Wind Energy Conversion System with Integrated Power Smoothing
Capability based on an EVT-coupled Flywheel
MARCELO G. CENDOYA, SANTIAGO A. VERNE, MARÍA I. VALLA,
PEDRO E. BATTAIOTTO
Instituto de Investigaciones en Electrónica, Control y Procesamiento de Señales (LEICI),
UNLP - CONICET - CIC PBA,
Departamento de Electrotecnia, Facultad de Ingeniería,
Universidad Nacional de La Plata,
Calle 1 y 47, La Plata (B1900TAG), Buenos Aires,
ARGENTINA
Abstract: - This paper presents a wind energy conversion system that integrates, in a compact form, both
generation and power smoothing functions. The system is based on a vertical axis wind turbine (VAWT) and a
low-speed flywheel that are coupled by an electric variable transmission (EVT). The EVT performs a dual
function by simultaneously injecting power to the grid and regulating the power flow to the flywheel to smooth
the power fluctuation caused by wind turbulences. The main advantages of the presented generation/smoothing
topology are integrating two machines (generator and motor) in the same physical structure and the reduced
power demanded by the converter that feeds the flywheel’s rotor. A theoretical analysis is carried out to
determine an appropriate control law and flywheel sizing to satisfy a desired attenuation of the power
fluctuation. Simulation results show an effective smoothing action, as the power fluctuation is reduced to an
adequate level. In addition, the validity of the flywheel sizing procedure is verified.
Keywords: - Vertical Axis Wind Turbines; Power Smoothing; Wind Turbulence; Low-speed Flywheel; Electric
Variable Transmission; Weak Grid; Renewable Distributed Generation; Wind Energy
Conversion.
Received: March 11, 2023. Revised: December 19, 2023. Accepted: February 7, 2024. Published: March 27, 2024.
1 Introduction
The southern region of Argentina, widely known as
"Patagonia," has several remote villages dedicated
to agricultural, livestock, mining, and oil activities.
As is well known, the availability of abundant
electrical energy is a key driver for the development
and growth of such productions. While numerous
rural settlements have access to electricity grids,
these lines are typically weak and can only provide
a limited amount of power to meet current demand.
Moreover, due to their low population, renewal or
expansion of these grids is unlikely to occur, as the
energy distribution companies would need to absorb
the high costs involved with little economic return.
Fortunately, this region of the country has
significant wind energy potential, [1], [2], which has
led to numerous wind farm projects. However, these
farms, typically of large energy production, are
designed to supply major power consumption
centers rather than small rural villages. As a result,
to increase the available electrical power,
inhabitants of these villages must individually turn
to the utilization of diesel generators. This entails
undesirable environmental pollution and high
operating costs, as the elevated fuel prices are
compounded by the cost of transportation from
remote locations. This scenario motivates the search
for alternative solutions, and in this context, the
application of Renewable Distributed Generation
(RDG) seems appropriate, [3]. In RDG, small-scale
generation close to consumption points replaces
large and distant generation plants, eliminating the
need for long and expensive transmission lines and
associated equipment.
Due to its remarkable characteristics in terms of
high energy yield and control flexibility, the three-
bladed horizontal-axis wind turbine (HAWT) is the
most used type of turbine and leads the wind power
industry, especially in the high-power segment.
However, it has some drawbacks. For instance,
turbine rotor and conversion machinery must be
installed in a nacelle on the top of a tower of large
height involving high logistic costs. Also, the
nacelle requires a yaw system that constantly faces
the turbine to the incoming wind. For low and
WSEAS TRANSACTIONS on POWER SYSTEMS
DOI: 10.37394/232016.2024.19.9
Marcelo G. Cendoya, Santiago A. Verne,
María I. Valla, Pedro E. Battaiotto
E-ISSN: 2224-350X
68
Volume 19, 2024
medium-power applications, the vertical-axis wind
turbine (VAWT) has recently gained renewed
interest from researchers and manufacturers because
it does not require a tower or orientation
mechanism, and the electric generator is located at
ground level, [4], [5]. Among VAWTs, the
Darrieus-H appears as the most attractive due to its
structural simplicity compared to the traditional
Darrieus. Additionally, recent studies have allowed
the development of modern design methods, notably
increasing its efficiency to around 40%, [6].
Due to the fluctuating nature of the wind
resource, one of the main challenges in the wind
conversion industry is to enhance power quality by
delivering smooth and constant power, [7]. If not, a
fluctuating power injected into the grid introduces
disturbances that can lead to instabilities of voltage
and frequency, and this worsens as the generated
power is on the order of the power capacity of the
grid (high penetration). Power fluctuations are
originated by rapid variations of wind speed which
lead to different effects on the conversion chain.
Very rapid variations in wind speed generally do not
cause serious disturbances as their effect is
mitigated by the turbine's inertia. However, changes
in wind speed defined as "turbulence" are not fast
enough for the turbine's inertia to filter the effect.
Therefore, it could produce significant changes in
the generator's rotation speed (or torque) leading to
variations in the electrical power output of the
generator, as well as voltage variations at the point
of connection.
One method to mitigate the power fluctuations
generated by a turbine due to wind turbulence is
using a short-term (low capacity) energy storage
system. Among different technologies
(supercapacitors, batteries, etc.), the flywheel
emerges as one of the most attractive options due to
its high power density, high efficiency, and long
lifespan, [8], [9]. Particularly, low-speed flywheels
are characterized by constructional simplicity, low
maintenance, and cost since (unlike high-speed
flywheels) they are made using conventional
materials, use standard bearings, and do not require
a safety container, [10]. A typical flywheel-based
power smoothing system consists of a dedicated
driving machine with a grid-connected electronic
converter. The converter transfers all the flywheel
mechanical power to the grid. A control loop adjusts
this power to compensate for fluctuations. The
machine that moves the flywheel is mechanically
separated from the machine that, driven by the wind
turbine, acts as a generator.
Electrical Variable Transmissions (EVT) are
devices based on electrical machines that allow
controlling the transmitted mechanical power
between two shafts, which rotate at different speeds,
[11]. EVTs are widely used in electric vehicles,
[12], but are rarely employed in other applications,
exceptions can be seen in [13], [14], [15]. In [13],
[14] an EVT is used to couple a variable-speed
HAWT with a grid-connected fixed-speed
synchronous generator. In [15], a fixed-speed micro-
hydro turbine is coupled through an EVT to a
variable-speed centrifugal pump for irrigation.
In this work, an EVT is used to integrate both
functions: power generation and smoothing in a
compact form. A VAWT is coupled to the input
shaft of the EVT such that its associated rotor and
the stator function as a grid-connected induction
generator (fixed-speed "Danish Concept" scheme).
The second rotor associated with the output shaft of
the EVT, is fed by a variable frequency drive
(VFD), which is used as the driving machine of a
low-speed flywheel. The flywheel operates at
variable speed and is controlled to minimize
fluctuations in the power injected into the grid
caused by wind turbulence. Thus, an integrated
wind energy conversion and power smoothing
structure is proposed, specially designed to be used
as an RDG system in rural areas with access to
electrical utility through a weak grid. Compared to
the typical and previously mentioned scheme, the
two main advantages of the presented
generation/smoothing topology are: (a) the
integration of the two machines into a single
physical structure resulting in a more compact,
simple, and economical system and (b) the electrical
power handled by the converter that feeds the
flywheel driver is smaller than the one required if it
were operated with a separate machine.
2 System Description and Problem
Statement
Figure 1 shows a diagram of the proposed wind
power conversion system with integrated smoothing
capability. It consists of a VAWT, a multiplier
gearbox, an EVT, and a low-speed flywheel.
Figure 2 shows the internal structure of the EVT
[11]. There is a stator winding, a cup-shaped rotor
(rotor 1), and a wound rotor (rotor 2). Shaft 1 drives
the cup-shaped rotor which has an inner face and an
outer face. The outer face has a squirrel cage, so the
stator and the outer face of rotor 1 can be considered
as an induction machine. On the other hand, the
inner face of rotor 1 has a permanent magnet
arrangement which generates a rotating magnetic
field around rotor 2. In this way, the machine has
WSEAS TRANSACTIONS on POWER SYSTEMS
DOI: 10.37394/232016.2024.19.9
Marcelo G. Cendoya, Santiago A. Verne,
María I. Valla, Pedro E. Battaiotto
E-ISSN: 2224-350X
69
Volume 19, 2024
two mechanical ports and two electrical ports, such
that electrical and mechanical power can flow from
one to the other depending on the electrical
excitations of both windings and the relative speeds
and torques that are exerted on both shafts.
Fig. 1: Integrated generation/smoothing system
diagram
Fig. 2: EVT internal structure
In this proposal, the high-speed shaft of the
gearbox is coupled to shaft 1 of the EVT. The stator
is directly connected to a weak grid, such that rotor
1 and the stator act as an induction generator. This
configuration causes the turbine speed to stay within
a small range close to the synchronous speed
imposed by the grid frequency. The flywheel is
driven by rotor 2 of the EVT whose winding is
connected to the output of the VFD, also fed by the
grid. The VFD regulates the torque of rotor 2 to
establish its idle speed and regulate the power flow
between shafts.
2.1 EVT Stator Power Fluctuation Caused
by Wind Turbulence
The VAWT is coupled to shaft 1 of the EVT
through a multiplier gearbox with a speed ratio a, so
that the rotation speed of shaft 1 of the EVT is ω1:
(1)
Where ωT is the speed of the turbine. The speed
ω1 varies according to Newton’s law of rotational
dynamics (2):
(2)
Where: JT1: Total inertia of shaft 1:
with J1 inertia moment of rotor 1 and JT the
inertia of the turbine itself. TT1: torque exerted by
the turbine referred to shaft 1. Tem1: electromagnetic
torque between rotor 1 and the stator of the EVT.
Tem2: electromagnetic torque between rotor 1 and
rotor 2 of the EVT.
The torque exerted by the VAWT referred to
shaft 1 of EVT is given by (3), [16]:
(3)
Where ρ: air density, A: capture area (A=D.H,
turbine diameter D times height H), v: wind speed
and CT: torque coefficient (depends on ω1 and v).
Tem1 in (2), is the torque that arises from the
interaction of the rotating magnetic field established
by the stator winding with the bars of the outer face
of rotor 1. Since this part of the EVT configures a
traditional induction machine, for Tem1 to occur,
rotor 1 must rotate at a different speed from the
magnetic field produced by the stator (there is a
slip) which rotates at the synchronous speed ωs:
(4)
Where fe1 is the electrical frequency of the
stator, fixed by the grid and p1 is the number of pole
pairs of the stator winding. Considering that rotor 1
operates at a speed close to synchronous speed (low
slip), it is possible to calculate Tem1 through a linear
approximation given by (5), [17]:
(5)
Where KG is the slope of the torque-speed
characteristic whose value depends on the amplitude
of the flux in the air gap (determined by the ratio
Vs/fe1) and the resistance of rotor bars. Neglecting
losses in the stator winding of the EVT, the
electrical power injected to the grid (Ps) can be
calculated as:
(6)
WSEAS TRANSACTIONS on POWER SYSTEMS
DOI: 10.37394/232016.2024.19.9
Marcelo G. Cendoya, Santiago A. Verne,
María I. Valla, Pedro E. Battaiotto
E-ISSN: 2224-350X
70
Volume 19, 2024
Taking a time span of less than 10 minutes, the
wind speed v reaching the VAWT can be
decomposed as a mean value vo and a turbulence of
relatively rapid variation vt [18]:
(7)
Turbulence is random and is mainly caused by
ground friction and vertical wind displacement due
to thermal heating. The repetition period of
turbulence, corresponding to the peak in the high-
frequency zone of the Van der Hoven spectrum, is
on the order of one minute, [18]. As for the
turbulence amplitude, this can be estimated using
the IEC61400-2 standard, which provides a Normal
Turbulence Model. In many practical cases, it is
sufficient to consider that the turbulence intensity is
equal to 20% of the mean wind speed, as suggested
by DNV GL in 1993, [18].
Although turbulence is a complex phenomenon
and cannot be exactly represented by deterministic
equations, recent studies propose representing it as a
sinusoidal variation to facilitate the analysis, design,
and simulation results comparison, [19], [20]. Wind
variation can thus be described as (8):
(8)
Where Vt and ft are the amplitude and frequency
of the equivalent turbulence, respectively.
The torque of the VAWT resulting from
introducing (8) into (3) is composed of a constant
component and a variable component due to the
turbulence. The nonlinear dependence of TT1 on v
produces harmonics of ft on the turbine torque but to
simplify the system analysis, controller design, and
the sizing the flywheel inertia, only the fundamental
component is considered here. Therefore, it is
assumed that this component exhibits a cosine
variation (9):
(9)
Where TT10 is the mean torque, TT1p and fp are
the amplitude and frequency of the torque variation
caused by turbulence, respectively.
To analyze power fluctuations on the stator of
the EVT using (6), the variation of ω1 produced by
(9) must be evaluated, as Tem1 depends on ω1 (5).
The relationship between the variable component of
TT1 and ω1 is evaluated assuming that Tem2=0. Then,
a transfer function G1(s) can be derived from (2):
(10)
G1(s) has a low-frequency gain G10=1/KG and
exhibits a real pole. This means that the VAWT
behaves like an inertial low-pass filter with a cutoff
frequency f1 given by (11):
(11)
Therefore, the magnitude of the variation of ω1
caused by TT1p will depend on the relation fp /f1. For
instance, if fp>>f1, the variations of ω1 will be small
due to the attenuation produced by the moment of
inertia JT1. On the contrary, if fp<<f1, this
attenuation is very small, and then ω1 can be
calculated using G10 and (9) as:
(12)
Where ω10 is the mean component and ω1p is the
variation due to wind turbulence.
If fp <<f1, Tem1 and TT are almost equal, as the
reaction torque produced by the moment of inertia
of the VAWT is very small. Therefore, neglecting
second-order terms, the power injected to the grid
by the stator of the EVT can be approximated as:
(13)
In (13) PS0 is the mean power and PSp the
variable power due to the wind turbulence.
3 Compensation of Power Fluctuations
In the previous section, it was assumed that the
electromagnetic torque Tem2 (which arises from the
interaction between the field of the permanent
magnets located on the inner face of rotor 1 and the
winding of rotor 2) has remained null. However,
Tem2 can be adjusted to any desired value through
the VFD. Equation (2) indicates that if the VFD is
properly managed, it is possible to neutralize the
effect that TT1p has over ω1. If Tem2 is made equal to:
(14)
The speed ω1 will not vary in the presence of
the considered turbulence. This means that Tem1 will
not vary either and therefore, Ps will not be
disturbed. The drawback of applying (14) is that TT1
should be measured, which is costly and complex.
However, it is possible to find the appropriate value
of Tem2 if a controller is used to minimize
fluctuations in Ps (Psp). Psp is thus obtained by
WSEAS TRANSACTIONS on POWER SYSTEMS
DOI: 10.37394/232016.2024.19.9
Marcelo G. Cendoya, Santiago A. Verne,
María I. Valla, Pedro E. Battaiotto
E-ISSN: 2224-350X
71
Volume 19, 2024
subtracting the mean value of Ps (Pso) obtained, in
turn, through low-pass filtering:
(15)
Then, Tem2p is derived:
(16)
Here KC is the gain of the controller and Pspref is
set to 0. To ensure that the flywheel is always
capable of absorbing/delivering energy, its idle
speed must be set in an intermediate value. Thus, a
proportional speed control loop is used (17):
(17)
Where KC0 is the controller gain and ω2ref is the
reference speed. It is convenient to set ω2ref=ω10, the
speed of shaft 1 of the EVT corresponding to the
mean value of the wind speed. The closed-loop
transfer function between ω2 and Tem20 has a pole at
frequency f2:
(18)
Therefore, KC0 must be adjusted such that the
frequency components of flywheel speed caused by
wind turbulence are outside the bandwidth of the
idle speed controller, that is f2<<fp.
Summarizing, to achieve the objectives already
mentioned, the reference value Tem2ref for the internal
torque controller of the VFD should consist of two
terms:
(19)
3.1 Sizing of the Flywheel
If the aerodynamic and bearing losses of shaft 2 are
negligible, the evolution of the rotation speed of the
flywheel is governed by (20):
(20)
Where JT2=J2+JF is the total moment of inertia
associated with shaft 2 of the EVT, with J2 inertia
moment of rotor 2 and JF the inertia of flywheel
itself.
If a perfect compensation for Ps is reached, it
means that the value of Tem2 given by (14) is applied
and the variations of shaft 1 speed are completely
canceled. Under the assumption that initially ω2=ω10
(considering an ideal idle speed control loop and
that this controller does not provide any corrective
action facing a fast speed change), the introduction
of (14) into (20) leads to the expression of the
instantaneous speed of the flywheel:
(21)
The first term in (21) represents the departure of
the flywheel speed concerning the idle speed due to
Tem2p.
The instantaneous power of the flywheel (pf)
can be obtained by multiplying (14) and (21):
(22)
In (22), the first term (p12) is the power
transmitted from shaft 1 to shaft 2 through
electromagnetic coupling Tem2. The second term
corresponds to the power provided by the VFD
(pVFD), which has twice the frequency of p12.
It is observed that the proposed compensation
technique cancels the power disturbance in the
stator of the EVT due to wind turbulence, deriving it
to the flywheel through the electromagnetic
coupling between both shafts. However, a residual
fraction of this disturbance reaches the grid through
the VFD. Therefore, an appropriate criterion for
sizing the moment of inertia of the flywheel is such
that the peak value of the second term in (22) is a
small fraction of the peak value of the first term. If
we call this fraction KP:
(23)
Neglecting J2 (As JF>>J2) the required moment
of inertia JF for the flywheel to comply with (23), is:
(24)
4 Simulation Results
To evaluate the performance of the control scheme
for attenuating power fluctuations in the weak grid,
a computational model of the system was
developed. This model is based on the equations
presented in Section 2 in which the main
characteristics and parameter values of each part of
the model are detailed below.
WSEAS TRANSACTIONS on POWER SYSTEMS
DOI: 10.37394/232016.2024.19.9
Marcelo G. Cendoya, Santiago A. Verne,
María I. Valla, Pedro E. Battaiotto
E-ISSN: 2224-350X
72
Volume 19, 2024
4.1 VAWT
Three straight-blade Darrieus-H VAWT with a
height of H=11.8m and rotation radius of R=10m
(capture area A=236m2), fixed pitch. Nominal power
PT=100kW for a wind speed v=12m/s. The torque
coefficient CT in (3) is calculated with (25):
(25)
Where λi is calculated as (26):
(26)
In (25) and (26), β is the pitch angle (β=0 in our
case), and λ=(ωTR)/v is the ratio of the tangential
blade speed to the wind speed. The numerical values
of the coefficients in (25) are: c1=0.4332, c2=116,
c3=0.4, c4=5, c5=21, c6=0.0128. The moment of
inertia of the turbine is JT=52000kgm2. The turbine
is coupled to shaft 1 of the EVT through a speed
multiplier gearbox with a ratio of a=38.75. This
leads to a rotation speed of shaft 1 of the EVT of
1550RPM for turbine operation in rated conditions.
4.2 EVT
Stator winding: 4 poles, 380V/50Hz, rated power
PN=100kW, synchronous speed ns=1500RPM.
KG=117.67Nm/(r/s). The rotor of shaft 2 has 4 poles
winding, 380V/50Hz, and is fed by a 25kW VFD.
4.3 Flywheel
Using the sizing procedure of Section 2 and
adopting KP=1/5, applying (24) results in
JF=48Kgm2. Given the low speed of the flywheel,
its mechanical and aerodynamic losses are
considered negligible.
In a preliminary step, a simulation was
conducted to observe how wind turbulence affects
the behavior of the VAWT, with the power
fluctuation attenuation controller disabled. The
results are shown in Figure 3. Figure 3(a) shows the
wind speed profile used in the simulations. Starting
from a stable wind speed of v=10m/s, after the
turbine reached its rated speed, at t=105s, a
sinusoidal turbulence with an amplitude of 2m/s
appears (representing a turbulence intensity of 20%,
[18]). The frequency of turbulence is 16.66mHz,
corresponding to the peak of the Van der Hoven
spectrum (approximately 1 cycle per minute). In
Figure 3(b), the variation of the turbine torque
(referred to as shaft 1 of the EVT) is shown, and it is
observed that its shape is also sinusoidal with the
same frequency as the turbulence, as indicated in
Section 2. Using (11), the frequency of the
mechanical pole of shaft 1 dynamics can be
calculated, resulting in f1=0.54Hz. Since the
frequency of the turbulence is much lower than the
pole frequency f1, nearly all the resulting torque
disturbance of Figure 3(b) is directly transferred to
the electromagnetic torque Tem1 (Figure 3(d)).
Evidence of this fact is the small value of the torque
with which the inertia reacts, Figure 3(c). After the
start-up transient, the turbine speed stabilizes, and
shaft 1 of the EVT reaches approximately
n1=1525RPM (Figure 3(e)). Due to the steep slope
(KG) of the torque-speed characteristic of the
generator composed by the EVT stator and bars on
the outer face of rotor 1, the appearance of
turbulence causes a very small speed variation
around this value, approximately +/-25RPM.
The preliminary simulation confirms that the
turbine's inertia has a negligible filtering effect on
wind turbulence and the need to add a system to
reduce the Tem1 variations is justified.
This is why the entry into operation of the
proposed smoothing system is enabled to attenuate
the power fluctuations caused by wind turbulence,
and the results are presented in Figure 4, Figure 5
and Figure 6.
In Figure 4, the dynamic behavior of the
mechanical variables associated with the flywheel is
shown. The flywheel's accelerating/decelerating
torque is controlled by the VFD feeding the inner
rotor windings. As the dynamic of this loop is
determined by short electrical time constants, the
torque applied to the flywheel (shaft 2 of the EVT)
almost instantaneously copies the torque reference
signal. This reference consists of two components:
Tem20 generated by the flywheel’s idle speed control
loop, and Tem2p generated by the control loop for the
power smoothing in the EVT stator. Initially, to
bring the flywheel into operation, only the slow-
speed control loop for idle speed is enabled. This
loop has a speed reference n2ref consisting of a step
from 0 to 1525RPM at t=50s, as shown in Figure
4(a). A proportional gain KCo=0.3 for the idle speed
closed loop is used and produces a torque reference
Tem20 according to (17), which is shown in Figure
4(c). Once the flywheel speed is stabilized (Figure
4(b), t=1025s) the power smoothing loop is enabled.
From this moment, the torque reference Tem2 has an
additional component Tem2p calculated through (16)
with a controller gain KC=−0.35. This loop produces
a Tem2p that is sinusoidal with no DC component
(Figure 4(d)) resembling the desired torque given in
(14) for ideal conditions. Tem2p produces fluctuations
in the flywheel's speed which varies between 1000
and 2000RPM around its idle speed (Figure 4(b)).
WSEAS TRANSACTIONS on POWER SYSTEMS
DOI: 10.37394/232016.2024.19.9
Marcelo G. Cendoya, Santiago A. Verne,
María I. Valla, Pedro E. Battaiotto
E-ISSN: 2224-350X
73
Volume 19, 2024
On the other hand, it can be observed that the
control loop for the idle speed of the flywheel is
indeed slow, presenting a very small gain at the
frequency of the turbulence (16.66mHz) since the
frequency of the pole f2 calculated using (18) is
1mHz. Therefore, due to its limited bandwidth, it
exerts minimal corrective action when facing rapid
speed changes, resulting in a small ripple in Tem20, as
shown in Figure 4(c). Since mechanical and
aerodynamic losses are not considered, the total
Tem2=Tem20+Tem2p constitutes flywheel acceleration
torque, Tjf in Figure 4(e).
Fig. 3: Effect of wind turbulence on turbine variables (values referred to shaft 1 of the EVT and compensation
disabled)
Fig. 4: Mechanical variables of the flywheel (shaft 2 of the EVT and compensation enabled from t=1025s)
WSEAS TRANSACTIONS on POWER SYSTEMS
DOI: 10.37394/232016.2024.19.9
Marcelo G. Cendoya, Santiago A. Verne,
María I. Valla, Pedro E. Battaiotto
E-ISSN: 2224-350X
74
Volume 19, 2024
Fig. 5: Mechanical variables on shaft 1 of the EVT (compensation enabled from t=1025s)
Fig. 6: Power at different points of the system (compensation enabled from t=1025s)
Figure 5 shows the effect on the variables
related to shaft 1 of the EVT when the proposed
power smoothing system is activated. It can be seen
in Figure 5(d), that the speed of shaft 1 (n1) stops
showing significant variations after t=1025s
because the variations in turbine torque caused by
turbulence (Figure 5(a), referred to as shaft 1 of the
EVT) are compensated by equal and opposite
variations in Tem2, as depicted in Figure 5(b). Since
the electromagnetic torque, Tem1 only depends on
the shaft 1 speed, it also ceases to undergo
significant variations after t=1025s, as illustrated in
Figure 5(c).
In Figure 6, the effect of the smoothing
controller on the power at various points of the
system is shown. When the controller is enabled
(starting from t=1025s), mechanical power is
transferred from shaft 1 to shaft 2 through the air
gap that exists between the internal face of rotor 1
and rotor 2. This power, P12 (Figure 6(b)) is the
product of Tem2 and the rotational speed of rotor 1.
As a result, the power in the stator of the EVT, Ps
(Figure 6(a)) becomes almost constant from that
moment onwards (with a value of around 60kW,
corresponding to the average wind speed of 10m/s),
as its variations are redirected towards rotor 2 of
the EVT. The product of Tem2 and shaft 2 speed n2
is the flywheel power Pf (Figure 6(c)). Pf is
composed of P12 and the power that is managed by
the VFD. Neglecting losses in the winding of the
inner rotor, the power contributed by the VFD PVFD
(Figure 6(d)) can be calculated as the product of
Tem2 and the speed difference between shaft 1 and
shaft 2. It is observed that PVFD has twice the
frequency of P12. Finally, in Figure 6(e), the power
delivered to the weak grid is shown. Before the
appearance of turbulence, 60kW corresponding to a
constant wind speed of 10m/s is injected. From
WSEAS TRANSACTIONS on POWER SYSTEMS
DOI: 10.37394/232016.2024.19.9
Marcelo G. Cendoya, Santiago A. Verne,
María I. Valla, Pedro E. Battaiotto
E-ISSN: 2224-350X
75
Volume 19, 2024
t=105s to t=1025s a power fluctuation due to wind
turbulence is observed, varying between 20kW to
100kW. This large disturbance in the power (40kW
peak value) is injected by the stator of the EVT
because the smoothing control system is disabled.
Upon enabling the control system, from t=1025s,
the initial mean power (60kW) remains unchanged,
but the power fluctuation is significantly reduced as
the observed new peak value is slightly above 7kW.
This corresponds to an attenuation factor of about 5
times, the value that has been adopted in the
flywheel sizing stage. This residual power
fluctuation is constituted only by the power
managed by the VFD, as variations in Ps disappear
because are transferred to the flywheel.
5 Conclusions
In this study, we propose a wind energy conversion
system that integrates, in a compact form, two
functions: power generation and power smoothing.
The system has been specifically designed for rural
areas in southern Patagonia, Argentina, where
challenging environmental conditions and abundant
wind resources are prevalent.
The most prominent features of the proposed
system are:
Owing to simplicity and robustness, a VAWT
Darrieus-H type has been selected as the wind-
to-mechanical energy conversion device.
A low-speed flywheel is employed as a short-
term energy storage device to smooth power
fluctuations caused by wind turbulences. This
type of flywheel does not require sophisticated
materials, special bearings, or a safety
container.
The flywheel is coupled to the VAWT through
an EVT.
The EVT integrates both the generator and the
motor to drive the flywheel in a single
electrical machine, resulting in a more
compact, simple, and economical system than
the traditional scheme based on separated
machines.
The chosen EVT is based on induction
machines, for reliability and robustness.
The way the EVT windings are powered and
connected is unusual. Only one VFD is
required, which is used to feed the inner rotor
of the EVT. The stator winding of the EVT is
directly connected to the grid. This constitutes
a simple and low-cost alternative.
As a significant fraction of the fluctuating
power (generated by the wind turbulence) is
transmitted through the electromagnetic
coupling between both rotors of the EVT, the
VFD-rated power is reduced in comparison
with a standard independent flywheel module.
The proposed control system is
straightforward and for its practical
implementation a device with low computing
power and the measurement of electrical
variables are only required. Any simple and
commercially available PLC (Programmable
Logic Controller) would be more than
adequate.
The results of the conducted simulations have
been encouraging, showing a significant reduction
of the fluctuation in the injected power into the
weak grid (around five times in the studied
example). Additionally, they validate the
theoretical analysis of the system, especially the
method for sizing the flywheel to achieve a desired
fluctuation attenuation factor.
In future work, an attempt will be made to find
an alternative control strategy to achieve a higher
level of attenuation of the grid power fluctuation
without the need to increase the size of the
flywheel. On the other hand, it is planned to use
controllers based on advanced techniques which
can enhance the system’s dynamic performance.
Acknowledgement:
The authors express their gratitude to UNLP,
CONICET, CIC, and ANPCyT, since without the
support of these institutions, this work would not
have been possible.
References:
[1] Q. Tang, J. Wu, J. Xiao, F. Zhou and X. Wu,
"A Case Study of Renewable Energy
Resources Assessment Results in Argentina",
2021 IEEE 4th International Electrical and
Energy Conference (CIEEC), Wuhan, China,
2021, pp. 1-5, DOI:
10.1109/CIEEC50170.2021.9510993.
[2] Ruschetti, C., Verucchi, C., Bossio, G.,
García, G. and Meira, M. (2019), “Design of
a wind turbine generator for rural
applications”. IET Electric Power
Applications, 13: 379-384, DOI: 10.1049/iet-
epa.2018.5734.
[3] Y. Y. Adajah, S. Thomas, M. S. Haruna and
S. O. Anaza, "Distributed Generation (DG):
A Review," 2021 1st International
Conference on Multidisciplinary Engineering
and Applied Science (ICMEAS), Abuja,
WSEAS TRANSACTIONS on POWER SYSTEMS
DOI: 10.37394/232016.2024.19.9
Marcelo G. Cendoya, Santiago A. Verne,
María I. Valla, Pedro E. Battaiotto
E-ISSN: 2224-350X
76
Volume 19, 2024
Nigeria, 2021, pp. 1-5, DOI:
10.1109/ICMEAS52683.2021.9692353.
[4] Johari, Muhd Khudri & Jalil, Muhamad &
Shariff, Mohammad. (2018). “Comparison of
horizontal axis wind turbine (HAWT) and
vertical axis wind turbine (VAWT)”.
International Journal of Engineering &
Technology, 7 (4.13). 74-80, DOI:
10.14419/ijet.v7i4.13.21333.
[5] Sun, J., Huang, D., “Impact of trailing edge
jet on the performance of a vertical axis wind
turbine”, Journal of Mechanical Science and
Technology, Vol. 37, No. 3, pp. 1301-1309,
2023, DOI: 10.1007/s12206-023-0216-0,
2023.
[6] Alessandro Bianchini, Giovanni Ferrara,
Lorenzo Ferrari, “Design guidelines for H-
Darrieus wind turbines: Optimization of the
annual energy yield”, Energy Conversion and
Management, Vol. 89, 2015, Pages 690-707,
ISSN: 0196-8904, DOI:
10.1016/j.enconman.2014.10.038.
[7] Xing Luo, Jihong Wang, Mark Dooner,
Jonathan Clarke, “Overview of current
development in electrical energy storage
technologies and the application potential in
power system operation”, Applied Energy,
Vol.137, pp. 511-536, 2015, DOI:
10.1016/j.apenergy.2014.09.081.
[8] M. Nadour, A. Essadki and T. Nasser,
"Power Smoothing Control of DFIG Based
Wind Turbine using Flywheel Energy
Storage System," 2020 International
Conference on Electrical and Information
Technologies (ICEIT), Rabat, Morocco,
2020, pp. 1-7, DOI:
10.1109/ICEIT48248.2020.9113213.
[9] Andrew Hutchinson, Daniel T. Gladwin,
“Optimisation of a wind power site through
utilisation of flywheel energy storage
technology”, Energy Reports, Volume 6,
Supplement 5, 2020, Pages 259-265, ISSN
2352-4847, DOI:
10.1016/j.egyr.2020.03.032.
[10] R. Sebastián, R. Peña-Alzola, “Control and
simulation of a flywheel energy storage for a
wind diesel power system”, International
Journal of Electrical Power & Energy
Systems, Volume 64, 2015, Pages 1049-1056,
ISSN: 0142-0615, DOI:
10.1016/j.ijepes.2014.08.017.
[11] Hoeijmakers, M.J. and Ferreira J.A., ‘‘The
Electric Variable Transmission’’, IEEE
Trans. on Ind. Appl., Vol. 42, no. 4, 2006, pp.
1092–1100, DOI: 10.1109/TIA.2006.877736.
[12] Q. Xu, F. Wang, X. Zhang and S. Cui,
"Research on the Efficiency Optimization
Control of the Regenerative Braking System
of Hybrid Electrical Vehicle Based on
Electrical Variable Transmission," in IEEE
Access, vol. 7, pp. 116823-116834, 2019,
DOI: 10.1109/ACCESS.2019.2936370.
[13] Y. Zhu and C. Cai, "Current Source
Converter Based Control Strategies for the
DPF-WECS", 21st International Conference
on Electrical Machines and Systems
(ICEMS), Jeju, Korea (South), 2018, pp.
1105-1109, DOI:
10.23919/ICEMS.2018.8549187.
[14] X. Sun, M. Cheng, Y. Zhu and L. Xu,
"Application of Electrical Variable
Transmission in Wind Power Generation
System," in IEEE Transactions on Industry
Applications, vol. 49, no. 3, pp. 1299-1307,
May-June 2013, DOI:
10.1109/TIA.2013.2253079.
[15] Marcelo G. Cendoya, Santiago A. Verne,
Pedro E. Battaiotto, "Novel Pumping System
based on Micro-Hydro Turbine and
Centrifugal Pump coupled using EVT
Induction Machine. Application in Rural
Area Irrigation," International Journal of
Electrical Engineering and Computer
Science, vol. 5, pp. 22-32, 2023, DOI:
10.37394/232027.2023.5.4.
[16] Chen, JS, Chen, Z, Biswas, S, Miau, J, &
Hsieh, C. "Torque and Power Coefficients of
a Vertical Axis Wind Turbine with Optimal
Pitch Control." Proceedings of the ASME
2010 Power Conference. Chicago, Illinois,
USA. July 13–15, 2010. pp. 655-662, DOI:
10.1115/POWER2010-27224.
[17] B. K. Bose, Modern Power Electronics and
AC Drives 1st Edition, Pearson Education
India, January 2015, ISBN-10: 9332557551.
[18] Burton, T., Jenkins, N., Bossanyi, E., Sharpe,
D. & Graham, M., Wind Energy Handbook
3rd Ed. John Wiley & Sons Ltd (2021),
ISBN: 978-1-119-45109-9.
[19] Wang, X.; Liu, Y.; Wang, L.; Ding, L.; Hu,
H. “Numerical Study of Nacelle Wind Speed
Characteristics of a Horizontal Axis Wind
Turbine under Time-Varying Flow”.
Energies 2019, 12, 3993, DOI:
10.3390/en12203993.
[20] Lianjun Zhou, Minghui Yin, Xuekun Sun,
Dandan Song, “Maximum power point
tracking control of wind turbines based on
equivalent sinusoidal wind”, Electric Power
Systems Research, Vol. 223, 2023, 109534,
WSEAS TRANSACTIONS on POWER SYSTEMS
DOI: 10.37394/232016.2024.19.9
Marcelo G. Cendoya, Santiago A. Verne,
María I. Valla, Pedro E. Battaiotto
E-ISSN: 2224-350X
77
Volume 19, 2024
ISSN 0378-7796, DOI:
10.1016/j.epsr.2023.109534.
Contribution of Individual Authors to the
Creation of a Scientific Article (Ghostwriting
Policy)
- Marcelo G. Cendoya has carried out the
theoretical analysis of the system. He has
developed the mathematical model and has
carried out the computer simulations. He has
participated in the conception of the system
topology. He has collaborated with the writing
and revision of the manuscript.
- Santiago Verne has conducted the research on
potential advantages of using vertical axis wind
turbines in southern Patagonia, Argentina. He has
also collaborated in the study of the EVT
machine in wind energy conversion applications.
He has participated in the conception of the
system topology. He did the writing and editing
of the manuscript. He has made the figures.
- María Inés Valla has organized and coordinated
the tasks that each member of the research group
had to perform. She has also carried out the
revision and correction of the manuscript.
- Pedro E. Battaiotto has collaborated in the
conception of the system topology and the
analysis of the EVT operation inside the system.
He has also collaborated in the development of
the system control scheme.
Sources of Funding for Research Presented in a
Scientific Article or Scientific Article Itself
UNLP Proyecto 11/I255 “Electrónica de Potencia
y Sistemas de Control Avanzado Aplicados a
Fuentes de Energía Alternativas”. 1/2020 -
12/2023.
CONICET PIP 112-2020-0102801CO “Control
Avanzado y Electrónica de Potencia Aplicados a la
Optimización de Sistemas Basados en Energías No
Convencionales”. 12/2021 – 12/2024.
ANPCyT PICT N°2018-03747 “Control,
Conflict of Interest
The authors have no conflicts of interest to declare.
Creative Commons Attribution License 4.0
(Attribution 4.0 International, CC BY 4.0)
This article is published under the terms of the
Creative Commons Attribution License 4.0
https://creativecommons.org/licenses/by/4.0/deed.e
n_US.
WSEAS TRANSACTIONS on POWER SYSTEMS
DOI: 10.37394/232016.2024.19.9
Marcelo G. Cendoya, Santiago A. Verne,
María I. Valla, Pedro E. Battaiotto
E-ISSN: 2224-350X
78
Volume 19, 2024
