J Electr Eng Technol.

2018; 13(2): 659-668 ISSN(Print) 1975-0102

http://doi.org/10.5370/JEET.2018.13.2.659 ISSN(Online) 2093-7423

Techno-Economic Optimization of a Grid-Connected Hybrid Energy

System Considering Voltage Fluctuation

Samia Saib†, Ahmed Gherbi*, Abdelhamid Kaabeche** and Ramazan Bayindir***

Abstract – This paper proposes an optimization approach of a grid-connected photovoltaic and wind
hybrid energy system including energy storage considering voltage fluctuation in the electricity grid. A
techno-economic analysis is carried out in order to minimize the size of hybrid system by considering
the benefit-cost. Lithium-ion battery type is used for both managing the electricity selling to the grid
and reducing voltage fluctuation. A new technique is developed to limit the voltage perturbation
caused by the solar irradiance and the wind speed through determining the state-of-charge of battery
for every hour of a day. Improved particle swarm optimization (PSO) methods, referred to as FC-
VACPSO which combines Fast Convergence Particle Swarm Optimization (FCPSO) method and
Variable Acceleration Coefficient Based Particle Swarm Optimization (VACPSO) method are used to
solve the optimization problem. A comparative study has been performed between standard PSO
method and PSO based methods to extract the best size with the benefit cost. A sensitivity analysis has
been studied for different kinds and costs of batteries, by considering variable and constant state-of-
charge of battery. The simulations, performed under Matlab environment, yield good results using the
FC-VACPSO method regarding the convergence and the benefit cost of the hybrid system.

Keywords: Hybrid energy systems, Optimization, Storage system, State-of-charge, Voltage fluctuation

1. Introduction particularly when the amount of solar and wind power

integration increases or the grid is not strong enough to
Nowadays, renewable energy has gained widespread handle rapid changes in generation levels. Hence, a
acceptance in several areas, predominantly photovoltaic combination of these two sources improves overall energy
(PV) and wind systems as grid-connected or autonomous output, especially when they are connected to the grid. A
system modes. Integration of storage energy battery to proper optimization is required to ensure having optimal
these hybrid systems is very important and necessary for number and size of PV and wind turbine. A review of
storing energy and assuring system stability. Hybrid optimization of hybrid renewable energy system with more
renewable energy systems (HRES) are designed for the focus on wind and PV systems was carried out in [4,5]
electrical power generation using a combination of a using different optimization methods, i.e., conventional,
number of power generation components e.g. wind turbine, unconventional, hybrid and software tools used in optimum
PV and/or other conventional generators along with storage sizing of hybrid wind-solar systems. Reference [6]
batteries. HRES capture the best features of each energy discussed the effect of voltage fluctuation of a grid-
resource and can provide “grid-quality” electricity as connected wind farm. Stored energy by vanadium redox
well as improving the overall economy and reliability of flow type of battery was used and connected with the
renewable power generation to supply its load [1]. To wind farm to control the stability and the power quality
precisely sizing the different devices of HRES, simulations of the grid. Simulation and experimental results have
of the system under real operating conditions, such as demonstrated that the injected power to the grid is stable
appropriate weather, insolation, wind speed and loads, are and does not fluctuate even the wind power was fluctuated.
necessary [2-4]. Nacer et al. [7] presented a techno-economic optimization
Solar and wind powers are naturally intermittent. They of a solar/wind hybrid energy system connected to a grid
can create technical challenges to the grid power supply, for a cattle farm located in the north of Algerian desert.
Using HOMER software, optimization of the net present
† Corresponding Author: Laboratory of Automatics Dept. of Electrical
Engineering, Ferhat Abbas Setif1 University, Setif, Algeria. cost of hybrid system was done and three scenarios were
([email protected]) proposed to estimate system configuration relating its
* Laboratory of Automatics, Dept. of Electrical Engineering, Ferhat effect on techno-economic study of the farm. Kaabeche
Abbas Setif1 University, Setif, Algeria. ([email protected]).
** Renewable Energy Development Center, Bouzareah, Algiers, Algeria. and Ibtiouen [8] improved an optimal sizing design
([email protected]). through a repeated process to minimize the cost and the
*** Department of Electrical and Electronics Engineering, Gazi University, size of autonomous PV/wind/diesel/battery hybrid system.
Ankara, Turkey. ([email protected]).
Received: December 1, 2016; Accepted: December 31, 2017
Simulation results have demonstrated that the hybrid

659
Copyright ⓒ The Korean Institute of Electrical Engineers
This is an Open-Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/
licenses/by-nc/3.0/) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.
 Techno-Economic Optimization of a Grid-Connected Hybrid Energy System Considering Voltage Fluctuation

PV/wind/diesel/battery system is an economical and

reliable system than solar/wind/battery system and diesel
generator alone. In reference [9], the authors studied an
optimal control and sizing of stand-alone wind/PV/diesel
system including battery storage energy. Particle swarm
optimization (PSO) method was used to solve the
optimization problem and compared with other improved
based PSO algorithms. The obtained results were evaluated
through an economic factor, the system cost, to treat the
real-world system. Recently, Durairasan et al [10]
proposed a new hybrid technique for locating and sizing of
renewable energy like wind and PV in power system. Fig. 1. Schematic diagram of grid-connected HRES
The proposed method consists on the combination of
performances of both the Biography Based Optimization
(BBO) and PSO techniques. These techniques were used 3. Modeling of Hybrid System
for optimizing the optimum location and capacity of the
DG sources for radial distribution network. Initially, the 3.1 Photovoltaic system
availability of these sources was analyzed in the 24 hours
using ANN technique. After that, the voltage, power and Electric power, produced by the PV system, depends on
power loss were analyzed in the normal and faulty the solar irradiance affected by weather conditions. It is
conditions. The optimal location and capacity of the DG computed using the following expression [11, 12]:
sources were initially determined by using BBO, then the
BBO input parameters were classified into the sub GT
Ppv = Pmax,ref ´ é1 + g (Tc - 25)ùû
GT ,ref ë
parameters and allowed as the PSO input. A PSO (1)
algorithm was used to find the optimum location and
capacity of DG. Finally, the best results were taken among
the techniques. where GT is the solar irradiance in the current time step
This paper presents a new optimization approach of a (Wh/m2), GT ,ref is the solar irradiance at reference of
grid-connected HRES in which the battery energy storage 1000Wh/m2, Ppv is the output power of PV array (kW),
Pmax,ref is the maximum power at GT ,ref of PV array (kW);
system (BESS) is used for both managing the electricity
g is the temperature coefficient of maximum power, taken
selling to the grid, and preventing voltage fluctuation in the
grid. Two strategy cases are considered: storage and selling to be -0.0035 °C; and T C is the ambient temperature (°C).
energy. PSO metaheuristic method is used to solve the
optimization problem, and then compared with other PSO 3.2 Wind system
based methods. The simulation is performed under Matlab
environment. Data of wind speed and solar irradiation The output electric power of a wind turbine can be
are extracted from the site of the Renewable Energy expressed by the following set of equations [13-15]:
Development Center (CDER) in Bouzaréah City, situated
at (36°48’N, 3°1’E, 345m), in north of Algeria. ì
ï
ï0 vw < vci , vw > vco
ï 3
2. Grid-connected Hybrid System Overview ï æv -v ö
Pw = í Pw,max ´ çç w ci ÷÷ vci £ vw < v p (2)
ï è v p - vci ø
A schematic diagram of grid-connected PV/wind hybrid ï
ïP æ Pf0 - Pw,max ö
ï w,max ( w p ) çç vco - v p
system with battery storage energy is illustrated in Fig. 1. + v -v ÷÷ v p £ vw £ v f 0
The overall hybrid system is linked to a common DC bus î è ø
through a converter DC/DC for the PV array and AC/DC
for the wind system in order to obtain the maximum power where Pw ,max , Pf 0 are the output powers at rated and cut-
point. Another DC/DC converter is used for the battery out speeds (kW); Pw is the output power of wind turbine
system to regulate the DC bus voltage. An inverter DC/AC (kW) ; v ci , v co , v p are respectively cut-in, cut-out and
is connected between the interface of DC bus and the grid. rated wind speeds (m/s); v w is the wind speed at the height
The operation strategy system consists to inject power to of the wind generator hub hhub , estimated by:
the grid. The battery energy storage system (BESS) is used
a
to charge and store energy produced by the PV/wind hybrid æh ö
energy system and to discharge this energy in order to sell vw = vw,meas ç hub ÷ (3)
it to the utility. è hmeas ø

660 │ J Electr Eng Technol.2018; 13(2): 659-668

 Samia Saib, Ahmed Gherbi, Abdelhamid Kaabeche and Ramazan Bayindir

v w ,meas is the wind speed measured at the reference 5. System Strategy and Energy Management
height hmeas , (m/s); a depends on temperature, time of
day, season and pressure , and is taken to be 0.14 [16]. The operation strategies are considered to store and sell
The power, generated by the hybrid system, is given by: energy by the battery system and to estimate the state of
charge of battery for each hour in order to reduce voltage
æP ö æ Ppr ö perturbation.
Phyb = N w ç wr ÷÷ Pw + N pv çç ÷÷ Ppv (4)
çP
è gw ø è Pgp ø
5.1 Storage strategy
where Nw, Npv are respectively the numbers of wind and PV
systems; Pwr , Ppr , Pgw , Pgp are rated power and generated In this strategy, the producer of hybrid (PV/wind) energy
power of wind generator and PV array in kW. system charges the BESS, when the buying prices of the
micro-grid are decreased than the desired price (Cgmax=
3.3 Battery energy storage system 0.3$/kWh). The power of battery charging can be
computed as follows:
A BESS is used for managing the energy between the
HRES and the micro-grid. It stores energy, consumes it SOCmax ( t + 1) - SOC ( t )
Pch ,lim ( t ) = Cdb N b (9)
when the grid needs, reduces voltage fluctuation and h
ensures system stabilization. To achieve this last benefit, Pch ( t ) = Pa ( t ) (10)
an average power, in kW, of a HRES is calculated for
n observations by: Pch (t ) = min ( Pch ( t ) , Pch ,lim ( t ) ) (11)

1 n
Pa = å ( Phyb )i
n i=1
(5) When the state-of-charge of BESS is lower than
SOC min (t +1) , the power of charge is calculated by:

The energy of the battery, in kWh, can be written as: SOCmin (t + 1) - SOC (t )
Pch ,min (t ) = Cdb N b (12)
h
E = ( Phyb - Pa ) Dt (6)
Pch (t ) = max ( Pch (t ) , Pch ,min ( t ) ) (13)
The minimum and maximum values of the battery’s Pch (t ) = min ( Pch (t ) , N b Cdb cch ) (14)
state-of-charge are given by:
Then, the battery’s state-of-charge can be computed as
Emin
SOCmin = follows:
N b Cdb
(7)
æ ö h Pch (t )
SOCmax = 1 - çç Emax ÷÷ SOC ( t + 1) = SOC ( t ) + (15)
N C
è b db ø Cdb N b

where Nb is the number of batteries; Cdb is the battery 5.2 Sell strategy
capacity (kWh) and E max = max( E ); E min = min( E ) .
This strategy consists to sell the stored energy in the
batteries to the micro-grid, when the buying prices of the
4. Voltage Fluctuation micro-grid are increased compared the price required
(Cgmax=0.3$/kWh). The expression of power discharge by
The use of PV and wind systems may cause a voltage battery can be given by:
fluctuation since they are influenced by weather conditions.
The voltage perturbation can lead to instability of the Pdch ,lim (t ) = Cdb N bh ( Soc ( t ) - Socmin (t + 1) ) (16)
entire grid. An estimation of voltage change through the
power line is computed, in percent (%), by the following Pdch ( t ) = Pdch ,lim ( t ) (17)
equation:
When the state-of-charge SOC(t) is higher than
R DP + X DQ SOC max (t + 1) , the discharge power is determined by:
dv = ´ 100% (8)
V2
Pdch ,min ( t ) = Cdb N b
( Soc ( t ) - Soc ( t + 1) )
max
(18)
where R is the resistance of grid ( W ), X is the inductive h
reactance of grid ( W ), DP and DQ are active and
reactive power variations (kW), V is voltage of the grid (V). Pdch ( t ) = max ( Pdch ( t ) , Pdch ,min ( t ) ) (19)

http://www.jeet.or.kr │ 661
 Techno-Economic Optimization of a Grid-Connected Hybrid Energy System Considering Voltage Fluctuation

Pdch ( t ) = min ( Pdch ( t ) , N b Cdb cdch ) (20) With C g (t ) the buying prices of the grid for each hour
during a day.
With cch , c dch are the limit of the charge and discharge The benefit cost can be determined by:
capacity of the battery bank, h is the charge/discharge
efficiency of the battery bank take to be 0.9. CB = NPC - CG (27)
The state-of-charge of battery can be determined by:
The buying electricity prices of the grid are presented in
Pdch (t ) [24]. The solution of the optimization problem considers
SOC ( t + 1) = SOC ( t ) - (21)
Cdb N bh the following constraints:

The total output power of the PV/wind/battery hybrid cch .Cdb £ Cdb £ cdch .Cdb (28)
system can be computed by: SOCmin ( t ) < SOC ( t ) < SOCmax ( t ) (29)

Ptot ( t ) = Pa ( t ) + Pdch ( t ) - Pch ( t ) SOC ( t + 1) < SOC ( t ) (30)

(22)
Ni > 0 (31)
5.3 Economic optimization
The flow chart of the energy management between the
(PV/wind) hybrid system and the micro grid through the
An economic optimization has been applied, for the
battery is illustrated in Fig. 2.
hybrid PV/wind/battery energy system, to select the optimal
sizing with saving a benefit cost of system through selling
energy to the micro-grid. The suggested objective function
based on the capital cost, replacement and maintenance
cost and the grid cost during 20 years of life, is given by
[17, 18]:
k
NPC = å N i ´ ( CCi + K ´ RCi + PWA ´ MCi ) (23)
i

Where NPC is the net present cost of the system

components, k is number of the renewable energy sources,
Ni is the capacity of the ith unit in (kW); CC, RC, MC are
respectively the capital, replacement and maintenance
costs of the components. Replacement cost represents the
exchange of cost for each component in its end of life,
with K defines the constant of this cost and is evaluated by
[19, 20]:
L1
1
K =å (24)
(1 + i )
k ´L2
i =1
r

With L1, L2 are respectively the number of the life span

and replacement of HRES for the life span of the project,
ir is the real interest rate and is equal to 6%, PWA is a
coefficient applied to convert the total cost of operation
and maintenance to the present cost and computed as Fig. 2. Flowchart of the techno-economic optimization
mentioned in the following equation, for L the lifetime of
the project, [20-23]:
6. Optimization method
(1 + i ) - 1
L

PWA = r
(25) A metaheuristic method is applied to solve the problem
i (1 + i )
L
r r
of optimization, in order to minimize the size of HRES and
gain cost. A particle swarm optimization algorithm is used
The grid cost can be formulated by: to simulate this problem since it provides a good result
and execution time compared to the other metaheuristic
CG = Ptot ( t ) ´ C g ( t ) (26) methods. PSO algorithm uses a number of particles,

662 │ J Electr Eng Technol.2018; 13(2): 659-668

 Samia Saib, Ahmed Gherbi, Abdelhamid Kaabeche and Ramazan Bayindir

which move through the solution space, and are evaluated best particle position in the swarm and the ith particle’s
according to some fitness criterion after each step to get an position.
optimum [25]. The concept of swarm involves a several The inertia weight w is used to allow a particle to
units that are able of interacting with every other solution balance between global and local detections. In general, the
in a complex behavior, and the intelligence suggests that inertia weight w is calculated according to the following
this approach is successful. equation:
In this work, other PSO based methods, such as Fast
Convergence (FCPSO), Variable Acceleration Coefficient æ w - wmin ö
w = wmax - ç max ÷ iter (34)
(VACPSO) and FC-VACPSO, are used and compared with è itermax ø
standard PSO technique in order to select the best method
providing the best result for the optimization problem. Where iter is the current number of iterations and
itermax is the maximum number of iterations.
6.1 Standard particle swarm optimization
6.2 Fast convergence based particle swarm optimi-
In PSO algorithm, each random solution of the studied
zation
problem corresponds to an artificial particle that moves
together with its kinds in the super-organism. Particles in
To improve the PSO performance, this method introduces
a population adapt by returning stochastically toward
a new parameter, called particle mean dimension (Pmd).
previously successful regions in the search space, and are
The basic PSO can converge rapidly, but apt to drop into
influenced by the successes of their topological neighbors.
local minimum solution easily. Then, to solve this problem,
In other words, each individual particle represents a
a study has been made considering the following
potential solution which moves its position in the search
improvements.
space and updates its velocity according to its own flying
When the particle updates from the gth generation to
experience and that of its neighbor’s, aiming for a better
(g+1)th generation, through pursue the Pbest ,i and G best ,
position for itself at the next move. The PSO algorithm is
the particle can follow the Pmd i which is selected from the
presented as follows:
particles search. The new parameter Pmd i of ith particle
1. Each particle i has a current position in search space Xi, expression is generated by the following equation:
a current velocity Vi and its best position in search space
P i. Pmd i = (x
i1
+ xi 2 + ............. + xiD ) D (35)
2. The individual best position corresponds to the position
in search space, where particle i presents the smallest
Where D is the dimension of particles in the search
error as determined by the objective function (fitness).
space and x is the particles swarm, the velocity is
3. The global best position represents the position yielding
calculated by the following expression [26].
the lowest error among the entire Pi’s.

To reach the optimal point, particles must update their (

Vi k +1 = w.Vi k + c1 .rand1 Pbest ,i - X ik )
next displacement according to their own velocities. The (36)
+ c2 .rand 2 ( Gbest - X k
) + c .rand ( Pmd - X ik )
velocities and positions are defined by: i 3 3 i

Vi k +1 = w.Vi k + c1 .rand1 ( Pbest ,i - X ik ) + c2 .rand 2 ( Gbest - X ik ) Where c3 is the average best coefficient, with
(c 1
+ c 2 + c 3 ) ³ 4 ; rand 1 , rand 2 and rand 3 are random
(32) numbers between 0 and 1.
X ik +1 = X ik + Vi k +1 (33) Then, including Pmd i into the velocity expression,
Pbest ,i , G best and Pmd i give information to the next
Where V i k is the velocity of particle at kth iteration; generation together and increase the number of information.
X ik is the current position of particle at kth iteration; w Therefore, it is possible to get rapidly the optimal solution.
is the inertia weight , c1 and c2 are factors acceleration At the same time, the suggested weight factor of Pmd i is
terms, which pull every particle towards the Pbest ,i and weak, which is equivalent to disturbance information and
G best positions, rand 1 and rand 2 are two random functions increases the diversity of particles. G best is used to enhance
in the range of [0,1]; Pbest ,i is the best previous experience the convergence. The parameter Pmd i can move the
of i-th particle that is recorded and G best is the best particle particles to a better position and decrease the attraction of
among the entire population. the G best to local minima.
Eq. (32) is used for computing the ith particle’s velocity
considering the particle’s previous velocity and the 6.3 PSO based time varying acceleration coefficients
distance between the particle’s best previous and the current
position. The second part defining the distance between the The PSO technique with time varying inertia weight

http://www.jeet.or.kr │ 663
 Techno-Economic Optimization of a Grid-Connected Hybrid Energy System Considering Voltage Fluctuation

provides good solution location at a significantly fast rate.

The concept of time varying acceleration coefficients
ì
(
ïc1 = c1i + c1 f - c1i
ï
) iteritermax
(TVAC) based PSO is to improve and converge towards
ïï iter
the global optima in the optimization formulation in the
swarm. This is achieved by time varying the acceleration
(
íc2 = c2i + c2 f - c2i ) itermax
(39)
ï
coefficients c1 and c2 in such a manner that the cognitive ï
components are reduced while the social component is (
ïc3 = c3i + c3 f - c3i
ïî
) iter
iter
max
increased in the search space. The expressions of the
accelerations coefficients are:
The inertia weight, computed through Eq. (38) and the
PSO algorithm parameters for each method, are shown in
ì
( iter
)
ïc1 = c1i + c1 f - c1i iter
ï
Table 1.
max The parameters of each component of the hybrid energy
í (37)
system, with the different types of battery bank used in this
ïî 2 2i (
ïc = c + c - c
2f 2i
iter
)
itermax paper are presented in Table 2.

Where c 1i , c1f are the initial and final the social

acceleration coefficients respectively, and c 2i , c 2f are the
7. Simulation Results and Discussion
initial and final cognitive coefficients respectively. The
inertia weight is formulated as follows [27]: In this study, standard PSO and improved PSO methods
are applied for simulating the problem formulation of
HRES connected to the utility grid including BESS with
æ iter - iter ö
w = wmin + (wmax - wmin ) ç max ÷ (38) the aim to obtain an optimum with a better solution.
è itermax ø Simulations are performed using Matlab software. Data of
solar irradiation and wind speed for the year 2014 are used
6.4 PSO based fast convergence with time varying
acceleration coefficients

This novel modified PSO method is proposed in order to

enhance the solution quality and robustness of standard
PSO. It introduces a new parameter named ( Pmd ) in the
velocity formula by considering a time varying of
acceleration coefficients c1, c2 and c3 in Eq. (36). The
acceleration coefficients are varied according the following
formulas: Fig. 3. Wind speed profile for one day

Table 1. Parameters of PSO algorithms

Population = 30 Iteration =50
PSO & wmin wmax c1 c2 c3
FCPSO 0.1 0.9 1.5 1.5 1.5
wmin wmax c1f c1i c2f c2i
VACPSO
0.4 0.9 0.5 2.5 2.5 0.5
VACPSO wmin wmax c1f c1i c2f c2i c3f c3i
& FC-VA
CPSO 0.1 0.9 0.5 2.5 2.5 0.5 1.5 0.5
Fig. 4. Solar irradiation profile for one day

Table 2. Parameters of individual components of HRES

Battery types
PV array Wind turbine Lithium-ion Lead acid Ni-Cd VRB
Type Sanyo HIP 225-HDE1
Pmax, kW 2.55 6.5
Rated power 10kW 60kW 10.8kWh
Capital cost 2000$/kW 3000 $/kW 1500 $/kWh 300 $/kWh 1200 $/kWh 600 $/kWh
Replacement cost 0 0 1500 $/kWh 300 $/kWh 1200 $/kWh 600 $/kWh
Maintenance cost 20 $/kW 90 $/kW 50 $/kWh 10 $/kWh 40 $/kWh 20 $/kWh
Life of time (year) 20 20 15 5 20 10

664 │ J Electr Eng Technol.2018; 13(2): 659-668

 Samia Saib, Ahmed Gherbi, Abdelhamid Kaabeche and Ramazan Bayindir

Fig. 5. Total energy of (PV/wind) hybrid system for one

day Fig. 9. Convergence of the metaheuristic methods

Table 3. Sizing and cost of hybrid system

Benefit cost
Method Npv Nw Nb
(MUS$)
PSO 31 11 08 51.3
FCPSO 15 23 20 52
VACPSO 18 20 28 53.0
FC-VACPSO 24 19 24 55

Table 4. Variable state of charge – Case 1-

Fig. 6. Voltage change without connection of battery Battery Price benefit cost
Npv Nw Nb
type ($/kWh) (MUS$)
1500 22 08 60 38.3
Li-ion
1500x5 04 46 96 77.0
1200 11 21 44 49.8
Ni-Cd
1200x5 09 09 32 27.0
600 25 22 56 69.0
VRB
600x5 05 33 28 62.8
300 49 05 44 60.6
Lead-acid
300x5 06 09 40 26.6

Fig. 7. Voltage change with connection of battery Table 5. Constant state of charge – Case 2
Battery Price benefit cost
Npv Nw Nb
type ($/kWh) (MUS$)
1500 22 08 60 36.1
Li-ion
1500x5 04 46 96 79.0
1200 05 16 08 34.5
Ni-Cd
1200x5 09 09 32 26.4
600 26 22 56 67.0
VRB
600x5 05 34 28 63.6
300 31 25 08 76.0
Lead-acid
300x5 46 05 44 52.0

Fig. 8. State-of-charge of the battery benefit cost of system is illustrated in the Table 3.

in this study. Profiles of solar irradiance and of wind speed 7.1 Sensitivity analysis
for one day in January month are chosen and applied in this
simulation. A sensitivity analysis has been performed about the type
The performance of metaheuristic method, such as PSO and cost of the battery. Four types of batteries are used
algorithm, cannot be evaluated by a single run due to the such as lead acid, nickel cadmium (Ni-Cd), lithium-ion
inherent randomness involved in the optimization process. (Li-ion) and vanadium redox flow (VRB) [28, 29] as
Therefore, the robustness of each PSO algorithm is shown in the Table 2. Then, these prices have been increased
evaluated based on many different runs. The algorithm is five times and applied to the optimization process with the
robust when it capable to produce consistence results. The aim to evaluate the sensitivity of use of the battery. In this
convergence of the above described metaheuristic methods state, two study cases have been considered about the state-
is well shown in Fig. 9. of-charge. In case 1, the state-of-charge has been taken
The number of optimal sizing of hybrid energy system variable which is calculated for each hour. In case 2, the
of PV panel, wind turbine and battery system with the state-of-charge is taken to be constant with a maximal and

http://www.jeet.or.kr │ 665
 Techno-Economic Optimization of a Grid-Connected Hybrid Energy System Considering Voltage Fluctuation

minimal value equal to 1 and 0 respectively. The obtained that the minimal and maximal state-of-charge are calculated
results show the difference between both cases about the for each hour for reducing the voltage fluctuation.
sizing of hybrid system and the benefit cost as mentioned
in the tables 4 and 5.
8. Conclusion
7.2 Discussions
In this paper, a techno-economic optimization has been
Profiles of the wind speed and solar irradiance are treated on a hybrid (PV/wind) system, with a BESS
illustrated in Figs 3 and 4 for one day with a time of five connected to the micro-grid. The optimization study has
minutes per five minutes, and chosen under agreeable been performed by considering the voltage fluctuation of
weather condition. The total energy of hybrid (PV/wind) the hybrid system caused by the solar irradiation and wind
system for one day is illustrated in Fig. 5. From this figure, speed. The insertion of the BESS to the hybrid system has
a voltage perturbation appears. In this case, a study has a significant benefit as it reduces voltage fluctuation, stores
been achieved for the (PV/wind) energy system with and energy and discharges it when the grid needs. Overall,
without connection of a battery during one day with five these advantages have been confirmed and demonstrated in
minutes per five minutes of time. The voltage change, in the simulation results. A new technique has been done by
percent, computed using Eq. (8) is illustrated in Figs. 6 and the battery system about storing and selling energy to the
7. It can be noted that the integration of lithium-ion type grid, and defining the state-of-charge of battery for each
battery with the PV/wind hybrid energy system is very hour for one day in order to reduce the voltage perturbation.
important and has a great advantage to reduce the voltage The simulation results show that Li-ion battery type has a
fluctuation as depicted in Fig. 7. good performance in the optimization study, and is better
The state-of-charge of battery has been calculated for than the other battery types. To get the best size and gain of
each hour of one day and the result has shown that it varies the cost of the hybrid system, it is desirable to select a
during the twenty-four hours depending on the charge reasonable price as well as a battery type.
and discharge of battery in order to reduce the voltage A metaheuristic method has been applied to solve the
perturbation as mentioned in Fig. 8. Fig. 9 presents the optimization problem for the hybrid (PV/Wind/Battery)
convergence of PSO compared with the improved PSO connected to the micro-grid. A comparative study of basic
algorithms. It can be noted that the improved FC-VACPSO PSO and improved PSO algorithms, termed as FCPSO,
method gives a better solution regarding the convergence VACPSO and FC-VACPSO have solved the optimization
and the cost (55 MUS$) of the hybrid PV/wind/battery problem of the hybrid system. The simulation results
energy system than other methods as shown in Table 3. demonstrate clearly that the improved FC-VACPSO
The performance of this method is well proven as it method gives rise to a better convergence. The optimal
produces a better solution than PSO, FCPSO and solution of the studied system is selected using the FC-
VACPSO techniques. VACPSO algorithm. This method proves its efficiency and
A sensitivity study has been considered about the sizing it is better than PSO, FCPSO and VACPSO methods.
and the benefit cost of the (PV/Wind/Battery) hybrid The proposed strategy management applied in the
system. Taking into account different prices and types of optimization process proves its effectiveness for the hybrid
batteries, the simulation results show that the first case is renewable energy system (PV/Wind/Battery) which gives a
better than the second one regarding the sizing and the profitable system and it offers a benefit cost during its
benefit cost of the system, especially for the Li-ion battery lifetime.
type. Selected Li-ion battery type and the first case in this
study determine their efficiency and are perfect for the
sensitivity analysis and for the optimization study. The Acknowledgements
increase of the benefit cost depends on the rise of the sizing
of the hybrid system mostly the number of batteries. The work reported in this paper has been developed
Whenever the price of the Li-ion battery type is high, within the Laboratory of Automatics and it has been
number of sizing of the battery and the benefit cost supported by the CNEPRU university research project No.
increase, and vice versa for the other types. A01L07UN190120120019, conducted at the department of
It can be concluded that the variation of the system electrical engineering, University of Setif, Algeria. This
sizing depends on the battery price, and the increase of the support is gratefully acknowledged.
benefit cost depends on the rise of number of the hybrid
system components typically the number of batteries.
Therefore, in order to obtain an optimal size with an References
advantageous cost of the (PV/Wind/Battery) energy system,
or for another system with a BESS, it is preferred to choose [1] M. Nayeripour, M. Kheshti, Renewable Energy.
a reasonable price with type of battery. It should be noted Trends and Applications, Ed. InTech, Croatia, 2011.

666 │ J Electr Eng Technol.2018; 13(2): 659-668

 Samia Saib, Ahmed Gherbi, Abdelhamid Kaabeche and Ramazan Bayindir

[2] Akbar Maleki, Alireza Askarzadeh, “Artificial bee off grid hybrid PV/wind/diesel system with battery
swarm optimization for optimum sizing of a stan storage,” J Engin. Appl Scien, vol. 6, no. 4, pp. 23-39,
d-alone PV/WT/FC hybrid system considering LP 2014.
SP concept,” J. Sol Energy, vol. 107, pp. 227-235, [16] H. Yang, L. Lu, W. Zhou, “A Novel optimization
2014. sizing model for hybrid solar-wind power generation
[3] Yiwei Ma, Ping Yang, Hongxia Guo and Yuewu system,” J. Sol Energy, vol. 81, pp. 76-84, 2007.
Wang, “Dynamic economic dispatch and control of a [17] Hamid Hassan Zadehfard, Seyed Masoud Moghaddas
stand-alone micro-grid in Dongao Island,” J Electr. Tafreshi, Seyed Mehdi Hakimi, “Optimization of
Eng Technol, vol. 10, no. 4, pp. 1432-1440, 2015. grid connected micro grid consisting of PV/FC/UC
[4] Bhandari, B., Lee, K., Lee, G. et al, “Optimization of with considered frequency control,” J Elec Engin.
hybrid renewable energy power systems,” A review: Comp. Scien, pp. 1-16, 2015.
Int. J.Precision. Eng and Manuf.-Green Technol, vol. [18] Azadeh Kamjoo, Alireza Maheri, Ghanim A Putrus,
2, no. 1, pp. 99-112, Jan 2015. “Chance constrained programming using non-
[5] Shakti Singh, Muesh Singh and S. C. Kaushik “A Gaussian joint distribution function in design of
review on optimization techniques for sizing of solar- standalone hybrid renewable energy systems,” J.
wind hybrid energy systems,” Int. Journal. Green Energy, vol. 66, pp. 677-688, 2014.
Energy, vol. 13, issue 15, pp. 1564-1578, 2016. [19] S. Mina Mirbagheri, S. Mohsen Mirhassani, Hazlie
[6] B. Ge, W. Wang, D. Bi, C.B. Rogers, F. Z. Peng, A.T. Mokhlis, “Techno-economic optimization for a
de Almeida, Haitham Abu Rub, “Energy storage hybrid PV/Wind/battery stand-alone system using
system based power control for grid connected wind imperialistic competitive algorithm,” J Chem. Environ
power farm,” J. Elec Power. Energy Systs, vol. 44, Engin, vol. 5, no. 2, pp. 105-110, 2014.
pp. 115-122, 2013. [20] Saeid Lotfi Trazouei, Farid Lotfi Trazouei, Mohammad
[7] T. Nacer, A. Hamidat, O. Nadjemi, “Techno- Ghiamy, “Optimal design of a hybrid solar/wind/
economic impacts analysis of a hybrid grid connected diesel power system for rural electri-fication using
energy system applied for a cattle farm,” J. Energy imperialist competitive algorithm,” J. Renewable.
Proc, vol. 75, pp. 963-968, 2015. Energy Res, vol. 3, no. 2, 2013.
[8] A. Kaabeche, R. Ibtiouen, “Techno-economic optimi- [21] S. Gangwar, D. Bhanja, A. Biswas, “Cost reliability
zation of hybrid photovoltaic/wind/diesel/battery and sensitivity of a standalone hybrid renewable
generation in a stand-alone power system,” J. Sol energy system - A case study on a lecture building
Energy, vol. 103, pp. 171-182, 2014. with low load factor,” J. Ren. Sust Energy, vol. 7,
[9] Ce Shang, Dipti Srinivasan, Thomas Reindl, “An 2015.
improved particle swarm optimization algorithm [22] A. Hiendro, R. Kurnianto, M. Rajagukguk, Y. M
applied to battery sizing for stand-alone hybrid power Simanjuntak, “Techno-economic analysis of photo
systems,” J. Electrical Power. Energy systems, vol. voltaic/wind hybrid system for onshore/remote area
74, pp. 104-117, 2016. in Indonesia,” J. Energy, vol. 59, pp. 652-657, 2013.
[10] M. Durairasan, A. Kalaiselvan, Habeebullah Sait [23] Hai Lan, Shuli Wen, Ying Yi Hong, David. C. Yu,
“Hybrid Technique for Locating and Sizing of Lijun Zhang, “Optimal sizing of hybrid PV/diesel/
Renewable Energy Resources in Power System,” J battery in ship power system,” J. Appl Energy, vol.
Electr Eng. Technol.; vol. 12, no. 1, pp. 161-172, 2017. 158, pp. 26-34, 2015.
[11] E. Skoplaki, J. A. Palyvos, “On the temperature [24] C. Shuaixun, H. B. Gooi, M. Wang, “Sizing of
dependence of photovoltaic module electrical perfor- energy storage for microgrids,” IEEE Trans. on
mance: A review of efficiency/power correlations,” J. Smart Grid, vol. 3, no. 1, 2012.
Sol Energy, vol. 83, pp. 614-624, 2009. [25] Saber Arabi. Nowdeh, Abbas Rajabi. Ghahvanieh,
[12] H. Fathima, K. Palanisamy, “Optimized sizing, Saheb Khanbdal, “PV/FC/wind hybrid system optimal
selection, and economic analysis of battery energy sizing using PSO modified algorithm,” Bulletin
storage for grid connected wind /PV hybrid system,” Polytechnic Inst. Tech. University Gheorghe Asachi
J Model. Sim Engin, pp. 1-16, 2015. Iasi, vol. 4, pp. 49-66, 2012.
[13] Fatemeh Jahanbani, Gholam H. Riahy, “Optimum [26] A. Sahu, S. Kumar Panigrahi, S. Pattnaik, “Fast
design of a hybrid renewable energy system,” Convergence Particle Swarm Optimization for
Renewable. Energy. Trends and applications, 2011. Functions Optimization,” J. Proc. Technol, vol. 4, pp.
[14] Rachid Belfkira, Lu Zhang, George Barakat, “Optimal 319-324, 2012.
sizing study of hybrid wind/PV/diesel power [27] Krishna T.C., Manjaree P, Laxmi S, “Particle swarm
generation unit,” J.Sol Energy, vol. 85, pp. 100-110, optimization with time varying acceleration coefficients
2011. for non-convex economic power dispatch,” Elec
[15] Akbar Maleki, Mehran Ameri, Fathollah pourfayaz, Power. Energy Syst, vol. 31, pp. 249-257, 2009.
“Techno-economic analysis and optimal design of an [28] Xin Qiu, Tu A. Nguyen, Joe. D. Guggenberger, M.L.

http://www.jeet.or.kr │ 667
 Techno-Economic Optimization of a Grid-Connected Hybrid Energy System Considering Voltage Fluctuation

Crow, A.C. Elmore, “A field validated model of a Ramazan Bayindir He was born in
vanadium redox flow battery for micro grids,” IEEE Ankara Turkey, on 1970. He completed
Trans. Smart Grid, vol.5, no.4, pp.1592-1601, 2014. his undergraduate studies at Gazi
[29] M. C. Falvo, L. Martirano, D. Sbordone, E. Bocci, University, Electrical Education Dep.
“Technologies for smart grids: a brief review,” Int. in 1992 and received his M.Sc. and
Conf. On Environmental and Electrical Engineering, Ph.D. degrees from the Institute of
Poland, 05-08 May 2013. Science & Technology, Gazi University
in 1998 and in 2002 respectively. Dr.
Ramazan BAYINDIR worked as a Research Assistant,
Assistance Professor in Electrical Education Department at
Samia Saib She received the Licentiate Gazi University. He became an Assoc. Prof. at Electrical &
and Master degrees from department of Electronics Engineering Dep. at Gazi University. Currently,
electrical engineering, Setif, Algeria, in he is working as a Professor at the same department. His
2009 and 2011, respectively. She is main expertise field include renewable energy sources,
currently a Ph.D. student at the Faculty microgrids, distributed generation, smart grid applications
of Technology, University of Setif, Her and industrial automation with programmable logic
research interests include hybrid (PV, controller (PLC).
wind) systems, energy storage system
(ESS), optimization theory and its applications, and
renewable energies.

Ahmed Gherbi He was born in Setif,

Algeria in 1967. He received the
engineer degree, Magister and PhD
degrees from the department of
Electrical Engineering, Setif, Algeria,
in 1990, 1995 and 2006 respectively.
Currently, he is Professor at the
department of electrotechnics, Faculty
of Technology, Ferhat Abbas Setif1 University. His current
research interests are energy storage system, renewable
energy systems, power system stability and optimization.

Abdelhamid Kaabeche He was born

in Algiers, Algeria, on January 03,
1967. He studied at the Polytechnic
School of El Harrach Algiers and
received the Ph.D. degree in electrical
engineering from the same School in
2012. Since 1993 he worked in the
Development Center of Renewable
Energy (CDER) located at Algiers as a Researcher in the
field of renewable energy sources. His interest field
includes power electronics, electrical machines, electrical
networks, renewable energy sources, and electrical aspects.
Abdelhamid-Kaabeche is a member of the commission of
the national innovation prize.

668 │ J Electr Eng Technol.2018; 13(2): 659-668