Journal of Energy Storage 45 (2022) 103613

Contents lists available at ScienceDirect

Journal of Energy Storage

journal homepage: www.elsevier.com/locate/est

A power management control and optimization of a wind turbine with

battery storage system
Saloua Belaid , Djamila Rekioua *, Adel Oubelaid , Djamel Ziane , Toufik Rekioua
Laboratoire de Technologie Industrielle et de l’information (LTII), Faculté de Technologie, Université de Bejaia, 06000 Bejaia, Algérie

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

Keywords: Battery storage systems are an important alternative to compensate for wind turbine irregularities. This paper
Wind turbine contributes to the feasibility of a wind energy installation with battery storage. In order to manage these different
Battery storage power sources, a power management control (PMC) strategy is developed and connected to the proposed two-
Hybrid maximum power point tracking
level MPPT controller. PMC provides an efficient optimal operation of two MPPT algorithms (torque and
State of charge
Batteries stress
speed) and ensures minimal stress on the used storage batteries. A complete system with better performances is
obtained by combining the advantages of both MPPT and PMC proposed methods. The obtained results highlight
the positive impact of MPPT hybridization on the reduction of usage of battery storage system. The different
wind turbine parameters were identified in our research laboratory. All these parameters have been used in
simulation models in order to obtain more realistic mathematical models and the closest to the empirical data.
The study was performed under Matlab/Simulink. The obtained results with the different comparisons are
presented to show the best algorithm to be used in a wind energy system (WES) in terms of power, efficiency,
response time, and effect on the battery state of charge.

whether a microcontroller is needed (nature of analog or digital cir­

cuits), difficulty of implementation with complex algorithms, conver­
1. Introduction
gence time (tracking factor), efficiency, price (more or less expensive),
the sensors number of to be provided, independence or no of system
Due to the different advantages of wind energy systems (WES) with
characteristics, number of input variables, stability, ....The most
battery storage, a great interest is attributed to them [1–3]. In addition
commonly used algorithms are the following: Tip Speed Ratio (TSR),
to their ability to provide continuous energy regardless of load and
Hill Climbing Search (HCS) [10], Optimal Torque Control (OTC) [11],
metrological variations, hybrid energy systems can manage various
Fuzzy Logic Controller (FLC) [12], Power Signal Feedback (PSF), Ge­
sources in a smart way by using power management control strategies
netic Algorithm(GA), Artificial Neural Network (ANN) based controller;
(PMC) which satisfies the load demand and optimizes battery state of
Adaptive Neuro-Fuzzy Inference System (ANFIS), Particle Swarm Opti­
charge [4–6].
mization (PSO),...… The HCS algorithm is the most popular approach
The most known WES drawback is the output power that depends on
because of its ease of implementation. It compares the previously
the wind speed. Therefore, it is not easy to keep the maximum wind
delivered power with that after the disturbance. In the PSF method, a
turbine power output for all wind speed conditions [7–9]. Various MPPT
reference power signal is generated to obtain the optimal power PTb-opt.
approaches have been investigated to track the maximum power point of
The OTC method adjusts the generator torque Tg to its optimal value
the wind turbine [10–12]. They all have the objective of maximizing
Tem-opt at different wind speeds. In FLC, inputs controllers are power
power. However, each of them differs from the others based on different
variation (ΔPTb) and speed variation (ΔωTb) and the output is the
characteristics such as precision (accurate or no), tracking process,

Abbreviations: AC, Alternate Current; ANN, Artificial Neural Network; ANFIS, Adaptive Neuro-Fuzzy Inference System; DAS, Data Acquisition System; DC, Direct
Current; EMC, Energy Management Control; EV, Electric Vehicle; FLC, Fuzzy Logic Controller; FOC, Field-oriented control; GA, Genetic Algorithm; HCS, Hill Climb
search; Hyb(HCS/OTC), Hybrid HCS & OTC; Hyb (HCS/FLC), Hybrid HCS & FLC; Hyb(OTC/FLC), Hybrid OTC& FLC; MPP, Maximum power point; MPPT, Maximum
power point tracking; OTC, Optimal torque control; PMC, Power Management Control; PMSG, Permanent magnet synchronous generator; P&O, Perturb. and
Observe; PSO, Particle Swarm Optimization; SOC, State of charge; SOCmin, Minimum state of charge; SOCmax, Maximum state of charge; TSR, Tip speed ratio; WES,
Wind Energy System.
* Corresponding author.
E-mail address: djamila.rekioua@univ-bejaia.dz (D. Rekioua).

https://doi.org/10.1016/j.est.2021.103613
Received 26 April 2021; Received in revised form 8 October 2021; Accepted 6 November 2021
Available online 7 December 2021
2352-152X/© 2021 Elsevier Ltd. All rights reserved.
S. Belaid et al. Journal of Energy Storage 45 (2022) 103613

Nomenclature Tem− opt,best Selected best optimal electromagnetic torque, N.m

Tr Response time, s
Cp-opt Maximum power coefficient va,b, c Machine phase voltages, V
ea,b,c Induced f.e.m in the stator phase windings, V VBat Terminal battery voltage, V
ed,eq Direct and quadratic magnetomotive force, V Vd, Vq Direct and quadratic stator voltage, V
EBat Open circuit voltage of the battery, V VDC-bus DC bus voltage, V
ia,b, c Current phase voltages, A Vstart Starting wind speed, m/s
IBat Battery current, A Vwind Wind speeds, m/s
Id, Iq Direct and quadratic stator currents, A
Lc Cyclic inductance, H Greek letters
Rs Resistance of the stator winding, Ω ηMPPT MPPT efficiency
nBat Series cells, λ Tip speed ratio
P Number of pole pairs, λopt Optimal Tip speed ratio
PBat Battery power, W ΔωTb Speed variation, rad/s
PTb-best Selected best optimized turbine power, W ΔωTb,ref Reference speed variation, rad/s
PTb− opt Optimized power turbine, W ΔPTb Power variation, W
Pload Load power, W Φf Permanent magnetic flux produced by the rotor magnets,
Popt Maximal power, W Wb
PTb Turbine power, W ω Rotor speed, rad/s
RTb Radius, m ωopt1, ωopt2 , ……., ωoptk the different rotational speed values of each
RBat Battery internal resistance, Ω MPPT method (k=1,6), rad/s
Tg Generator torque, N.m ωTb Turbine rotational speed, rad/s
Tem-opt Optimal electromagnetic torque, N.m ωTb− opt Optimal Turbine rotational speed, rad/s
Tem− opt1, Tem− opt2 , …, Tem− optk are different electromagnetic torque ωTb− opt,best Selected best optimal turbine rotational speed, rad/s
values of each MPPT method (k=1,6), N.m

Fig. 1. Proposed system architecture.

reference speed variation (ΔωTb,ref). In order to converge to the optimal renewable energy systems has been conducted. The authors in [15] have
point, the rules will depend on the variations of power and speed. presented different methods to utilize excess energy in renewable sys­
Different architectures of the multi-source system were suggested tems. Different methods were presented to improve the operation
with different power management controls (PMC) methods. Some of without additional cost. In [16], energy management control (EMC) is
them are logic-based and others are intelligent-based. The latter are developed using a predictive control strategy and applied to a wind/PV
more interesting, especially for stand-alone applications [13–15] and turbine with battery storage. This method achieves optimal values and
[16–18]. All PMC strategies are based on the concept of controlling the the overall cost has been reduced. In another study [17], artificial in­
output power of each source and protecting the storage systems used telligence methods are widely applied to the supervision of renewable
[19–21]. Authors in [13] synthesized the most important supervisory energy systems. For example, in [18], the authors propose a review of
controls and different energy management methods used. In another energy management for photovoltaic installations. The similar approach
study [14], an extensive review of energy management methods in has also been applied in [24–26] and [27–29]. The PMC developed in

2
S. Belaid et al. Journal of Energy Storage 45 (2022) 103613

this publication; takes into account all input variables considering the
power optimization by a hybrid MPPT method. The supervision used
allows controlling the different output powers, protecting the storage
system and controlling the DC voltage.
This paper contributes to the feasibility of a wind energy system with
a battery storage and equipped with a two-level MPPT controller. It
achieves an efficient operation of both MPPT algorithms to obtain an
optimal performance level of wind power system and a minimal stress
on the battery of the studied system. This new and improved controller
level is based on a mathematical optimization method. A PMC was also
applied. It is based on control different powers (the power supplied by
the wind generator (PTb), the power supplied or required by the battery
for compensation or recharge respectively (PBat) and the power needed
by the load (PLoad). The objectives of the PMC based are, on the one
hand, to satisfy the load power demand and in the other hand to
maintain the state of charge of the battery bank to prevent blackout and
to extend the batteries life. For wind power maximization, HCS, OTC
and FLC methods were first investigated, and then, due to the advan­
tages and disadvantages of each MPPT method, three hybrid algorithms
were proposed. The combination of HCS and OTC method (Hyb(HCS/
OTC)), then the combination of HCS and FLC (Hyb(HCS/FLC)) and
finally the combination of OTC and FLC (Hyb(OTC/FLC)). The best al­
gorithm to choose is based on different criteria (power, time response,
efficiency, effect on state of charge of battery…). This not only maxi­
mizes wind power through faster responses, but also puts less strain on
the batteries thus increasing their life. To keep the voltage DC bus
constant whatever the wind speeds variations, field-oriented control
Fig. 2. Installed wind turbine system.
(FOC) based on hysteresis current was used. This PMC strategy is con­
nected to a proposed hybrid maximum power point tracking (MPPT)
Table 1 algorithm. A complete system incorporating the advantages of the MPPT
Wind turbine parameters. method and the PMC is thus produced resulting in better performance of
Nominal power PN 900 W the studied system. The different wind turbine parameters were identi­
Resistance of the stator winding RS 0.49 Ω fied at the LTII of the University of Bejaia (Algeria). Wind speed mea­
Stator inductance LS 0.0016 H surement was performed during a whole day using a data acquisition
Number of pole pairs P 5 system (DAS). All these parameters were used in simulation models in
Flux Φf 0.148 Wb order to obtain the most realistic mathematical models and the closest to
Turbine radius RTb 1.05 m the experiment. The results obtained in MATLAB/Simulink are pre­
TotalinertiaJ 0.016 kg/m2 sented and analyzed. The proposed PMC with Hybrid MPPT algorithm
1
Viscous friction coefficient f 0.0001 N.m.s rad− (Hyb(HCS/FLC)) has been applied to a WECS with battery storage,

Fig. 3. DC bus regulation-based field-oriented control.

3
S. Belaid et al. Journal of Energy Storage 45 (2022) 103613

Fig. 4. Battery equivalent circuit model.

Fig. 5. Studied system representation.

2. System description

The architecture of the studied system (Fig. 1) comprises a wind

turbine connected to PMSG, a rectifier, DC/DC converter, batteries
storage, a load and power management control unit to manage the
different powers. For wind power maximization, three hybrid ap­
proaches have been developed. Field-oriented control (FOC) is used for
DC bus regulation.

3. Wind turbine parameters identification

The wind turbine (1) intalled in our Laboratory is shown in Fig. 2. It

delivers a peak power of 900 Watts at 12.5 m/s. It is combined with a
Whisper controller (2) and 24-Volt battery system (3).
The parameters of the wind turbine used in this paper were identified
Fig. 6. Equivalent circuit of a generator phase. as it is shown in Table 1.

4. System modeling
under two wind speed profiles. The first one is used to show the different
performances (the power coefficient curves, tip speed ratio, mechanical 4.1. Wind turbine model
power, rotational speeds and electromagnetic torque). But in order to
compare the different methods in terms of efficiency, power and effect The wind turbine tip speed ratio (TSR) is defined as the ratio between
on the state of charge of the battery, a step profile was chosen. This the rotational speed of the tip of a blade ωTb .RTb and the actual wind
allows making the different calculations at constant speeds during time speed V wind [22].
intervals. So, the custom profile is chosen, in purpose, to illustrate and
compare the transient dynamics response and efficiency. λ=
ωTb .RTb
(1)
The obtained results point out the important role that hybridization V wind
and maximum power point tracking algorithms play in optimizing bat­ The mechanical power PTb is expressed using Eq. (2):
tery usage and lifetime in hybrid storage systems.

4
S. Belaid et al. Journal of Energy Storage 45 (2022) 103613

Fig. 7. Duty cycle evolution versus time.

Fig. 8. The principle of the P&O MPPT algorithm.

Fig. 9. Optimal torque control principle.

5
S. Belaid et al. Journal of Energy Storage 45 (2022) 103613

Fig. 10. Fuzzy logic controller operation principle.

4.2. PMSM model

Table 2
Fuzzy inference rules. Permanent magnet synchronous machine is a very interesting solu­
ΔωTb ΔPTb tion for stand-alone wind power applications because of its high effi­
ciency and important torque- mass ratio and the absence of necessity for
BN MN SN Z SP MP BP
an excitation circuit. Its main drawback is its high cost compared to
BN BN BN BN BN MN SN ZE induction machines [22,23].
MN BN BN BN MN SN ZE SP
The voltage and electromagnetic torque equations are given in (a,b,
SN BN BN MN SN ZE SP MP
Z BN MN SN ZE SP MP BP c) frame by:
SP MN SN ZE SP MP BP BP ⎡ ⎤ ⎡ ⎤ ⎡ ⎤ ⎡ ⎤
MP SN ZE SP MP BP BP BP
va ia ia ea
⎢ ⎥ ⎢ ⎥ ⎢ ⎥ ⎢ ⎥
BP ZE SP MP BP BP BP BP ⎢ vb ⎥ = Rs ⎢ ib ⎥ + Ls d ⎢ ib ⎥ + ⎢ eb ⎥
⎣ ⎦ ⎣ ⎦ dt ⎣ ⎦ ⎣ ⎦
With: BN: Big Negative, BP: Big Positive, MN: Middle Negative, MP: Middle vc ic ic ec (4)
Positive, SN: Small Negative, SP: Small Positive, Z: Zero
P
Tem = (ea .ia + eb .ib + ec .ic )
ω
1
PTb = .Cp .ρ.π.R2Tb .V 3wind (2)
2 Where: va,b,c are machine phase voltages, ia,b,c are the three phase cur­
The power coefficient Cp has a unique maximum Cp− opt that corre­ rents, Ls is the inductance, Rs represent stator winding resistance, ea,b,c is
sponds to a maximum power as given in Eq. (3) the induced f.e.m in the stator phase windings, P the number of pole
pairs, ω is the rotor speed (ω=P.Ω).
ωTb−
λopt =
opt
(3) Eq. (4) can be expressed in the (d,q) frame as follow [22]:
V wind

6
S. Belaid et al. Journal of Energy Storage 45 (2022) 103613

Fig. 11. Mathematical optimization method for A 2-level MPPT controller.

Fig. 12. Wind speed profile.

7
S. Belaid et al. Journal of Energy Storage 45 (2022) 103613

Fig. 15. Power coefficient Cp.

[ ] [ ][ ] [ ] [ ]
Vd Rs − ω.Ls Id d Id ed 4.3. Storage sytem modeling
= + Ls +
Vq ω.Ls Rs Iq dt Iq eq (5) The used battery model is shown in Fig. 4. This model it character­
Tem =
P(
ea .Iq
) ized by a variable resistance in series setting with a voltage source
ω representing the electromotive force. For nb cells in series, the battery
voltage can be written as [27,28]:
Where: ed and eq are the direct and quadratic magnetomotive forces, Id,
Iq are respectively the direct and quadratic stator currents, Vd, Vq V Bat = nBat .EBat ± nBat .RBat .IBat (7)
represent direct and quadratic stator voltage, Ls is the stator inductance,
Rs is the stator resistance and ω is the rotational speed. Where:V Bat is the terminal battery voltage, EBat Open circuit voltage,
To avoid harmful voltage fluctuations caused mainly by abrupt and RBat battery internal resistance, IBat the battery current and nBat series
large wind speeds variations, field oriented control (FOC) based hys­ cells.
teresis current control is used to ensure DC bus voltage regulation and is
shown in Fig. 3. Direct and quadrature reference currents are given by 5. Hybrid MPPT algorithms
Eq. (6) shown below. It is worth noticing that Idref is set to zero for
decoupling and torque maximization. Over the years, several MPPT strategies have been introduced for
⎧ tracking the wind turbine maximum power [5–11]. In our work, HCS,
⎨ Iqref = 2.T eref
⎪ OTC and FLC methods were combined for sake of increasing system
3.ϕf .P.ωTb (6) effeciency and reducing the stress applied on the storage batteries. To
⎪
⎩
Idref = 0 invistigate the positive impact of hybridization on enhancing the overal
system performance, HCS and OTC methods were first combined Hyb
(HCS/OTC)). After that, HCS and FLC methods were gathered (Hyb
(HCS/FLC)). Finally, OTC and FLC methods were simultaneously used

8
S. Belaid et al. Journal of Energy Storage 45 (2022) 103613

Fig. 16. Tip speed ratio.

(Hyb (OTC/FLC)). ( )2
V eff π2 1 − α
Req = = RLoad (11)
5.1. Optimal duty cycle calculation Ieff 18 α
The provided electrical power by the generator is given by:
In order to find the optimal duty cycle expression, a theoretical ( )
analysis of the system represented in Fig. 5 is performed. p.Φeff ω
Pele = 3Req .I2eff = 3Req ( )2 (12)
Neglecting the various losses and assuming system ideality (unity Rs + Req + (Ls .ω)2
efficiency), the electric generated power is equal the power received by
the battery (Pele=PBat). Mathematically, this last sentence is expressed as Thus, it is obtained as a function of the duty cycle and the rotational
follow: speed:
⎛ ⎞
V 2Bat V2
P = = DC (8) ( )⎜ ⎟
ele
RLoad RDC π2 1 − α ⎜ (K.Φ.ω)2 ⎟
Pele = ⎜[
⎜ ( )2 ] 2
⎟
⎟ (13)
6 α ⎝ π2 ⎠
With: R
18 Load
1− α
α + Rs + (Ls .ω)2

( )2
1− α
RDC = RLoad (9) The optimal duty cycle is obtained when:dPdαele = 0
α
( )([ ] )
V DC 18.V eff dPele π2 1+ α − 2α2 (A+Rs )2 +(Ls .ω)2 − 2(A+Rs )A
RDC = = 2 (10) dα
=−
3
(K.Φ.ω)RLoad
α3 ( )2
IDC π .Ieff (A+Rs )2 +(Ls .ω)2

The equivalent resistance seen from the generator is written as =0

(Fig. 6): (14)

9
S. Belaid et al. Journal of Energy Storage 45 (2022) 103613

Fig. 17. Mechanical power.

10
S. Belaid et al. Journal of Energy Storage 45 (2022) 103613

Fig. 18. Rotational speeds.

With: ⎧ ( )
⎪
⎪
⎪ 1 Cp.ρ.π.R5Tb
( )2 ⎪
⎨ P Tb = .ω3Tb
π 2
1− α 2 λ5
A= RLoad ( ) (16)
18 α ⎪
⎪ 1 Cp.ρ.π.R5Tb
⎪
⎪ .ω2Tb
⎩ T em = 3
The accepted solution of this equation is: 2 λ

1 Optimal rotational speedand torque are given as:

αopt = √̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
√̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅ (15) ⎧
1
1 + π2 R18Load (Ls .ω)2 + Rs 2 ⎪ 2 3
⎨ PTb− opt = 2.Cp− opt. ρ.π.RTb V wind
⎪
(17)
The graphical representation of the optimal duty cycle derived above ⎪
⎪
⎩ ωopt =
V wind. λopt
is depicted in (Fig 7). RTb
Thus, the optimal electromagnetic torque will be:
5.2. Hill climbing search (HCS) algorithm ( )
1 Cp− opt. ρ.π.R5Tb
T em− opt = . ω2opt (18)
Hill climbing search (HCS) aims to climb the PTb (ωTb) curve in the 2 λ3Tb
power increase direction (PTb), varying the rotational speed periodically
The optimal torque control principle is graphically illustrated in
with a small incremental step (Fig. 8) to reduce oscillations around the
Fig. 9.
maximum power point (MPP). The HCS algorithm compares the previ­
ously delivered power with that after the disturbance. The method is
simple, but this has the disadvantage of less efficient and has a poor 5.4. Fuzzy logic control algorithm
directionality in case of sudden changes in wind speeds.
The FLC method is used to find the operating point corresponding to
5.3. Optimal torque control (OTC) the optimal power for different wind speeds. The advantage of this
method is that it can be used with low-precision input values and does
Optimal torque control will adjust the generator torque to its opti­ not require high accuracy. The inputs are power variation (ΔPTb) and
mum at various wind speeds[4]. However, and the characteristics of the speed variation (ΔωTb). The output is reference voltage variation (ΔωTb,
turbine must be known (Cp-opt and λopt). ref). .The principle of operation of the fuzzy controller bloc is described in
The power and electromagnetic torque are written as: Fig. 10.

11
S. Belaid et al. Journal of Energy Storage 45 (2022) 103613

Fig. 19. Electromagnetic torque.

12
S. Belaid et al. Journal of Energy Storage 45 (2022) 103613

Fig. 20. Voltage battery.

The different linguistic rules depend on the variations of power ΔPTb With: ωopt1, ωopt2 , ……., ωoptk are the different rotational speed values of
and voltage fluctuations ΔωTb each MPPT method (k=1,6), T em− opt1, T em− opt2 , …, T em− optk are different
{
ΔωTb = ωTb (k) − ωTb (k − 1) electromagnetic torque values of each MPPT method (k=1,6), PTb-best is
(19) the selected best optimized turbine power, ωTb− opt is the optimal turbine
ΔPTb = PTb (k) − PTb (k − 1)
rotational speed, ωTb− opt,best is the selected best optimal turbine rota­
Where PTb(k) and ωTB (k) are respectively the output turbine power and tional speed, T em− opt,best is the selected best optimal electromagnetic
its rotational speed at sampled instants (k) and ΔωTb,ref (k) the instant of torque, Tem-opt is the optimal electromagnetic torque.
reference speed. These inputs and output variables are expressed in In our work, three hybrid MPPT algorithms have been investigated.
terms of linguistic variables (table 2.) The first hybrid control strategy combines HCS and OTC algorithms
(Hyb(HCS/OTC)), the second hybrid control strategy gathers HCS and
FLC algorithms (Hyb(HCS/FLC)). The last hybrid method considered is
5.5. Proposed hybrid MPPT algorithms
the combination of OTC and FLC algorithms (Hyb(OTC/FLC)).

A 2-level controller that governs the hybrid MPPT approach is shown

in Fig. 11. It guarantees an efficient operation of the two MPPT algo­ 5.6. Simulation results of different MPPT algorithms
rithms operating at an optimal performance level (torque and speed) of
the wind system and a minimal stress on the battery of the studied global Simulations are performed under Matlab/Simulink environment
system. This high level controller is based on a mathematical optimi­ using the two days wind speed profile shown in Fig. 12.
zation method as it is shown in (Fig. 11). The first stage allows the choice The sizing of the proposed system depends mainly on the site loca­
of the different optimal values of each MPPT algorithm (Eq. (20)). The tion that dictates the average wind speed, the turbine orientation and
second stage provides the global best value of rotational speed and the average energy consumption of the application. The studied system
electromagnetic torque as highlighted in (Eq. (21)). includes a 900 W wind turbine and two batteries (100 Ah, 12 V) in
{ { } parallel. The following Figs. 13–17 represent, respectively, the power
ωTb− opt = {ωTb− opt1 , ωTb− opt2 , ……., ωTb− optk }
(20) coefficient curves, tip speed ratio, mechanical power, rotational speeds
T em− opt = T em− opt1 , T em− opt2 , ……., T em− optk
and electromagnetic torque using the different MPPTs used for the wind
{
ωTb−
(
= max( ωTb−
) turbine and simulated under the same wind speed profile. It can be
opt,best opt )
(21) noticed in Fig. 15 that both curves (CP and λ) reach their optimal values
T em− opt,best = max T em− opt
quite quickly (Cp-opt =0.48, λopt=8.1) whatever what the wind speed
PTb− best = ωTb− opt,best .T em− opt,best (22) values are. The different zooms for each curve allow us to see the
response and the performance of each MPPT method. It is noticed that

13
S. Belaid et al. Journal of Energy Storage 45 (2022) 103613

Fig. 21. Power battery.

hybrid control methods give the best results in terms of power, elec­ It is noticed that compared to the other methods, the Hyb(HCS/FLC)
tromagnetic torque and speed (Figs. 17–19) and especially the Hyb method gives the highest power values for all wind speeds which are
(HCS/OTC) method and the Hyb(HCS/FLC). The batteries are initially ranging from 4 to 12 m/s.
charged. The SOCmin is defined at 30% and the SOCmaxat90%. Voltage The efficiency of each method has been calculated and summarized
and power battery are represented in Figs. 18 and 19 and the state of in Table 4. It is calculated by the following equation:
charge in Fig. 20. It is seen that the battery voltage remains around its
ΔPMPPT = PTb− − PTb (23)
reference voltage of 24 V for the hybrid methods (Fig. 18). The less stress opt

on the battery when using the hybrid methods causes a reduction in the PTb PTb − ΔPMPPT ΔPMPPT
discharge power of the battery (Fig. 19). In Fig. 22, it is noticed at ηMPPT (%) = x100 = =1− (24)
PTb− opt PTb− opt PTb− opt
startup (Zoom1), the HCS/OTC and HCS/FLC hybrid methods do not
stress the battery much and their SOC stays around the SOCmax of 90%, Where: PTb− opt is the optimized power turbine and PTb is the turbine
contrary to the OTC method which brings the SOC down to 79% but power without optimization and ΔPMPPT is the increase of generated
hybridizing of the OTC with the FLC, will be increased to 85%. During a turbine power.
period of time, at a given wind speed (Zoom2), it is seen that the hybrid In fact, there was an increase in the power generated by the MPPT
methods keep the SOC around 90% unlike the HCS and OTC methods. compared to the system without MPPT (Fig. 25). That is why the effi­
To conclude, the simulation results reveal that the proposed hybrid ciency of the proposed MPPT methods achieves high values.
methods provide better results than the standard non-hybrid methods. It As for the response time, a comparison of the different methods is
is also observed that when using hybrid methods, the SOC remains summarized in Table 5.
around a maximum value of 90% unlike other methods such as OTC and Response time Tr is the necessary time for a system to go from 10% to
HCS. The best level of the SOC is the one of Hyb(HCS/FLC). 95 % of its steady state value. In general wind profiles are in permanent
In order to compare the different methods in terms of transient dy­ variation and don’t settle to a steady state value. This is why another
namics, efficiency, power and effect on the state of charge of the battery, custom speed profile was added to highlight the system responsiveness
a step profile was chosen (Fig. 23). using the different MPPT approaches. It is clear that the hybrid MPPT
The turbine power waveform using the different MPPT methods is methods respond faster than the non hybrid methods whatever the wind
shown in Fig. 24. speeds, with always a slight superiority for the Hyb(HCS/FLC) method.
In order to calculate the different power for each MPPT method, In terms of state of charge (Fig. 26.), it can be seen that at startup
different zooms were taken for each time interval which represents a (Zoom1), the HCS/OTC and HCS/FLC hybrid methods keep their state of
given constant speed. The different power values are reported in the charge between a SOCmax of 90% and a SOCmin of 50%. When using non
following table (Table 3). hybrid methods, battery SOC reaches a SOCmin of 40%. Hence, one can

14
S. Belaid et al. Journal of Energy Storage 45 (2022) 103613

Fig. 22. Battery state of charge.

Fig. 23. Wind speed step profile.

15
S. Belaid et al. Journal of Energy Storage 45 (2022) 103613

Fig. 24. Turbine powers.

Table 3
Evaluation of the different powers.
Vwind(m/s) PTb (W)

Without MPPT HCS OTC FLC Hyb(HCS/OTC) Hyb(HCS/FLC) Hyb(OTC/FLC)

10 888.10 971.70 1005.00 1002.00 999,30 1012.00 1008.00

7 303.80 337.30 343,60 345.40 343,70 347.70 345.30
4 56.29 63.73 63,58 63.68 64,07 64.83 63.76
9 647.00 711.12 732.00 728.60 733,40 738.40 732.90
12 1536.00 1666.00 1738.00 1746.00 1740,37 1751.00 1745.00

Table 4
MPPT Efficiency of each control strategy.
Vwind(m/ ηMPPT (%)
s)
HCS OTC FLC Hyb(HCS/ Hyb(HCS/ Hyb(OTC/
OTC) FLC) FLC)

10 91,40 88,37 88,63 88,87 87,76 88,11

7 90,07 88,42 87,96 88,39 87,37 87,98
4 88,33 88,53 88,40 87,86 86,83 88,28
9 90,98 88,39 88,80 88,22 87,62 88,28
12 86,19 82,62 82,25 82,51 82,01 82,29

Fig. 25. Generated turbine and its increasing a of each MPPT method. (a)
Generated turbine powers. (b) Increased generated turbine power.

16
S. Belaid et al. Journal of Energy Storage 45 (2022) 103613

Table. 5
Comparison of the different methods in terms of response time.
Vwind(m/s) Tr (s)

Without MPPT HCS OTC FLC Hyb(HCS/OTC) Hyb(HCS/FLC) Hyb(OTC/FLC)

10 5.7100 0.5800 0.6320 0.3830 0.1000 0.0160 0.3650

7 0.2970 0.0460 0.2090 0.2200 0.2670 0.0900 0.2700
4 0.8080 0.0580 0.7810 0.3010 0.6030 0.1200 0.3780
9 0.0380 0.0050 0.0520 0.0310 0.0350 0.0015 0.0350
12 0.0240 0.0040 0.0330 0.0220 0.0150 0.0010 0.0200

Fig. 26. State of charge under chosen step profile.

• -All of the proposed methods were effective in recovering the

maximum amount of wind energy and were able to achieve the
Table 6 optimal power coefficient at all times.
Different values of SOC of the different studied methods. • -The proposed hybrid methods are more efficient than the other non
Vwind(m/ SOC(%) hybrid methods in terms of power, electromagnetic torque, speed
s) and efficiency.
HCS OTC FLC Hyb Hyb Hyb
(HCS/ (HCS/ (OTC/ • -The Hyb(HCS/FLC) method shows a better performance in terms of
OTC) FLC) FLC) power and response time at various wind speeds. In addition to that,
10 90.040 90.040 90.070 90.040 90.185 90,040 it has enabled the reduction stress applied on storage batteries.
7 80.200 79.500 82.880 89.110 90.176 84.720 • It can be deduced that Hyb(HCS/FLC) method offers often the best
4 40.020 40.000 43.350 50.720 51.510 48.550 performances, so this method will be used in the next part of our
9 90.010 90.010 90.060 90.100 90.106 90.080
12 90.090 90.070 90.050 90.090 90.096 90.070
work.

6. Proposed power management control

A power management control (PMC) is proposed to provide moder­

conclude that the hybridization resulted in saving 10% of battery SOC.
ate power sources utilization. It determines the different operating
The different SOC values during the different time intervals have
modes of the wind/battery system according to the wind speeds
been reported in Table 6.
(Fig. 27).
From the presented simulation results, the following conclusions can
The algorithm identified six different operating modes. The different
be made:
powers considered are:

17
S. Belaid et al. Journal of Energy Storage 45 (2022) 103613

Fig. 27. Overall bloc diagram of the studied system.

Table 7
Status of the switches and the different modes.
Description ΔP(W) Powers (W) State of Switches Modes
batteries

Disconnected mode: The wind speed is below Vstart, the system stops and the load is disconnected to ΔP = 0 PLoad = 0 ———– K1=0 Mode 1
preserve the stored power. PTb = 0 K2=0 (M1)
PBat = 0 K3=0
K4=0
Windmode: The power delivered by the generator is equal to load power. ΔP = 0 PLoad = PTb ———– K1=0 Mode 2
PBat = 0 K2=0 (M2)
K3=0
K4=1
Compensation mode: The wind turbine power is insufficient to supply the load, energy compensation is ΔP < 0 PLoad = PTb + PBat Discharging K1=0 Mode 3
essential. The lack is provided by the storage (SOCmin=30%). SOC > SOCmin K2=0 (M3)
K3=1
K4=1
Excessmode: The power delivered by the generator is greater than the load power and the SOC is greater ΔP > 0 PTb > PLoad Charging K1=0 Mode 4
than the SOCmax, the battery is disconnected, the excess power is sent to d a dump load SOC ≥ SOCmax K2=1 (M4)
K3=0
K4=1
Battery discharged mode: The power supplied by the wind turbine is lower than the load power and the ΔP < 0 PTb ≤ PLoad Discharging K1=1 Mode 5
SOC is lower than the SOCmin, the load is disconnected and the power supplied by the wind turbine is PLoad = 0 K2=0 (M5)
directed to the batteries for charging. K3=0
PBat = − PTb
K4=0
SOC < SOCmin
Battery chargemode: The wind turbine power is greater than the load power, the excess energy is ΔP > 0 PTb > PLoad Charging K1=1 Mode 6
directed to the battery for charging as long as the state of charge is below the SOCmax. SOC < SOCmax K2=0 (M6)
K3=0
K4=1

Fig. 28. Different switches.

18
S. Belaid et al. Journal of Energy Storage 45 (2022) 103613

Fig. 29. Proposed PMC flowchart.

19
S. Belaid et al. Journal of Energy Storage 45 (2022) 103613

Fig. 30. Obtained powers waveforms with the different switches.

20
S. Belaid et al. Journal of Energy Storage 45 (2022) 103613

• the power supplied by the wind generator (PTb),

• the power supplied or required by the battery for compensation or

670

760

M6
>0
90

1
0
0
1
recharge respectively (PBat),
• and the power required by the load (Pload).

0 =0
670

670

M2
0
0
1
1
The available power is expressed as follows:

870,4
200,4
(25)
670
ΔP = PTb − PLoad

M6
>0
1
0
0
1
Mathematically, this value can be negative, positive or zero. If it is
701,4
positive, the load will be provided by the wind turbine, and the batteries
701,4

M5
>0 are being charged. On the other hand, if it is negative, the wind turbine
0

1
0
0
0
power is either zero or still insufficient to supply the load. In this case,
the batteries will assume the relay and supply the load, either by them or
-339,3
330,7

in compensation with the power of the wind turbine.

670

M3
<0
0
0
1
1
From the management algorithm, six different operating modes can
be identified. The different powers involved are: the power supplied by
334,5
334,5

the wind generator (PTb), the power supplied or required by the battery
M5
>0
0

1
0
0
0

for compensation or recharge respectively (PBat) and the power required

by the load (Pload). The proposed power management control strategy
-123,5

presented in Table 7.
546,5
670

M3
<0

The different switches are defined as follows (Fig. 28). The flowchart
0
0
1
1

of the proposed PMC is given in Figs.29 and 30 shows the power

waveforms of the different sources with the different switches. Fig. 30
180
180

M1
>0
0

0
0
0
0

represents obtained powers waveforms with the different switches. The

different powers are the power supplied by the wind generator PTb (in
574,8
-95,2

green color), the power supplied or required by the battery for

670

M3
<0
0
0
1
1

compensation or recharge respectively PBat (in red color) and the power
required by the load Pload (in sky blue color). In this figure, it is repre­
715,5
45,5
670

sented the different powers under two different profiles to show the
M6
>0
1
0
0
1

effectiveness of the proposed PMC to both speed profiles and to

demonstrate that the proposed PMC doesn’t depend on the type of the
620,5
-49,5
670

speed profile.
M3
<0
0
0
1
1

A variable load profile ranging from 0 to a maximum peak power of

694.4 W has been chosen in purpose to operate all the switches and to
692,4

694,4
2 >0

M6

highlight explicitly the different modes that can be ensured by the

1
0
0
1

proposed power management strategy.

0 =0
670

670

Table 8 presents the results of the different power calculations

M2
0
1
0
1

performed.
When ΔP<0 we have:
693,7
23,7
670

PBat = ΔP = PLoad − PTb < 0 that corresponds to Mode 3 (M3)

M6
>0
1
0
0
1

Or PLoad = 0 PBat = ΔP = − PTb In this case, the battery is charging

and this corresponds to Mode 5 (M5)
652,4
-17,6
670

M3

When ΔP=0 we have:

<0
0
0
1
1

PLoad = 0 that corresponds to Mode 1 (M1) or PBat = ΔP = PLoad − PTb

=0 which corresponds to Mode 2 (M2)
688,6

690,6

M4
1.4
>0

When ΔP>0 we have:

0
1
0
1

PTb > PLoad so PBat = ΔP > 0 the battery is charged and this corre­
694,4

sponds to Mode 4 (M4) or: PTb > PLoad so PBat = ΔP > 0 the battery is
24,4
670

M6
>0

not charged, which is corresponds to Mode 6 (M6)

1
0
0
1

Fig. 31 shows the different developed powers versus the different

694,4

692,4
-2 <0

operating modes. One can notice that all the six scheduled modes have
M3
0
0
1
1

appeared which reflects the effectiveness of the developed power

management strategy.
1129

1129
0 =0

M2

In the above figure, The developed turbine power PTb and battery
0
1
0
1

power PBat are respectively highlighted in green and red colors. Load
1150

power is represented in light blue color. One can notice that due to wind
671

479

M6
>0
1
0
0
1

turbine power maximization, battery power usage is reduced. Negative

battery power corresponds to its charge and the positive value to its
-670

discharge.
670

M3
<0
0

0
0
1
1

It can be observed that with the chosen profile, all the modes have
appeared (M1, M2,…). This proves that the proposed PMC has well-
Power calculations.

0 =0

M1

managed the different sources. Also, a good sizing was performed and
0

0
0
0
0

this is why batteries have not been requested too much. Moreover, with
Pload (W)

the proposed hybrid MPPT method Hyb(HCS/FLC), there has been a

ΔP=Pbat
PTb (W)
Table 8

Modes
(W)

increase in wind turbine power, so less stress on the batteries.

K1
K2
K3
K4

21
S. Belaid et al. Journal of Energy Storage 45 (2022) 103613

Fig. 31. Battery and turbine developed powers versus operating modes.

7. Conclusion References

The contribution of this work is the combination of the proposed [1] J.F. Manwell, J.G. McGowan, A.L. Rogers, Wind Energy Explained: Theory, Design
and Application, Willey Edition, 2010. ISBN: 978-0-470-01500-1 DOI:10.1002/
hybrid MPPT and the proposed power management strategy for the sake 9781119994367.
of enhancing the performances of the wind energy system with storage [2] S.O Amrouche, D. Rekioua, T. Rekioua, Overview of energy storage in renewable
battery. This paper shows the importance of MPPT hybridization. energy systems (2016), in: Proceedings of 2015 IEEE International Renewable and
Sustainable Energy Conference, IRSEC, 2015, https://doi.org/10.1109/
Hybrid MPPT selects the best control technique combination that pro­ IRSEC.2015.7454988 art. no. 7454988.
vides the maximum power value. This approach results in stress battery [3] S.R. Salkuti, C.M. Jung, Comparative analysis of storage techniques for a grid with
minimization. Two distinct wind speed profiles are used to test the renewable energy sources, Int. J. Eng. Technol. (UAE) 7 (3) (2018) 970–976. DOI:
10.14419/ijet.v7i3.12728.
effectiveness of the proposed strategy. The obtained results show, [4] M.M. Hossain, H. Ali, Futre research directions for the wind turbine generator
clearly, a better performance of Hybrid (Hyb(HCS/FLC)) method in system, Renewable Sustainable Energy Rev. 49 (2015) 481–489, https://doi.org/
terms of power, response time, efficiency and effect on the battery state 10.1016/j.rser.2015.04.126, september.
[5] P. Yadav, R. Kumar, S. Panda, C. Chang, An intelligent tuned harmony search
of charge level. The simulation results point out that the system per­
algorithm for optimization, Inf. Sci. (NY) 196 (2012) 47–72, https://doi.org/
formance is not affected by wind speed variations. The integration of the 10.1016/j.ins.2011.12.035.
proposed control in a water pumping system is a promising idea that is to [6] D. Karaboga, B. Basturk, A powerful and efficient algorithm for numeric
be considered as a future work. optimization: Artificial Bee Colony(ABC) algorithm, J. Global Optim. 39 (2007)
459–471, https://doi.org/10.1007/s10898-007-9149-x.
[7] A.R. Mehrabian, C. Lucas, A novel numerical optimization algorithm inspired from
CRediT authorship contribution statement weed colonization, Ecol. Inform. 1 (2006) 355–366, https://doi.org/10.1016/j.
ecoinf.2006.07.003.
[8] D. Kumar, K. Chatterjee, A review of conventional and advanced MPPT algorithms
Saloua Belaid: Investigation, Methodology, Software. Djamila for wind energy systems, Renewable Sustainable Energy Rev. 55 (2016) 957–970,
Rekioua: Methodology, Supervision, Writing – original draft. Adel https://doi.org/10.1016/j.rser.2015.11.013.
Oubelaid: Data curation, Formal analysis. Djamel Ziane: Investigation, [9] O. Elbeji, M. Hannachi, M. Benhamed, Artificial neural network-based sensorless
control of wind energy conversion system driving a permanent magnet
Visualization. Toufik Rekioua: Writing – review & editing, Funding synchronous generator, Wind Eng. (A) 45 (3) (2021), https://doi.org/10.1177/
acquisition. 0309524X20903252.
[10] S.H. Moon, B.G. Park, J.W. Kim, J.M. Kim, Maximum power-point tracking control
using perturb and observe algorithm for tidal current generation system, Int. J.
Declaration of Competing Interest Precis. Eng. Manuf. - Green Technol. 7 (4) (2020) 849–858. DOI:10.1007/s40684-
019-00162-w.
The authors declare that they have no known competing financial [11] Z. Cui, S.Li L.Song, Maximum power point tracking strategy for a new wind power
system and its design details, IEEE Trans. Energy Convers. 8969 (c) (2017)
interests or personal relationships that could have appeared to influence
1063–1071. DOI:10.1109/TEC.2017.2694008.
the work reported in this paper. [12] M.A. Soliman, H. Hasanien, H.Z. Azazi, E.E. El-kholy, S.A. Mahmoud, An adaptive
fuzzy logic control strategy for performance enhancement of a grid-connected
Acknowledgement PMSG-based wind turbine, IEEE Trans. Ind. Inf. (2019), 1-12018 DOI:https://doi.
org/10.1109/TII.2018.2875922.
[13] C. Wang, Z. Zhang, S.Gholami Farkoush, Modeling and analysis of a microgrid
This work has been supported by the « Direction Générale de la considering the uncertainty in renewable energy resources, energy storage systems
Recherche Scientifique et du Développement Technologique (DGRST) and demand management in electrical retail market, J. Energy Storage (2020),
102111, https://doi.org/10.1016/j.est.2020.102111.
(Algeria) ». [14] A. Ur Rahman, S.S. Zehra, H. Armghan, Fuzzy supertwisting sliding mode-based
energy management and control of hybrid energy storage system in electric vehicle
considering fuel economy, J. Energy Storage 24 (2021), 102468, https://doi.org/
10.1016/j.est.2021.102468.

22
S. Belaid et al. Journal of Energy Storage 45 (2022) 103613

[15] M.T. Hussain, N. Bin Sulaiman, M. Jabir, Optimal management strategies to solve Proceedings 2, 2003, pp. 861–866, https://doi.org/10.1109/PTC.2003.1304660,
issues of grid having electric vehicles (EV): a review, J. Energy Storage 33 (2021), art. no.
102114, https://doi.org/10.1016/j.est.2020.102114. [23] D. Rekioua, T. Rekioua, DSP-controlled direct torque control of induction machines
[16] L. Meng, E.R. Sanseverino, A. Luna, T. Dragicevic, J.C. Vasquez, J.M. Guerrero, based on modulated hysteresis control, in: Proceedings of the International
Microgrid supervisory controllers and energy management systems: a literature Conference on Microelectronics, ICM, art. no. 5418603, 2009, pp. 378–381. DOI:
review, Renewable Sustainable Energy Rev. 60 (2016) 1263–1273, https://doi: 10.1109/ICM.2009.5418603.
10.1016/j.rser.2016.03.003. [24] D. Rekioua, Energy management for PV installations, Adv. Renew. Energies Power
[17] L. Olatomiwa, S. Mekhilef, M.S. Ismail, M. Moghavvemi, Energy management Technol. 1 (2018) 349–369, https://doi.org/10.1016/B978-0-12-812959-3.00011-
strategies in hybrid renewable energy systems: a review, Renew. Sustain. Energy 3.
Rev. 62 (2016) 821–835. DOI: https://doi:10.1016/j.rser.2016.05.040. [25] C. Serir, D. Rekioua, N. Mezzai, S. Bacha, Supervisor control and optimization of
[18] M.S. Ismail, M. Moghavvemi, T.M.I Mahlia, K.M. Muttaqi, S. Moghavvemi, multi-sources pumping system with battery storage, Int. J. Hydrogen Energy 41
Effective utilization of excess energy in standalone hybrid renewable energy (45) (2016) 20974–20986. DOI: https://doi.org/10.1016/j.ijhydene.2016.05.096.
systems for improving comfort ability and reducing cost of energy: a review and [26] Z. Mokrani, D. Rekioua, T. Rekioua, Modeling, control and power management of
analysis, Renew. Sustain. Energy Rev. 42 (2015) 726–734, https://doi.org/ hybrid photovoltaic fuel cells with battery bank supplying electric vehicle, Int. J.
10.1016/j.rser.2014.10.051. Hydrogen Energy 39 (27) (2014) 15178–15187. DOI: https://doi.org/10.1016/j.
[19] M. Farhadi, O. Mohammed, Adaptive energy management in redundant hybrid DC ijhydene.2014.03.215.
micro grid for pulse load mitigation, IEEE Trans. Smart Grid 6 (2015) 54–62, [27] N. Mebarki, T. Rekioua, Z. Mokrani, D. Rekioua, S. Bacha, PEM fuel cell/battery
https://doi.org/10.1109/TSG.2014.2347253. storage system supplying electric vehicle, Int. J. Hydrogen Energy 41 (45) (2016)
[20] S. Abedi, A. Alimardani, G. Gharehpetian, G. Riahy, S. Hosseinian, 20993–21005. DOI: https://doi.org/10.1016/j.ijhydene.2016.05.208.
A comprehensive method for optimal power management and design of hybrid [28] N.Benyahia S.Tamalouzt, T. Rekioua, D. Rekioua, R. Abdessemed, Performances
RES-based autonomous energy systems, Renew. Sustain. Energy Rev. 16 (3) (2012) analysis of WT-DFIG with PV and fuel cell hybrid power sources system associated
1577–1587, https://doi.org/10.1016/j.rser.2011.11.030. with hydrogen storage hybrid energy system, Int. J. Hydrogen Energy 41 (45)
[21] B. Robyns, A. Davigny, C. Saudemont, Methodologies for supervision of hybrid (2016) 21006–21021. DOI: https://doi.org/10.1016/j.ijhydene.2016.06.163.
energy sources based on storage systems – a survey, Math. Comput. Simul. 91 [29] H. Hassani, F. Zaouche, D. Rekioua, S. Belaid, S. Bacha, Feasibility of a standalone
(2013) 52–71, https://doi.org/10.1016/j.matcom.2012.06.014. photovoltaic/battery system with hydrogen production, J. Energy Storage 31
[22] T. Rekioua, D. Rekioua Direct torque control strategy of permanent magnet (2020), 101644, https://doi.org/10.1016/j.est.2020.101644. October.
synchronous machines, in: 2003 IEEE Bologna PowerTech - Conference

23