Proc. of the Sixth Intl. Conf.

Advances in Civil, Structural and Mechanical Engineering - CSM 2018

Copyright © Institute of Research Engineers and Doctors, USA. All rights reserved.
ISBN: 978-1-63248-150-4 doi: 10.15224/978-1-63248-150-4-34

Numerical Modelling of large diameter monopiles

supporting OWT-Head stiffness assessment in Non-
homogeneous Soils
AISSA Mohammed Hemza1, AMAR BOUZID Djillali2

Abstract—Many countries are now planning to build new wind The design is based on the p-y method where the pile is
farms with high capacity up to 5MW. Consequently, the size of the simulated as a beam and the soil as series of elastic springs,
foundation increased. This kind of structures are subjected to this was described by Reese et al. (1977)[3] and O’Neill &
fatigue damage from environmental loading mainly due to wind Murchison (1983)[4] who tested full scale piles with small
and waves as well as from cyclic loading imposed through the
rotational frequency (1P) through mass and aerodynamic
diameter (0.61m). In this topic, a recent report shows the
imbalances and from the blade passing frequency (3P) of the wind limitations of using conventional models for Offshore Wind
turbine which makes their behavior dynamically very sensitive. Turbine foundations (TRB, 2011)[5]. This report stated that
That is why natural frequency must be determined with accuracy in the diameter and slenderness are much more than
order to avoid the resonance of the system. experienced for oil and gaz platform foundations and
This paper presents a numerical procedure which combines the consequently this can reduce the life of the Wind Turbines.
finite element method based on the 2D discretization in the radial Also many researches approve that the famous p-y curves
plane and the Fourier series based on expansion of displacements method is not conservative for the design of the Offshore
and forces in the circumferential direction. Firstly analytical Wind Turbines foundations, for exemple: Lesny &
expressions of stiffnesses of foundations with large diameter
embedded in soils with different stiffness profile are established.
Wiemann (2006)[6], Augustesen et al. (2009)[7], Andersen
Secondly and in order to check the accuracy of the proposed et al. (2012)[8], Harte et al. (2012)[9] and Swagata and
formulas, a mathematical model approach based on non- Sumanta (2014)[10].
dimensional parameters is used to calculate the natural frequency
taking into account the soil structure interaction (SSI) compared
with measured frequency in five wind farms selected from the
litterature.
Keywords—Offshore Wind Turbines; Semi-analytical FE
analysis; DNV/Risø; Monopiles foundations; Natural frequency.

I. Introduction :
A serious drive to develop Offshore Wind Energy sector
is known in the last 20 years. Most Offshore Wind
Converters are founded in Monopiles. These kinds of
foundations are made of steel or concrete with a diameter
( ) ranging between 4m and 7m, and embedded length
( ) less than 30m. Inhence, they are subjected to lateral
loading ( ) and overturning moment ( ) due to wind and
waves (Figure 1).
The design procedures for Monopiles foundations are
based on the experience gained from the oil and gaz field, on
which the American Petroleum Institute (API) method[1]
has been based.
This semi-empirical approach was then included in the
recommended practice of several Offshore Wind Turbines
design codes (DNV/Risø[2] for exemple) and already used
to simulate the lateral response of monopiles foundations.
Fig. 1 Loads acting on Offshore Wind Turbine Structure.
1
PhD student
The Offshore Wind Turbines experienced during their life
Department of Material Engineering, Faculty of Sciences and Technology,
University of Medea service a high dynamic loading cyclic in nature between 106
Algeria and 108 cycles. The main external excitations are:
2
Professor (1) Environmental loading due to wind and waves. The
Department of Civil Engineering, Faculty of Technology, University Saad
Dahleb of Blida predominant wave frequency is generally 0.1 Hz.
Algeria (2) Rotor loading at a frequency which is referred to as
1P

62
 Proc. of the Sixth Intl. Conf. Advances in Civil, Structural and Mechanical Engineering - CSM 2018
Copyright © Institute of Research Engineers and Doctors, USA. All rights reserved.
ISBN: 978-1-63248-150-4 doi: 10.15224/978-1-63248-150-4-34

(3) The blade passing frequency is a forced loading Where ̅ , ̅ and ̅ are the amplitudes of nodal loading on the
generated from the effect of wind deficiency that first harmonic. The load system displacements of Eq. (2)
occurs at each passes of the blade through the will reduce to:
shadow of the tower. It refers to as 2P for two =̅ , v= ̅ , w= ̅ (4)
blades and 3P for three blades.
The natural frequency must be designed to avoid III. Soil stiffness and FE mesh
considered in the present
resonance between (1P) and (3P) (Figure 2). So the soil
stiffness must be determined accurately from the soil data
for a good estimation of the natural frequency of the system.
The main aim of this paper is to propose new design
analysis:
formulas of stiffness coefficients by a semi-analytical In our analysis we considered a cylindrical pile with
approach using the finite element method (Amar Bouzid et is the diameter of the monopile, and it’s length , Young’s
al. 2004)[11]. Two soil profiles are considered: Gibson’s Modulus . is the horizontal load acting on the head of
soil and parabolic soil, the nature of the interface between the pile and is the corresponding overturning moment.
the soil and the pile is taken into account: rough or The soil is described by its Young’s modulus and
S=smooth (Amar Bouzid and Vermeer, 2007)[12]. Poisson’s ratio . In the parametric study we assume the
Thereafter, the accuracy of the obtained formulae is stiffness linear increasing with depth which is a typical
checked using a method based on Euler-Bernoulli Beam- consideration for sand or Gibson’s soil profile (Scott,
Column with elastic end supports as reported by Adhikari 1981)[19] and the stiffness increasing with square root of
and Bhattacharya (2012)[13] and (Laszlo Arany et al. depth for parabolic soil profile. This consideration of soil
2014)[14]. non-homogeneity has proved to be more realistic in many
practical cases where the effective stresses increase with
II. Semi Analytical Method depth. The soil modulus is usually taken to have a power
Background: law variation with depth (Booker et al. 1985)[20] as
expressed by the equation:
This so called semi analytical FE approach was proposed ( ) ( ⁄ ) (5)
first by WILSON (1965)[15] for FE analysis of axisymmetric Where is the soil modulus at a depth z equal to the
structures loaded non-axisymmetrically and later used by monopile diameter and is an exponent that varies
many authors (Cook et al. (1989 for instance)[16]. The main between zero and one. Equation (5) includes a soil with a
idea of this method is to use Fourier Series to resolve three parabolic variation of stiffness with depth for and a
dimensional problems as a two dimensional harmonic model Gibson’s soil for
and superposing each term result. We can found many The monopile has a diameter =3m with variable
applications of Semi-Analytical Approach in practical cases
slenderness ratio ⁄ ranging between 1 to 15. The
(For example Kim et al. 1994 [17]; Zienkiewicz and
relative rigidity of the pile/soil ratio is taken between 10 and
Taylor[18]). This method finds its applications in the present
problem. 106. The flexural rigidity of the pile ( ) =
The nodal loads applied to the monopile can be expanded in 79521564.04kN.m2 and =0.25. The poisson’s ration of the
Fourier series as: soil is taken equal to 0.499. A Fortran program named
̅ ∑ ̅ ̿ Pile-Joint has been used. The model consists of 1326 eight-
̅ ̅ ̿ noded quadrilateral elements, whose 200 elements for
∑ (1)
modeling the monopile and 1126 for the surrounding soil
̅ ∑ ̅ ̿ (Figure 3). To ensure a good accuracy and because the
Where R, Z and T are the radial, axial and circumferential diameter of the monopile is large we refine the mesh in the
components respectively with respect to the θ = 0 as plane vicinity of the monopile and around the interface area. The
of symmetry. distance under the monopile tip is one length .
The displacement can be expressed in the form of Fourier In the horizontal direction, the distance is taken equal to
series: 40 times the radius of the monopile. The interface is taken
=∑ ̅ +∑ ̿ into account for this study. We have to cases: Rough
=∑ ̅ +∑ ̿ (2) Interface (normal rigidity = and the shear rigidity
=∑ ̅ +∑ ̿ = ) and Smooth Interface (normal rigidity =
where ̅ , ̅ , ̅ are the amplitudes of displacements that and the shear rigidity = ).
are symmetric with respect to the ( ) plane and ̿ , ̿ ,
̿ are the amplitudes of displacement that are antisymmetric A. Results of the Stiffness Analysis:
with respect to the ( ) plane, is harmonic number,
Various criteria for rigid or flexible behavior have been
and is the total number of harmonic terms considered in
suggested in the literature, for example Poulos and Hull
the series.
(1989)[21]. The results of the stiffness coefficients
For a Monopile subjected to a lateral and/or an overturning
moment, only the second term for i=1 survives, because this ( ) have been plotted as function of ⁄ and
loading has a plane of symmetry. In this situation the compared with existing solutions proposed by Higgins et al.
components of loading in Eq. (1) reduce to: (2013)[22] in Table 1 and Table 2.
= ̅ , = ̅ , =̅ (3)

63
 Proc. of the Sixth Intl. Conf. Advances in Civil, Structural and Mechanical Engineering - CSM 2018
Copyright © Institute of Research Engineers and Doctors, USA. All rights reserved.
ISBN: 978-1-63248-150-4 doi: 10.15224/978-1-63248-150-4-34

Fig. 2 Dyanmic design approach showing the forcing frequency as function of power spectral density.

B. Discussion of stiffness results:

The obtained results show a good agreement with
Higgins et al. In Gibson’s soil profile for rough interface
case. Indeed, the interface state which has a great effect on
the results above in both Gibson’s soil and Parabolic soil
profiles.

IV. Natural frequency

assessment:
In the design approach the support structure modeled by
static springs which lead to an independency of the stiffness
coefficients on the natural frequency of the system, that why
the Soil Structure Interaction (SSI) is very important in any
dynamic analysis of the system. S. Adhikari & S.
Bhattacharya (2012)[13] reported the equation of motion of
the beam and includes with analytical resolution non-
dimensional parameters of the foundation stiffness (In this
case we find just and terms).
Fig. 3 Geometric model taken in this study. Laszlo Arany et al. (2014)[14] give a coupled stiffness term
( ) and prove by mean of sensitive analysis that the
The Table 3 shows the stiffness equations obtained in the
determination of the natural frequency according to two
present analysis for the parabolic soil case. For this specific
terms is not sufficient.
case, expressions for monopile head stiffness were no
For our analysis we choose six Offshore Wind Turbines
localized in the literature.
situated in the North Sea and we compare results of the
The results of the stiffness coefficients ( ) computed natural frequency determined by two
have been tabulated as function of ⁄ and compared approaches(the approach used in the present study with the
with existing solutions proposed by Higgins et al. equations founded in the DNV/Risø Guideline) with the
(2013)[22] in Table I and Table II. . measured one. Table IV and Table V gives the obtained
The Table III shows the stiffness equations obtained in results from this dynamic analysis:
the present analysis in the parabolic soil case.

64
 Proc. of the Sixth Intl. Conf. Advances in Civil, Structural and Mechanical Engineering - CSM 2018
Copyright © Institute of Research Engineers and Doctors, USA. All rights reserved.
ISBN: 978-1-63248-150-4 doi: 10.15224/978-1-63248-150-4-34
TABLE I. Stiffness equations obtained in the present analysis in Gibson’s soil profile.

Relative Stiffness Rough Interface Smooth Interface

Term
.6944 .7482
Relative Horizontal 𝐾𝐻 ⁄𝐸𝑠𝐷 𝐷 = .647 (𝐿𝑝 ⁄𝐷𝑝 ) 𝐾𝐻 ⁄𝐸𝑠𝐷 𝐷 = . 4 (𝐿𝑝 ⁄𝐷𝑝 )
Stiffness
Relative Moment 𝐾𝑀 ⁄𝐸𝑠𝐷 𝐷3 = . 48(𝐿𝑝 ⁄𝐷𝑝 )3.633 𝐾𝑀 ⁄𝐸𝑠𝐷 𝐷3 = .8 5 (𝐿𝑝 ⁄𝐷𝑝 )3.6858
Stiffness
Relative Coupling 𝐾𝑀𝐻 ⁄𝐸𝑠𝐷 𝐷2 =− . 889(𝐿𝑝 ⁄𝐷𝑝 )2.687 𝐾𝑀𝐻 ⁄𝐸𝑠𝐷 𝐷2 =−=− .8967(𝐿𝑝 ⁄𝐷𝑝 )2.73 6

Stiffness

TABLE II. Stiffness equations proposed by Higgins et al. (2013) in Gibson’s soil profile.

Relative Stiffness Term Rough Interface

Relative Horizontal Stiffness 𝐾𝐻 ⁄𝐸𝑠𝐷 𝐷 = .9 6 (𝐿𝑝 ⁄𝐷𝑝 )2. 4 5
Relative Moment Stiffness 𝐾𝑀 ⁄𝐸𝑠𝐷 𝐷3 = .66 5(𝐿𝑝 ⁄𝐷𝑝 )3.94 7
Relative Coupling Stiffness 𝐾𝑀𝐻 ⁄𝐸𝑠𝐷 𝐷2 =− .6 44(𝐿𝑝 ⁄𝐷𝑝 )3. 6 8

TABLE III. Stiffness equations obtained in the present analysis in Parabolic soil profile.

Relative Stiffness Rough Interface Smooth Interface

Term
.9958 . 675
Relative Horizontal 𝐾𝐻 ⁄𝐸𝑠𝐷 𝐷 = .83 (𝐿𝑝 ⁄𝐷𝑝 ) 𝐾𝐻 ⁄𝐸𝑠𝐷 𝐷 = . 555(𝐿𝑝 ⁄𝐷𝑝 )
Stiffness
Relative Moment 𝐾𝑀 ⁄𝐸𝑠𝐷 𝐷3 =3.9374(𝐿𝑝 ⁄𝐷𝑝 )2.57 7 𝐾𝑀 ⁄𝐸𝑠𝐷 𝐷3 = .5554(𝐿𝑝 ⁄𝐷𝑝 )2.6863
Stiffness
Relative Coupling 𝐾𝑀𝐻 ⁄𝐸𝑠𝐷 𝐷2 =−=− .94 (𝐿𝑝 ⁄𝐷𝑝 ) .7824
𝐾𝑀𝐻 ⁄𝐸𝑠𝐷 𝐷2 =−=− . 997(𝐿𝑝 ⁄𝐷𝑝 ) .8649

Stiffness
TABLE IV. Computed and Measured Natural Frequency of various Wind Turbines in Gibson’s soil profile.

Wind Turbine Computed Computed Computed Measured

Name Frequency Frequency Frequency Frequency (Hz)
with Rough with with
Interface Smooth DNV/Risø
Interface (Hz)
Walney 1 0.3454 0.3446 0.3342 0.3500
Lely A2 0.7613 0.7608 0.7358 0.6300
North Hoyle 0.4481 0.4481 0.4258 0.3500
Irene Vorrink 0.5509 0.5505 0.5297 0.5400-0.5600
Sheringham Shoal 0.4996 0.4989 0.4648 0.8500-0.9600
Kentish Flats 0.6282 0.6273 0.5460 0.8500-0.9600

TABLE V. Computed and Measured Natural Frequency of various Wind Turbines in Parabolic soil profile.

Wind Turbine Name Computed Frequency Computed Frequency Computed Measured

with Rough Interface with Smooth Interface Frequency with Frequency
DNV/Risø (Hz) (Hz)
Walney 1 0.3454 0.3445 0.3315 0.3500
Lely A2 0.7609 0.7603 0.7302 0.6300
North Hoyle 0.4480 0.4480 0.4212 0.3500
Irene Vorrink 0.5506 0.5501 0.5251 0.5400-
0.5600
Sheringham Shoal 0.4990 0.4981 0.4574 0.8500-
0.9600
Kentish Flats 0.6272 0.6259 0.5302 0.8500-
0.9600
65
 Proc. of the Sixth Intl. Conf. Advances in Civil, Structural and Mechanical Engineering - CSM 2018
Copyright © Institute of Research Engineers and Doctors, USA. All rights reserved.
ISBN: 978-1-63248-150-4 doi: 10.15224/978-1-63248-150-4-34

[6] K. Lesny; J. Wiemann. ―Finite Element Modeling of Large Diameter

A. Discussion of the dynamic results: Monopoles for Offshore Wind Energy Converters‖ 2006,
A preliminary analysis of the results above shows that GeoCongress.
the difference between the natural frequency in the two [7] A.H. Augustesen, K.T.Brødbaek, M. Møller; S.P.H. Sørensen, L.B.
Ibsen, T.S. Pedersen and L. Andersen. ―Numerical Modelling of
profile is very smaller. Furthermore, the interface state has Large Diameter Steel Piles at Horns Rev‖. Proceedings of the
no effect on the natural frequency of the system. In Twelfith International Conference on Civil, Structure &
comparison with the measured natural frequency the error is Environmental Engineering Computing. 2009, Civil-Comp Press.
admissible in Walney 1 and Irene Vorrink Wind farms, but [8] L.V. Andersen, M.J. Vahdatirad, M.T. Sichani and J.D. Sørensen.
―Natural Frequencies of Wind Turbines on Monopile Foundations in
for example in Kentish flats the error is more than 60% Clayey Soils– A Probabilistic Approach‖. Computers & Geotechnics
which is not allowable. 2012, 43, 1-11.
In all cases, the DNV/Risø approach underestimates the [9] M. Harte, B. Basu and S.R.K. Nielsen. ―Dynamic Analysis of Wind
natural frequency (Walney 1, Irene Vorrink, Sheringham Turbines Including Soil-Structure Interaction‖. Engineering Structures
Shoal and Kentish flats cases) or overestimates the natural 2012, 45, 509-518.
frequency (Lely A2 and North Hoyle cases). Returning to [10] S. Bisoi, S. Haldar. ―Dynamic Analysis of Offshore Wind Turbine in
Clay Considering Soil-Monopile-Tower Interaction‖. Soil Dynamics
our results we remark that the computed natural frequency & Eartquake Engineering 2014, 63, 19-35.
with the present approach is closer to those measured in four [11] Dj. Amar Bouzid, B.Tiliouine and P.A. Vermeer. ―Exact Formulation
cases: Walney 1, Irene Vorrink, Sheringham Shaol and of Interface Stiffness Matrix for Axisymmetric Bodies under Non-
Kentish flats Wind turbines. Axisymmetric Loading‖. Computers & Geotechnics 2004, (31) 75-87.
[12] Dj. Amar Bouzid, Pieter A. Vermeer. ―Effect of Interface
Conclusion: Characteristics on the Influence Coefficients of an Embedded Circular
This paper focused on the monopile head stiffness fo Footing under Horizontal and Moment Loading‖. Geotech Geol Eng
2007, 25:487-497.
Offshore Wind Turbinres using a numerical analysis. Analy
[13] S. Adhikari, S. Bhattacharya. ―Dynamic Analysis of Wind Turbine
tical expressions were proposed both Gibson’s soil and Towers on Flexible Foundations‖. Shock & Vibration 2012, (19) 37-
parabolic soil profiles with different interface states and 56.
validated with existing solutions [14] L. Arany, S. Bhattacharya, S.J.Hogan and J. Macdonald. ―Dynamic
the dynamic behavior of piles. Most monopiles are sensitive Soil-Structure Interaction Issues of Offshore Wind Turbines‖.
to dynamic/cyclic loading are those supporting Offshore Proceedings of the 9th International Conference on Structural
Dynamics, EURODYN 2014. Porto, Portugal, 30 June-2 July 2014.
Wind Turbines in a great depth which are caracterised by
[15] EL. Wilson. ―Structural Analysis of Axisymmetric Solids. J Am Inst
large diameter. Aeronaut Astronaut‖ 1965 .3(12):2269–74.
To check the applicability of the present study equations [16] R.D. Cook, D.S. Malkus, M.E. Plesha. ―Concepts and applications of
we use it in the calculation of the natural frequency of Wind finite element analysis‖, 3rd edition, John Wiley (1989).
Turbines structures and compared with the DNV/Risø
Guideline’s equations and measured data. The results of the [17] D.S. Kim, A. Bouchalkha, J.M. Jacob, J.F. Zhou, J.J. Song and J.F.
natural frequency of Wind Turbines structures and Klem. Phys. Rev. Lett. 1994, 72, 1572.
compared with the p-y equations and measured data. [18] O. C. Zienkiewicz ; R. L. Taylor. (2000). ―The Finite Element
Method: The basis‖. Volume 1, Fifth edition. Butterworth-
In a hand, the results of our study show a well accuracy Heinemann, 2000.
with the measured natural frequency contrary to the [19] R.F. Scott. ―Foundation Analysis‖. Englewood Cliffs, NJ : Prentice-
DNV/Risø results (Walney 1 and Irren Vorrink), this is due Hall, 1981.
to the rigid behavior of these piles, which leads to the [20] J.R. Booker, N.P. Baalam and E.H. Davis. ―The behavior of an elastic
useless of the equations founded in the famous DNV/Risø non-homogenous half-space‖, Int. J. Numer. Anal. Meth.
Guideline sited previously. Geomech.1985, 353-367
In the other hand, the error of the present method is [21] H.G. Poulos, T.Hull. ―The Role of Analytical Geomechanics in
Foundation Engineering‖. Foundation Engineering: Current Principles
much higher in the others case study (The error in Kentish & Practices.1989, 2, 1578–1606.
flats is more than 60% of the measured natural frequency). [22] W. Higgins, Celio Vasquez, Dipanjan Basu and D. V. Griffiths.
This is explained by the flexible behavior of the foundations, ―Elastic Solutions for Laterally Loaded Piles‖. Journal of
it means that the length of these piles exceeds the critical Geotechnical & Geoenvironmental Engineering 2013, ASCE.
length. So this make the applicability of our formulas
outside of them range.

References
[1] API, Recommended Practice for Planning, Design and Constructing
Fixed Offshore Platforms―Working Stress Design, 1997. American
Petroleum Institute.
[2] Det Norske Veritas (DNV). Offshore Standard DNV-OS-J101:
Design of Offshore Wind Turbine Structure, 2007.
[3] C. Lymon. Reese. ―Laterally Loaded Piles: Program Documentation‖.
Journal of the Geotechnical Engineering Division, 1977, 287-305.
[4] M.W. O’Neill; J.M. Murchinson. ―An Evaluation of p-y relationships
in Sand‖. 1983,. A report to the American Petroleum Institut.

[5] TRB. ―Structural Integrity of Offshore Wind Turbines‖.

Transportation Research Board, Speacial Report, 2011. 305, p.57.

66