% when the radial stiffness changed from

Fig. 8. Comparison of load-displacement curves.

5.1. Effects of radial stiffness

5.1.1. Load-displacement curves

The effect of radial stiffness on the load-displacement curve of GCs
under axial load is illustrated in Fig. 11. It can be seen that the load- Fig. 10. Dimensions of the GCs.

Fig. 9. Comparison of longitudinal strain at shear keys.

6
T. Chen et al. Thin-Walled Structures 154 (2020) 106880

Table 3
Parameters design of the GCs.
Specimen No Pile Sleeve tg (mm) Lg (m) S (mm) Radial stiffness k Length to diameter ratio Lg/Dp Height to space ratio h/s

Dp (mm) tp (mm) DTp (mm) tTp (mm)

GC-K-1 1420 24 1500 12 28 2272 160 0.010 1.6 0.027

GC-K-2 1400 38 0.012
GC-K-3 1360 58 0.015
GC-K-4 1330 73 0.018
GC-H-1 1400 24 1500 12 38 2240 100 0.012 1.6 0.040
GC-H-2 150 0.027
GC-H-3 200 0.020
GC-H-4 250 0.016
GC-L-1 1400 24 1500 12 38 1400 160 0.012 1.0 0.027
GC-L-2 1820 1.3
GC-L-3 2240 1.6
GC-L-4 2800 2.0

Table 5
Energy dissipation coefficient.
Specimen Elastic energy Plastic energy Total energy energy
No dissipation dissipation dissipation dissipation
zone zone Etot (kN⋅mm) coefficient
Eel (kN⋅mm) Epl (kN⋅mm) ηE
GC-K-1 21,857 23,050 44,907 1.05
GC-K-2 21,941 26,043 47,784 1.20
GC-K-3 23,183 28,980 52,163 1.25
GC-K-4 24,174 31,634 55,808 1.31

data in the tables, the ductility coefficient and the energy dissipation
coefficient of the GCs increase with the increase in the radial stiffness.
The ductility coefficient and energy dissipation coefficient is enhanced
by 26% and 25%, respectively, as the radial stiffness varies from 0.01 to
0.018. It shows that the increase in radial stiffness is beneficial to
improving ductility and energy dissipation capacity of GCs.

5.1.4. Nominal average stress

Fig. 11. Influence of radial stiffness on the load-displacement curve. The change in the nominal average stress in the GC due to the change
in radial stiffness is presented in Fig. 12. The results indicate that the
0.010 to 0.018. The reason for this phenomenon is that the increase of maximum Tresca stress σTresca-grout and the nominal average stress σN of
the radial stiffness results in an increase in the constraint of the pile grout layer tend to decrease with the increase of radial stiffness. The
(sleeve) tube on the grout, which improves the friction between the pile predicted maximum Tresca stress σTresca-grout and nominal average stress
(sleeve) tube and grout, and also strengthens the bearing capacity of σ N decrease by 33.7% and 29.5%, respectively, when the radial stiffness
diagonally compressed struts. increases from 0.010 to 0.018. The corresponding stress correlation
factors are 2.43, 2.82, 3.28, and 3.46. In contrast, the increase in radial
5.1.2. Interface transfer strength stiffness has no significant effect on the maximum Mises stress σ Mises-tube
When GCs achieve ultimate bearing capacity, the interface transfer of the steel tube.
strength of four different radial stiffness specimens are 1.96 MPa, 2.02
MPa, 2.14 MPa and 2.23 MPa. The increase in interface transfer strength 5.1.5. Grout material contact pressure
with the radial stiffness is due to the increased constraint of the pile Due to the symmetry of the configurations and axial loading condi­
(sleeve) tube on the grout, which can strengthen the friction and inter­ tion of the GC, contact pressure is uniform along the end of the grout.
locking force between the pile (sleeve) tube and grout. The curves between axial displacement and contact pressure of different
radial stiffness are presented in Fig. 13 to show the influence of radial
5.1.3. Ductility and energy dissipation stiffness on contact pressure at the end of GCs. Observing the curve, the
Table 4 and Table 5 depict the ductility coefficients and energy trend in the contact pressure-axial displacement curve is similar to the
dissipation coefficients for different radial stiffnesses. According to the load-displacement curve, both have obvious ascending and descending

Table 4
Ductility coefficient.
Specimen Yield Peak Ultimate Ductility coefficient
No μ
Py (kN) Δy (mm) Pmax (kN) Δmax (mm) 0.85Pmax (kN) Δu (mm)

GC-K-1 14,783 2.76 19,819 4.01 16,846 6.46 2.34

GC-K-2 15,993 2.86 20,463 4.18 17,394 7.25 2.53
GC-K-3 16,670 2.79 21,690 4.06 18,437 7.77 2.78
GC-K-4 17,084 2.72 22,577 4.28 19,190 8.01 2.94

Note: Δy,Δmax and Δu represent the displacements of the yield point, peak point and 0.85Pmax after peak load, respectively, as shown in Fig. 4.

7
T. Chen et al. Thin-Walled Structures 154 (2020) 106880

Fig. 12. The effect of radial stiffness on the stress of GCs.

Fig. 14. Maximum contact pressure.

Fig. 13. Contact pressure-displacement curve.

Fig. 15. Influence of h/s on load-displacement curve.
branches. Special attention should be paid to the nominal contact
pressure of specimens GC-K-1 and GC-K-2 as they show a sudden drop compression struts formed between the shear keys. Thereby, the axial
and then increase when the ultimate bearing capacity is achieved. This is ultimate bearing capacity of the GC is reduced.
because the grout is crushed and cracked, and diagonal compression
struts are formed, which results in the redistribution of stress in the 5.2.2. Interface transfer strength
grout layer. When GCs achieve their ultimate bearing capacity, the interface
The contact pressure peaks of the GC-K-1~GC-K-4 GCs are 1.17 MPa, transfer strength of four different radial stiffness specimens are 2.21
1.38 MPa, 1.77 Mpa, and 1.89 MPa, respectively, as shown in Fig. 14. As MPa, 2.11 MPa, 2.05 MPa and 1.95 MPa. The interface transfer strength
can be seen, the increases in radial stiffness lead to an increase in the increases as the h/s increases. The interface transfer strength increased
peak contact pressure at the end of the GC. The peak contact pressure by about 13% when the h/s increased from 0.016 to 0.04.
increased by about 61% when the radial stiffness changed from 0.010 to
0.018. 5.2.3. Ductility and energy dissipation
In this section, the ductility and energy dissipation of GCs with
5.2. Effects of h/s different h/s are investigated. The ductility coefficients and energy
dissipation coefficients for different h/s are presented in Table 6 and
5.2.1. Load-displacement curves Table 7, respectively. Observing the tables, the ductility coefficient and
Fig. 15 shows the effect of the h/s on the load-displacement curve. It the energy dissipation coefficient of the GC decrease with the increase in
can be seen that the ultimate bearing capacities increase with the in­ the h/s. The predicted ductility and energy dissipation coefficient
crease in the h/s. When the h/s is changed from 0.016 to 0.04, the axial decrease by 12% and 7%, respectively, when the h/s increases from 0.04
ultimate bearing capacity of the GCs is increased by about 13%. The to 0.016. The results indicate that the increase of the h/s negatively
effect of the h/s on the ultimate bearing capacity of the GCs is mainly affects the improvement of ductility and energy dissipation capacity of
achieved by changing the effective number of diagonal compression GCs.
struts. As the h/s decreases, the spacing between the shear keys in­
creases, which results in a reduction in the number of diagonal

8
T. Chen et al. Thin-Walled Structures 154 (2020) 106880

Table 6
Influence of h/s on ductility coefficient.
Specimen Yield Peak Ultimate Ductility coefficient
No μ
Py (kN) Δy (mm) Pmax (kN) Δmax (mm) 0.85Pmax (kN) Δu (mm)

GC-H-1 16,791 2.88 22,387 4.28 19,029 7.09 2.46

GC-H-2 16,010 2.87 21,347 4.35 18,145 7.42 2.58
GC-H-3 15,549 2.91 20,732 4.46 17,622 7.81 2.71
GC-H-4 14,792 2.98 19,723 4.55 16,765 8.18 2.75

Table 7
Influence of h/s on energy dissipation coefficient.
Specimen Elastic energy Plastic energy Total energy Energy
No dissipation dissipation dissipation dissipation
zone zone Etot (kN⋅mm) coefficient
Eel (kN⋅mm) Epl (kN⋅mm) ηE
GC-H-1 25,015 28,193 53,207 1.13
GC-H-2 23,596 28,258 51,854 1.20
GC-H-3 21,393 26,158 47,551 1.22
GC-H-4 20,972 25,446 46,418 1.21

5.2.4. Nominal average stress

This section depicts the results of studying the influence of the h/s on
the maximum stress of GCs. As shown in Fig. 16, the maximum Tresca
stress and the nominal average stress σ N of the grout layer tend to
decrease with the increase of the h/s. When the h/s increased from 0.016
to 0.04, the maximum Tresca stress of the grout layer decreased by
37.4%, while the nominal average stress σN decreased by 61.1%. When
the h/s varies from 0.016 to 0.04, the corresponding stress correlation
factors are 1.52, 1.93, 2.81, 3.93, respectively. Fig. 17. Contact pressure-displacement curve.

5.2.5. Grout material contact pressure

Fig. 17 reveals the effect of the h/s on the contact pressure at the end
of the grout connections. The curves of test pieces GC-H-1 to GC-H-4 are
relatively close, and all of them have a sudden drop after reaching the
ultimate contact pressure. The peak contact pressures of the GC-H-1 to
GC-H-4 specimens are 1.05 MPa, 0.97 MPa, 0.94 Mpa, and 1.09 MPa,
respectively, as shown in Fig. 18. It indicates that the h/s has no sig­
nificant effect on the contact pressure at the end of GCs.

5.3. Effects of Lg/Dp

5.3.1. Load-displacement curves

In Fig. 19, the load-displacement curve is illustrated for models of
different Lg/Dp. As depicted in Fig. 19, the increase of the Lg/Dp will
increase the axial ultimate bearing capacity of GCs. This is because of the

Fig. 18. Maximum contact pressure.

growth of Lg/Dp, which increases the contact area between the grout and
pile (sleeve) tube. On the other hand, the increase in Lg/Dp increases the
number of shear keys, which leads to an increase in the number of di­
agonal compression struts. The ultimate bearing capacities of the test
pieces GC-L-1 to GC-L-4 are 16,566 kN, 17,869 kN, 18,907 kN and
19,673 kN, respectively. The axial ultimate bearing capacity of the GC is
enhanced by about 19% when Lg/Dp is changed from 1.0 to 2.0.

5.3.2. Interface transfer strength

When GCs achieve the ultimate bearing capacity, the interface
transfer strength of four different radial stiffness specimens GC-L-1 to
GC-L-4 are 2.69 MPa, 2.23 MPa, 1.92 MPa and 1.59 MPa, respectively.
Fig. 16. Effect of h/s on the stress of GCs. As the Lg/Dp is increased from 1.0 to 2.0, the axial ultimate load capacity

9
T. Chen et al. Thin-Walled Structures 154 (2020) 106880

Table 9
The influence of the Lg/Dp on the energy dissipation coefficient.
Specimen Elastic energy Plastic energy Total energy Energy
No dissipation dissipation dissipation dissipation
zone zone Etot (kN⋅mm) coefficient
Eel (kN⋅mm) Epl (kN⋅mm) ηE
GC-L-1 15,072 25,309 40,381 1.68
GC-L-2 19,067 28,622 47,689 1.51
GC-L-3 23,070 30,763 53,833 1.33
GC-L-4 25,892 33,070 58,963 1.28

Fig. 19. The influence of the Lg/Dp on the load-displacement curve.

of the GC is increased by 19.0%, while the interface transfer strength is

reduced by 69.0%. The contact area between the pile (sleeve) tube and
grout increases linearly with the Lg/Dp, but the increase in the axial
Fig. 20. The effect of the Lg/Dp on the stress of GCs.
bearing capacity does not increase linearly, resulting in the interface
transfer strength decreasing. Therefore, it is uneconomical to increase
the ultimate bearing capacity of GCs by increasing the Lg/Dp.

5.3.3. Ductility and energy dissipation

Table 8 and Table 9 show ductility coefficients and energy dissipa­
tion coefficients for different Lg/Dp, respectively. Observing the tables,
when the Lg/Dp increased from 1.0 to 2.0, the ductility coefficient and
energy dissipation coefficient of pieces GC-L-1 to GC-L-4 decreased by
29.0% and 31.0%, respectively. This indicates that the increase in the
Lg/Dp negatively affects the ductility and energy dissipation capacity of
GCs.

5.3.4. Nominal average stress

The change of nominal average stress in the GC due to the change of
the Lg/Dp is revealed in Fig. 20. The results indicate that the maximum
Tresca stress σTresca-grout and the nominal average stress σ N of the grout
layer tend to decrease with the increase in the Lg/Dp. The predicted
maximum Tresca stress σTresca-grout and nominal average stress σ N de­
creases by 47.8% and 46.9%, respectively, when the Lg/Dp increases
from 1.0 to 2.0. In contrast, the increase in radial stiffness has no sig­
nificant effect on the maximum Mises stress σmises-tube of the steel tube. Fig. 21. Contact pressure-displacement curve.
The corresponding stress correlation factors of pieces GC-L-1 to GC-L-4
are 1.50, 2.16, 2.63, and 2.80, respectively. shown in Fig. 22. As the Lg/Dp increases, the peak contact pressure de­
creases. The aspect ratio varies from 1.0 to 2.0, and the peak contact
5.3.5. Grout material contact pressure pressure at the end of the GC is reduced by about 32%.
Fig. 21 shows the effect of the Lg/Dp on the contact pressure at the
end of the GCs. Similar to the GC-H specimens, the GC-L-1 to GC-L-4 6. Conclusions
specimens showed a sudden drop after reaching the peak contact pres­
sure, and the corresponding contact pressure peaks of GC-L-1 to GC-L-4 We constructed 3D finite element models to study the mechanical
are 1.43 MPa, 1.17 MPa, 1.06 Mpa, and 0.97 MPa, respectively, as behavior of large-diameter GCs with high-strength grout under axial

Table 8
The influence of the Lg/Dp on the ductility coefficient.
Specimen Yield Peak Ultimate Ductility coefficient
No μ
Py (kN) Δy (mm) Pmax (kN) Δmax (mm) 0.85Pmax (kN) Δu (mm)

GC-L-1 12,425 2.48 16,566 3.97 14,081 7.01 2.82

GC-L-2 13,402 2.81 17,869 4.49 15,189 7.39 2.63
GC-L-3 14,181 3.16 18,907 4.96 16,071 7.68 2.43
GC-L-4 14,755 3.49 19,673 5.29 16,722 7.59 2.18

10
T. Chen et al. Thin-Walled Structures 154 (2020) 106880

tube and the nominal average stress σ N. The calculation results

show that the stress correlation coefficient ηcor of the GCs varies
between 1.5 and 4. This calculation result can provide an
important reference value for practical engineering in the pre­
liminary design stage.

Declaration of competing interest

The authors declared that they have no conflicts of interest to this

work.

CRediT authorship contribution statement

Tao Chen: Conceptualization, Methodology, Software, Supervision,

Funding acquisition. Chengcheng Cao: Data curation, Writing - original
draft, Writing - review & editing, Investigation. Chihai Zhang: Visual­
ization, Investigation, Software, Validation, Formal analysis. Xian
Wang: Investigation, Writing - review & editing. Ke Chen: Investiga­
tion, Writing - review & editing. Guokai Yuan: Investigation, Writing -
Fig. 22. Maximum contact pressure. review & editing.

loading. The finite element model was validated by comparison with the Acknowledgement
results of published literature. It was proved that the finite element
model could accurately capture the failure mode, load-displacement Special Fund for Promoting Economic Development in Guangdong
curve, and stress distribution of the GC under axial loading. Subse­ Province: Research on real-time assessment of strength and fatigue of
quently, a parametric analysis was carried out to study the effects of offshore wind support structures based on big data (Contract of
radial stiffness, h/s of shear keys, and Lg/Dp on the axial force behavior Guangdong Natural Resources Department [2019]019)
of the GC. Based on the research results, the following conclusions can
be drawn: References

[1] H. Wu, L. Zhang, Contrastive analysis of transient potential of wind turbine

(1) The finite element model proposed in this paper can correctly onshore and offshore in lightning strokes, in: 2015 IEEE Adv. Inf. Technol.
reflect the mechanical behavior of GCs under axial loading and Electron. Autom. Control Conf., IEEE, Chongqing, China, 2015, pp. 1206–1209,
can be used for the further investigation of the mechanical https://doi.org/10.1109/IAEAC.2015.7428751.
[2] Global Wind Energy Council, Global Wind Report, 2018, 2019, https://gwec.net/
behavior of GCs. wp-content/uploads/2019/04/GWEC-Global-Wind-Report-2018.pdf.
(2) The increase in radial stiffness can strengthen the friction and [3] C.W. Zheng, C.Y. Li, J. Pan, M.Y. Liu, L.L. Xia, An overview of global ocean wind
mechanical interlock between the pile (sleeve) tube and the energy resource evaluations, Renew. Sustain. Energy Rev. 53 (2016) 1240–1251,
https://doi.org/10.1016/j.rser.2015.09.063.
grout, which leads to a 14% improvement of the ultimate bearing [4] A. Rahim, P. Sparrevik, A. Mirdamadi, Structural health monitoring for offshore
capacity of the GC as the radial stiffness increased from 0.01 to wind turbine towers and foundations, in: Offshore Technol. Conf., 2018, pp. 1–18,
0.18. At the same time, as the radial stiffness increases, the https://doi.org/10.4043/29041-ms.
[5] M. Seidel, Substructures for offshore wind turbines current trends and
restraining effect of the pile tube and sleeve tube on the grout developments, Festschrift Peter Schaumann (2014) 363–368, https://doi.org/
layer increases, and the compressive strength and ductility of the 10.2314/GBV:77999762X.
grout will be improved. Therefore, with the radial stiffness in­ [6] F. Selot, D. Fraile, G. Brindley, Offshore Wind in Europe-Key Trends and Statistics
2018, Wind Eur, 2019, pp. 1–38, https://doi.org/10.1016/S1471-0846(02)80021-
creases, the ductility and energy dissipation capacity of the GCs X.
increase, and the end pressure of the grout increases also. [7] X. Wang, T. Chen, Q. Zhao, G. Yuan, J. Liu, Fatigue evaluation of grouted
(3) The h/s of the shear keys affects the number of diagonal connections under bending moment in offshore wind turbines based on ABAQUS
scripting interface, Int. J. Steel Struct. 16 (2016) 1149–1159, https://doi.org/
compression struts, which is important for providing axial
10.1007/s13296-016-0050-7.
bearing capacity. Therefore, with the increase in the h/s, the ul­ [8] C.J. Billington, G.H.G. Lewis, The strength of large diameter grouted connections,
timate bearing capacity and interface transfer strength of the GCs in: Offshore Technol. Conf., Offshore Technology Conference, Houston, Texas,
are improved. On the contrary, the ductility coefficient and 1978, pp. 291–297, https://doi.org/10.4043/3083-ms.
[9] D.I. Karsan, N.W. Krahl, New API equation for grouted pile-to-structure
nominal average stress of the steel tube decreased by 12% and connections, in: Offshore Technol. Conf., Offshore Tachnology Conference,
61%, respectively, when the h/s increased from 0.016 to 0.04. Houston, Texas, 1984, pp. 49–54.
However, it has little effect on the maximum contact pressure of [10] T. Yamasaki, M. Hara, C. Takahashi, Static and dynamic tests on cement-grouted
pipe-to-pipe connections, in: Offshore Technol. Conf., Offshore Technology
grout. Conference, Houston, Texas, 1980, pp. 467–475, https://doi.org/10.4043/3790-
(4) Raising the Lg/Dp of the GCs can increase the contact area be­ ms.
tween the steel tube and grout, which results in a slight increase [11] W.B. Lamport, J.O. Jirsa, J.A. Yura, Strength and behavior of grouted pile-to-sleeve
connections, J. Struct. Eng. 117 (1991) 2477–2498.
in the ultimate bearing capacity of GCs. When the Lg/Dp is [12] W. Aritenang, A.S. Elnashai, P.J. Dowling, B.C. Carroll, Failure mechanisms of
increased from 1.0 to 2.0, the axial ultimate load capacity is weld-beaded grouted pile/sleeve connections, Mar. Struct. 3 (1990) 391–417,
increased by 19.0%, while the interface transfer strength is https://doi.org/10.1016/0951-8339(90)90011-F.
[13] W. Aritenang, A.S. Elnashai, P.J. Dowling, Analysis-based design equations for
reduced by 69.0%. From this point of view, it is uneconomical to
composite tubular connections, Eng. Struct. 14 (1992) 195–204, https://doi.org/
increase the Lg/Dp of GCs to increase the ultimate bearing ca­ 10.1016/0141-0296(92)90031-K.
pacity. At the same time, increasing the aspect ratio negatively [14] L.F. Boswell, C. D’Mello, The experimental behaviour of grouted connections for
construction and repair of offshore platforms, in: Behav. Offshore Struct., 1985,
affected the ductility of GCs, but it can reduce the contact pres­
pp. 835–844.
sure at the end of the grout. [15] N.W. Krahl, D.I. Karsan, Axial strength of grouted pile-to-sleeve connections,
(5) The stress correlation coefficient ηcor can be used to characterize J. Struct. Eng. 111 (1985) 889–905, https://doi.org/10.1061/(asce)0733-9445
the relationship between the maximum Mises stress of the steel (1985)111:4(889).

11
T. Chen et al. Thin-Walled Structures 154 (2020) 106880

[16] A.B. Sele, A.S. Veritec, H.B. Kje�

oy, Background for the new design equations for Thin-Walled Struct. 132 (2018) 522–536, https://doi.org/10.1016/j.
grouted connections in the DNV draft rules for fixed offshore structures, in: tws.2018.09.025.
Offshore Technol. Conf., Houston, Texas, 1989, pp. 463–474, https://doi.org/ [34] J.H. Lee, D.H. Won, Y.J. Jeong, S.H. Kim, Y.J. Kang, Interfacial shear behavior of a
10.4043/6163-ms. high-strength pile to sleeve grouted connection, Eng. Struct. 151 (2017) 704–723,
[17] Z. Wang, Y. Zhang, F. Chen, G. Wang, L. Wang, J. Jiang, Axial bearing capacity of https://doi.org/10.1016/j.engstruct.2017.08.035.
large-diameter grouted connections analyzed by means of a simplified double shear [35] L.F. Boswell, C. D’Mello, The fatigue strength of grouted repaired tubular
test, Construct. Build. Mater. 134 (2017) 245–253, https://doi.org/10.1016/j. members, in: Offshore Technol. Conf., Offshore Technology Conference, 1986,
conbuildmat.2016.11.123. pp. 147–152, https://doi.org/10.4043/5307-ms.
[18] P. Schaumann, J. Henneberg, A. Raba, Axially loaded grouted connections in [36] F. Wilke, Load Bearing Behavior of Grouted Joints Subjected to Predominant
offshore conditions using ordinary portland cement, Proc. 12th Int. Conf. Adv. Bending, Shaker Verlag Gmbh, Germa, 2013.
Steel-Concrete Compos. Struct. (2018) 541–547, https://doi.org/10.4995/ [37] T. Chen, X. Wang, X.L. Gu, Q. Zhao, G.K. Yuan, J.C. Liu, Axial compression tests of
asccs2018.2018.6944. grouted connections in jacket and monopile offshore wind turbine structures, Eng.
[19] S. Anders, L. Lohaus, Optimized high-performance concrete in grouted Struct. 196 (2019) 1–19, https://doi.org/10.1016/j.engstruct.2019.109330.
connections, Tailor Made Concr. Struct. Stoelhorst. (2008) 369–374, https://doi. [38] T. Chen, X. Wang, Q. Zhao, G. Yuan, A numerical investigation on grouted
org/10.1201/9781439828410.ch63. connections for offshore wind turbines under combined loads, J. Mar. Eng.
[20] I. Lotsberg, Structural mechanics for design of grouted connections in monopile Technol. 18 (2019) 134–143, https://doi.org/10.1080/20464177.2018.1507443.
wind turbine structures, Mar. Struct. 32 (2013) 113–135, https://doi.org/10.1016/ [39] DNV(Det Norske Veritas), DNVGL-ST-0126 Support Structures for Wind Turbines,
j.marstruc.2013.03.001. 2018. http://www.dnvgl.com.
[21] Z. Li, T. Chen, X. Wang, Behavior of flat grouted connections subjected to lateral [40] J. Lubliner, J. Oliver, S. Oller, E. Onate, A plastic-damage model for concrete, Int.
pressure and vertical load, Construct. Build. Mater. 212 (2019) 329–341, https:// J. Solid Struct. 25 (1989) 299–326.
doi.org/10.1016/j.conbuildmat.2019.04.005. [41] J. Lee, G.L. Fenves, Plastic-damage model for cyclic loading of concrete structures,
[22] M.S. Andersen, P. Petersen, Structural design of grouted connection in offshore J. Eng. Mech. 124 (1998) 892–900.
steel monopile foundations, Glob. Wind. (2004) 1–13. [42] J.G. Nie, Y.H. Wang, Comparison study of constitutive model of concrete in
[23] L.P. Nielsen, Finite element analysis of large diameter grouted connections, in: ABAQUS for static analysis of structures, Gongcheng Lixue/Engineering Mech. 30
ASME 2007 26th Int. Conf. Offshore Mech. Arct. Eng., American Society of (2013) 59–67þ82, https://doi.org/10.6052/j.issn.1000-4750.2011.07.0420 (in
Mechanical Engineers Digital Collection, San Diego, California, USA, 2007, Chinese).
pp. 449–457, https://doi.org/10.1115/omae2007-29199. [43] Z. Guo, Principles of Concrete Structures, Tsinghua University Press, 1999 (in
[24] P. Schaumann, A. Bechtel, S. Lochte-Holtgreven, Fatigue performance of grouted Chinese).
joints for offshore wind energy converters in deeper waters, in: Twent. Int. Offshore [44] X. Ｒen, J. Li, Calculation of concrete damage and plastic deformation, Build.
Polar Eng. Conf., International Society of Offshore and Polar Engineers, Beijing, Struct. 45 (2015) 29–31, https://doi.org/10.19701/j.jzjg.2015.02.007 (in
China, 2010, pp. 672–679. Chinese).
[25] T. L€ohning, M. Voßbeck, M. Kelm, Analysis of grouted connections for offshore [45] D. Darwin, C.W. Dolan, A.H. Nilson, Design of Concrete Structures, 2016. New
wind turbines, in: Proc. Inst. Civ. Eng. - Energy, 2013, pp. 153–161, https://doi. York, USA.
org/10.1680/ener.12.00009. [46] G.L. Dnv, Analysis of Grouted Connections Using the Finite Element Method, 2016.
[26] Y.H. Lin, C.H. Tsai, C.J. Lian, H.J. Lin, H.H. Huang, Design and analysis for Norway.
improved grout joint of offshore monopile foundation, in: Ocean. 2014 - TAIPEI, [47] C.J. Billington, I.E. Tebbett, The basis for new design formulae for grouted jacket to
2014, pp. 1–6, https://doi.org/10.1109/OCEANS-TAIPEI.2014.6964382. pile connections, in: Offshore Technol. Conf., Offshore Technology Conference,
[27] G. Prakhya, C. Zhang, N. Harding, Grouted connections for monopiles-limits for 1980, pp. 449–458, https://doi.org/10.4043/3788-ms.
large wind turbines, Struct. Eng. 90 (2012) 30–45. [48] API Rp2A-WSD, Planning, Designing, and Constructing Fixed Offshore Platforms-
[28] S. Marion, A. Johansen, G. Solland, T. Nybø, Testing of jacket pile sleeve grouted Working Stress Design, American Petroleum Institute, Houston, 2014.
connections exposed to shear forces and bending moments, Mar. Struct. 59 (2018) [49] R. Park, Ductility evaluation from laboratory and analytical testing, in: Proc. 9th
401–422, https://doi.org/10.1016/j.marstruc.2018.02.008. World Conf. Earthq. Eng., Tokyo-Kyoto Japan, 1988, pp. 605–616. http://www.iit
[29] T. Chen, X. Wang, Q. Zhao, G. Yuan, A numerical investigation on grouted k.ac.in/nicee/wcee/article/9_vol8_605.pdf.
connections for offshore wind turbines under combined loads, J. Mar. Eng. [50] Chinese Building Standard Committee, National Standard of the People’s Republic
Technol. 18 (2019) 134–146, https://doi.org/10.1080/20464177.2018.1507443. of China: Specification for Seismic Test of Buildings (JGJ/T 101-2015), China
[30] T. Chen, Z. Xia, X. Wang, Q. Zhao, G. Yuan, J. Liu, Experimental study on grouted Architecture & Building Press, Beijing, China, 2015.
connections under static lateral loading with various axial load ratios, Eng. Struct. [51] T. Zhou, Y. Jia, M. Xu, X. Wang, Z. Chen, Experimental study on the seismic
176 (2018) 801–811, https://doi.org/10.1016/j.engstruct.2018.09.056. performance of L-shaped column composed of concrete-filled steel tubes frame
[31] E. Lozano-Minguez, A.J. Kolios, F.P. Brennan, Multi-criteria assessment of offshore structures, J. Constr. Steel Res. 114 (2015) 77–88, https://doi.org/10.1016/j.
wind turbine support structures, Renew. Energy 36 (2011) 2831–2837, https://doi. jcsr.2015.07.009.
org/10.1016/j.renene.2011.04.020. [52] D. Jiang, C. Xiao, T. Chen, Y. Zhang, Experimental study of high-strength concrete-
[32] P. Schaumann, F. Wilke, Design of large diameter hybrid connections grouted with steel plate composite shear walls, Appl. Sci. 9 (2019) 2820, https://doi.org/
high performance concrete, in: Seventeenth Int. Offshore Polar Eng. Conf., 10.3390/app9142820.
International Society of Offshore and Polar Engineers, Lisbon, Portugal, 2007, [53] Y. Zheng, T. Usami, Ductility evaluation procedure for thin-walled steel structures,
pp. 340–347. J. Struct. Eng. 126 (2000) 1312–1319. J. Struct. Eng. 2000.126:1312-1319.
[33] T. Chen, Z. Li, X. Wang, G. Yuan, J. Liu, Experimental study on ultimate bending
performance of grouted connections in offshore wind turbine support structures,

12