Energy Storage and Saving 3 (2024) 295–304

Contents lists available at ScienceDirect

Energy Storage and Saving

journal homepage: www.keaipublishing.com/en/journals/energy-storage-and-saving/

Research article

A hybrid fuzzy logic-based MPPT algorithm for PMSG-based variable speed

wind energy conversion system on a smart grid
S. Karthikeyan a,*, C. Ramakrishnan b
a
Department of ICE, SNS College of Technology (Autonomous), Coimbatore, 641035, India
b
Department of EEE, SNS College of Technology (Autonomous), Coimbatore, 641035, India

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

Keywords: Recently, wind power has gained popularity as a sustainable energy source. Wind energy conversion systems
Fuzzy logic controller (FLC) (WECSs) can accept fixed speed and variable speed (VS) operations. VS-WECSs are preferable to conventional
Maximum power point tracking (MPPT) WECSs because of their higher electricity collection capacity. Maximum power point tracking (MPPT) systems
Permanent magnet synchronous generator
are essential for maximizing the efficiency of wind energy generation in wind turbine (WT) installations linked to
(PMSG)
power grids. This study introduces a hybrid fuzzy logic controller-based MPPT (FLC-MPPT) for WTs connected to
Wind energy conversion system (WECS)
permanent magnet synchronous generators (PMSGs) to accurately determine the maximum power output of
WTs. This study employs a three-phase back-to-back converter to link a PMSG to a utility grid. The reference
signals for pulse width modulation controllers comprise two-phase system currents. It constructs a converter that
can transfer electrical energy in both directions using insulated-gate bipolar transistor technology and is powered
by a battery. The machine-side converter uses model predictive control for the present control loop. Given the
generator’s susceptibility to changes in wind conditions, this factor is of the utmost importance. A WT simulation
was conducted using MATLAB/Simulink and an FLC methodology was employed. The model used a PMSG.
Measurements of rotor speed, power, induced voltage, and current were taken in relation to variations in wind
speed. The simulation results show that the FLC-MPPT can maximize the output power over a wide range of wind
speeds with a higher efficiency of 92.6% and performance of 95.7%.

1. Introduction driving. These benefits have boosted their usefulness in large-scale

offshore applications. WECSs require an effective MPP tracking
Wind generation systems (WGSs) are the most common wind-energy (MPPT) algorithm to maximize the wind power [5,6].
conversion systems. These systems can employ variable-speed genera- Maintaining a steady voltage throughout the load using MPPT can
tors. Highly efficient wind turbines (WTs) can adapt their speeds to wind maximize wind power. The Perturb-and-Observe (P&O) MPPT approach
conditions. A power converter is required to monitor the energy flow is popular owing to its simplicity and efficacy [7]. However, the P&O
and track the maximum power point (MPP) changeover in this turbine. approach produces poor power output because it creates significant
WTs with changing rotation speeds can adjust to wind velocity [1,2]. fluctuations and fails to follow the greatest power point owing to the
Wind energy conversion system (WECS) make use of permanent magnet high nonlinearity of the wind speed. The proportional integral deriva-
synchronous generator (PMSG) has been quickly adopted by large WT tive (PID) control approach is another popular control technique [8].
generators in variable-speed WGSs. Advances have been made in this The PID control is easy to implement. Because its settings are arbitrarily
field. WECSs optimize wind energy production at various velocities [3]. selected, the PID control reduces the overall system efficiency. The fuzzy
This study maximizes the power output within the prescribed voltage logic controller (FLC) is used to circumvent this issue and maximize
range by monitoring the MPPs at varying wind speeds and modifying the wind power extraction [9]. The FLC can monitor the nonlinearity of the
rotational speed of a WECS using a variable speed controller. This im- system and provide the highest possible output based on the available
proves grid support, fault management, performance, operational costs, wind. Unlike the direct current (DC)-DC converter, which produces the
and energy density [4]. PMSG-based WECS is more efficient than direct duty cycle as an output, the FLC uses the DC voltage and current flowing

Peer review under responsibility of Xi’an Jiaotong University.

* Corresponding author.
E-mail address: skarthigp@gmail.com (S. Karthikeyan).

https://doi.org/10.1016/j.enss.2024.08.001
Received 2 April 2024; Received in revised form 17 July 2024; Accepted 23 August 2024
Available online 26 September 2024
2772-6835/© 2024 The Authors. Published by Elsevier B.V. on behalf of KeAi Communications Co. Ltd. This is an open access article under the CC BY-NC-ND
license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
S. Karthikeyan and C. Ramakrishnan Energy Storage and Saving 3 (2024) 295–304

through the load as inputs [10]. recorded a peak value of 0.6 radian⋅s− 1, the system delivered a minimal
A doubly-fed induction generator (DFIG) is the most robust option steady-state error of roughly 0.01 radian⋅s− 1 upon parameter
for WECS in the commercial sector. Nevertheless, PMSGs are increas- modification.
ingly used in WECSs because of their superior power density, efficiency, The ant colony optimization (ACO) was introduced by Mrabet et al.
and dependability [11]. Wind power can be connected to a power grid [22] to maximize the WECS power. Five performance indicators (inte-
through a power electronic interface. The grid can be supplied with grated squared error, integrated absolute error, integrated time absolute
wind energy using a back-to-back converter comprising a machine-side error, integrated time squared error, and mean squared error) were used
converter (MSC) and grid-side converter (GSC) [12]. An efficient MSC in the ACO technique to select the best settings (Kp, Ki, and Kd). The
controller allows the greatest extraction of electricity at different wind results demonstrate the power produced, DC bus voltage, electromag-
speeds. Model predictive control (MPC) and FLC are two examples of netic torque, and currents, and the ability of the WT to effectively
sophisticated control methods that can improve the WECS performance monitor the required MPP.
[13]. The development of digital microcontrollers has led to an increase Kumar and Vijayakumar [23] presented a wind-energy conversion
in the use of MPC to control power electronic converters. Although MPC system-based DC nanogrid. The WECS architecture implements MPPT
is computationally intensive, its primary benefit is that it can handle by integrating a diode bridge rectifier (DBR) and a DC-DC converter. A
systems with nonlinearities and many variables [14]. cost-effective and durable self-excited induction generator powers this
However, the generators are not the only factor influencing WT design. The dynamic performance of the proposed DC nanogrid topology
performance; the pulse width modulation (PWM) converters linked to was investigated to understand the selection process of a DC-DC con-
the generators and how they are managed also play a role [15]. MSC is verter with a DBR for various DC grid voltage situations.
used to describe the converter and it is attached to the electric machine. Shutari et al. [24] aimed to optimize the performance of
GSC is used to describe the converter and it is also connected to the grid. variable-speed WTs (VSWTs) by presenting a new version of the MPPT
The MSC controls the electric machines, whereas the GSC integrates the algorithm that uses parabolic prediction methods (PPTs). The
WTs into the grid. A decoupled d-q control technique using a PPT-MPPT method estimates the parabolic WT output curve and de-
voltage-regulated PWM scheme is the primary control strategy of a GSC termines its MPP using parabolic convex functions. According to the
and is extensively utilized in wind power and various other renewable simulation results, the proposed PPT-MPPT algorithm improved the
energy systems [16,17]. VSWT system more effectively than the alternatives in terms of tracking
A buck converter connected the wind generator to the DC load. The efficiency and extracted power.
PMSG produces alternating current (AC), which is transformed into DC Jain et al. [25] proposed a grid-integrated power converter using a
using a diode-controlled rectifier. This process eliminates the ripple WECS coupled to a variable-speed PMSG. Proportional-integral (PI)
inherent in the AC component. To further reduce the ripple caused by controller-based feedback control designs a controlled DC-DC converter.
nonlinearity, a smoothing capacitor is placed across the rectifier [18]. Monitoring the output voltage and adjusting the switch duty ratio
The FLC monitors the voltage and current outputs to create an optimal maintains the DC-link voltage, a grid-connected variable load receives
duty cycle for the functioning of the converter when wind power fluc- the inverter output via PWM.
tuates. Therefore, the available wind speed determines the maximum A squirrel-cage induction generator (SCIG)-based variable-speed
power generated [19]. The optimal method for controlling the MPPT in WECS was modeled and studied in the context of fault circumstances
a standalone WECS was examined in this study. This study compares the [26]. A dual current controller was designed to independently manage
performance of the traditional PID controller and the FLC when the wind the positive and negative sequence currents to avoid double-frequency
speed varies. One advantage of the proposed control technique is its oscillations in the DC-link voltage owing to asymmetrical faults. This
enhanced capacity to harness wind power to its full potential. The effi- controller limited the fault current to 1.5 pu and increased the DC-link
ciencies of the controllers can be compared by examining the converter’s voltage and generator speed to 1.2 pu.
output power [20]. A WECS with efficient robust adaptive perturb and observe MPPT
The main contributions of the study are as follows: algorithms is described in Ref. [27]. It calculates the adaptive ratio at
each operational point, monitors the power fluctuations, perturbs the
(a) This study provides a hybridized FLC and an MPPT of WESs from rotor speed with an appropriate step size, and compares the actual and
WTs and PMSGs. ideal speeds. It calculates the relationship between the extracted me-
(b) A back-to-back converter with an MSC and GSC converts wind chanical power, rotor speed, perturbing step size, and operational wind
energy into a power grid via a PMSG and a WT. speed at high-inertia WTs to eliminate tracking and direction errors
(c) Next, MPC, which relies on the generator inertia energy, was used during rapid.Table 1 [21–27] shows the advantages and limitations of
to improve the WECS voltage ride-through capacity. the existing work.
(d) The proposed hybridized FLC-MPPT technique improves the
WECS efficiency and monitors the MPP quickly with reduced 3. Proposed methodology
constant-state fluctuations for high-resistance PMSGs compared
with existing PID controllers. Based on the speed control requirements, two primary types of
WECSs can be identified: fixed-speed and variable-speed turbines.
This framework serves as the foundation for future research. The However, much interest has been piqued in variable-speed WECS that
WECS is the primary focus of Section 1. Section 2 reviews the relevant are operable in the MPPT mode. There are two main types of drive
literature. Section 3 describes the proposed fuzzy-based MPPT turbines: direct and indirect turbines. PMSGs are a practical choice when
controller. Section 4 presents the experimental results. Section 5 sum- a large-scale power converter system is required to connect a generator
marizes prospective actions and future perspectives of this research. to a grid or load.
Fig. 1 shows the schematic of the three-phase WECS. A fully
2. Related survey controlled BTBC interface with a shared DC connection was used to
inject the collected wind power into the grid. The BTBC incorporates an
The fuzzy fractional-order robust control (FFORC) was proposed by MSC to ensure that the WT is harnessed to its full potential. To manage
Ullah et al. [21] for wind energy conversion. A fuzzy structure performs the DC link voltage and maintain unity power factor operation, a GSC
a dynamic correction of the intermittently controlled strength to avoid was also incorporated.
buzzing occurrences in the stimulation signal and to keep the In WTs, the most common types of variable-speed applications
closed-loop approach resilient. Although the saturation function include DFIGs, PMSGs, or direct-driven PMSGs. Each WT uses an AC/

296
S. Karthikeyan and C. Ramakrishnan Energy Storage and Saving 3 (2024) 295–304

Table 1 WECS. Efficient irregular controllers may be more effective than PID
Inference from existing work. controllers in handling unknown and external impacts owing to the
Ref. Method name Advantages Limitations nonlinear characteristics of the WECS.
The authors suggested using a PMSG-based PWM to power a grid-
[21] Fuzzy fractional The data indicate that The complexity of the
order robust FFORC provides minimal provided controller is connected WECS. It is thought that wind generator control approaches
control (FFORC) chattering. further increased will switch to MPPT instead of PID controllers. PWM provides respon-
because the stated siveness, stability, and dynamic performance; however, it causes sub-
control strategy stantial fluctuations in electromechanical power, nonlinear stator up-to-
necessitates greater
state vector estimation.
date direct grid current, and injected reactive and active energy. The
[22] Ant colony The simulation results MOACO solutions must performance of the MPPT controller is assessed using simulations.
optimization show that the WT may be durable and Despite specific issues such as imprecise tracking, it was discovered that
(ACO) effectively follow the dependable under it worked effectively with fluctuating dynamic responses and wind
required MPP. different situations to
conditions. Real-time implementation has become more complicated
work consistently.
[23] Diode bridge The capacitor bank Integrating converters owing to lengthy training procedures.
rectifier (DBR) parallel to the generator with differing dynamic
supplies SCIG reactive properties might make 3.1. Turbine model
power. system stability
difficult.
[24] Parabolic To achieve ongoing Maintenance and
The generated power Kn and torque of the turbine Fn can be calcu-
prediction methods convergence, the monitoring may be lated using Eqs. (1) and (2), respectively.
(PPT) algorithm changes the needed to keep the
concavity and optimum system running
1
Kn = ρπQ2 b3 Ek (λ) (1)
sections of the anticipated smoothly, increasing 2
parabola curve operational
systematically. complexity. Kn 1
Fn = = ρπ Q3 D2 Ek (2)
[25] Proportional- The PMSG’s variable If the power produced Ω 2λ
integral PI power output is rectified is less than the three-
controller-based by a three-phase diode phase load needs, the The variable ρ is the ambient air volume, Q is the extent spanned by
feedback control rectifier and a closed loop additional power is the rotor blades, and b is the pitch angle, which is defined in Eqs. (1) and
DC-DC converter before absorbed from the grid. (2). λ represents the turbine’s angle of rotation for the wind speed D in
synchronizing with the
Eq. (3):
grid.
[26] Dual current The system must manage DC link voltage and ΩQ
controller (DCC) WECS or communication generator speed may λ= (3)
D
connection failures increase no more than
without affecting 1.2 pu; however, fault Ek , the turbine’s aerodynamic effectiveness is a function of the spe-
performance. current cannot exceed cific speed Ω and the blade angle. It varies from one turbine to another
1.5 pu.
[27] Robust adaptive Among DPC methods, The algorithm must be
and is typically provided by the manufacturer, which may be roughly
perturb and sensorless P&O MPPT system-fault-resistant. characterized mathematically.
observe (RA-PO) algorithm tracks optimal
rotational speed by being 3.2. Generator model
more responsive to rotor
speed and power
fluctuations. Induction generators (IGs) are the generators of choice for wind-
energy conversion. Dvr , Dsr , Dvl , Dsl and xvr , xsr , xvs , xsl refer to the d-q
Note: WT: wind turbine; MPP: maximum power point; MPPT: MPP tracking;
axis machine voltages, and the mechanics constitute the dynamic model
SCIG: squirrel-cage induction generator; PMSG: permanent magnet synchronous
generator; WECS: wind energy conversion systems; DC: direct current; DPC:
of the IG, also known as the Park model. Eqs. (4)− (7) for electricity are
direct power controller; P&O: perturb-and-observe. v
Dvr = Qr xvr + ϕ − ωr ϕsr (4)
vt vr
DC/AC PWM frequency converter for control and grid connections. In a
v
PMSG WT, the converter may be connected to either generator or grid Dsr = Qr xsr + ϕ − ωr ϕvr (5)
vt sr
(Fig. 1). Two self-commutated pulse width modulators on the machine
side and one on the grid side comprise a standard frequency converter v
Dvl = Qs xvs + ϕ − ωs ϕsl (6)
with a DC voltage connection in the middle. MSCs and GSCs can work as vt vs
efficiently as feasible by creating a DC connection with the central
v
capacitor and then separating their duties. Modern designs of variable- Dsl = Qr xsl + ϕ − ωr ϕvl (7)
vt sl
speed WTs use decoupled d-q vector control techniques to indepen-
dently operate two PWM converters, which are stacked on top of each Here, ϕvr , ϕsr , ϕvl , ϕsl are the flux linkages for Eqs. (8)− (11),
other. In most cases, the direct-axis current component of a converter is respectively:
used to achieve a single control goal. By contrast, the quadrature axis
ϕvr = Ar xvr + Pqs xvl (8)
current component was used to achieve another control goal.
For PMSGs with full-capacity power converters, an indirect drive ϕsr = Ar xsr + Pqs xvl (9)
turbine is that uses a gearbox to adjust the low-speed turbine to the high-
speed generator is available. Because of their direct coupling with tur- ϕvl = Al xvl + Pqs xvr (10)
bines, small-scale WTs that operate PMSG have recently become
increasingly popular. PMSG have high power density and are light- ϕsl = Al xsl + Pqs xsr (11)
weight, efficient, and self-excited. A grid-connected WECS with a vari-
able speed was introduced in this study. A PMSG is primarily generated where Ar and Al are the base electrical frequencies and Pqs is the
by a WT, which is the major component. A power electronics intercon- externally applied mechanical torque. Here, S vtv Ω is a representation of
nection and resistor and inductor (RL) filter link the electrical grid to the the mechanical equation for IG and electromagnetic torque, Fw is the

297
S. Karthikeyan and C. Ramakrishnan Energy Storage and Saving 3 (2024) 295–304

Fig. 1. Fuzzy logic controller (FLC) & maximum power point tracking (MPPT)-based wind energy conversion system (WECS) with permanent magnet synchronous
generators (PMSGs). AC: alternating current; DC: direct current; PWM: pulse width modulation; RL: resistor and inductor.

angular frequency of rotor, and VΩ is the angular velocity of reference controls the motor speed, direct stator current, polynomial rotor elec-
frame, as shown in Eqs. (12) and (13): tricity, and electrostatic torque using the FLC-MPPT. Conventional
v electrical torque is generated as an input using the MPPT method. In
S Ω = Fw − Fq − VΩ (12) addition, the amounts of reactive and active power transmitted to the
vt
grid are controlled by the FLC-MPPT of the GSC. Voltage and frequency
Fq = k(ϕvr xsr − ϕsr xvr ) (13) can also be modified to ensure that the power supply is sufficiently
secure.
The WT must operate on the MPP for the WECS to operate efficiently.
3.3. Principles of fuzzy logic Maximizing power production from the WT is the goal of MSC. Hence,
the functioning of the MPP was ensured by applying intelligent control
Fuzzy controls include the intuition and expertise of human opera- to the MSC. The WT output power can be calculated based on the power
tors (and sometimes designers and researchers). R, the inference rela- coefficient Cp. In addition, the rotor speed determines the Cp, which is
tion, is derived from a knowledge base consisting of a database and dependent on the tip speed ratio λ. To attain the best possible λ, which
rules. The input and output variables are described in the database using results in the greatest power coefficient, the MSC is adjusted so that the
fuzzy sets. In their simplest form, the rules are based on a control method rotor speed follows a reference speed that has been computed.
for the overall system. Fuzzy conditional assertions included in profes- Two control loops comprise the control system of the MSC. MPC
sional expertise or IF-THEN rules are familiar sources from which they allows the outer loop to act as a speed controller and the inner loop as a
are produced. current controller. To obtain the q-axis current reference value, the
M(x) : If i1 is G1 and i2 is G2 …and ix is Gx THEN T is F(x) , x = 1, …, n speed controller subtracts the actual rotor speed from the ideal rotor
speed (Fig. 2). The following formulas for MPC can be obtained by
(14) sx
referring to Eqs. (15) and (16) and solving for stsf :
Assuming that i1 , i2 , …ix are all elements of Gx , then M(x) holds. If x =
1, …, n, the array of input parameters is i1 , i2 , …ix , the value that is not sxsf Wf He 1
=− xsf + φf xef + Qsf (15)
changed is T, the set of guidelines is n, the number of personal factors is st Hs Hs Hs
x, and the fuzzy sets are (G1 , G2 , … Gx ) in Eq. (14). A fuzzy controller
takes a control system rule base as input, uses the input signal condition
to decide which rules to fire, and determines the successful control ac-
tions and fuzzy variables produced.
Additionally, even when dealing with significant levels of model
uncertainty, PID controllers combined with fuzzy logic technology can
provide high-quality tracking. To counteract the harmonic output of the
nonlinear system, the authors investigated active grid-interfacing in-
verters with PI linear and FLCs. Fuzzification mitigates the irregularities
and perturbations in several ways; representing complicated, nonlinear
input-output interactions may help. A control system can identify non-
linearities that standard approaches cannot. Fuzzy logic can handle
imprecise or inadequate information. Uncertainties and disruptions may
cause unexpected behavior in conventional control systems. Fuzzy logic
uses fuzzy sets to characterize the input and output variables and the
reason for the degrees of membership to overcome uncertainty. Fuzzy
logic allows the controller to adapt to changing operating conditions in
real time, thereby ensuring system stability and performance. The FLC-
MPPT controls motor speed, direct stator current, polynomial rotor
electricity, and electrostatic torque in the PMSG-side converter. MPPT
generates conventional electrical torque. Voltage and frequency can also
be adjusted to safeguard the power supply. The PMSG-side converter Fig. 2. Model predictive control (MPC) scheme for electrical power inverters.

298
S. Karthikeyan and C. Ramakrishnan Energy Storage and Saving 3 (2024) 295–304

sxef Wf Hs ψ 1 constitute the DC/AC conversion are interfaced with a bidirectional

=− xef − φ xsf − φ + Qef (16)
st He He f He f He converter. This helps transform the electricity generated by the solar PV
array into an alternative current that can be sent back into the grid. The
A model in discrete time by discretizing Eqs. (15) and (16) using an
integrated circuit tailored to the requirements of the control unit was a
Rf sampling time sampler, and then the Euler approximation. The for-
field-programmable gate array. Fig. 3 shows the correlation between the
ward Euler approximation was substituted for (sx ), resulting in Eq. (17).
st DC-link voltage and the current.
sx x(g + 1) − x(g) vzvc
= (17) xck = Cvc = xvcf − xvc (19)
st Rf vt
Fig. 2 depicts the overarching structure of converting power elec- Eq. (19) represents the relationship among the inductance of the
tronics using MPC. For each of the n changes in condition, electricity, capacitor in a DC link (Cvc ), DC link power (xvc ), energy (xvcf ) moving via
power, and actuation, the model uses the observed variables D(L) to the DC connection from the generator-side adapter, and electricity (xvc )
make predictions about the controlled variables D(L
̃ + 1). The best entering the system. If the wind speed is less than the PM generator’s
activation E, for the conversion is then selected after evaluation using a maximum, it automatically switches to the MPPT mode. The next step in
cost estimator that considers the reference values (D* (L +1)) and design developing a well-thought-out system for managing an engine is to
limitations. The standard structure of cost variable r under reduction employ a simple model for engine-side converters. The DC connection
constraints is shown in Eq. (18). received electrical power that was directly proportional to the squared
[ ] [ ] wind speed.
r= D ̃ 1 (L + 1) − D* (L + 1) + λ1 D2 (L + 1) − D* (L + 1)
1 2
[ ] (18) Ksc = zvc xvcf = pυ3 (20)
+ … + λn D ̃ n (L + 1) − D* (L + 1)
n

where p is the WT system-related factor in Eq. (20). The injection current

where λ represents the relative importance of each objective, each po-
xvcf can be represented as
tential switching state is evaluated for its impact on the cost function r,
and the state with the lowest value is retained for future use. Fig. 3 il- pυ3
xvcf = (21)
lustrates the power converter, which is applicable to any configuration zvc
and phase count. The generic load depicted can represent a powered
In Eq. (21), the input current of the DC connection, xvcf , affects the
device, power grid, or an active or passive load.
voltage variations in the DC link. The back-to-back PWM converter is a
The components that comprise a WECS are listed below: the WT, IG,
crucial component of a WECS because it transfers power from the
semiconductor rectifier gadgets, filtering, and PWM-driven converter.
generator to the grid. The resulting energy of the grid-side conversion
The power source was generated from the physical power of a WT using
should match the input power of the engine, assuming that the DC-link
a three-phase incentive motor. Electricity was sent directly to the elec-
voltage remains constant and excluding losses in the electronic devices.
trical system. The process of controlling the rotor voltages, which in turn
regulates the amount of energy extracted and sent to the grid, is shown Ksc = Kfc (22)
in Fig. 3. An inverter based on PWM was proposed to maintain photo-
voltaic (PV) and grid systems in an appropriate balance. The DC bus where Eq. (22) provides the instantaneous power of the generator-side
receives the output power from the boost converter and is immediately converter and the GSC, which are denoted as Ksc and Kfc , respectively.
linked to the load that manages the DC power. The DC variables that

Fig. 3. Permanent magnet synchronous generator (PMSG)-based wind system with fuzzy logic controller (FLC) for wind turbines. PWM: pulse width modulation;
REG: renewable energy generation.

299
S. Karthikeyan and C. Ramakrishnan Energy Storage and Saving 3 (2024) 295–304

The electricity transferred from the engine to the network converters is creation of the fundamental boost converter.
affected in real time by the randomness of the wind speed of the WT Blade angle controllers are the most advanced and reliable control-
takes in nature. The DC-link capacitor is then used as a power buffer, and lers for the mechanical power conversion stages. The controller changed
its instantaneous power is supplied using Eq. (23). the blade angle to manage the aerodynamic power of the turbine system.
vzvc MPPT involves operating the blade angle controller to maintain the
Kc = zvc xck = Cvvc (23) power at its rated value, even during strong wind gusts. The pitch angles
vt
of WTs are controlled using various approaches, one of which is the FLC
where xck denotes the current passing through the capacitor linked to algorithm.
DC. Eq. (23) shows that the current passing through the DC-link Nevertheless, the operation of the topology can be modified to
capacitor depends on the variation in the momentary energy. The function as a converter in one or two quadrants. These converters are
power flowing between the generator-side converters and GSCs causes projected to operate continuously in the conduction mode. Using an FLC
an uneven flow of energy during dynamic WECS regulation, which in ensures that the smart grid (SG) gets power properly and securely
turn causes the DC-link voltage to vary. The current balance can be (Fig. 4). MPPT increases the regulation of the charges and discharges.
calculated using Eq. (24): Finally, MPPT helps obtain maximum juice from the SG system. A
possible major increase in system utilization can result from the MPPT
pυ3 vzvc
xvc = xvcf − xck = − Cvc (24) controller. There are several potential real-life applications.
zvc vt A DC-DC converter is essential for maximum power factor correction
When Eqs. (20) and (21) are used, the following expression is ob- (MPPT) systems. The example presented in this study uses a DC-DC
tained for a connection that connects the DC voltage link of the con- converter to control the DC output voltage from an unregulated DC
verter and its AC-side current: input. Using MPPT algorithms, the duty cycle was controlled such that
the WT model operated at the MPP. The parasitic resistance of an
pυ3 vzvc inductor increases to produce ohmic losses, and this sum is represented
xa PA + xb PB + xc PC = − C (25)
zvc vt as the series resistance. In the continuous-conduction mode of a boost
Eq. (25) shows that DC connection voltage oscillations may influence converter, there are two states in each commutation cycle. It can operate
the grid-side electricity waveforms. When the FLC-MPPT runs, the DC- by turning the diode and circuit on and off. When the switch was turned
link voltage ripples are more significant because the wind turbulence on, the voltage across the diode became negative, causing the switch to
causes dynamic power changes. Therefore, the DC-link voltage man- turn off. The diode continued to function even in the absence of elec-
agement approach has become an essential concern for the WECS, as it tricity. This operation is controlled by the duty-cycle signal, which is a
will lead to a decline in grid power quality. WECS has recently used PWM, whose normalized values may be anywhere from 0 to 1.
several DC-link voltage management techniques. The following sections The FLC follows the MPP to set the generator reference speed when
detail and demonstrate DC-link voltage control techniques based on the wind speed changes. The FLC estimates the power output changes
fuzzy logic using the MATLAB/Simulink environment. In the proposed due to changes in the generator reference speed. The search continued
hybridized fuzzy-MPPT control algorithm, the tracing approach was run until the output power increased in each step. Even when the search
concurrently with the wind sources to maximize the power extraction. direction is inverted, the output power decreases with speed. An FLC is
Fig. 4 shows the model of the WT system connected to the grid and the invaluable for monitoring MPPs under unexpected or fast-changing

Fig. 4. Fuzzy logic controller (FLC)-maximum power point tracking (MPPT) for wind energy conversion system with permanent magnet synchronous generator
(PMSG). MPC: model predictive control; PWM: pulse width modulation; DC: direct current.

300
S. Karthikeyan and C. Ramakrishnan Energy Storage and Saving 3 (2024) 295–304

wind conditions. In response to wind speed and direction, the controller speed. Generators produce most power at their rated wind speeds [28].
determines the ideal setting and applies the maximum force. The addi- To avoid harm, the power production stops at the cut-out wind speed.
tion of the preceding speed instruction to the current reference speed Manufacturers provide power curves for ideal weather and terrain
creates a new reference speed by merging the control input signal var- conditions. Turbine location, air density, wind velocity dispersion, wind
iables. To obtain maximum power, it is necessary to maintain the WT direction, mechanical and control issues, and measurement inaccuracies
torque at its optimal value; measuring the wind speed is unnecessary. may affect the power curves. WTs seldom operate perfectly.

B*y = Nopt Ω2r (26)

Eq. (26) shows that the value of the WT reference torque is repre- 4.1. DC-DC converter ’s output based on PID & FLC
sented by B* . Eq. (27) was used to estimate the WT torque ( B
̂ y ).
y
Fig. 5 shows that the WT was exposed to an irradiance profile over
Vref time, and its current and voltage properties dictated its active load
(27) power production. There is perfect synchronization between the d-q-axis
̂ y = Vz Kw
B load current references, changes in the grid active and reactive powers,
Ωr
DC voltage input to the inverter, and load voltage and current injected
where the rectifier’s output voltage (Vz ) and current (Kw ) are defined. into the utility grid from the WT system at the connection point.
The battery charging start time μ allows the battery’s system-on-chip Using the given wind-speed data, Fig. 5 shows the voltage, current,
√̅̅̅̅̅̅ and power waveforms generated by the wind system. The wind system
(SoC) σ 2π to determine the required power. To determine the SoC,
and the voltage, current, and power output of the DC/DC boost
one must be aware of the average distance driven per minute. Eq. (28)
can be used to calculate the standard distance between the grids.

1 − (ln(h)− μ)2
e(h) = √̅̅̅̅̅̅r 2σ 2 h>0 (28)
hσ 2π
Eq. (29) makes it easy to compute the battery SoC using h.
⎧
⎨ 0
h > ED
SoC = ED − h (29)
⎩ × 100% h ≤ ED
ED

ED represents the entire electrical spectrum and is the threshold beyond

which the power will not start. Eq. (30) can forecast the charging
duration if the SoC and battery powers are known.
Ebat × (1 − SoC) × SOS
gh = (30)
ηe × M
The charge type (M) and its efficacy (h) affect the charging time (gh ),
as shown in Eq. (30). Therefore, MPPT is a reasonable choice for con-
trolling and storing surplus power. In recent decades, innovative
methods, including adaptive controller fuzzy logic, have been used to
regulate the angle of WT blades. The ever-changing wind speed profiles
affect several elements, including the generator rotor speed, mechanical
torque, and wind system dynamics; hence, fuzzy logic is considered the
best way to control the blade angles of WTs. Wind system errors were
transformed into fuzzy codes, which were then used to construct lan-
guage variables. These errors include numerical inaccuracies and rotor
speed issues. One advantage of fuzzy controllers is that fuzzy linguistic
variables can learn new methods to track parameter changes in WT
systems. This allows them to adapt to the changing conditions. Because
of this benefit, FLCs are well suited for handling unpredictable and rapid
changes in wind-speed profiles that occur throughout the year. To
demonstrate the uniqueness of the fuzzy controllers, the authors
extensively compared them with the conventional PID controller.

4. Results and discussions

This study primarily focuses on strategies for arranging the input and
output characteristics of an FLC. The PMSG was directly connected to
the modeled 25 kW WT. Because the direct connection does not require a
gearbox, the mechanical rotational speed of the PMSG is equal to the WT
rotor speed. Building a grid-linked system requires a back-to-back con-
verter with a DC connection. To validate the validity and effectiveness of
the proposed algorithm, the WECS was simulated using the FLC-MPPT
algorithm in MATLAB/Simulink software. This research continued by Fig. 5. The voltage, current, and power waveforms generated by the wind
comparing the outcomes with the performance of the standard adaptive system. (a) Load voltage of fuzzy logic controller (FLC); (b) load current of FLC;
PID algorithm under the same operational settings. (c) load power of FLC; (d) load voltage of proportional integral derivative (PID);
Dataset description: The turbine generates electricity at a cut-in wind (e) load current of PID; (f) load power of PID.

301
S. Karthikeyan and C. Ramakrishnan Energy Storage and Saving 3 (2024) 295–304

converter are shown in Fig. 5, with the help of Eqs. (20)− (25).
Table 2 lists the total voltage, current, and power outputs of the wind
system using the boost converter for various wind speeds. For instance,
at 3 m⋅s− 1 wind speed and simulation time period of 0–0.15 s, the
average power value attained is 1,432 W. Similarly, at 8 m⋅s− 1 wind
speed and simulation time period of 0.45–0.60 s, the average power
value attained is 1,559 W. We obtain an average power of 1,528 W
within the range of 0.60 to 0.75 s, when the available wind speed is 3
m⋅s− 1 (the rated wind velocity of the system), as shown in the graph
showing that the output voltage of the wind system is dependent on the
available wind speed. It generates an average of 1,336 W for 0.75 to 0.9 s
when all sources are turned on.

4.2. Comparison analysis of load voltage, load current and power

The WT is exposed to an irradiance profile over time, and its load

current and voltage properties dictate its active-load power production.
This indicates that the DC voltage input to the inverter, changes in active
and reactive power on the grid side, current reference on the d-axis, and
voltage and current injected into the utility grid from the WT system at
the connection point are all in phase.
The mechanical power of a WT represents two distinct MPPT ap-
proaches, as shown in Fig. 6. The results demonstrate that the FLC is
better than the PID because of its ability to swiftly adapt controller
settings in response to changes in system dynamics, such as sudden
changes in wind speed. The simulation results for the proposed
controller under varying loads are shown in Fig. 6. The DC load was
increased from zero to double its initial value in 30 s. Regardless of the
amount of current directly sent to the load, the distortion-free voltage
did not fluctuate. However, the current waveform instantly stabilized.
Fig. 6. Comparison of (a) load current, (b) load voltage, and (c) power. FLC:
Fig. 6 also shows the FLC compared with the PID controller. The DC-
fuzzy logic controller; PID: proportional integral derivative.
DC converter cannot function without a costly, high-rated semi-
conductor switch, and the PID controller alerts the user to the possibility
of transient voltage and current overrun, which might damage the under ideal circumstances. The effectiveness of the system, based on the
switch. Compared with the PID controller, the output waves of the FLC- characteristics of the wind speed profile, is briefly examined in Fig. 7.
based system were purer and less distorted. In addition, whereas the Fig. 7 shows the overall efficiency of the hybrid FLC-MPPT system,
PID-based converter creates a distorted output current, the FLC creates a which is higher than those of FFORC and fuzzy logic self-adaptive
more uniform current wave with less loss. approach using vector control (FLSA-VC), which are the two
competing WT control algorithms. In contrast to FFORC (54.2%) and
FLSA-VC (72.4%), the FLC method boosts the system efficiency to
4.3. Efficiency of the proposed system 92.6%. The test results showed that the proposed simulator could
operate continuously for long periods under different loads and wind
This section demonstrates the viability of the FLC methodology by speeds. The proposed emulation outperforms the simulated WT’s nom-
comparing it with conventional methods for the same WT configuration, inal power by a wide margin, at 96%, considering the accuracy of the
while also introducing a novel evaluation based on the overall efficiency new MPPT method and proposed emulation model.
of the system. This formula was used to study the efficiency of the PMSG
wind-producing system, as indicated in Eq. (31).
∫s 4.4. Performance of FLC
Kd
κsy = ∫ 0s × 100% (31)
0 th
K
A comparison of the output voltages of the MPPT controllers is
∫ s shown in Fig. 8. When running continuously, the FLC-based MPPT
where κsy represents the efficiency, Kd represents the active power of controller of the WT maintained the voltage at its highest steady levels.
0
∫s Fig. 8 shows how each MPPT controller is stacked in terms of power
the grid, and Kth represents the theoretical mechanical power drawn output. Assessing their efficiency is useful by utilizing Eqs. (26)–(30),
0

Table 2
Simulation results.
Configuration Simulation time (s) Wind speed (m⋅s− 1) Voltage (V) Current (A) Power (W)

FLC PID FLC PID FLC PID

1 0–0.15 3 133.8 129.9 11.3 10.96 1,432 1,424

2 0.15–0.30 9 126.4 123.8 10.63 10.42 1,341 1,290
3 0.30–0.45 3 135 130.5 11.32 10.95 1,528 1,429
4 0.45–0.60 8 136.1 123.8 11.47 10.41 1,559 1,289
5 0.60–0.75 3 135.3 130.5 11.32 10.94 1,528 1,429
6 0.75–0.9 9 126.5 123.8 10.59 10.39 1,336 1,286

Note: FLC: fuzzy logic controller; PID: proportional integral derivative.

302
S. Karthikeyan and C. Ramakrishnan Energy Storage and Saving 3 (2024) 295–304

Fig. 8. Performance of fuzzy logic controller (FLC)-maximum power point

Fig. 7. Efficiency of the proposed system. tracking (MPPT). FFORC: fuzzy fractional-order robust control; FLSA-VC: fuzzy
logic self-adaptive approach using vector control.
and it is possible to determine how different FLC-based MPPT controllers
fare in terms of their performance. Among the three controllers, the conversion chain using fuzzy logic.
greatest power output was observed for the FLC. With a high perfor- Various wind speeds were used for system-wide testing. The simu-
mance ratio of 95.7% compared with other controllers, the hybridized lation results show that the FLC can maximize wind power, manage the
FLC-MPPT controller uses this wind speed model to compute the amount of power transmitted to the grid by modifying the rotor volt-
generator power, pitch angle, drive-train torsional torque, and produc- ages, and achieve outstanding controller performance. The results were
tion speed. The FLC regulates the speed, power, and rotational efficiency more consistent, less skewed, and less variable for all step wind speeds.
of a turbine. As shown in Fig. 8, the proposed FLC-MPPT operates as The simulation results showed that the FLC-based MPPT had a higher
advertised. The WT system can achieve maximum power with a tracking efficiency of 92.6% and a performance of 95.7%. Consequently, the
accuracy of up to 95.7% and a tracking speed of <0.3 s, as shown in developed system can work well for grid-based WECS. Using the pro-
Fig. 8. When used in a constant-voltage system, the FLC can stabilize the posed Fuzzy-MPPT controller model, this study aims to determine the
voltage within 2.7% with time to set a point of <0.4 s. best way to generate electricity from renewable sources, such as wind.
The experimental results indicated a longer settling time when the Before the model was utilized for grid-connected WTs, it was developed
wind speed changed, which contradicted the conclusions. For the most into a prototype for testing and validation.
part, this is due to the implementation of wind speed variations. A large
sudden change in the value of the variables was made to change the CRediT author statement
wind speed in the simulation. In this experimental configuration, instead
of using a step-change simulation, the wind speed was varied linearly by S. Karthikeyan: Data curation, Conceptualization. C. Ram-
slowly increasing or decreasing the blower speed of the wind tunnel via akrishnan: Funding acquisition, Formal analysis.
an input keyboard. This disparity may be understood by comparing the
settling time data obtained from the simulations and experiments. Declaration of competing interest

5. Conclusion The authors declare that there are no conflicts of interest.

By accurately recording the highest power point, the FLC technique References
optimizes the input and output membership parameters of the fuzzy
system, allowing it to attain the maximum power output in the WECS. [1] M.A. Hannan, A.Q. Al-Shetwi, M.S. Mollik, et al., Wind energy conversions,
Finding the absolute duty value for the converter inverter switching may controls, and applications: a review for sustainable technologies and directions,
Sustainability 15 (2023) 3986.
be possible with the help of an optimized fuzzy system, allowing optimal [2] I.H. Panhwar, K. Ahmed, M. Seyedmahmoudian, et al., Mitigating power
power generation even when the wind speed varies. This paper reports fluctuations for energy storage in wind energy conversion system using
the results based on the efficiency of the intended system’s outgoing supercapacitors, IEEE Access 8 (2020) 189747–189760.
[3] M.S. Nazir, Y.Q. Wang, M. Bilal, et al., Comparison of small-scale wind energy
reaction, which was achieved using the controller recommended in conversion systems: economic indexes, Clean Technol 2 (2020) 144–155.
MATLAB/Simulink. The efficiency analysis results of the Fuzzy-MPPT [4] M. Vardia, N. Priyadarshi, I. Ali, et al., Design of wind energy conversion system
are shown with the PMSG’s real-time electrical current and voltage under different fault conditions. Advances in Greener Energy Technologies,
Springer, Singapore, 2020, pp. 707–716.
data, average output voltage and current patterns, output power pattern, [5] Y.A. Eldahab, N.H. Saad, A. Zekry, Assessing wind energy conversion systems
DC connection voltage waves, and generation speed. The FLC-MPPT based on newly developed wind turbine emulator, Int. J. Smart Grid (2020)
control system results demonstrated superior MPP performance and 139–148.
[6] B.P. Ganthia, S.K. Barik, B. Nayak, Wind Turbines in Energy Conversion system:
more efficient wind energy tracking than the non-optimized system. This Types & techniques. Renewable Energy and Future Power Systems, Springer,
study proposes a novel MPPT approach for a PMSG drive WECS that uses Singapore, 2021, pp. 199–217.
FLC. The simulation results show that the controller can monitor the [7] A. Azam, A. Ahmed, H. Wang, et al., Knowledge structure and research progress in
wind power generation (WPG) from 2005 to 2020 using CiteSpace based
MPPT, exhibit excellent performance with little movement, and react
scientometric analysis, J. Clean. Prod. 295 (2021) 126496.
quickly to changes in conditions with only one input to the controller [8] P.J. dos Santos Neto, T.A. dos Santos Barros, E.H. Catata, et al., Grid-connected
structure. Subsequently, a system can be built to control the proposed SRG interfaced with bidirectional DC-DC converter in WECS, IEEE Trans. Energy
Convers. 36 (2021) 3261–3270.

303
S. Karthikeyan and C. Ramakrishnan Energy Storage and Saving 3 (2024) 295–304

[9] C. Wanigasekara, A. Swain, D. Almakhles, et al., Design of delta–sigma-based PID [19] E. Elbouchikhi, G. Feld, Y. Amirat, et al., Design and experimental implementation
controller for networked wind energy conversion systems, IEEE Trans. Ind. Appl. of a wind energy conversion platform with education and research capabilities,
58 (2022) 879–889. Comput. Electr. Eng. 85 (2020) 106661.
[10] J.G. González-Hernández, R. Salas-Cabrera, R. Vázquez-Bautista, et al., A novel [20] M. Hermassi, S. Krim, Y. Kraiem, et al., Design of vector control strategies based on
MPPT PI discrete reverse-acting controller for a wind energy conversion system, fuzzy gain scheduling PID controllers for a grid-connected wind energy conversion
Renew. Energy 178 (2021) 904–915. system: hardware FPGA-in-the-loop verification, Electronics 12 (2023) 1419.
[11] M.P. Shadakshara Swamy, P. Rakshith, K.S. Varchasvi, et al., MPPT controlled [21] N. Ullah, I. Sami, M.S. Chowdhury, et al., Artificial intelligence integrated
wind energy conversion system(WECS) supplying DC micro-grid, in: 2020 Third fractional order control of doubly fed induction generator-based wind energy
International Conference on Smart Systems and Inventive Technology (ICSSIT), system, IEEE Access 9 (2021) 5734–5748.
IEEE, Tirunelveli, India, 2020, pp. 449–454. [22] N. Mrabet, C. Benzazah, A. El Akkary, et al., Multi-objective ant colony
[12] O. Cornea, D. Hulea, N. Muntean, et al., Step-down switched-inductor hybrid DC- optimization for enhancing the maximum power of variable-speed wind turbines
DC converter for small power wind energy conversion systems with hybrid storage, based on PMSG, Int. J. Energy Convers. IRECON 11 (2023) 195.
IEEE Access 8 (2020) 136092–136107. [23] S. Kumar, K. Vijayakumar, Simulation and experimental comparative analysis of
[13] Y. Xu, Z. Wang, Z.X. Zou, et al., Voltage-fed isolated matrix-type AC/DC converter the DC-DC converter topologies for wind driven SEIG fed DC nanogrid, Electr.
for wind energy conversion system, IEEE Trans. Ind. Electron. 69 (2022) Power Syst. Res. 181 (2020) 106196.
13056–13068. [24] H. Shutari, T. Ibrahim, N. Bin Mohd Nor, et al., Development of a novel efficient
[14] A.A. Chhipą, P. Chakrabarti, V. Bolshev, et al., Modeling and control strategy of maximum power extraction technique for grid-tied VSWT system, IEEE Access 10
wind energy conversion system with grid-connected doubly-fed induction (2022) 101922–101935.
generator, Energies 15 (2022) 6694. [25] A. Jain, S. Trivedi, P. Sharma, et al., Simulation study of permanent magnet
[15] M. Elsisi, M.Q. Tran, K. Mahmoud, et al., Robust design of ANFIS-based blade pitch synchronous generator (PMSG) connected to variable speed wind energy
controller for wind energy conversion systems against wind speed fluctuations, conversion system (WECS, in: 2017 International Conference on Trends in
IEEE Access 9 (2021) 37894–37904. Electronics and Informatics (ICEI), IEEE, Tirunelveli, India, 2017, pp. 966–970.
[16] L. Xing, Q. Wei, Y.W. Li, A PWM current-source converter-based wind energy [26] H. Ahuja, P. Kumar, A novel approach for coordinated operation of variable speed
conversion system, IEEE Trans. Power Electron. 39 (2024) 2787–2797. wind energy conversion in smart grid applications, Comput. Electr. Eng. 77 (2019)
[17] E.H. Dursun, A.A. Kulaksiz, MPPT control of PMSG based small-scale wind energy 72–87.
conversion system connected to DC-bus, Int. J. Emerg. Electr. Power Syst. 21 [27] H.H.H. Mousa, A.R. Youssef, E.E.M. Mohamed, Study of robust adaptive step-sizes
(2020) 20190188. P&O MPPT algorithm for high-inertia WT with direct-driven multiphase PMSG, Int.
[18] B. Sarsembayev, K. Suleimenov, B. Mirzagalikova, et al., SDRE-based integral Trans. Electr. Energy Syst. 29 (2019) e12090.
sliding mode control for wind energy conversion systems, IEEE Access 8 (2020) [28] Wind power curve modeling, 2023. https://www.kaggle.com/code/winternguye
51100–51113. n/wind-power-curve-modeling. (Accessed 1 October 2023).

304