Content uploaded by Lieven Vandevelde
Author content
All content in this area was uploaded by Lieven Vandevelde on Jan 12, 2024
Content may be subject to copyright.
Available via license: CC BY-NC-ND 4.0
Content may be subject to copyright.
Received 18 December 2023, accepted 25 December 2023, date of publication 5 January 2024,
date of current version 11 January 2024.
Digital Object Identifier 10.1109/ACCESS.2024.3350439
Load-Aware Operation Strategy for Wind
Turbines Participating in the Joint Day-Ahead
Energy and Reserve Market
NARENDER SINGH 1,2, SEYYED AHMAD HOSSEINI 3, (Member, IEEE),
JEROEN D. M. DE KOONING 1,2, (Senior Member, IEEE), FRANÇOIS VALLÉE 3, (Member, IEEE),
AND LIEVEN VANDEVELDE 1,2, (Senior Member, IEEE)
1Department of Electromechanical, Systems and Metal Engineering, Faculty of Engineering and Architecture, Ghent University, 9052 Ghent, Belgium
2FlandersMake@UGent—MIRO Core Laboratory, 3001 Flanders Make, Belgium
3Power Systems and Market Research Group, Electrical Power Engineering Unit (EPEU), University of Mons, 7000 Mons, Belgium
Corresponding author: Narender Singh (narender.singh@ugent.be)
This work was supported by the Frame of the BEOWIND Project funded by the Energy Transition Fund of the Belgian Federal
Government.
ABSTRACT Wind power has emerged as a clean alternative to traditional power production, with a
significant increase in its installed capacity observed over the past decade. In numerous regions, wind power
producers are now afforded the opportunity to participate in day-ahead energy and reserve markets. In this
context, the wind turbines can provide ancillary services such as frequency containment reserve (FCR).
However, the provision of ancillary services is known to affect the physical loading on a wind turbine. This
loading on different parts of the wind turbine can possibly result in sub-optimal performance leading to a
reduced net power output or even a faster degradation of the wind turbine. On the other hand, no or low
participation of a wind turbine in ancillary services market will lead to a lesser revenue on a long term.
Moreover, low participation of wind turbines in the ancillary market will eventually limit the amount of
wind power, since the ancillary services are needed to stabilise the grid and must then be provided by other
energy sources. Addressing this challenge requires a holistic method to gauge both load and revenue for
wind power producers (WPP), thus enabling them to make informed decisions. This study firstly presents
a method of calculating major loads on the wind turbine. Then, a load-aware optimisation method of wind
power scheduling in the joint day-ahead energy and reserve market (JERM) is proposed that provides WPPs,
an ability to strike a balance between revenue and the physical loading of wind turbine.
INDEX TERMS Ancillary services, frequency containment reserve, wind turbines, wind turbine loading,
wind turbine operation strategy.
I. INTRODUCTION
In 2022, 77.6 GW of new wind power installations were
added globally. This sizeable increment brought the global
installed wind energy capacity to 906.2 GW. The record
installations from European nations, Sweden, Finland and
Poland brought the total installed wind capacity in Europe in
2022 to 255.5 GW [1]. According to the International Energy
The associate editor coordinating the review of this manuscript and
approving it for publication was Ton Duc Do .
Agency, the aim to achieve 61 % of total electricity generation
from renewable energy by 2030 will only be possible with
a threefold increase in renewable energy installations. The
biggest contribution will be from wind and solar energy [2].
The net zero emissions by 2050 scenario requires a wind
power generation of 7400 TWh per year by 2030 [3]. Only the
future holds the answer to whether such monumental increase
in renewable, and especially in wind power will be achieved.
Nevertheless, a significant growth is inevitable and necessary.
The energy markets around the world are being reshaped
VOLUME 12, 2024
2024 The Authors. This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 License.
For more information, see https://creativecommons.org/licenses/by-nc-nd/4.0/ 5309
N. Singh et al.: Load-Aware Operation Strategy for Wind Turbines
due to the escalating share of wind energy in the global
power mix. The reserve market landscape is also evolving
due to these changes, as grid codes now allow wind power
producer’s (WPP) to participate in the provision of ancillary
services [4].
As a result, extensive studies have been conducted within
this domain. The possibility of temporary power over-
production in variable speed wind turbines is explored in [5].
In [6], an active control of wind turbine for ancillary services
using pitch and torque control methods is explored. The
study presented in [7], presents a wind turbine participation
in ancillary services market by the means of rotational
kinetic energy. A grid frequency stabilisation strategy using
inertial control of the wind turbine incorporating energy
storage systems is presented in [8]. The present studies
are not only confined to frequency support service. In [9],
a method of voltage control from a power plant is presented.
The capability of wind farms to provide reactive power
ancillary services is explained in [10] and [11]. Another
study presented in [12] explores the coordination strategy of
large wind farms for voltage support by the means of high
converter capacity. An ancillary services provision method
by the means of coordination between wind turbine and
electrolysis systems is presented in [13]. The evolution of
fast acting control systems for wind turbines has enabled
the integration of wind farms into the ancillary services
market [14]. In [15], a control system is presented that
is capable of following the grid frequency with minimal
error. An advanced converter control maintaining a transient
frequency stability is presented in [16]. The study presented
in [17], proposes a coordinated control strategy to efficiently
utilise energy in permanent magnet synchronous generators
(PMSG). A hybrid control strategy of frequency based power
point tracking method is presented in [18]. In [19] control
system capability for frequency containment reserve (FCR)
and fast frequency response (FFR) is shown. A strategy
incorporating wake control for optimised operation of wind
farms providing FCR is proposed in [20]. These control
techniques have made it possible for the wind turbines to
have a deeper integration in ancillary services market. Crucial
grid frequency services such as FCR and FFR can greatly
benefit from these enhanced control capabilities of the wind
turbines to swiftly respond to the grid frequency fluctuations.
Grid services essential to maintain the grid stability such as
synchronous inertial response, enhanced frequency response
and fast post-fault active power recovery can also benefit from
increased wind energy participation in the ancillary services
market [21].
Studies point to an increase in the revenue of WPPs as a
result of participating in the ancillary service market. A study
exploring the advanced bidding strategy dedicated to optimal
dispatch of WPP in the JERM is presented in [22]. The
revenue generated using this technique has shown an increase
in the revenue of WPP. A data-driven probabilistic energy and
reserve bidding approach for wind turbines participating in
reserve market has also concluded similar results [23].
Although the ancillary services market is a lucrative avenue
for the WPPs, the physical loading of the wind turbine may
be affected by these control techniques. There has been
some research regarding the physical loading of the wind
turbine. The impact of loads on composite wind turbine
blades is studied in [24]. Loads on wind turbine blades
are studied by using finite element analysis in [25]. Load
identification of a wind turbine tower using Kalman filtering
techniques is presented in [26]. Dynamic analysis of offshore
wind turbine tower subjected to wind and wave loading
is shown in [27]. There are also studies that explore the
global physical loading of the wind turbine [28],[29],[30].
Amidst all these studies, there exists limited literature on
the effect of ancillary services provision on the loading
of the wind turbine components. In [31], a study about
the dynamic frequency control considering wind turbine
fatigue is presented. The effects of power reserve control
on the structural loading of wind turbines are discussed
in [32]. A control design to reduce the drive-train, blades and
tower mechanical stresses of a wind turbine participating in
the grid primary frequency regulation is presented in [33].
A non-linear virtual inertia control of wind turbines in order
to enhance primary frequency response and suppressing
drivetrain torsional oscillations is presented in [34]. A study
presented in [35] studies the effect of primary reserve
provision on the main bearing loads. These studies point to
the different impacts of ancillary services provision on the
wind turbine. However, none of the existing studies aims to
combine these qualitatively to provide a metric for the overall
loading of the wind turbine. Moreover, the economical aspect
of the loading is not taken into account when considering the
wind turbine participation in JERM. Hence, there is a need
to quantify the overall major loads on a wind turbine, and
moreover, to study the impact of ancillary service provision
on these loads while integrating the techno-economic impact
on the JERM participation.
To this end, this study proposes a load-aware optimisation
method of wind power bidding in the JERM. The purpose
of the developed method is not only to maximise the profit
of the WPP but also to create an optimal balance between
the net market revenue and the physical loads on the wind
turbine. Firstly, a methodology is developed to calculate
the physical loading on different components of the wind
turbine, namely, the main bearing, the blades, the shaft and
the tower. These loads are calculated for different reserve
market bids. These loads are then used as an input into
an optimisation problem that generates energy and reserve
market bids for profit maximisation of the WPP while taking
into account the wind turbine loading. This results in a load-
aware operation strategy compromising economic profit with
structural loading, which is the main novel contribution of this
work.
This article is organised as follows: Section II describes
the model and data used in this study. Section III presents the
methodology used to calculate the loads on the wind turbine
as well as the optimisation strategy. In Section IV, results
5310 VOLUME 12, 2024
N. Singh et al.: Load-Aware Operation Strategy for Wind Turbines
from this study are discussed. The conclusion of this study
is derived in Section V.
II. MODELS AND DATA
The models used in the study include the wind turbine, the
PMSG and the control systems used for torque and pitch
control of the wind turbine. Wind data is used to simulate
these models for different cases. These models and data are
detailed in this section.
A. WIND TURBINE AND WIND FIELD DESIGN
This study employs the NREL 5 MW offshore wind turbine
model [36]. The model has been chosen for its high
correspondence on the REpower 5M machine, which is
a widely used wind turbine. The NREL 5 MW model
is generally recognised and used as a benchmark in the
field of wind energy. Its design and characteristics closely
align with the real-world offshore wind turbines, offering
a realistic representation for simulations and studies. The
model incorporates design information gathered from various
sources, including published documents from reputable
turbine manufacturers. To develop and simulate the model,
a wind turbine simulator called FAST v8 is used, which
was developed by the National Renewable Energy Lab-
oratory [37]. FAST v8 features several different modules
including aerodynamics, structural loads, electrical system
and hydrodynamic loads. To have a greater control on
the control mechanism for this study, the electrical and
pitch systems have been separately developed in Simulink.
Moreover, the hydrodynamic loads module has not been
included in this study as the focus of this study is on analysing
the impact of dynamic control of the electric power on the
turbine’s structural loading. The primary characteristics of the
wind turbine are listed in the Table 1.
TABLE 1. Wind turbine properties.
The specified wind turbine operates in distinct regions,
each of which is determined by the free-flow wind speed
measured on the rotor of the turbine. These regions, depicted
in Figure 1, are based on the wind speed at the turbine
rotor. In Figure 1, Region 1 corresponds to wind speeds
below 3 m/s, which is below the cut-in speed of the wind
turbine. In this region, no torque is generated, and no power
is extracted. In Region 2, the wind turbine functions in
maximum power point tracking mode. Region 2.5 is the
transition zone between Regions 2 and 3. Finally, in Region
3, the wind speeds exceed the nominal wind speed, and
control methods limit the torque, speed, and power to prevent
overloading of the drivetrain components. If the wind speed
exceeds 25 m/s, the wind turbine enters the cut-out region,
and no power is generated.
The wind data used for the simulations in this research
work are generated by using TurbSim [38]. Two different
cases have been simulated with the mean wind speed of 6 m/s
and turbulence intensities of 5 % and 20 %, as presented in
Figure 2and Figure 3.
FIGURE 1. Power curve of a 5 MW wind turbine.
FIGURE 2. Wind profile with 5% turbulence intensity.
FIGURE 3. Wind profile with 20% turbulence intensity.
B. PERMANENT MAGNET SYNCHRONOUS GENERATOR
The study employs a PMSG. To accurately control the
generator’s performance, a rotating direct, quadrature (d,q)
reference frame representation of the generator is used.
Table 2provides a comprehensive list of the generator
parameters used in developing the model [39],[40]. It is
worth noting that PMSGs have become increasingly popular
in recent years due to their superior efficiency, high power
density, and low maintenance requirements. Furthermore,
modern wind turbines are increasingly using PMSGs, as they
can operate efficiently in variable speed conditions and allow
direct-drive operation without gearbox.
VOLUME 12, 2024 5311
N. Singh et al.: Load-Aware Operation Strategy for Wind Turbines
TABLE 2. Generator properties.
C. CONTROL
The different control schemes used in the study are detailed
in this section.
1) TORQUE CONTROL
The effectiveness of this study relies heavily on the accurate
control of torque in the PMSG. To achieve this, Field Oriented
Control (FOC) is employed in Simulink. In addition to FOC,
other control strategies such as model predictive control and
sliding mode control have been proposed for torque control
in PMSG-based wind turbines [41],[42],[43]. However,
FOC remains one of the most popular control strategies due
to its simplicity and robustness [44],[45]. FOC regulates
the quadrature current component ˆ
iqproportionally to the
torque setpoint while keeping the direct current component
ˆ
idat zero for field orientation. The torque control scheme
is depicted in Figure 4, with the Proportional Integral (PI)
controller’s proportional and integral gains set at 0.0001. All
symbols represent the numerical values of the quantities in
SI units. At each time step, the PI controller generates a
current signal based on a comparison between the reference
power ˆ
Pand actual power P. The current controller ensures
that the desired current is generated to control the torque
effectively. The reference power is determined by the grid
frequency and the amount of contracted ancillary reserve.
The Clarke-Park transformation is used to transform the
three-phase abc current signals into the two-phase dq current
components [46]. The resulting dq components are then used
for torque control. The inverse Clarke-Park Transformation
is subsequently applied to convert the dq components back
to three-phase abc signals for modulation [47]. The PI
controller is configured to minimise the tracking error
by minimising overshoot and settling time. Although the
controller’s performance varies with changing wind speeds
and the rate of change of the reference power, it is capable of
tracking the reference power with minimal error. It is worth
noting that accurate torque control is crucial for PMSGs to
operate efficiently and deliver the desired power output.
2) PITCH CONTROL
The Simulink-developed pitch control system is integrated
with the wind turbine model in FAST to ensure smooth
operation. During each iteration of the simulation, the pitch
control system transmits a pitch command to each of the
three blades of the wind turbine. Figure 5illustrates the block
diagram of the pitch control system, which employs a PI
controller with proportional and integral gains of 206.3 and
FIGURE 4. Field oriented control.
25, respectively. The PI controller continuously compares
the reference rotor speed, ref , a preset parameter, with the
actual rotor speed, , at each time step. The gain scheduling
technique utilised a dimensionless gain correction factor, G,
defined as in (1). The pitch angle, θ, plays a critical role in the
gain scheduling process, and a tuning parameter, θd, is set to
0.055 radians to optimise its performance [48].
G=1.6
1+θ
2θd(1)
FIGURE 5. Pitch control of the wind turbine blades.
III. METHODOLOGY
The methodology used in this study is developed for a wind
turbine participating in JERM, where bids are submitted
day-ahead for energy and reserve markets by the WPP. The
methodology is outlined in Figure 6. With the inputs of wind
profile, grid frequency, model parameters and the reserve
market bid, software-in-the-loop simulations are performed
to calculate the loads on different sections of the wind
turbine. As an output from these simulations a mapping
between the reserve bid and total loading is achieved.
Incorporating the derived loading map as a penalty term into
the WPP’s scheduling objective function, our proposed model
strategically optimizes the WPP profit while mitigating wind
turbine loads. The optimal decision variables of the model
yield reserve and energy market bids. Additionally, the
revenues and real-time compensations are also calculated.
The following subsections present the calculation methods in
detail.
A. LOAD CALCUL ATION
The total load on the wind turbine, referred to as Lconsists
of 4 different loads. The RMS values of these loads from
simulations for a time period of 300 s for each case are used.
5312 VOLUME 12, 2024
N. Singh et al.: Load-Aware Operation Strategy for Wind Turbines
FIGURE 6. Methodology.
These loads associated with the main bearing, blades, shaft
and the tower are presented in the following subsections.
1) BEARING LOAD
The bearing load calculations are performed based on the
method presented in [35]. The forces acting on the main
bearing of the wind turbine are, axial force Fa(in N), lateral
force Fvand vertical force Fl. Figure 7shows these three
forces acting on the wind turbine. Fais calculated as the
average of the 3 blade root forces Fb1,Fb2and Fb3as in (2).
The radial force Fr(in N) is calculated using Fland Fv
as in (3). The root mean square (RMS) of the two forces
are then used to calculate the dynamic equivalent force Lbr
acting on the main bearing as in (4). Here, tand Trepresent
the time variable and the total duration of the simulation,
respectively. bxand byare dimensionless empirical factors
for load calculations in a spherical roller bearing.
Fa=Fb1+Fb2+Fb3
3(2)
Fr=qF2
l+F2
v(3)
Lbr =s1
TZT
0hbxFr(t)+byFa(t)i2dt (4)
2) BLADE LOAD
In the wind turbine model used for the simulations, each of
the three blades are divided into 9 spans along the length of
the blade. Each of these spans are subjected to forces directed
along x, y and z axis. These forces and the division of the
wind turbine blade in 9 spans is represented in Figure 7. The
equivalent of the 3 forces on each span is calculated as per (5),
where nis the number of blade (from 1 to 3) and mis the
number of the span (from 1 to 9). The total blade load Lbl is
then calculated as the RMS of the individual forces on each
FIGURE 7. Forces acting on the wind turbine.
span of each blade as in (6). Note that the time dependence
(t) is omitted to not overload the equations.
Fbln,spm=qFx2
bln,spm+Fy2
bln,spm+Fz2
bln,spm(5)
Lbl =v
u
u
u
t
1
TZT
0 9
X
m=1
Fbl1,spm+Fbl2,spm+Fbl3,spm!2
dt
(6)
VOLUME 12, 2024 5313
N. Singh et al.: Load-Aware Operation Strategy for Wind Turbines
3) SHAFT LOAD
The shaft load comprises of rotating shaft bending moment
at the shaft’s strain gauge along y and z axis (in Nm). The
bending moments are then divided by the distance from rotor
apex to calculate the two equivalent forces Fysh and Fzsh (in
N) along y and z axis, respectively. These coordinates follow
the similar plane as Fland Fv, as shown in Figure 7. Finally,
the total shaft load Lsh is calculated using the RMS of the
equivalent forces as shown in (7).
Lsh =s1
TZT
0hFy2
sh(t)+Fz2
sh(t)idt (7)
4) TOWER LOAD
The tower loads are divided into two categories, the loads
associated with the main structure of the tower LTM and the
loads at the base of the tower LTB. For the calculation of LTM ,
the tower is divided into 9 sections. Each section of the tower
is subjected to forces FxTM ,FyTM and FzTM along x, y and z
axis, respectively. The section of the wind turbine tower and
associated forces are presented in Figure 7. The equivalent
force at each of the sections is calculated as in (8), where k
represents the section of the tower from 1 to 9. LTM is then
calculated as the sum of loads on each of the section as in (9).
The tower base loads FxTB,FxTB and FxTB (in N) are directed
along x, y and z axis respectively. The equivalent load on the
tower base, LTB is calculated as in (10). Finally, the total tower
load Ltwr is calculated as the sum of tower main and base
loads as in (11).
LTM (k)=qFx2
TM (k)+Fy2
TM (k)+Fz2
TM (k) (8)
LTM =v
u
u
u
t
1
TZT
0 9
X
k=1
LTM (k,t)!2
dt (9)
LTB =s1
TZT
0hFx2
TB(t)+Fy2
TB(t)+Fz2
TB(t)idt (10)
Ltwr =LTM +LTB (11)
5) EQUIVALENT LOAD
The total load on the wind turbine Lis calculated as weighted
sum the four individually calculated loads as in (12). Here, a,
b,cand dare the weight factors associated with the loads.
To determine the values of these weight factors, informed
by their economic implications, a base case simulation is
conducted over a duration of 3000 s. The weight factors based
on the base case are setup such that each individual term
on the right hand side of (12) is equivalent to 0.25. Hence,
making the base value of Lequal to 1. This arbitrary choice
is made to demonstrate the functioning of the methodology.
However, these are adjustable factors and can be tuned by the
wind farm operator based on the economic costs associated
with each of these components.
L=aLbr +bLbl +cLsh +dLtwr (12)
B. OPTIMISATION
The optimisation strategy presented in [22] proposed a wind
power scheduling framework that accounts for the revenue
stream from both day ahead and real-time stages of the
energy and reserve markets. However, in this strategy the
loads acting on the wind turbine due to the optimal bids in
JERM are not accounted for. For this study, the optimisation
strategy has been improved to include the effect of wind
turbine loading. The optimisation is performed in Matlab,
using the Gurobi optimiser [49]. Based on the optimisation,
optimal decisions are made to maximise the revenue of
the WPP while taking the wind turbine physical loading
into account. The objective function is stated in (13). Here,
Las defined in (12) is the factor based on the mapping
generated from the load calculations. The day-ahead bids
related to the energy and reserve markets are, respectively,
shown by Embid and Rmbid.αis a dimensionless variable
weight factor associated with the load L.λsp and λcap are
the spot market and the reserve capacity prices, respectively.
λBU and λBD are imbalance prices for surplus and deficit,
respectively. λcp is the unavailability penalty price for the
reserve. πωis the probability of occurrence of the scenario.
δtis the market time unit equal to 1 hour. /ωrepresents
the scenario index/set. 1Puω,1Pdωare the positive and
negative deviation of injected power at scenario ω.1Rdωis
the deviation of available capacity margin from the offered
bid at scenario ω.
J= −αL+λspEmbid1t+λcapRmbid
+X
ω∈
πω1tλBU1Puω−λBD1Pdω−λcp1Rdω
(13)
The loading constraint used for the optimisation is shown
in (14). This constraint takes into account the impact of
increasing Rmbid on the overall loading of the wind turbine.
The choice of using only reserve market bid is based on the
fact that the most influential factor of the wind turbine loading
is the amount of reserve power provided since it has to be
injected to the network dynamically thus affecting the load
on the wind turbine. The values of p1and p2are calculated as
shown in (15) and (16), where nrefers to the number of data
points simulated related to α.
L=p1Rmbid +p2(14)
p1=nPn
i=1(RmbidiLi)−Pn
i=1RmbidiPn
i=1Li
nPn
i=1Rm2
bidi−Pn
i=1Rmbidi2(15)
p2=Pn
i=1Li−p1Pn
i=1Rmbidi
n(16)
The other constraints associated with the optimisation are
listed in (17)-(26). Here, Qcis the amount of the offered bids
for the day-ahead market. Pω,Rωand Qωare the delivered
power to the energy market, available reserve power and
the total available wind power at scenario ω.Mis a large
positive constant for mixed integer programming. δis the
5314 VOLUME 12, 2024
N. Singh et al.: Load-Aware Operation Strategy for Wind Turbines
binary variable associated with the sufficient power capacity.
Embid +Rmbid =Qc(17)
Pω+Rω=Qω(18)
Embid −Pω=1Pdω−1Puω(19)
Rmbid −Rω≤1Rdω(20)
Qω−Rmbid−Mδ≤0 (21)
Qω−Rmbid +M(1 −δ)≥0 (22)
Rω≤Rmbid (23)
Rω≤Qω(24)
Rω≥Rmbid−M(1 −δ) (25)
Rω≥Qω−Mδ(26)
C. REVENUE CALCULATION
The revenue calculations for a wind turbine participating
in JERM are presented in this section. The total rev-
enue consists of energy and reserve market revenues. All
the revenue calculations are performed for a duration of
1 hour.
The combined revenue from energy market and imbal-
ance settlement, REI is calculated as shown in the (27).
It accounts for the spot market price, the optimal reserve
power bid, the positive and negative deviations of injected
power and the respective imbalance prices for surplus and
deficit.
REI =λspEmbid1t+X
ω∈
πω1t(λBU 1Puω−λBD1Pdω)
(27)
The revenue generated by WPP in the day-ahead reserve
market, RDR is determined by multiplying the reserve
capacity price by the optimal reserve power bid, as shown
in (28).
RDR =λcapRmbid (28)
In (29), unavailability penalty price for the reserve and
deviation of available capacity margin from the offered bid
are used to calculate RPB, the penalty incurred by WPP during
the balancing stage.
RPB =X
ω∈
πω1t(−λcp1Rdω) (29)
The total revenue from participating in the reserve market
and balancing stage, RRB, as shown in (30), encompasses
both the reserve capacity price and penalties. It takes into
account the optimal reserve power bid, the unavailability
penalty price, and the deviation of available capacity margin
from the offered bid.
RRB =λcapRmbid +X
ω∈
πω1t(−λcp1Rdω) (30)
The overall profit Ris the sum of various revenue
components, including revenue from the energy market,
reserve market, and imbalance settlement as shown in (31).
R=λspEmbid1t+λcapRmbid
+X
ω∈
πω1tλBU 1Puω−λBD1Pdω−λcp 1Rdω
(31)
IV. RESULTS AND DISCUSSION
A first set of simulations is performed to find a mapping
function between the reserve market bid and the loading of
the wind turbine. The reserve bid is varied in the range of
0 MW to 0.75 MW in steps of 0.05 MW. These simulations
are performed with a 6 m/s wind profile with 5 % and 20 %
turbulence as shown in Section II-A. The other inputs to
the coupled model of wind turbine and generator are, grid
frequency, model parameters and the reserve market bid,
as presented in Figure 6. For each of these simulations, the
loads on different parts of the wind turbine are analysed.
Figure 8shows the forces in the base case which is simulated
to define the base values of the weight factors associated
with each of the 4 loads, as explained in Section III-A5.
This simulation is performed for a duration of 3000 s.
Figure 8 (a) and (b) show the axial and radial forces
on the wind turbine bearing. Figure 8 (c) and (d) show
the loads on the span 1 and span 9 of the blade. Fig-
ure 8 (e) and (f) show the load on shaft along the y and z axis.
Figure 8 (g) and (h) show the main load and base load of the
tower.
Table 3and Table 4present the results from the simulations
in terms of the normalised loading for the cases of 5 % and
20 % wind turbulence intensity. Here, Lbr ,Lbl ,Lsh,Ltwr and
Lrepresent the bearing, blades, shaft, tower and the total
loading respectively. The data in Table 3and Table 4, while
demonstrating a clear increasing trend in loading with respect
to Rmbid, also reveal a predominantly linear characteristic.
This validates our decision to employ linear regression as
detailed in Section III-B, through (14) -(16). As Rmbid
increases from 0 to 0.75 MW, there is a noticeable rise in
each of the loading terms. For the 5 % turbulence case, Lbr
exhibits a consistent increasing progression with its value
increasing from 0.471 for no reserve bid to 1.027 for the
highest reserve bid in the range. For the case of 20 %
turbulence, this rise is from 0.378 to 0.897. Similarly, Lbl
shows a consistent increment from 0.403 to 0.828 for 5 %
turbulence case and 0.335 to 0.726 for the 20 % turbulence
case, as Rmbid increases. In contrast, for the 5 % turbulence
case, Lsh and Ltwr demonstrate a relatively marginal increase
from 0.236 to 0.246 and 0.25 to 0.251 across the same Rmbid
range, respectively. Notably, Lsh stabilises from Rmbid value
of 0.25 MW onwards. A similar trend is observed for the
case of 20 % wind turbulence intensity. With the maximum
influence coming from Lbr and Lbl ,Ltotal shows a consistent
upward trend. The increased Lis due to the increased control
action that in turn effects the forces acting on different
components of the wind turbine. This dataset establishes
the relationship between Rmbid and the corresponding load
VOLUME 12, 2024 5315
N. Singh et al.: Load-Aware Operation Strategy for Wind Turbines
FIGURE 8. Forces acting of different wind turbine components.
variations, indicating a trend of load increase with the rise in
Rmbid. A corresponding graph between the cumulative Land
Rmbid is shown in Figure 9. It can be observed that the red
curve representing the Lfor 20 % turbulence case is slightly
lower compared to the 5 % turbulence case, shown by the blue
curve. This difference is attributed to the relative difference
in control actions in the two cases.
FIGURE 9. Reserve market bid versus Loading.
The next set of simulations are performed by the opti-
misation model (13)-(26), to evaluate the optimal market
bids in order to balance the wind turbine loads and the
WPP’s revenue. The values of λsp,λcap,λBU ,λBD and
λcp for revenue related calculations are e33, e34, e30,
e35 and e45, respectively. These market rates are in a
similar and comparable range as in the related literature [22],
[23] and several European electricity markets, such as
in Denmark, Norway, and Belgium [50],[51]. In these
simulations of a time period of 1 hour each, different
scenarios are studied in terms of α, which is the weight factor
associated with L. The intent is to analyse the impact of
weightage (α) of Lon the revenue and the optimised bids.
Table 5and Table 6present a synopsis of the reserve market
bids and expected revenue from JERM for the two simulated
cases of 5 % and 20 % wind turbulence intensity, respectively.
An examination of the data shows the impact of changing α
on the optimal bid and various presented revenues. Firstly,
it is evident that Rmbid decreases with the increasing α.
For the 5 % turbulence case, the maximum reserve bid of
0.60 MW is offered when αis zero. This bid is the maximum
possible reserve market bid, equivalent to a bid when the
Lis not taken into account for the optimisation. As the
value of αgradually increases, a decreasing trend is observed
in the Rmbid. Eventually, for αvalues 75 and higher, the
corresponding Rmbid is zero. For the case of 5 % turbulence
intensity, the maximum Rmbid of 0.28 MW is observed when
αis zero. It can be noticed that, Rmbid offered for the case of
20 % turbulence intensity is lower in comparison to th 5 %
turbulence intensity case. This is due to the fact that with a
higher level of turbulence in the wind, more fluctuations in
the wind are present that eventually lead to more instances
of lower wind speed. The optimisation algorithm developed
for this study takes these uncertainties into account for the
calculation of the optimal Rmbid.
The revenues associated with each scenario and the total
profit are presented in Table 5and Table 6. The revenue
5316 VOLUME 12, 2024
N. Singh et al.: Load-Aware Operation Strategy for Wind Turbines
TABLE 3. Normalised loading associated with the changing Rmbid [MW], wind turbulence intensity 5 %.
TABLE 4. Normalised loading associated with the changing Rmbid [MW], wind turbulence intensity 20 %.
from energy market and imbalance settlements, REI shows
proportionate changes to the values of α. For the 5 %
turbulence case, the value of REI rises from e5.02 to e24.52,
as αincreases up to 75. Whereas, for the case of 20 %
turbulence, REI increases from e18.63 to e27.49. The
growth indicates that the increasing values of αpositively
influence the revenue from the energy market. This is because
with higher values of α, a greater portion of energy is
designated for Embid, while a lower amount is set aside for
Rmbid. This occurs due to the negative term Lin the objective
function being directly influenced by Rmbid. Expanding on
this, it can be observed that for both the turbulence case,
the revenue for the reserve market, RRB is highest for the
lower values of αand gradually decreases as the value of
αincreases. However, it should be noted that the RRB is
the less dominant mode of revenue for the case of 20 %
turbulence as compared to the 5 % turbulence case. This is
due to the optimisation algorithm that accounts for the reserve
unavailability. RRB is a combined sum of the revenue in the
reserve market RDR and the penalty in balancing stage RPB.
RDR is bound with αand thereby by Rmbid. Similar is the case
for RPB. As αgrows, the amount of reserve offered reduces.
As a result, due to the lower WPP’s deviations from the
planned capacities, lower penalties are incurred. The overall
profit of the wind turbine participation in JERM, the total
revenue, Ris also presented in Table 5and Table 6. The values
of Rinclude the revenues from energy and reserve markets
along with the penalties due to the unavailability. It can be
seen here that the maximum profit is earned when the value of
TABLE 5. Revenues as a function of α, wind turbulence intensity 5 %.
αis zero, indicating that the impact of loading, Lis not taken
into account. The least revenue is observed for the higher
values of α.
Another key observation is made by identifying the knee
point in the data trend. This is achieved by calculating the
difference between consecutive data points of Rmbid and
R. The allocated reserve bid Rmbid and total revenue R,
as presented in Table 5, pinpoint a knee point at the αvalue
of 75, as illustrated in Figure 10 and Figure 11. This suggests
VOLUME 12, 2024 5317
N. Singh et al.: Load-Aware Operation Strategy for Wind Turbines
TABLE 6. Revenues as a function of α, wind turbulence intensity 20 %.
FIGURE 10. Variation in Rmbid with respect to α.
FIGURE 11. Variation in R with respect to α.
that decision-makers can prioritize the loading of the wind
turbine over a broad range without sacrificing profit, up to
an αvalue of 75. However, if the emphasis on the wind
turbine loading becomes significantly pronounced, the rate of
change in revenue (or the potential for revenue loss) becomes
steeper, rather than gradual. It should be noted that the values
presented here are for a single wind turbine, for 1 hour and
for a low wind speed scenario. The cumulative difference in
the profit for a longer duration and for an entire farm can be
highly significant in the interest of WPPs.
V. CONCLUSION
This study presents a novel method of wind turbine
participation in JERM using a load-aware profit maximising
approach. A methodology is developed to aggregate the
major loads of the wind turbine and present it as a single
unit. The loading factor generated from this methodology
acknowledges the loads on the blade, main bearing, shaft and
the tower of the wind turbine. The aggregate loading is then
used as an input to the optimisation problem that, given the
constraints, maximises the WPP’s profit while minimising
the physical loading of the wind turbine. The results of this
study indicate that the physical loading of the wind turbine
can be effectively modeled as a function of the reserve bid.
It is observed that when considering the wind turbine loading
in wind power scheduling, a lower reserve bid is achieved as
the loading weight increases. Similarly, as the loading weight
rises, the market revenue decreases. This is because, unlike
traditional scheduling models that overlook the wind turbine
loading and its hidden costs, our model implicitly accounts
for the costs associated with the wind turbine physical health.
From the simulations it is also observed that the level of
turbulence intensity of the wind also plays a role in the overall
loading of the wind turbine. This is due to the fact that the
control actions must be adapted to match the changing wind
speed, which in turn reflects on the overall loading. Another
crucial consideration for decision-makers is that the profit
and loss remains minimal across a broad range of loading
weights before experiencing a significant drop with only a
slight change in loading weight. Therefore, decision-makers
should identify the knee point in their data, as demonstrated
in this study, to ensure a profit that does not compromise
turbine health significantly. The weight factors a,b,cand
dassociated with the loads as defined in the study, provide
the WPP a useful tool at their disposal. These factors can be
tuned as per the costs and condition of these wind turbine
components. Additionally, the factor αassociated with the
physical loading in the bidding optimisation is an adjustable
factor that can be tuned to balance the revenue and the loading
as per the requirement of the WPP. In this way the WPPs
always have a trade-off option to create a balance between the
physical load on the wind turbine and the monetary benefit.
REFERENCES
[1] Global Wind Energy Council. (2023). Global Wind Report.
Accessed: Apr. 25, 2023. [Online]. Available: https://gwec.net/
globalwindreport2023/
[2] (2022). World Energy Outlook. Accessed: Apr. 25, 2023. [Online].
Available: https://www.iea.org/reports/world-energy-outlook-2022
[3] IEA. (2022). Wind Electricity. IEA, Paris. Accessed: Apr. 25, 2023.
[Online]. Available: https://www.iea.org/reports/wind-electricity
[4] (2019). Technical Requirements for the Connection of Generating Stations
to the Hydro-Qubec Transmission System. Accessed: Sep. 1, 2022.
[Online]. Available: http://www.hydroquebec.com/
5318 VOLUME 12, 2024
N. Singh et al.: Load-Aware Operation Strategy for Wind Turbines
[5] G. C. Tarnowski, P. C. Kjar, P. E. Sorensen, and J. Ostergaard, ‘‘Variable
speed wind turbines capability for temporary over-production,’’ in Proc.
IEEE Power Energy Soc. Gen. Meeting, Calgary, AB, Canada, Jul. 2009,
pp. 1–7, doi: 10.1109/PES.2009.5275387.
[6] J. Aho, P. Fleming, and L. Y. Pao, ‘‘Active power control of wind
turbines for ancillary services: A comparison of pitch and torque control
methodologies,’’ in Proc. Amer. Control Conf. (ACC), Boston, MA, USA,
Jul. 2016, pp. 1407–1412, doi: 10.1109/ACC.2016.7525114.
[7] A. Žertek, G. Verbic, and M. Pantoš, ‘‘Participation of DFIG wind turbines
in frequency control ancillary service by optimized rotational kinetic
energy,’’ in Proc. 7th Int. Conf. Eur. Energy Market, Madrid, Spain,
Jun. 2010, pp. 1–6, doi: 10.1109/EEM.2010.5558666.
[8] C. Nikolakakos, U. Mushtaq, P. Palensky, and M. Cvetkovic, ‘‘Improving
frequency stability with inertial and primary frequency response via
DFIG wind turbines equipped with energy storage system,’’ in Proc.
IEEE PES Innov. Smart Grid Technol. Eur. (ISGT-Europe), The Hague,
The Netherlands, Oct. 2020, pp. 1156–1160, doi: 10.1109/ISGT-
Europe47291.2020.9248807.
[9] H. Karbouj and Z. H. Rather, ‘‘Voltage control ancillary service from wind
power plant,’’ IEEE Trans. Sustain. Energy, vol. 10, no. 2, pp. 759–767,
Apr. 2019, doi: 10.1109/TSTE.2018.2846696.
[10] N. R. Ullah, K. Bhattacharya, and T. Thiringer, ‘‘Wind farms as reactive
power ancillary service providers—Technical and economic issues,’’ IEEE
Trans. Energy Convers., vol. 24, no. 3, pp. 661–672, Sep. 2009, doi:
10.1109/TEC.2008.2008957.
[11] E. Gatavi, A. Hellany, M. Nagrial, and J. Rizk, ‘‘An integrated
reactive power control strategy for improving low voltage ride-through
capability,’’ Chin. J. Electr. Eng., vol. 5, no. 4, pp. 1–14, Dec. 2019, doi:
10.23919/CJEE.2019.000022.
[12] Z. Dong, Z. Li, L. Du, Y. Liu, and Z. Ding, ‘‘Coordination strategy
of large-scale DFIG-based wind farm for voltage support with high
converter capacity utilization,’’ IEEE Trans. Sustain. Energy, vol. 12, no. 2,
pp. 1416–1425, Apr. 2021, doi: 10.1109/TSTE.2020.3047273.
[13] X. Cheng, J. Lin, F. Liu, Y. Qiu, Y. Song, J. Li, and S. Wu, ‘‘A coordinated
frequency regulation and bidding method for wind-electrolysis joint sys-
tems participating within ancillary services markets,’’ IEEE Trans. Sustain.
Energy, Jul. 3, 2023, early access, doi: 10.1109/TSTE.2022.3233062.
[14] S. A. Hosseini, J.-F. Toubeau, N. Singh, J. D. M. De Kooning,
N. Kayedpour, G. Crevecoeur, Z. De Grève, and L. Vandevelde, ‘‘Impact
of fast wind fluctuations on the profit of a wind power producer jointly
trading in energy and reserve markets,’’ in Proc. 9th Renew. Power Gener.
Conf., Mar. 2021, pp. 240–245, doi: 10.1049/Icp.2021.1386.
[15] N. Singh, J. D. M. De Kooning, and L. Vandevelde, ‘‘Simulation
of the primary frequency control pre-qualification test for a 5 MW
wind turbine,’’ in Proc. IEEE/PES Transmiss. Distribution Conf. Expo.,
Chicago, IL, USA, Oct. 2020, pp. 1–5, doi: 10.1109/TD39804.2020.
9299921.
[16] J. Dakovic, P. Ilak, T. Baskarad, M. Krpan, and I. Kuzle, ‘‘Effectiveness
of wind turbine fast frequency response control on electrically distanced
active power disturbance mitigation,’’ in Proc. Medit. Conf. Power Gener.,
Transmiss., Distribution Energy Convers. (MEDPOWER), Dubrovnik,
Croatia, Nov. 2018, pp. 1–7, doi: 10.1049/cp.2018.1923.
[17] H. Kim, J. Lee, J. Lee, and G. Jang, ‘‘Novel coordinated control strategy of
BESS and PMSG-WTG for fast frequency response,’’ Appl. Sci., vol. 11,
no. 9, p. 3874, Apr. 2021, doi: 10.3390/app11093874.
[18] X. Zhao, Z. Lin, B. Fu, and S. Gong, ‘‘Research on frequency control
method for micro-grid with a hybrid approach of FFR-OPPT and pitch
angle of wind turbine,’’ Int. J. Electr. Power Energy Syst., vol. 127,
May 2021, Art. no. 106670, doi: 10.1016/j.ijepes.2020.106670.
[19] N. Singh, J. D. M. De Kooning, and L. Vandevelde, ‘‘Dynamic wake
analysis of a wind turbine providing frequency support services,’’ IET
Renew. Power Gener., vol. 16, no. 9, pp. 1853–1865, Jul. 2022, doi:
10.1049/rpg2.12455.
[20] N. Kayedpour, J. D. M. De Kooning, A. E. Samani, F. Kayedpour,
L. Vandevelde, and G. Crevecoeur, ‘‘An optimal wind farm operation
strategy for the provision of frequency containment reserve incorporating
active wake control,’’ IEEE Trans. Sustain. Energy, Jan. 1, 2024, early
access, doi: 10.1109/TSTE.2023.3288130.
[21] Energy Storage Applications Summary. (2020). European Association
for Storage of Energy. Accessed: May 6, 2022. [Online]. Available:
https://ease-storage.eu/wp-content/uploads/2020/06/ES-Applications-
Summary.pdf
[22] S. A. Hosseini, J.-F. Toubeau, Z. De Grève, and F. Vallée, ‘‘An advanced
day-ahead bidding strategy for wind power producers considering
confidence level on the real-time reserve provision,’’ Appl. Energy,
vol. 280, Dec. 2020, Art. no. 115973, doi: 10.1016/j.apenergy.2020.
115973.
[23] S. A. Hosseini, J.-F. Toubeau, N. Amjady, and F. Vallée, ‘‘Day-ahead
wind power temporal distribution forecasting with high resolution,’’ IEEE
Trans. Power Syst., Jul. 17, 2023, early access, doi: 10.1109/TPWRS.2023.
3295915.
[24] A. S. Verma, J. Yan, W. Hu, Z. Jiang, W. Shi, and J. J. E. Teuwen, ‘‘A review
of impact loads on composite wind turbine blades: Impact threats
and classification,’’ Renew. Sustain. Energy Rev., vol. 178, May 2023,
Art. no. 113261, doi: 10.1016/j.rser.2023.113261.
[25] R. P. Tavares, V. Bouwman, and W. Van Paepegem, ‘‘Finite element
analysis of wind turbine blades subjected to torsional loads: Shell vs solid
elements,’’ Compos. Struct., vol. 280, Jan. 2022, Art. no. 114905, doi:
10.1016/j.compstruct.2021.114905.
[26] D. Wei, D. Li, T. Jiang, P. Lyu, and X. Song, ‘‘Load identification
of a 2.5 MW wind turbine tower using Kalman filtering techniques
and BDS data,’’ Eng. Struct., vol. 281, Apr. 2023, Art. no. 115763, doi:
10.1016/j.engstruct.2023.115763.
[27] Y. Hu, J. Yang, C. Baniotopoulos, X. Wang, and X. Deng, ‘‘Dynamic
analysis of offshore steel wind turbine towers subjected to wind, wave and
current loading during construction,’’ Ocean Eng., vol. 216, Nov. 2020,
Art. no. 108084, doi: 10.1016/j.oceaneng.2020.108084.
[28] S. Wang, T. Moan, and Z. Gao, ‘‘Methodology for global structural load
effect analysis of the semi-submersible hull of floating wind turbines
under still water, wind, and wave loads,’’ Mar. Struct., vol. 91, Sep. 2023,
Art. no. 103463, doi: 10.1016/j.marstruc.2023.103463.
[29] S. Wang, Y. Xing, R. Balakrishna, W. Shi, and X. Xu, ‘‘Design,
local structural stress, and global dynamic response analysis
of a steel semi-submersible hull for a 10-MW floating wind
turbine,’’ Eng. Struct., vol. 291, Sep. 2023, Art. no. 116474, doi:
10.1016/j.engstruct.2023.116474.
[30] N. Beganovic, J. Njiri, and D. Soffker, ‘‘Reduction of structural loads
in wind turbines based on an adapted control strategy concerning online
fatigue damage evaluation models,’’ Energies, vol. 11, no. 12, p. 3429,
Dec. 2018, doi: 10.3390/en11123429.
[31] X. Wang, Y. Wang, and Y. Liu, ‘‘Dynamic load frequency control for
high-penetration wind power considering wind turbine fatigue load,’’ Int.
J. Electr. Power Energy Syst., vol. 117, May 2020, Art. no. 105696, doi:
10.1016/j.ijepes.2019.105696.
[32] P. A. Fleming, J. Aho, A. Buckspan, E. Ela, Y. Zhang, V. Gevorgian,
A. Scholbrock, L. Pao, and R. Damiani, ‘‘Effects of power reserve
control on wind turbine structural loading,’’ Wind Energy, vol. 19, no. 3,
pp. 453–469, Mar. 2016, doi: 10.1002/we.1844.
[33] H. Camblong, I. Vechiu, A. Etxeberria, and M. I. Martínez, ‘‘Wind
turbine mechanical stresses reduction and contribution to frequency
regulation,’’ Control Eng. Pract., vol. 30, pp. 140–149, Sep. 2014, doi:
10.1016/j.conengprac.2014.03.007.
[34] B. Liu, J. Zhao, Q. Huang, F. Milano, Y. Zhang, and W. Hu,
‘‘Nonlinear virtual inertia control of WTGs for enhancing primary
frequency response and suppressing drivetrain torsional oscillations,’’
IEEE Trans. Power Syst., vol. 36, no. 5, pp. 4102–4113, Sep. 2021, doi:
10.1109/TPWRS.2021.3055262.
[35] N. Singh, D. Boruah, J. D. M. De Kooning, W. De Waele, and
L. Vandevelde, ‘‘Impact assessment of dynamic loading induced by the
provision of frequency containment reserve on the main bearing lifetime
of a wind turbine,’’ Energies, vol. 16, no. 6, p. 2851, Mar. 2023, doi:
10.3390/en16062851.
[36] J. Jonkman, S. Butterfield, W. Musial, and G. Scott. (2009). Definition
of a 5-MW Reference Wind Turbine for Offshore System Development.
[Online]. Available: https://www.nrel.gov/docs/fy09osti/38060.pdf
[37] J. Jonkman. (2016). FAST (Version V8) [Software]. NREL. [Online].
Available: https://www.nrel.gov/wind/nwtc/fastv8.html
[38] NREL. Accessed: Nov. 11, 2023. [Online]. Available:
https://www.nrel.gov/wind/nwtc/turbsim.html
[39] J. D. M. De Kooning, J. Van de Vyver, B. Meersman, and L. Vandevelde,
‘‘Maximum efficiency current waveforms for a PMSM including
iron losses and armature reaction,’’ in Proc. 22nd Int. Conf. Electr.
Mach. (ICEM), Lausanne, Switzerland, Sep. 2016, pp. 1875–1881, doi:
10.1109/ICELMACH.2016.7732779.
VOLUME 12, 2024 5319
N. Singh et al.: Load-Aware Operation Strategy for Wind Turbines
[40] L. Sethuraman and K. L. Dykes, ‘‘GeneratorSE: A sizing tool for
variable-speed wind turbine generators,’’ Nat. Renew. Energy Lab.
(NREL), Golden, CO, USA, Tech. Rep. NREL/TP-5000-66462, 2017, doi:
10.2172/1395455.
[41] M. E. Zarei, D. Ramírez, M. Prodanovic, and G. M. Arana, ‘‘Model
predictive control for PMSG-based wind turbines with overmodulation and
adjustable dynamic response time,’’ IEEE Trans. Ind. Electron., vol. 69,
no. 2, pp. 1573–1585, Feb. 2022, doi: 10.1109/TIE.2021.3057021.
[42] A.-R. Youssef, E. E. M. Mohamed, and A. I. M. Ali, ‘‘Model predictive
control for grid-tie wind-energy conversion system based PMSG,’’ in Proc.
Int. Conf. Innov. Trends Comput. Eng. (ITCE), Aswan, Egypt, Feb. 2018,
pp. 467–472, doi: 10.1109/ITCE.2018.8316668.
[43] S. M. Mozayan, M. Saad, H. Vahedi, H. Fortin-Blanchette, and M. Soltani,
‘‘Sliding mode control of PMSG wind turbine based on enhanced
exponential reaching law,’’ IEEE Trans. Ind. Electron., vol. 63, no. 10,
pp. 6148–6159, Oct. 2016, doi: 10.1109/TIE.2016.2570718.
[44] C. Busca, A.-I. Stan, T. Stanciu, and D. I. Stroe, ‘‘Control of permanent
magnet synchronous generator for large wind turbines,’’ in Proc. IEEE
Int. Symp. Ind. Electron., Bari, Italy, Jul. 2010, pp. 3871–3876, doi:
10.1109/ISIE.2010.5637628.
[45] H. Salime, B. Bossoufi, S. Motahhir, and Y. El Mourabit, ‘‘A novel
combined FFOC-DPC control for wind turbine based on the permanent
magnet synchronous generator,’’ Energy Rep., vol. 9, pp. 3204–3221,
Dec. 2023, doi: 10.1016/j.egyr.2023.02.012.
[46] S. B. Crary, ‘‘Two-reaction theory of synchronous machines,’’ Electr. Eng.,
vol. 56, no. 1, pp. 27–31, Jan. 1937.
[47] P. Krause, O. Wasynczuk, S. D. Sudhoff, and S. Pekarek, Analysis of
Electric Machinery and Drive Systems. Piscatawy, NJ, USA: Wiley, 2013.
[48] A. E. Samani, J. D. M. De Kooning, N. Kayedpour, N. Singh, and
L. Vandevelde, ‘‘The impact of pitch-to-stall and pitch-to-feather control
on the structural loads and the pitch mechanism of a wind turbine,’’
Energies, vol. 13, no. 17, p. 4503, Sep. 2020, doi: 10.3390/en13174503.
[49] Gurobi Optimization. (2023). Gurobi Optimizer Reference Manual. LLC.
[Online]. Available: https://www.gurobi.com
[50] Datasets. (2021). Energinet. Accessed: Oct. 5, 2023. [Online]. Available:
https://www.energidataservice.dk/search
[51] ENTSO-E Transparency Platform. Accessed: Oct. 5, 2023. [Online].
Available: https://transparency.entsoe.eu/
NARENDER SINGH received the Master of
Science degree in electric power engineering from
the KTH Royal Institute of Technology, Sweden,
and the Ph.D. degree, in 2023. He joined the
Department of Electromechanical, Systems and
Metal Engineering, Faculty of Engineering and
Architecture, Ghent University, Belgium, in 2019.
His research interests include renewable energy
and power systems studies.
SEYYED AHMAD HOSSEINI (Member, IEEE)
received the Ph.D. degree in electrical engineering
from the Power Systems and Markets Research
Group, University of Mons, Mons, Belgium,
in 2023. He is currently a Postdoctoral Researcher
with the Power Systems and Markets Research
Group. Furthermore, he was an Adjunct Research
Fellow with Federation University, Australia,
in 2023. His research interests include machine
learning, integration of wind power into electricity
markets, and decision-making under uncertainty.
JEROEN D. M. DE KOONING (Senior Member,
IEEE) was born in Kapellen, Belgium, in 1987.
He received the Master of Science and Ph.D.
degrees in electromechanical engineering from
Ghent University, Belgium, in 2010 and 2015,
respectively. Since 2019, he has been an Assistant
Professor with the Department of Electromechan-
ical, Systems and Metal Engineering, Faculty of
Engineering and Architecture, Ghent University.
In 2022, he was a Visiting Professor with the
Lappeenranta University of Technology, Finland. He conducted research on
current waveform shaping techniques for permanent magnet synchronous
machines and optimal control and design of renewable energy systems. His
current research interests include modelling, optimization and control of
mechatronic systems, drivetrains, and manufacturing machines in an industry
4.0 context, with a particular interest in digital twins. He is a member of
the FlandersMake@UGent MIRO Core Laboratory, Dynamical Systems and
Control (DySC) Research Group.
FRANÇOIS VALLÉE (Member, IEEE) received
the degree in civil and electrical engineering
and the Ph.D. degree in electrical engineering
from the Faculty of Engineering, University of
Mons, Belgium, in 2003 and 2009, respectively.
He is currently a Professor and the Leader of the
‘‘Power Systems and Markets Research Group,’’
University of Mons. His research interests include
PV and wind generation modeling for electrical
system reliability studies in presence of dispersed
generation. His Ph.D. work has been awarded by the SRBE/KBVE Robert
Sinave Award, in 2010.
LIEVEN VANDEVELDE (Senior Member, IEEE)
was born in Eeklo, Belgium, in 1968. He received
the Graduate degree in electrical and mechani-
cal engineering (main subject: electrical power
engineering) and the Ph.D. degree from Ghent
University, in 1992 and 1997, respectively. Since
then, he has been with the Electrical Energy
Laboratory (EELAB), Department of Electrome-
chanical, Systems and Metal Engineering, Ghent
University. He has conducted research in various
domains of electrical power engineering, inter alia electrical machines, and
(computational) electromagnetics. Since 2004, he has been a member of
the Professorial Staff and has been coordinating the research on electric
power systems with EELAB. In this research, renewable energy and its
integration in electric power systems play a prominent role. He is a member
of EnerGhentIC, the interdisciplinary community of Ghent University
researchers working on the energy challenge.
5320 VOLUME 12, 2024
