You are on page 1of 17

Ocean Engineering 248 (2022) 110798

Contents lists available at ScienceDirect

Ocean Engineering
journal homepage: www.elsevier.com/locate/oceaneng

Load utilisation (LU) ratio of monopiles supporting offshore wind turbines:


Formulation and examples from European Wind Farms
Muhammad Aleem a, Subhamoy Bhattacharya a, *, Liang Cui a, Sadra Amani a,
Abdel Rahman Salem b, Saleh Jalbi c
a
Department of Civil and Environmental Engineering, University of Surrey, Guildford, GU2 7XH, United Kingdom
b
Arcadis Consulting, London, CR0 1EA, UK
c
Sea & Land Project Engineering, London, SE1 1UN, UK

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

Keywords: Offshore wind farms are currently being constructed worldwide, and most of the Wind Turbine Generator (WTG)
Offshore wind turbines structures are supported on single large-diameter steel piles, commonly known as monopile. One of the chal­
Soil structure interaction lenging design aspects is predicting the long-term deformation of the foundation and, in particular, the accu­
Soil resisting capacity
mulation of rotation which is a complex Soil-Structure Interaction (SSI) problem. Accumulation of rotation
Load utilisation
requires the estimation of Load Utilisation (LU) ratio (i.e., ratio of the load-carrying capacity of the foundation to
Moment capacity
Lateral capacity the applied loads from wind and wave). Estimation of LU for monopile is not trivial due to the simultaneous
action of lateral load and moments and needs the introduction of interaction diagram concepts. This paper
proposes methodologies to obtain LU for monopiles using three types of methods: (a) Simplified method, which is
based on closed-form solution (where the load effects are uncoupled) and can be carried out using spreadsheets
or pocket calculators; (b) Standard method based on non-linear Winkler spring (also known as p-y method) where
the load effects are also uncoupled; (c) Advanced method, which uses Finite Element Method (where the load
effects are coupled). Examples of monopiles are taken from European Wind Farms covering different ground
profiles: Gunfleet Sands (clay profile), Walney-I (sandy profile), London Array-I (layered profile) and Barrow-II
(layered profile) sites are analysed using all three methods. It is hoped that the methodology will be helpful in the
design optimisation stage.

cost of the monopile. There are three variables in a monopile: embedded


length, diameter and wall thickness. The parametric study carried out
1. Introduction
and reported in Arany et al. (2017) showed that increasing the diameter
often provides the highest cost-benefit ratio. Increasing the diameter
1.1. Motivation and background
also results in other challenges associated with installation vessels and
handling equipment.
One of the expensive offshore wind turbine structure components is
Fig. 2 shows the various lateral loads on the wind turbine structure,
the foundation, and it costs about 25%–34% of the overall wind farm
causing the overturning moment. They are wind, wave and loads due to
cost, see Bhattacharya (2014). Numerous aspects must be considered
1P (rotor frequency) and 3P (blade passing frequency). The readers are
while selecting the foundation, such as ground conditions, operation and
referred to Bhattacharya (2019) for details of the loads and estimate
maintenance (O&M), and transportation & installation (T&I). Monopile
them.
is the most common form of foundation due to its simplicity in shape and
Fig. 3 shows the simplified combined wind and wave loading in a
ease of fabrication. Table 1 and Fig. 1 shows the future of wind turbines,
monopile following the work of Arany et al. (2015, 2017) and Jalbi et al.
and it is evident that as the rated power increases, the rotor diameter and
(2019). It is clear from Fig. 3 that the load is asymmetric, and one of the
the hub height will also increase. As a result, the overturning moment in
design issues is the prediction of long-term tilt during the design
the monopile will also increase. Research is underway to optimise the

* Corresponding author. Department of Civil and Environmental Engineering, Faculty of Engineering and Physical Sciences, University of Surrey, Guildford, Surrey,
GU2 7SU, United Kingdom.
E-mail addresses: m.aleem@surrey.ac.uk (M. Aleem), S.Bhattacharya@surrey.ac.uk (S. Bhattacharya), l.cui@surrey.ac.uk (L. Cui), sadra.amani@surrey.ac.uk
(S. Amani), abdelrahman.salem@arcadis.com (A.R. Salem), saleh.jalbi@seaandland.co.uk (S. Jalbi).

https://doi.org/10.1016/j.oceaneng.2022.110798
Received 10 August 2021; Received in revised form 14 January 2022; Accepted 6 February 2022
Available online 19 February 2022
0029-8018/© 2022 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-
nc-nd/4.0/).
M. Aleem et al. Ocean Engineering 248 (2022) 110798

Nomenclature and symbols NTM = Normal Turbulence Model


NSS = Normal Sea State
O&M = Operating and Maintenance NWH = Normal Wave Height
OWT = Offshore Wind Turbine EWH = Extreme Wave Height
T&I = Transportation and Installation ETM = Extreme Turbulence Model
1P = Rotor frequency ESS = Extreme Sea State
3P = Blade passing frequency EOG = Extreme Operating Gust
RPM = Revolution Per Minute LP = Pile length
RNA = Rotor Nacelle Assembly DP = Pile diameter
LU = Load Utilisation Su = Undrained shear strength
R/A = (Resistance/Actions) which is Load Utilisation Ratio ø= Diameter
BNWF = Beam on a Non-linear Winkler Foundation pu = Ultimate soil pressure
DLc = Design Load case γ′ = Soil’s submerged unit weight
SSI = Soil-Structure-Interaction Φ’ = Soil’s angle of friction
H= Lateral load Es= Soil young’s modulus
M= Moment load ′
σv = Effective vertical stress
Mmax = Maximum bending moment Kp = Passive pressure coefficient
Mmin = Minimum bending moment NP = Lateral bearing capacity
Mmean = Mean bending moment UR = Rated wind speed
ULS = Ultimate Limit State pu = Ultimate soil pressure
SLS = Serviceability Limit State tp = monopile thickness
FLS = Fatigue Limit State MC = Mohr-Coulomb
HR = Lateral resistance capacity HS = Hardening Soil
MR = Moment resistance capacity

states), the residual pile deflection at mudline level (initial + accumu­


Table 1 lated) is limited by 0.2 m (IEC, 2009). Numerous load cases are provided
Development of offshore wind turbines. by DNV, (2016) and IEC (2005) to serve a design life of approximately
Rated Power Tower Rotor Top Mass Manufacturer
25–30 years; however, only a few are relevant to foundation design, as
(MW) Height (m) Diameter (m) (tonnes) shown by Arany et al. (2017) and Bhattacharya (2019). Several load
cases are derived from those provided in DNV, (2016) and used in this
3.6 75 107 220 SIEMENS
5 85 116 340 AREVA study.
6 100 154 360 SIEMENS There are no codes of practice for predicting the tilt of an offshore
8 110 164 550 VESTAS wind turbine foundation under the action of millions of cycles of load.
9.5 110 174 585a MHI VESTAS However, the scaled model tests (1g and centrifuge) could analyse the
9.5 110 164 560a MHI VESTAS
10 105 164 570a VESTAS
tilt factors and to scale to small model tests to predict the prototype
12–14 135 220 800-1100a GE consequences is a challenging task. The readers are referred to a
15 138 236 1100-1250a Vestas comprehensive review in Arany et al. (2017) and Bhattacharya (2019)
16 141 242 1200-1300a MySE for the methods. The main understanding that emerged from the scaled
a
Estimated values. model tests, numerical work and back-analysis of case studies are as
follows:
lifetime.
Fig. 4 schematically shows the tilt together with the “desirable”, i.e., (1) The parameters that control the accumulation of tilt are the
acceptable design. Here, the loading is operational, causing accumula­ loading ratios: (Mmin/Mmax) and (Mmax/MR), where Mmin and
tion and episodes of extreme events such as storms, hurricanes, etc. Mmax are the minimum and maximum mudline bending moments
As previously mentioned, one of the main design aspects of monopile corresponding to the appropriate load combination of wind and
supported wind turbine is the Serviceability Limit State (SLS) and, in the wave. The readers are referred to Fig. 3 for the definition of
particular, the overall allowable tilt taking into account installation Mmin and Mmax. On the other hand, MR (Moment of Resistance of
tolerances. The current allowable tilt is typically 0.5–0.75◦ through the the pile) is the ultimate static moment capacity of the pile
lifetime (DNV, 2011). The readers are referred to Chapter 3 of Bhatta­ considering soil failure around the pile (i.e., geotechnical failure)
charya (2019) for a detailed discussion on the SLS requirements. One of and strictly not the structural failure of the pile through buckling
the calculations that engineers carry out from the point of view of cer­ or yielding.
tification or due diligence is the prediction of tilt for the entire lifetime of (2) Through a numerical study based on a 10-step monopile design
the structure, i.e., typically 25–30 years. methodology, Jalbi et al. (2019) assessed these load ratios (i.e.,
It is considered prudent to discuss acceptable levels of deformations Mmin/Mmax and Mmax/MR) for 15 operating wind turbines in
and long-term tilt. The initial tilt of the monopile due to installation European waters. The study showed that the load ratio varies
must be within the allowed limit (the current limit set in the DNV code is widely, ranging from extreme one-way loading to extreme
0.25◦ ). The accumulated tilt due to millions of load cycles during the two-way loading. The ratio of Mmax/MR ranged between 10% and
lifetime of the wind turbine has to be within the allowed limit (current 20% for normal operating conditions and occasionally reached
limit is 0.25–0.5◦ ) (DNV, 2011). Alternatively, the total tilt (i.e., initial 40% for extreme loading conditions.
+ accumulated at the nacelle level) is limited by 0.5–0.75◦ ). Alterna­
tively, as part of the performance requirement considerations (limit Typically, the ratio (Mmax/MR) is effectively the conventional defi­
nition of the inverse of factor of safety (FOS) under Ultimate Limit State

2
M. Aleem et al. Ocean Engineering 248 (2022) 110798

(ULS). However, the calculation of MR is not straightforward as


monopiles carries not only moment (M) but also lateral load (H) and
vertical load (V). The effect of vertical load can be conveniently dis­
counted as the bottom part of the pile provides the most axial resistance
(shaft resistance and end-bearing), and soil around the top part of the
pile has the most resistance to lateral load and moment. Therefore, the
effect of lateral load (H) must be considered while estimating the
moment carrying capacity as the same soil that resists the lateral loads
and moments in a monopile. The problem is similar to the interaction
curves in structural engineering, where the effect of axial, lateral and
moment-resisting capacity of sections are calculated. In the context of
monopiles, we introduce the term “Load Utilisation ratio”, as is explained
in the next section.
The aim of the paper is as follows:

(a) To develop procedures to obtain the Load Utilisation ratio, which


can be easily implemented in practice. Three types of methods
(Simplified, Standard and Advanced) are considered based on the
input data requirement (soil parameters) and computing re­
sources necessary.
(b) Demonstrate the application of the method through example case
studies from European Wind Farms.

1.2. Formulation for load utilisation (LU) ratio for monopiles

Various combinations of horizontal loads (H) and overturning


moment loads (M) may lead to the failure of a foundation system, and
the locus of such points will provide a 2D curve, as shown in Fig. 5. This
is similar to an Interaction diagram used in structural engineering and
can generate a Resistance curve (Eurocode concept of Resistance). Any
point inside the interaction diagram is SAFE, and these points can be
considered load Actions (Eurocode concept of Action). Different
methods can be used to draw these interaction diagrams:
Fig. 2. Realistic representation of the main four types of loads (Bhattacharya
et al., 2021).
(1) Simplified method based on Limit Equilibrium which can be
easily set up in a spreadsheet type program. This method is
described as a 10-min method in Bhattacharya (2019). (2) Standard Method based on Beam on Winkler Foundation or p-y
curve method where standard or bespoke p-y curves can also be

Fig. 1. Development of offshore wind turbines (OWTs).

3
M. Aleem et al. Ocean Engineering 248 (2022) 110798

definition of R and A is shown in Fig. 6 and using elementary geometry


one can arrive at the following expression.
√̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
(Hi MR HR )2 +(Mi MR HR )2
R (Mi HR +MR Hi )2
LU(i) = = FOS ​ (i) = √̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅ ̅ (1)
A M2 + H 2
i i

Arany et al. (2017) and Bhattacharya (2019) showed that five load

Fig. 3. Simplified time history of the mudline bending moments (Bhattacharya


et al., 2021).

implemented. This method can be described as 1-hour method in


Bhattacharya (2019).
(3) Advanced Method based on FEA where the curves generated
depends on the chosen failure criterion and the soil material
model. This method is described as a 1-week method as it needs
specialist knowledge of soil models.

Fig. 6 depicts the concept whereby the Load Utilisation ratio for ULS Fig. 5. Typical Moment-Horizontal force interaction curve (Aleem et al., 2020).
loads (i.e., Resistance divided by Actions or R/A, which is also suggested
in Eurocode concept (Eurocode 2, 2011). As mentioned before,
depending on the chosen soil model and the failure criteria, a yield
surface can be mapped, and therefore this is non-unique. It is relatively
straightforward to find the Moment Carrying Capacity (MR) of a foun­
dation system in the absence of any lateral load (i.e., a special case of H
= 0). Similarly, one can find Lateral Load Carrying Capacity (HR) of the
same foundation system in the absence of any moment (i.e., special case
of M = 0). For standard engineering design calculations, a linear line
joining MR and HR can be used as a simplified Yield Surface and is shown
as Line 2 (Idealized) in Fig. 6. Each design load case (DLc’s), i.e., the
combination of Hi and Mi, can be mapped in the H − M space. Fig. 6
shows a particular load case depicted by Hi and Mi, and a line may be
drawn connecting this point and the origin (shown as Line 1 in Fig. 6).
The Load Utilisation (LU) ratio is given by the ratio of the load car­
rying capacity of the foundation (R i.e. Resistance Magnitude) to the Applied
Load combination from wind and wave (A - Action) which can be visualised
in Fig. 6. This is essentially a geometrical representation of the safety
factor against ULS considering both lateral load and moment. Load
Utilisation (LU) ratio for a particular load case is also the notional Factor
Fig. 6. Profile of lateral and moment load resistance capacity.
of safety (FOS) under combined load. It can be defined as (R/A), and the

Fig. 4. Prediction of long-term tilt.

4
M. Aleem et al. Ocean Engineering 248 (2022) 110798

Table 2
Representative design environmental scenarios as load cases chosen for foundation design.
Name & description Wind Wave Alignment
model model

DLc-I Normal operational conditions NTM at 1-year Collinear


Wind and wave act in the same direction (no misalignment). UR (U-1) ESS
(W-1)
DLc-II Cut-out wind speed and extreme operating gust scenario Wind and wave act in the same direction (no misalignment). EOG at 50-year Collinear
Uout (U-4) EWH (W-
4)
DLc- Extreme wave load scenario ETM at 50-year Collinear
III Wind and wave act in the same direction (no misalignment). UR (U-2) EWH (W-
4)
DLc- Extreme wind load scenario EOG at 1-year Collinear
IV Wind and wave act in the same direction (no misalignment). UR (U-3) EWH
(W-2)
DLc-V Wind-wave misalignment scenario ETM at 50-year Misaligned at φ =
Same as E-2, except the wind and wave are misaligned at an angle of φ = 90◦ . The dynamic amplification is higher in the UR (U-2) EWH (W- 90◦
crosswind direction due to low aerodynamic damping. 4)

Fig. 7. Obtaining MR from analysis and schematic representation.

Fig. 8. Obtaining HR from analysis and schematic representation.

cases are necessary for conceptual foundation design, as shown in portions of the curve, and the intersection can be chosen as the MR. It
Table 2. may be noted that it is assumed that the pile will not yield structurally.
Similarly, another set of FEA analyses can be carried out whereby
1.3. Obtaining MR and HR only lateral loading is applied at the top of the foundation. The lateral
load-deflection relationship may be plotted schematically in Fig. 8,
The first step in estimating the utilisation ratio is to obtain the MR which is usually a curve. Tangents may be drawn in the initial and final
and HR, i.e. the capacity of the foundation. One of the three methods portions of the curve, and the intersection can be chosen as the HR. It
(Simplified, Standard or Advanced) can be used. If numerical analysis is may be noted that the intrinsic assumption is that the pile will not yield
used, one can use the definition of MR and HR as provided in Fig. 7 and structurally.
Fig. 8, respectively which is essentially double-tangent method. Brief The simplified method is based on a closed-form solution which can
details of numerical methods are provided in Appendix-A. be easily implemented in a spreadsheet type software for various ground
Beam on Winkler Foundation (also known as p-y analysis) or Finite profiles. These methods take less time to compute, and the judgment is
Element Analysis (FEA) analysis can be carried out whereby only needed for pressure distribution under the action of lateral loads. The
moment loading is applied at the top of the foundation. The moment- readers are referred to Fig. 9 for various types of proposed distribution
rotation relationship can be plotted as shown schematically in Fig. 7, based on the work of Broms B. (1964a, b), Reese et al. (1974), Fleming
which is usually a curve. Tangents may be drawn in the initial and final et al. (1992). These are adopted by API (2007) and DNV, (2016). Further

5
M. Aleem et al. Ocean Engineering 248 (2022) 110798

Fig. 9. Shows the different methods of pu distribution. (a) (Hansen, 1961) (b) (Broms, 1964a) (c) (Fleming et al., 1992) (d) (Reese et al., 1974) (e) (Prasad and Chari,
1999) (f) (Broms, 1964b) and (Jalbi et al., 2019).

Table 3
Geotechnical data for Walney-1 (Sandy Site).
Parameter Symbol Value Unit

Soil’s submerged unit weight γ′


8.5 kN/m3
Soil’s angle of internal friction Φ′ 33 ◦

Soil’s Young’s modulus Es 50 MPa


Length of pile Lp 30 m
Pile diameter Dp 6 m
Pile wall thickness tp 0.040 m

1.3.2. Example application 1: Walney Windfarm (sandy site)


The case study of monopile for Walney-I (Siemens SWT-3.6-107) is
taken to show the application of the methodology for the sandy soil
profile. Walney Offshore Wind Farm is located approximately 15 km
from the coastline of Walney Island, and the fetch is 60 km, as shown in
Fig. 10.
Table 3 provides geotechnical data, and Fig. 12 shows the lateral
earth pressure diagram whereas Table 4 tabulates the turbine data for
use for this example. The site is characterised by a mean wind speed of 9
m/s and a water depth of 21.5m. The peak wave frequency is approxi­
mately 0.197 Hz. Further details of the site can be found in Arany et al.
(2017).
The design load cases were calculated as discussed in Bhattacharya
(2019). Table 5 shows the design load cases for the Walney-I wind farm.
Fig. 10. Location for walney site (Ørsted, 2019).
Finite element software (PLAXIS-3D) was used to evaluate the rota­
tion and deflection of the foundation. In this analysis, Mohr-Coulomb
details of the Simplified Method are provided in Appendix A. and Hardening Soil constitutive models were used. The properties of
these models can be found in Table 3. In terms of the meshing method,
1.3.1. A note on the standard method 10-noded tetrahedral elements were assigned to the soil volume, and 6-
Standard method, also known as the p-y approach, uses non-linear node triangular elements were used for the pile. The interface between
springs and produces reliable results for the cases for which it is devel­ the plate and soil was modelled with double nodded elements. Offshore
oped, i.e., small diameter piles and few loading cycles. However, the wind turbine foundations are often subjected to combinations of verti­
method is not validated for large diameter piles and must be used with cal, horizontal and moment (V, H, and M) loads. Therefore, various
caution. Underprediction of foundation stiffness has been reported by probing tests were performed in finite element analyses to find the
Kallehave et al. (2012) and proposed an updated p-y curve formulation. combined M, and H load scenarios on the failure envelop. In each
Many researchers have recently worked on developing design method­ probing test, a moment smaller than MR was first applied.
ologies for the large diameter stockier monopiles with length to diam­ The lateral force was then gradually increased until failure occurred.
eter ratios typically in the range between 4 and 10, see Zdravković et al. Fig. 11 (a) shows the results obtained from the FE method (Mohr-
(2015) and the PISA method proposed by Byrne et al. (2020). The load Coulomb model and Hardening Soil model) using various combinations
cases used in this study conservatively represent typical foundation of lateral and moment resistance leading to failure of the whole system
loads and may serve as the basis for the conceptual design of founda­ and with polynomial fitting curves also included, which are more ac­
tions. However, detailed analysis for design optimisation and the final curate representations of yield curves. Fig. 11(a) shows that both the API
design may require addressing other load cases. These analyses require (2007) and Kallehave et al. (2012) approaches to determine the soil
detailed data about the site (wind, wave, current, geological, geotech­ stiffness are more reasonable than the simplified method and advanced
nical, bathymetry data, etc.) and the turbine (blade profiles, twist and FE modelling. It is also important to note that the simplified approach
chord distributions, lift and drag coefficient distributions, control pa­ provided reasonable results for the homogeneous sand profile. Fig. 11(a)
rameters and algorithms, drive train characteristics, generator charac­ shows the DLcs and the Yield Surfaces based on Standard and Advanced
teristics, tower geometry, etc.). Methods. The standard method is based on API (2007) and Kallehave
et al. (2012) p-y curves. Advanced Methods are based on Mohr-Coulomb
and Hardening Soil Models.

6
M. Aleem et al. Ocean Engineering 248 (2022) 110798

Fig. 11(a). Interaction curve for Walney-I.

Fig. 11(b). Load utilisation ratio for DLc-I.

Fig. 11(b) shows the Load Utilisation Ratio based on the DLc-I load
FOS ​ (i) = 3.1
case by considering the hardening soil yield surface. It may be noted that
the Yield surface used is based on a straight-line assumption joining MR Similar processes were carried out for other failure curves with
and HR. determined LU values listed in Table 6. It can be observed from Table 6
A typical calculation based on Equation (3) for the determination of that FE analysis with the Mohr-Coulomb model provides the upper
load utilisation is shown below (DLc-1): bound estimate, and API (2007) based calculations shows the lower
Using Line equation-1 (DLc-I) bound estimate. The reason is obvious as FE provides closest to reality.
√̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
2
̅
2
(6x1289x57.1) +(281.24x1289x57.1)
(281.24x57.1+1289x6)2 1.3.3. Example application 2: Gunfleet Sands (clay site)
LU(i) = FOS ​ (i) = √̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
281.242 + 6.02 The case study of Gunfleet sands (Siemens SWT-3.6-107) is used to
show the methodology for the clayey soil profile. The input parameters

7
M. Aleem et al. Ocean Engineering 248 (2022) 110798

Table 6
Moment and lateral resistance for Walney-I.
Computation Method Lateral Moment LU LU
Resistance Resistance for for
(HR) (MR) DLc-I DLc-II
(MN) (MNm) (− ) (− )

Simplified (Closed-form 57.1 1289 3.1 2.3


Solutions)
Standard API 58 1230 3 2.2
Method Kallehave 60.8 1390 3.3 2.5
Advanced Mohr- 64 1410 3.4 2.5
Method Coulomb
(Plaxis 3D) Hardening 60 1392 3.3 2.5
Soil

Table 7
Geotechnical Data for Gunfleet Sand (clayey site).
Parameter Symbol Value Unit

Soil’s submerged unit weight γ′ 8 kN/m3


Undrained shear strength Su 50 kN/m2
Soil’s Young’s modulus Es 2.50 MPa
Fig. 12. Lateral Earth Pressure for Walney-I (Simplified method). Length of pile Lp 38 m
Pile diameter Dp 5 m

Table 4
Turbine properties.
Rated power 3.6 MW

Cut-in wind speed 4 m/s


Rated wind speed 13.5 m/s
Cut-out wind speed 25 m/s
Rotor Diameter 107 m
Swept area 9000 m2
Hub height 80 m
Rotor (mass) 95 t
Nacelle (mass) 125 t
Length of tower 60 m
Diameter of tower (top and bottom) 3 and 5 m
Mass of tower 260 t
Rotational speed 5–13 rpm

Table 5
Design load cases for Walney-I.
Design Load Description Horizontal Overturning
Cases (DLc) Force (MN) Moment (MNm)

DLc-I Normal operational 2.73 99.56


conditions
DLc-II Extreme wave load 4.60 183.41
DLc-III Extreme gust at cut-out and 4.27 148.71
50-year EWH Fig. 13. Lateral Earth Pressure for Gunfleet Sands (Simplified method).
DLc-IV Extreme wind load 4.97 226.53
DLc-V Wind-wave misalignment 3.91 130.77
DLc-VI [ULS Maximum (E− 3) with the 6.70 305.82 method is in good agreement with the method proposed by Jeanjean,
load with load factor of safety of 1.35 (2009). In contrast, PLAXIS-3D showed stiffer behaviour as it considers
Load for ULS loading the boundary conditions, resistance from skin friction, end bearing, and
Factors] other factors and models the behaviour very close to reality. It should be
noted that this case study has only examined a single clay layer with
for the calculation shown in Table 7 were used. The horizontal and constant undrained shear strength.
moment-resisting capacity of the pile in clay can be computed using
Equations provided in the Appendix as a part of simplified method. 1.3.4. Example application 3: Barrow-II (layered soil)
Standard method is p-y and Jeanjean, (2009) is used. Barrow Offshore wind farm is considered to show the application in
Lateral earth pressure diagram is shown in Fig. 13 and the results are layered ground. The wind farm is situated in the East Irish Sea,
presented in Fig. 14 and summarised in Table 8. Fig. 14 indicates that approximately 4.3 miles away from southwest of Walney Island, near
the current API design guidelines significantly under-predicts the ca­ Barrow-in-Furness, covering 10 km2 4.75m diameter 30 m long
pacity when compared to other methods. However, being an empirical monopile supports a 120m long tower carrying a 3 MW vestas wind
method, its weaknesses are widely recognised within the industry and turbine with a rotor diameter of 90m in average water depth of 18m.
also discussed in the literature. It is interesting to note that simplified Lateral earth pressure distribution diagrams as discussed in Appendix A

8
M. Aleem et al. Ocean Engineering 248 (2022) 110798

Fig. 14. Interaction curve for Gunfleet sands.

Table 8
Moment and Lateral resistance for Gunfleet Sands.
Computation Method Lateral Moment LU LU
Resistance Resistance for for
(MN) (MNm) DLc-I DLc-
II

Simplified (Closed-form 34.4 812.2 2.8 2.1


Solution)
Standard API 18 480 1.6 1.2
Method Jeanjean 34 815 2.8 2.1
Advanced Mohr- 43.8 834 3 2.2
Method Coulomb
(PLAXIS- Hardening 41.9 898 3.2 2.3
3D) Soil

for multi-layered soils is plotted in Fig. 15. The readers are referred to
the work of Davisson and Gill (1963) and Gupta and Basu (2016) for
details of analysis.
The monopile is also modelled using standard method and PLAXIS-
3D. The results are shown in Fig. 16 are also summarised in Table 9. It
may be noted that advanced method provided upper bound estimate and
API method gives the lower bound estimate.

1.3.5. Example application 4: London Array-I (layered soil)


The final case study is that of London Array which is a wind farm
situated 12 miles from Essex, covering 100 km2 with 175 wind turbines
Fig. 15. Lateral Earth Pressure for Barrow-II (Simplified method).
with a total capacity of 630 MW. The ground profile consists of dense
sand, firm clay, stiff clay, dense sand, and stiff clay. The input param­
eters, including the monopile diameter and length of the pile, were taken 2. Discussion and conclusions
from Jalbi et al. (2019) and Aleem et al. (2020). Fig. 17 shows the lateral
earth pressure obtained using simplified method. Table 10 shows the Estimation of load carrying capacity (combined effect of moment and
moment and lateral resistance, and Fig. 18 shows the interaction curve. lateral load) is needed for optimisation of monopile and long term tilt
As with other cases, API provides the lower bound. calculation. Long term tilt of monopile foundation is directly correlated
to the well-known concept of Load Utilisation (LU) ratio which is
essentially a proxy for conventional Factor of Safety. LU is the ratio of
load-carrying capacity (defined by Resistance R or Capacity of the

9
M. Aleem et al. Ocean Engineering 248 (2022) 110798

Fig. 16. Interaction curve for Barrow-II.

Table 9
Moment and Lateral resistance for Barrow-II.
Computation Method Lateral Moment LU for LU for
Resistance Resistance DLc-I DLc-II
(HR) (MR)
(MN) (MNm)

Simplified (Closed-form Solution) 69.8 1934 7 5.2


Standard API 56 1550 5.6 4.2
Method Sorensen and 70 1990 7.1 5.3
Jeanjean
Advanced Mohr- 74 2000 7.3 5.4
(Plaxis, Coulomb
3D) Hardening 75 2034 7.4 5.5
Soil

foundation) to the applied load actions (defined by A which is a com­


bination of lateral loads and moments). LU ratio calculation involves
considering the combined action of lateral (H) and moment (M) and
closed-form solution does not exist for such a problem. As a result, a
pragmatic and conservative linear failure envelope is proposed in M-H
space by joining the decoupled MR (Moment Resistance in the absence of
Lateral Load) and HR (Lateral Resistance in the absence of Moment
loading) points. FEM analyses carried out in this study further shows Fig. 17. Lateral earth pressure for london Array-I.
that the failure envelop is a convex curve, and therefore the proposed
linear assumption is indeed conservative.
Table 10
Three methods have been used to carry out the LU ratio calculations
Moment and Lateral resistance for London Array-I.
and they are based on resource requirements (computational time,
Computation Method Lateral Moment LU for LU for
software needed and expertise) and the soil data necessary. Simplified
Resistance Resistance DLc-I DLc-II
method is based on a closed-form solution that can be implemented in (HR) (MR)
any spreadsheet program and is suitable for preliminary design and (MN) (MNm)
feasibility study. Standard method on the other hand is based on Winkler
Simplified (Closed-form Solution) 57 1001 2.2 1.6
Spring (or p-y approach) for which codes and standards are available Standard API 36 690 1.5 1.1
and conventional soil parameters are adequate. Advanced method is Method Sorensen and 55 1050 2.2 1.7
based on Finite Element, and detailed constitutive soil parameters are Jeanjean
Advanced Mohr- 70 1100 2.5 1.8
necessary. This paper provided a comparison of the three methods by
(Plaxis, Coulomb
taking example case studies from European Wind farms. The main 3D) Hardening 63 1204 2.6 1.9
conclusions from the study are that for the sandy site and clayey site, Soil

10
M. Aleem et al. Ocean Engineering 248 (2022) 110798

Fig. 18. Interaction curve for London Array-I.

standard method based on API p-y approach provides a lower bound Sadra Amani: Execution and writing. Abdel Rahman Salem: ORCID.
estimate of load-carrying capacity and that the simplified method is in Saleh Jalbi: Checking, Dr .
good agreement with the Advanced method.
Declaration of competing interest
CRediT authorship contribution statement
The authors declare that they have no known competing financial
Muhammad Aleem: Execution and writing. Subhamoy Bhatta­ interests or personal relationships that could have appeared to influence
charya: Conceptualization, and Resources, Prof . Liang Cui: and Dr. the work reported in this paper.

Appendix-A. Calculation of MR and HR using simplified methods

Lateral Load Transfer Mechanism

Pile foundations transfer the loads in axial and lateral directions from the superstructure to the neighbouring soil without significant deflection. It
is essential to understand the mechanism of load transfer for design and analysis. The transfer of axial load to the ground is done by base resistance and
shaft friction. Under lateral loading, the pile behaves as a transversely loaded beam and may undergo bending, rotation or translation (Fig. A-1).

Fig. A-1. Load transfer Mechanism of Laterally Loaded Piles

Pile can deform in different manners depending upon the relative pile-soil stiffness and the boundary conditions i.e. pile head constraint. Shorter

11
M. Aleem et al. Ocean Engineering 248 (2022) 110798

piles tends to rotate and translate but may not necessarily bend. On the other hand, flexible piles will bend under the applied load. The stiffness of the
pile depends on the soil stiffness, the length of the pile and cross-sectional properties. The ultimate lateral resistance is the combination of failure in
pile material or soil failure. The soil failure in laterally loaded piles is characterised by two failure mechanisms (Fig. A-2).

Fig. A-2. Effect of pile stiffness

Simplified approach

Broms (1964a) estimated the ultimate soil pressure, pu , applied on pile for cohesionless soil expressed as:

Equation A1

pu = 3σv Kp Dp

where
1 + sinφ’
Kp = ​ Equation A2
1 − sinφ’

σ v is the effective vertical stress in soil; Dp is the pile diameter. The formulation assumes that the active pressure behind the pile can be ignored, and

the resistance along the pile front is three times of Rankine passive pressure. The cross-sectional shape has no effect on the resistance, and the dis­
tribution of pressure with depth is linear and is dependent on the effective stress σ v recommended by Elson (1984). Broms (1964b) also suggested the

simplified distribution for cohesive soil that ignored the resistance between the ground surface and 1.5Dp depth. A constant value of 9SuDp is used
below that depth. Matlock et al., (1980) provided formulations for change in ultimate soil resistance as:
For z < Zcr
( )
Su z
Equation A3

pu = Dp 3Su + σ v + J
DP
For z > Zcr
pu = 9Su Dp Equation A4

⎛ ⎞
⎜ 6Dp ⎟
where. Zcr =⎜ ⎟
⎝γ′ Dp +J⎠
Su

Su is the undrained shear strength of soil; γ is the effective unit weight of soil; the recommended J value for stiff clay is 0.25 and 0.50 for soft clay.

The ultimate soil resistance value increases from 3Su to 9Su and remains constant throughout the depth, and this method is recommended by the API
(2007). Reese and Welch (1975) provided different pile strength variations in stiff clay below the water table, where soil resistance increases from 2Su
to 11Su. According to DNV, (2016) and API codes, the stiff clay properties are more brittle than soft clay. Hansen (1961) proposed an equation for both
sand and clay as:
( )
pu = Dp σ v Kq + Su EquationA5

12
M. Aleem et al. Ocean Engineering 248 (2022) 110798

Kq is the factor dependent on the angle of friction ф and depth; for sandy soils, the value of the value of Kq can be taken from Broms (1964a)
equation as 3Kp and can be used for layered soils. Reese et al. (1974) approximated the formulation, which considers the failure mechanisms and takes
the lesser soil wedge and plastic flow failure. Borgard and Matlock (1980) adjusted the empirical method derived from Reese et al. (1974), which
consider the failure mechanism with the increment in-depth as:
z > Zcr pu = C3 Dp γz Equation A6
( )
z < Zcr pu = C1 z + C2 Dp γz Equation A7

Reese et al. (1974) simplified the expression by grouping the terms dependent on ф, whereas C1 C2 and C3 are functions of ф (API, 2007)
reproduced here in Fig. A-3

Fig. A-3. C1 C2 and C3 as per API requirement (API, 2007).

In summary, various methods have been proposed to find out the response of laterally loaded piles making it complex to predict the actual value of
pu . The lateral resistance of the pile in clay is generally expressed in terms of lateral bearing capacity factor, Np , as
Pu
Np = Equation A8
cu Dp

The value of Np increases with the increase in depth, initially from a weaker value at the ground surface to the highest value at a certain depth and
remains constant after that, corresponding to plane strain soil movement around the pile. Various expressions have been proposed based by Matlock
et al. (1980), Reese and Welch (1975).

Laterally resistance of pile in sandy soil

The method is based on Broms (1964a) for cohesionless soil and Fig. A-4 is the problem statement. The pile is be assumed to rotate about a point (at
a distance z = h from top of the pile) and that the pile exhibits a rigid behaviour. Considering horizontal equilibrium, the following expression can be
obtained:
( 2 ) ( )
h2 L h2 L2
Equation A9
′ ′ ′
H = Dp Kγ − − DP Kγ = h2 − DP Kγ
2 2 2 2
Similarly, considering moment equilibrium and by taking moment about the top of the pile, we get Equation A-10.
( 3 ) ( 3 )
h3 L h3 L − 2h3
Equation A10
′ ′ ′
M= − Dp Kγ + − DP Kγ = DP Kγ
2 3 3 3
Taking h as a common variable, an interaction curve can be obtained as follows:
( )2 ( )3
1 3M H 1
− = + Equation A11
L2 Dp Kγ 2

2 2L3 Dp Kγ ′

13
M. Aleem et al. Ocean Engineering 248 (2022) 110798

Fig. A-4. Response of Rigid Pile in Sandy Soil

One can get MR (i.e. Moment Resistance in the absence of lateral load) by setting H as zero.
( √̅̅̅ )
2 − 1 L3 ′ ′
MR = √̅̅̅ DP Kγ = 0.09 Kγ DP L3
3 2
Similarly, HR can be obtained by setting moment (M) to zero.
(√ )
32 − 1 L2 L
EquationA13

H = HR = ​ DP Kγ h = √ ̅̅̅
2 3
2

Example of Walney-I

Geotechnical properties of the Walney-I site are given in Table 3. The diameter and length of the monopile is 30 and 6m, respectively. The
kN
submerged unit weight of the soil is 8.5 m3 and internal friction angle is ∅ = 33 . The value of K is taken as 3Kp .

1 + sin∅ 1 + sin 33
K =3× =3× = 10.2
1 − sin∅ 1 − sin 33
( ) ( 3 )
L3 − 2h3 ′ 30 − 2(21.2)3
MR = DP Kγ = × 6 × 10.2 × 8.50 = 1372 MNm ​
3 3

It may be noted that the pivot point for MR calculation is. h = √L̅̅2 = 21.2
HR can be estimated as follows:
( ) ( )
L2 ′ 302
HR = h2 − DP Kγ = 23.82 − × 6 × 10.20 × 8.50 = 60.6 MN
2 2

The pivot point for HR calculation is. h = √L


3̅̅
2
= 30̅̅
√3
2
= 23.8

Lateral resistance of monopile in clayey soil

Following Broms (1964b), the pressure distribution is shown in Fig. A-5, whereas h is assumed pivot point. It must be mentioned that this model
does not consider the reduced capacity in clays due to vertical loads.

14
M. Aleem et al. Ocean Engineering 248 (2022) 110798

Fig. A-5. Rigid pile in clayey soil

Considering horizontal equilibrium:


H = hDp Nc Su − (L − h)DP Nc Su = (2h − L)DP Nc Su EquationA14

Considering moment equilibrium and taking moment about the pivot point:

h2 (L − h)2
M + Hh = Dp Nc Su + Dp Nc Su EquationA15
2 2
Combining the above two equations, the following is obtained,
( )
Dp Nc Su L2 − 2h2
M= EquationA16
2
One can estimate MR when H = 0, which provides an
L2 L
MR = Dp Nc Su when ​ the ​ pivot ​ point ​ is h = EquationA17
4 2
One can estimate HR when M is zero (in Equation A-16) which gives the following:
(√̅̅̅ ) L
HR = 2 − 1 LDp Nc Su and ​ when ​ the ​ pivot ​ point ​ h = ​ √̅̅̅ EquationA18
2

Solved example of Gunfleet Sands

The geotechnical properties of Gunfleet Sands are given in Table 7, where the length and diameter of the pile is 38 and 5m, respectively, having the
undrained shear strength Su = 50 kN/m2 .
HR = (2(26.9) − 38) × 5 × 9 × 50 = 35.5 MN

L
h = √̅̅̅ = 26.9
2

382 − 2(19)2
MR = × 5 × 9 × 50 = 812.250 MNm
2
L
h= = 19
2
It may be noted that the above formulation is valid for uniform clay.

Example of a Layered Soil Profile

Fig. A-6 shows a sample pressure distribution for layered soils.

15
M. Aleem et al. Ocean Engineering 248 (2022) 110798

Fig. A-6. Example of the laterally loaded pile in a five-layers

APPENDIX-B. Standard and Advanced Methods of Analysis

Standard method (Winkler Approach)


This standard method, commonly known as Discrete Winkler Spring Approach (or p-y approach), has been widely adopted in the offshore industry
to estimate deflection/deformation/rotation of laterally loaded piles. Due to its simplicity to solve the complicated practical problems and ability to
analyse the behaviour of layered soils. This type of software is readily available at a reasonable price. In this method, foundation-soil interaction is
represented by a set of independent springs along the pile length. The stiffness of the springs is calculated using the p-y curves, which defines the
relationship between the p (load per unit length of the pile) and y (lateral displacement). The continuity of soil is ignored, see Fig. B-1).

Fig. B-1. Winkler Model for pile design [Adopted from PISA model]

The limitation of the Winkler Approach is that every soil spring responds independently and ignores the transfer of shear between the layers of soil.
As a result, there will be a significant contrast of stiffness.
Numerous methods have been published to predict the response of a single pile under lateral loading (Bhattacharya et al., 2021; Desia, 1975;
Hetényi, 1950; Krishnaveni et al, 2016; Li and Gong, 2008; Sun, 1994). Li and Gong (2008); Anders Hust Augustesen et al. (2009) developed some
equations using the fundamental basis of mechanics for the pile-soil system. The method considered both the fixed and free head piles in four-layered

16
M. Aleem et al. Ocean Engineering 248 (2022) 110798

soil with the advantage of predicting the preliminary response of laterally loaded piles.
Advanced method (FE Approach)
These are continuum-based, advanced 3D finite element packages, highly computationally demanding and can model any type of complex ge­
ometry. These packages are highly expensive and require a well-trained engineer to run the analysis. Finite element analysis PLAXIS 3D is used to carry
out the analysis because of several built-in soil models.

References Fleming, W.G.K., Weltman, A.J., Randolph, M.F., Elson, W.K., 1992. Piling Engineering,
Dairy Science & Technology. CRC Taylor & Francis Group.
Gupta, B.K., Basu, D., 2016. Response of laterally loaded rigid monopiles and poles in
Aleem, M., Demirci, H.E., Bhattacharya, S., 2020. Lateral and moment resisting capacity
multi-layered elastic soil. Can. Geotech. J. 53, 1281–1292. https://doi.org/10.1139/
of monopiles in layered soils. In: ICEESEN 2020. Kayseri, Turkey, pp. 19–21.
cgj-2015-0520.
Anders Hust Augustesen, K.T., Brødbæk, M., Møller, , Søren Peder Hyldal Sørensen, Bo
Hansen, J.B., 1961. The Ultimate Design of Piles against Transversal Loads 16.
Ibsen, Lars, Thomas Schmidt Pedersen, L.A., 2009. Numerical modelling of large-
Hetényi, M., 1950. A general solution for the bending of beams on an elastic foundation
diameter steel piles at horns rev. In: The International Conference on Civil,
of arbitrary continuity. J. Appl. Phys. https://doi.org/10.1063/1.1699420.
Structural and Environmental Engineering Computing - Funchal, Madeira, Portugal.
IEC, 2009. Wind Turbines - Part 3: design requirements for offshore wind turbines. Eur.
Civil-Comp Press, Funchal, Madeira, p. 14.
Comm. Electrotech. Stand. 2009, 1–132.
API, 2007. Recommended practice for planning, designing and constructing fixed
IEC, 2005. Iec 61400-12-1. Int. Electrotech. Comm. 2005, 179.
offshore Platforms- working stress design. API Recomm. Pract. 24, 242. WSD.
Jalbi, S., Arany, L., Salem, A., Cui, L., Bhattacharya, S., 2019. A method to predict the
Arany, L., Bhattacharya, S., Macdonald, J., Hogan, S.J., 2015. Simplified critical mudline
cyclic loading profiles (one-way or two-way) for monopile supported offshore wind
bending moment spectra of offshore wind turbine support structures. Wind Energy
turbines. Mar. Struct. 63, 65–83. https://doi.org/10.1016/j.marstruc.2018.09.002.
18, 2171–2197. https://doi.org/10.1002/we.1812.
Jeanjean, P., 2009. Re-assessment of P-Y curves for soft clays from centrifuge testing and
Arany, L., Bhattacharya, S., Macdonald, J., Hogan, S., 2017. Design of monopiles for
finite element modelling. In: All Days. OTC, Houston. https://doi.org/10.4043/
offshore wind turbines in 10 steps. Soil Dynam. Earthq. Eng. 92, 126–152. https://
20158-MS.
doi.org/10.1016/j.soildyn.2016.09.024.
Kallehave, K., leBlanc, C., Lingaard, M., 2012. Modification of the API P-Y formulation of
Bhattacharya, S., 2019. Design of Foundations for Offshore Wind Turbines, first ed. ed.
initial stiffness of sand. Conf. Pap. Sut. Osig. 465–472.
Wiley, Chichester. https://doi.org/10.1002/9781119128137.
Krishnaveni, B., Alluri, S.K.R., Murthy, M.V., 2016. Generation of p-y curves for large
Bhattacharya, S., 2014. Challenges in design of foundations for offshore wind turbines.
diameter monopiles through numerical modelling. Int. J. Res. Eng. Technol. 5,
Eng. Technol. Ref. https://doi.org/10.1049/etr.2014.0041.
379–388. https://doi.org/10.15623/ijret.2016.0507060.
Bhattacharya, S., Lombardi, D., Amani, S., Aleem, M., Prakhya, G., Adhikari, S.,
Li, R., Gong, J., 2008. Analysis of laterally loaded pile in layered soils. Electron. J.
Aliyu, A., Alexander, N., Wang, Y., Cui, L., Jalbi, S., Pakrashi, V., Li, W., Mendoza, J.,
Geotech. Eng. 13 (J.).
Vimalan, N., 2021. Physical modelling of offshore wind turbine foundations for TRL
Matlock, H., Ingram, W.B., Kelley, A.E., Bogard, D., 1980. Field tests of the lateral-load
(technology readiness level) studies. J. Mar. Sci. Eng. 9, 589. https://doi.org/
behavior of pile groups in soft clay. Offshore Technol. Conf. https://doi.org/
10.3390/jmse9060589.
10.4043/3871-MS.
Borgard, D., Matlock, H., 1980. Simplified Calculation of P-Y Curvesfor Laterally Loaded
Ørsted, 2019. Walney Offshore Wind Farm [WWW Document]. URL, 6.10.21. https://ors
Piles in Sand. Earth Technol. Corp. Inc.
tedcdn.azureedge.net/-/media/www/docs/corp/uk/updated-project-summaries
Broms, B.B., 1964a. Lateral resistance of piles in cohesive soils. J. Soil Mech. Found Div.
-06-19/190514walney-1–2-webaw.ashx?la=en&rev=42cdfa99bee6471eafce19a9
90, 27–63. https://doi.org/10.1061/jsfeaq.0000611.
7b07344a&hash=D461C0E55B40CF56592C440527B3E3A6.
Broms, B.B., 1964b. Lateral resistance of piles in cohesionless soils. J. Soil Mech. Found
Prasad, Y.V.S.N., Chari, T.R., 1999. Lateral capacity of model rigid piles in cohesionless
Div. 90, 123–156. https://doi.org/10.1061/JSFEAQ.0000614.
soils. Soils Found. 39, 21–29. https://doi.org/10.3208/sandf.39.2_21.
Byrne, B.W., Houlsby, G.T., Burd, H.J., Gavin, K.G., Igoe, D.J.P., Jardine, R.J., Martin, C.
Reese, L.C., Cox, W.R., Koop, F.D., 1974. Analysis of laterally loaded piles in sand.
M., McAdam, R.A., Potts, D.M., Taborda, D.M.G., Zdravković, L., 2020. PISA design
Offshore Technol. Conf. https://doi.org/10.4043/2080-MS.
model for monopiles for offshore wind turbines: application to a stiff glacial clay till.
Reese, L.C., Welch, R.C., 1975. Lateral loading of deep foundations in stiff clay. ASCE J.
Geotechnique 70, 1030–1047. https://doi.org/10.1680/jgeot.18.P.255.
Geotech. Eng. Div. 101, 633–649. https://doi.org/10.1016/0148-9062(75)90294-6.
Davisson, M.T., Gill, H.L., 1963. Laterally loaded piles in a layered soil system. J. Soil
Sun, K., 1994. A numerical method for laterally loaded piles. Comput. Geotech. https://
Mech. Found Div. 89, 63–94.
doi.org/10.1016/0266-352X(94)90011-6.
Desia, 1975. Numerical design-analysis for piles in sand. Int. J. Rock Mech. Min. Sci.
Zdravković, L., Taborda, D.M.G., Potts, D.M., Jardine, R.J., Sideri, M., Schroeder, F.C.,
Geomech. Abstr. https://doi.org/10.1016/0148-9062(75)92460-2.
Byrne, B.W., McAdam, R., Burd, H.J., Houlsby, G.T., Martin, C.M., Gavin, K.,
DNV, 2011. Design of Offshore Wind Turbine Structures DNV-OS-J101.
Doherty, P., Igoe, D., Muirwood, A., Kallehave, D., Skov Gretlund, J., 2015.
DNV, G.L., 2016. DNVGL-ST-0437: Loads and Site Conditions for Wind Turbines, vol.
Numerical modelling of large diameter piles under lateral loading for offshore wind
108. DNV GL - Stand.
applications. In: Frontiers in Offshore Geotechnics III - Proceedings of the 3rd
Elson, W.K., 1984. Design of Laterally-Loaded Piles.
International Symposium on Frontiers in Offshore Geotechnics. ISFOG. https://doi.
Eurocode 2, 2011. Actions on Structures.
org/10.1201/b18442-105, 2015.

17

You might also like