Energy Conversion and Management 126 (2016) 909–922

Contents lists available at ScienceDirect

Energy Conversion and Management

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

Optimal hybrid renewable energy design in autonomous system using

Modified Electric System Cascade Analysis and Homer software
Hassan Zahboune a,⇑, Smail Zouggar a, Goran Krajacic b, Petar Sabev Varbanov c, Mohammed Elhafyani a,
Elmostafa Ziani a
a
Laboratory of Electrical Engineering and Maintenance – LEEM, University Mohammed 1st, High School of Technology, Oujda, Morocco
b
Department of Energy, Power Engineering and Environment, University of Zagreb, Faculty of Mechanical Engineering and Naval Architecture, Ivana Lučića 5, 10000 Zagreb, Croatia
c
Faculty of Information Technology and Bionics, Pázmány Péter Catholic University, Práter u. 50/A, 1083 Budapest, Hungary

a r t i c l e i n f o a b s t r a c t

Article history: In this paper, a method for designing hybrid electricity generation systems is presented. It is based on the
Received 20 March 2016 Modified Electric System Cascade Analysis method. The Power Pinch analysis is used as a guideline for
Received in revised form 17 August 2016 development of an isolated power supply system, which consists of photovoltaic panels, wind turbines
Accepted 21 August 2016
and energy storage units. The design procedure uses a simulation model, developed using MATLAB/
SIMULINK and applies the developed algorithms for obtaining an optimal design. A validation of the
Modified Electric System Cascade Analysis method is performed by comparing the obtained results with
Keywords:
those from the Homer Pro software. The procedure takes as inputs hourly wind speed, solar radiation,
Standalone system
Hybrid system
demands, as well as cost data, for the generation and storage facilities. It is also applied to minimize
Pinch analysis the loss of power supply probability and to minimize the number of storage units. The algorithm has been
MESCA and Homer Pro demonstrated with a case study on a site in Oujda city, with daily electrical energy demand of 18.7 kWh,
resulting in a combination of photovoltaic panels, wind turbine and batteries at minimal cost. The results
from the Modified Electric System Cascade Analysis and HOMER Pro show that both tools successfully
identified the optimal solution with difference of 0.04% in produced energy, 5.4% in potential excess of
electricity and 0.07% in the cost of the energy.
Ó 2016 Elsevier Ltd. All rights reserved.

1. Introduction Mykines [4]. Due to the variable nature of supply on one hand and
variable demand on the other, in geographically limited areas these
In many places in the world, electricity is still a luxury com- systems are designed with some type of energy storage. Storage
modity. It is available just to those sites that can be connected to units certainly increase costs [5] but also add new complexity to
centralised grids or have sufficient resources to build standalone planning and designing processes. To cope with this problem many
power systems. Until recently, these systems were heavily depen- methods and tools were developed for optimising the selection of
dent on generators powered by imported fossil fuels [1]. However, components and designing the whole standalone systems [6].
the development of renewable energy technologies creates many One of the common tools for energy planning and optimisation
more opportunities to exploit locally available resources and pro- is HOMER [7]. It performs three main types of analysis: simulation,
vides power to small and most isolated sites. There are many inter- optimization, and sensitivity analysis [8]. It is commercial energy
esting examples of planning 100% renewable energy supply for planning software that is particularly suitable to design and opti-
island communities and applying hydrogen as an energy vector mise standalone energy systems. It has been tested on many cases
[2], Panayiotou et al. [3] demonstration and application examined – e.g. Shiroudi et al. in [9] used a HOMER computer model, to
the optimal design of a standalone PV system and a standalone determine the most economic design for a hybrid energy system
hybrid PV-Wind system, the locations are considered in Nicosia, for a site in Taleghan-Iran having a daily load of 5.5 kWh, which
Cyprus and Nice, France. Enevoldsen and Sovacool examined the would use only renewable energy sources. The configuration of
possibilities and challenges of implementing a 100% integrated the optimal hybrid system is selected based on the best compo-
standalone renewable energy systems on one of the Faroe Islands, nents and sizing with appropriate operating strategy to provide
an efficient, reliable and cost-effective system. Prasetyaningsari
⇑ Corresponding author. et al. [10] used HOMER for optimum sizing of electric power
E-mail address: [email protected] (H. Zahboune). facilities to support the electricity demand of fish pond aeration

http://dx.doi.org/10.1016/j.enconman.2016.08.061
0196-8904/Ó 2016 Elsevier Ltd. All rights reserved.
910 H. Zahboune et al. / Energy Conversion and Management 126 (2016) 909–922

Nomenclature

Symbols EBat,in initial battery according DOD (Wh)

T end of analysis to prevent Eacc (h) EBat total capacity of the batteries (Wh)
t time step (h) Ʌd number of days of autonomy (day)
i interest rate Vbat battery bank voltage (V)
L system life time(y) Ibat battery Nominal capacity (Ah)
Lbat battery life time(y) gwr wires losses’ factor (%)
LW wind turbine life time(y) Pinv the capacity of the inverter (W)
LPV PVP life time(y) ED deficit of energy supply during the time step (Wh)
CC,W initial capital cost of each wind turbine(€) gdc/ac = gac/dc inverter efficiency (%)
CM,W maintenance cost of each wind turbine (€) gdc/dc rectifier efficiency (%)
CI,W installation cost of each wind turbine (€) ED,T annual energy deficit(Wh)
CC,PV initial capital cost of each PVP (€) Eex,T annual Excess of the Energy(Wh)
CM,PV maintenance cost of each PVP (€) gch battery charge efficiency (%)
CI,PV installation cost of each PVP (€) gdis battery discharge efficiency (%)
CC,bat initial capital cost of each battery (€) Sp iteration of the ESCA algorithm with Eacc(t = 0) = 0 Wh
CI,bat installation cost of each battery (€) S0p iteration of the MESCA algorithm with Eacc(t = 0) > 0 Wh
CBOS balance-of-system cost(€) P iteration number
CTCS total Cost of the System(€)
CTACS total Annual Cost of the System (€) Subscripts
fw fraction of the wind turbines power (%) ESCA Electric System Cascade Analysis
fPV fraction of the PVPs power (%) MESCA Modified Electric System Cascade Analysis
EPV power generated by a single PVP during the time step LPSP loss of Power Supply Probability
(Wh) HES Hybrid Energy System
EW power generated by a single wind turbine during the PA Pinch Analysis
time step (Wh) PoPA Power Pinch Analysis
EL power (load) demand during the time step (Wh) PoCA Power Cascade Analysis
EPV,A total amount generation by single PVP (Wh) SCT Storage Cascade Table
EW,A total amount generation by single wind turbine (Wh) HCT Hybrid Cascade Table
EPV,T(t) power generated by all NPV PVPs during the time step DEG Distributed Energy Generation
(Wh) PV Photovoltaic
EW,T(t) power generated by all NW wind turbines during the WT Wind Turbine
time step (Wh) DC Direct current
NK number of the hybrid components used AC Alternative current
K the hybrid components used (PVP and WT) PVP Photovoltaic Panel
NPV number of PVPs (dimensionless) BOS Balance-Of-System
NW number of wind turbines (dimensionless) FEE Final Excess of the Energy(Wh)
NBat number of batteries (dimensionless) MFEE Maximum Final Excess of the Energy(Wh)
v wind speed (m/s) TACS Total Annual Cost of the System(€/y)
Pr rated power of the wind turbine (W) LCE Levelized Cost of Energy (€/kWh)
vc cut-in of the wind turbine (m/s) SOC State Of Charge (%)
vf cut-out of the wind turbine (m/s) SOCmin Minimum of SOC (%)
vr rated speed of the wind turbine (m/s) DOD Depth Of Discharge (%)
gPV efficiency of the installed PV module (%) AEP Annual energy product (Wh)
APV area of the installed PV module(m2) OC Operating Cost (€/y)
Rs average solar radiation (W/m2) IC Initial capital of all system (€)
Er(t) energy absorbed from or injected into the storage unit APVEP Annual PV energy product (Wh)
(Wh) AWTEP Annual WT energy product (Wh)
Eel,T(t) the total electric power produced by PVP and WT (Wh) AEE Annual Excess Electricity (Wh)
Chbat(t) energy charge in the batteries (Wh) RF Renewable Fraction (%)
Dbat(t) energy discharge from the batteries(Wh) ACPVS Annual Cost PV system (€)
Eacc(t) accumulation of electrical energy in the storage unit ACWTS Annual Cost WT system (€)
(Wh) ACBU Annual Cost of the batteries unit (€)
Eacc (t = 0) initial accumulation of electrical energy not according TAC Total annual Cost (€)
DOD (Wh)

system. Sen et al. [11] used HOMER to propose the best hybrid user interface and simplicity of use, HOMER provides other useful
technology combination for electricity generation from a mix of features to system designers, like PDF curves, plots [13] etc. It uses
renewable energy resources (small-scale hydropower, solar photo- different time steps to simulate and balance energy systems and
voltaic systems, wind turbines and bio-diesel generators) to satisfy search for optimal solution trough sensitivity analysis of many
the electrical needs of Palari, an off-grid remote village in the state options.
of Chhattisgarh, India, in a reliable way. Bekele et al. [12] used The major drawback of HOMER is that input variables should be
HOMER for optimisation and sensitivity analysis of small-scale inserted by the user or set as limits of sensitivity analysis. Exam-
Hydro/PV/WT based hybrid electric supply system (total popula- ples for these variables are number of wind turbines, solar photo-
tion of 107,710 within 23 villages). Beside its intuitive graphical voltaic panels, batteries etc. Only after providing those inputs, the
 H. Zahboune et al. / Energy Conversion and Management 126 (2016) 909–922 911

model calculates all the possible combinations and provides the The Modified Electric System Cascade Analysis (MESCA)
solutions [14]. In many cases, optimum search space is large within method is presented in this work, to optimise the design of the
many thousands possible solutions. After calculations of all the standalone hybrid solar-wind system for electrification of a remote
possibilities, the model ranks the solutions according to certain cri- area. The algorithm inputs are hourly wind speed, solar radiation
teria. Then the user can select optimal solution or check any other. and demands load as well as cost data for the generation and stor-
Another drawback is that HOMER does not guide user to certain age facilities. The Power Pinch Analysis (PoPA) is the guideline for
type of solution or problem areas such as Pinch point [22]. MESCA method and the power fractions (fw and fPV) are based on
There are also many other simulation tools similar to HOMER. the total annual cost of the system and the meteorological condi-
These are able to simulate standalone systems, with all compo- tions for each isolated area chosen.
nents and links between different parts of the system. H2RES In this work, the maximum allowable loss of power supply
model was used to simulate the isolated energy system of Malta probability (LPSP) and the Power Pinch Analysis are implemented
using hydrogen as an energy storage [16]. Like the Homer, H2RES together in the MESCA method, and The Final Excess of the Energy
uses hourly energy balancing through whole year to select the (FEE) is considered as the objective function, to optimise the sizing
optimal solution. It has additional energy storage options such as of a hybrid system. The MESCA method contains two algorithms:
pumped hydro or desalination as demand side measure. Both fea- the first algorithm will be sizing the standalone hybrid system
tures are very useful when calculating isolated energy systems for for T = 288 h. The second algorithm will correct the sizing for
islands [17]. However, in H2RES as in Homer users must insert the T = 8760 h, if the obtained LPSP is superior to the desired value of
input size of equipment and make sensitivity analysis to choose the LPSP by the user. Finally, to validate this method on a more
the optimal solution [18] Even some advanced energy planning realistic case the research is expanded to a whole year with hourly
models like EnergyPLAN [19] used for hourly simulation of stan- resolution and compared to optimized solution provided by
dalone and national energy systems do not guide users to select HOMER software.
optimal size of installed capacity. There are many optimisation In the second section, this paper describes the standalone hybrid
models and methods, that try to solve the problem of energy plan- power generation system configuration with necessary mathemat-
ning and security of supply [20]. They limit the search space to just ical model. Section 3 is going to shed light on the details of the new
feasible and most attractive solutions. Doing so is challenging algorithm illustrated with effective techniques applied. After meth-
when there are many type of solution coming from so many avail- ods, Section 4 provides the brief description of Homer software in
able types and sizes of production units, storage, energy carriers, order to give the full information to the readers. In Section 5, we
integration options like CHP [21] etc. describe a case study which has been selected because measure-
The goal of this work is to find a simple method that will guide ment data were available and we will use test site in laboratory
users to optimal solutions under given constraints, while providing to validate results of proposed methods, to prove that the MESCA
satisfactory wide selection of equipment. The advantage of the Pinch method can use any site in the world with required input data.
optimisation method [22] is to be used as it is done mostly in process The MESCA result will be compared with HOMER results of the case
industry. Nemet [23] presented a new graphical approach called - study and the most important results of the comparison will be in
The Retrofit Tracing Grid Diagram- which enables clear visualisation Section 6. Finally in Section 7, we provide most important conclu-
of the current state. The works of Klemeš presented in [24] show the sions from our results and give acknowledgement to the project
results of the three efficiently collaborating research groups, in the that allowed research on the topic possible.
last forty years [25], providing some tested and proven guidance
for use of Pinch method in future industrial applications.
2. The standalone hybrid power generation system
Wan Alwi et al. [26] presented the graphical power pinch
configuration and method
approach, in Process Integration to determine the minimum elec-
tricity targets for systems comprising hybrid renewable energy
A hybrid solar-wind power generation system consists of a pho-
sources. Rozali et al. [27] proposed two numerical techniques:
tovoltaic (PV) system, a wind power system, a battery bank, recti-
power cascade analysis, and storage cascade table, to determine
fiers, an inverter, and serves the load demand. The electric power is
(1) the minimum target for outsourced electricity, (2) the amount
generated by the PV modules and the wind turbines (WT). It is
of excess electricity for storage during start up and normal opera-
used to satisfy the power demand while the excess is used to
tions, (3) the amount of transferrable power, (4) the maximum
charge the battery bank. The energy sources are connected to the
storage (e.g. battery) capacity for the off-grid systems, (5) the
direct current (DC) bus via the generators. The energy losses during
amount of electricity transferrable to the grid, (6) the amount of
conversion, charging and discharging of the battery are taken into
outsourced electricity needed at each time interval and (7) the
account (see Fig. 1).
time interval where the maximum power demand occurs. The
The MESCA method is applicable, whatever the chosen architec-
new numerical methods [28] with consideration of the loss of
ture, just the converters efficiency in the equations has to be
power in hybrid power system during the conversion and transfer
were proposed by Rozali et al. The PoPA method with a mathemat-
ical model is applied to determine the amount of excess electricity
that has been applied as the indicator to perform the optimal sizing
DC bus
of a Hybrid Power Systems [29]. To identify the allocations of stor-
age and outsourced electricity, the new graphical PoPA tool called
the Outsourced and Storage Electricity Curve is used [30]. PV panels DC/DC Batteries
The Electric System Cascade Analysis (ESCA) method is used in
the design of PV systems [31] and Distributed Energy Generation
(DEG) system [32]. The innovative ESCA method was developed
Load
and tested [15] on hybrid power supply for site in Oujda. However Wind DC/AC
Demand
AC/DC
this solution has been designed for average system loads and sup- turbines
ply values which were presented in 24 h. It was just a demo of
capability of method to find solution for design the hybrid system
(PV/WT/Batteries). Fig. 1. Block diagram of standalone hybrid energy system.
912 H. Zahboune et al. / Energy Conversion and Management 126 (2016) 909–922

adjusted. The connection of all energy sources to a DC bus simpli- outputs of the PV array and wind turbine are more than the load
fies power management. The choice of the DC bus is based on the demand energy, the battery bank is charged Er(t) > 0 Wh, the rest
fractions of the energy generation from renewable sources (The of energy is used to charge the batteries ChBat.
energy production by the PV system and by the wind system are
62% and 38%, respectively of the total energy production.). ChBat ðtÞ ¼ Er ðtÞ  gch ð7Þ

When the load demand is greater than the energy generated by

2.1. Solar system model
the wind/PV system (Er(t) < 0 Wh), the amount of the energy dis-
charge from the batteries DBat(t) at hour t is obtained by:
PV module’s performance is highly influenced by the panel type,
solar irradiance and other weather variables such as temperature. DBat ðtÞ ¼ Er ðtÞ=gDis ð8Þ
Solar system model can produce electrical energy by using photo-
voltaic (PV) panels. Solar panels supply DC by converting solar The accumulation of electrical energy in the storage units is cal-
energy (hourly solar radiation) to electrical energy (electricity). culated by (9):
The current is connected directly to the AC bus through a DC/AC
inverter. The hourly energy output of the single PV generator dur- Eacc ðtÞ ¼ Eacc ðt  1Þ þ ChBat ðtÞ þ DBat ðtÞ ð9Þ
ing current time step is given by (1) [33]:
The number of the batteries is calculated by (10) and round up
EPV ðtÞ ¼ gpv  APV  Rs ðtÞ ð1Þ to the nearest integer:

where RS is the global solar radiation (Wh/m2); APV is the PV panels Eacc ðmaxÞ  Kd
receiving surface (m2) and gPV is the efficiency of PV generator NBat ¼ ð10Þ
IBat  VBat  DOD
(conversion) (%).
The total generated power by NPV of photovoltaic panels during Eacc(max): the maximum of accumulation of electrical energy in
the current time step (2): the storage unit (Wh).
Ʌd is the number of days of autonomy (Day).
EPV;T ðtÞ ¼ NPV  Eðm ðtÞ ð2Þ
where NPV is the number of PV panels. 2.4. Converters

2.2. Wind turbine model The use of converters is necessary because the PV panels and
wind turbine produce direct current and alternating current,
For a wind turbine, the power output characteristic can be spec- respectively. The converters can operate as a rectifier which con-
ified based on the cut-in wind speed vC, the power output increases verts AC to DC, and an inverter which converts DC to AC. Alterna-
linearly as the wind speed increases from vC to the rated wind tive currant is used to supply the AC load.
speed vR. The rated power PR is produced when the wind speed A power converter is required for a hybrid wind/PV/battery
varies from vR to the cut-out wind speed vF. Beyond this point power system to maintain the flow of energy between DC and AC
the wind turbine will be shut down for safety considerations. Then components [34]. The capacity of the inverter Pinv is given by (11):
the wind turbine power output can be simulated by (3):
8 vvC Pinv ¼ EL ðmaxÞ=gdc=ac ð11Þ
>
< Pr vvF v C < vðtÞ < vR
EW ðtÞ ¼ Pr
>
v R < vðtÞ < vF ð3Þ where EL(max) is the maximum of energy demand by Load system
: (Wh).
0 vðtÞ < vC or vðtÞ > vF In this work, it is assumed that the converter has an efficiency of
The overall produced power is determined by NW wind turbines 90% [35]. Other parameters of the converter can be specified by the
during the current time step (4): user, following the use of converter: Frequency limitation, the type
and characteristics of the board controls, the number of the AC
EW;T ðtÞ ¼ Nw  Ew ðtÞ ð4Þ
module and DC module, the input and the output voltage, fre-
where NW is the number of wind turbines. quency setting signal gain, torque boosting.
The total electric power produced Eel,T by both PV and Wind
generators is calculated by (5). 2.5. Optimization constraints

Eel;T ðtÞ ¼ Epv;T ðtÞ  gdc=dc þ Ew;T ðtÞ  gac=dc ð5Þ The maximum allowable loss of power supply probability, the
The net electricity surplus/deficit Er(t) is the difference between minimum of the total annual system cost and the Power Pinch
the global power generation and the load demand (EL(t)) during Analysis are implemented together in the MESCA method, to opti-
the current time step Er(t): mise the sizing of a hybrid system.

Er ðtÞ ¼ Eel;T ðtÞ  EL ðtÞ=gwr ð6Þ 2.5.1. Cost modelling

The Total Cost of the System, CTCS is calculated according of the
2.3. Storage battery model numbers of the PVP, wind turbine and batteries. It considers the
cost of components, maintenance, balance of system (BOS), and
During any time step, the excess power generated by the PV and the cost of installation. Then CTCS can be calculated by (12):
wind generators, after meeting the load demand, can be utilised to X
CTCS ¼ NK  ðCC;K þ CM;K þ C I;K Þ þ CBos þ C bat ; N ð12Þ
charge the battery, until the latter it is fully charged. Whereas the K2fPV;Wg
stored energy can be discharged, when energy generation is insuf-
ficient, the battery releases power to help covering the load The optimisation model uses the total annual cost of the System
requirements, until storage is exhausted. Then, the difference (TACS) as an objective function, according to (13).
between load demand energy and total energy generated, decides
CTACS ¼ Min½CTCS  CRFÞfw2½0;1 ð13Þ
whether battery is in charging or discharging state. When the total
 H. Zahboune et al. / Energy Conversion and Management 126 (2016) 909–922 913

CRF is the capital recovery factor, as determined in (14):

L
i  ð1 þ iÞ
CRF ¼ L
ð14Þ
ð1 þ iÞ  1
where L: System life time (y), i: discount rate (%).
The lifetime of the system is 25 y, but for the lifetime of each
battery is assumed to be 5 y. The present worth of batteries CBat,N
for all number of the batteries NBat is given by (15):

CBat;N ¼ CC;Bat
!
i i i i
 1þ 5
þ 10
þ 15
þ 20
ð1 þ iÞ ð1 þ iÞ ð1 þ iÞ ð1 þ iÞ
þ CI;Bat ð15Þ
The following formula (16) is used to calculate the levelized
cost of energy (LCE):

CTACS
LCE ¼ PT ð16Þ
t¼0 EL ðtÞ

2.5.2. Loss of power supply probability (LPSP)

LPSP stands for the probability of insufficient power when the
hybrid system (solar, wind and battery storage) is unable to meet
the load demand [24]. LPSP value of 1 means that the load will
never be satisfied and 0 means that the load will always be satis-
fied. Yang et al. [25] used LPSP to confirm a good optimisation of
the standalone PV–wind hybrid system using genetic algorithm
to minimize the annual cost. For a specified period only when
the renewable sources and the batteries are not able to satisfy
the load demand, the LPSP is defined as follows:
PT
ED ðtÞ
LPSP ¼ Pt¼0
T
ð17Þ
t¼0 EL ðtÞ

where ED(t) is the deficit of energy supply during the time step,
when the generated energy not satisfied to meet the power Fig. 2. Flowchart of the MESCA method with the four scenarios.
demands.
One of the objectives of the design procedure is to discover
combinations between the two energy resources (PV/WT) and  Time period of analysis.
the unit storage. The optimization procedure should give a LPSP  Hourly energy demand.
lower than or equal to the maximum allowed by the user.  Hourly wind speed and solar radiation.
 Type and characteristic of the PV panel.
3. Modified Electric System Cascade Analysis (MESCA)  Power generation capacity of a single solar PV module.
 Solar PV module efficiency.
Hybrid Energy Systems (HES) are designed to supply isolated  Wind turbine power capacity.
sites, but the instantaneous change in the weather, for example  Type and characteristic of the battery.
the variation in hourly solar radiation, wind speed and tempera-  Converters efficiency.
ture, makes harder the accurately predicting the output of the sys-  Hourly load demand.
tem. The Basis of these problems, Modified Electric System Cascade
Analysis method is used to design the HES Fig. 2. The second part of The share or fraction of each renewable source generating elec-
the algorithm checks the MESCA result during all the year tricity is decided in the first step of the MESCA method. The frac-
(T = 8760 h). If the design has not the desired value of Loss of tions should respect the minimum of total cost and satisfy the
Power Supply Probability (LPSP), the second algorithm presented load demand. The sum of all the fractions must be equal to 1 at
in Fig. 3 is used to find the optimal solution. any given time Eq. (18):
The theoretic basis for the MESCA method is the Pinch Analysis
f PV þ f W ¼ 1 ð18Þ
[22]. The validation of the method is done by comparing to the
optimal configuration obtained by Homer Pro, at technical and eco- The MESCA method is based on instantiating the hybrid Cascade
nomic level. The comparison result shows the feasibility of the new Table (HCT), where each row represents one time step of
method. the analysis and each column represents a system property.
Twelve columns (Tables 7 and 8) are used in the HCT. The aim of
3.1. MESCA method: first algorithm the first algorithm is to calculate number of the PVPs, the wind
turbines and the batteries. According to the algorithm depicted
According to [15], the first step of MESCA is to extract the rele- in Fig. 2, the batteries number is calculated at the end of the
vant data for analysis. The list of data includes: algorithm.
914 H. Zahboune et al. / Energy Conversion and Management 126 (2016) 909–922

Scenario 2 occurs when FEE is larger than the generated power

by a single wind turbine accounting the losses of the converter (AC/
DC), but too small to fit a PVP. This is because the generated power
by a single wind turbine is higher compared to the generated
power by a single wind PVPs. The remaining energy will be used
to resize the PVPs (scenario 3). Then the number of wind turbine
is rounded up to the nearest integer according to (22):
# "
½Eacc ðTÞ  Eacc ð0Þ
NWnew ¼ NW  ð22Þ
EW;A  gCh  gAC=DC

Scenario 3 occurs when FEE is larger than the generated power

by a single PVP accounting the losses of the converter (DC/DC), but
too small to accommodate a wind turbine. Then the number of
PVPs is rounded up to the nearest integer according to (23):
# "
½Eacc ðTÞ  Eacc ð0Þ
NPVnew ¼ NPV  ð23Þ
EPV;A  gCh  gDC=DC

Scenario 4 occurs when FEE is smaller than the generated power

by a single PVP accounting the losses of the converter (DC/DC).
Then the number of all power generators including PVPs and wind
turbines is fixed and considered final. For better precision, the area
of one the selected PV modules should be changed according to
(24):
# "
½Eacc ðTÞ  Eacc ð0Þ
Anew ¼ A  X ð24Þ
T
RsðtÞ  gPV  gCh  gDC=DC

The size of the battery can be calculated from column 13 of HCT.

Fig. 3. Flowchart of proposed method with controllable difference between the The largest value in the accumulation of electrical energy in the
desired and obtained LPSP.
storage unit is denoted as Eacc(max). It represents the largest elec-
tricity charge in the battery bank among all time steps, as required
by the demand. At the same time, this is the minimum battery
bank capacity in terms of demand.
The optimal design of the hybrid system should satisfy two
The obtained capacity value has to be increased, to account for
conditions during the variation of energy accumulation. In the
the Depth Of battery Discharge (DOD). The total energy in batteries
first condition, it must not have any negative value in Eacc(t), dur-
including DOD is calculated by (25):
ing their calculation, if negative value of the battery charge level
is encountered for any of the time steps, this indicates that power Eacc ðmaxÞ
EBat ¼ ð25Þ
has to be drawn from the battery. The most negative value among DOD
all steps is taken as the initial state of the batteries Eq. (19). Then, Then as a result the number of the batteries is calculated by (10)
the calculation of Eacc(t) is repeated, in order to obtain a feasible and round up to the nearest integer.
HCT, At t = T, the accumulated electrical energy Eacc(t = T) minus The initial battery state accounted DOD is calculated, to be the
the initial electrical energy Eacc(t = 0) is called the Final Excess sum of the initial battery charge from the first algorithm, and the
of Energy (FEE). Second condition respects the desired value battery remainder after maximum discharge:
of the Loss of Power Supply Probability (LPSPD). After completing
the calculation of Eacc(t), the energy accumulation at the end of EBat;in ¼ Eacc ðt ¼ 0Þ þ EBat  ð1  DODÞ ð26Þ
the design horizon represents the final surplus or deficit of
energy. 3.2. MESCA method: second algorithm
Eacc ðt ¼ 0Þ ¼  min Eacc ðtPinch Þ ð19Þ
The first algorithm result will be checked by calculating the
There are several scenarios affecting the number of power gen- obtained value of LPSP with the analysis time T = 8760 h, the differ-
erators. They are shown in Fig. 2, differ according to the best fit of ence between the desired value of the LPSP and obtained LPSP,
the combinations of the components within the gap formed by FEE. requires resizing energy resources according to the second algo-
Scenario 1 occurs when FEE is larger than or equal to the gener- rithm shown in Fig. 3.
ated power by a single wind turbine and a single PVP accounting Over one year, with T = 8760 h, the annual energy deficit is cal-
the losses of the converters. Then the numbers of PVPs and wind culated by (27):
turbines are rounded up according to the nearest integer, see
Eqs. (20) and (21). X
T¼8760

# " ED;T ¼ ED ðtÞ ð27Þ

½Eacc ðTÞ  Eacc ð0Þ  f W t¼0
NWnew ¼ NW  ð20Þ
EW;A  gCh  gAC=DC Scenario 1: the difference between ED,T and ED,min larger than or
equal to the generated power by a single wind turbine and a single
# "
½Eacc ðTÞ  Eacc ð0Þ  f PV PVP accounting the losses of the converters. Then the numbers of
NPVnew ¼ NPV  ð21Þ PVPs and wind turbines are rounded up to the nearest integer –
EPV;A  gCh  gDC=DC
Eqs. (28) and (29):
 H. Zahboune et al. / Energy Conversion and Management 126 (2016) 909–922 915

# "
EL;T  ðLPSP  LPSPD Þ  f W The goal of the optimisation process is to determine the optimal
NWnew ¼ NW  P ð28Þ
T EW;A  gCh  gAC=DC
value of each decision variable that interests the modeller. Possible
decision variables in HOMER include:
# "
EL;T  ðLPSP  LPSPD Þ  f PV  The size of the PV array.
NPVnew ¼ NPV  P ð29Þ
T EPV;A  gCh  gDC=DC
 The number of wind turbines.
 The number of batteries.
Scenario 2: if the difference between ED,T and ED,min larger than  The size of the AC–DC converter.
or equal to the generated power by a single wind turbine account-  The dispatch strategy (the set of rules governing how the sys-
ing the losses of the converters. Then the numbers of wind turbines tem operates)
is rounded up according to Eq. (30) to the nearest integer.
# " Homer has been extensively used as a sizing and optimisation
EL;T  ðLPSP  LPSPD Þ software tool. In this study, the hybrid power system consists of
NWnew ¼ NW  P ð30Þ
T EW;A  gCh  gAC=DC
PV panels, wind turbine, converters and battery storage. Homer
Pro compares the hourly electric energy demand to the hourly sup-
Scenario 3: if the difference between ED,T and ED,min larger than ply by the system. It also calculates the flows of energy to and fro
or equal to the generated power by a single PVP accounting the from each element of the system.
losses of the converters. Then the numbers of PVP is rounded up The model constraints include:
according to Eq. (31) to the nearest integer.
# "  Maximum annual capacity shortage.
EL;T  ðLPSP  LPSPD Þ  Minimum renewable fraction.
NPVnew ¼ NPV  P ð31Þ
T EPV;A  gCh  gDC=DC
The total net present cost (NPC) includes the costs of any com-
The Logical block diagram for battery energy management is ponent such as capital, replacement, operation and maintenance
shown in Fig. 4. and interest.

4. Homer software 5. Case study

HOMER (Hybrid Optimization Model for Electric Renewable) The power demands of the considered sites vary significantly
software has been developed by the U.S. National Renewable with time of the day, period of the year and type of the user. The
Energy Laboratory (NREL). It assists in designing of micropower variations of the renewable supplies and demands are partly pre-
systems and facilitates the comparison of power generation dictable with some changing in regular time intervals.
technologies across a wide range of applications. The developed methodology has been applied to design a stan-
The optimal system configuration determined by Homer Pro dalone hybrid PV/wind system, in order to supply power to resi-
satisfies the user-specified constraints at the lowest total net dential area located in Oujda, Morocco. The latitude and
present cost. Optimisation can help the modeller find the optimal longitude for the location are 34° 410 N and 1° 540 W. MATLAB/
system configuration out of many possibilities. SIMULINK environment is used in implementing and programming
the solution based on the proposed method.
In the case study, the wind and solar power are assumed to be
constant during each time step (1 h in this study). LPSP must be
equal to 0% for MESCA result (First algorithm). The case study
has average hourly data for each month (1 month ? 24 h)
throughout the year (T = 24 h * 12 = 288 h) for the hourly solar
radiation, wind speed and load demand. For T = 8760 h, the mini-
mum of the desired value of the LPSP is fixed to 2% (second
algorithm).
The technical and cost information of the selected PV module,
the wind turbine and the battery storage used in the case study
are summarized in Table 1.
The wind power curve of the 1 kW wind turbine model is illus-
trated in Fig. 5. The hourly variation of the load demand, wind
speed and solar radiation throughout the year are illustrated in
Fig. 15 and Fig. 16.

6. Results and discussion

The method validation is performed by the comparison

between results obtained by the method and the Homer Pro result.

6.1. MESCA result

The final result of the MESCA method is obtained in two steps:

the first step is to have the optimal dimensioning of the installation
Fig. 2, based on the monthly variation of wind speed, solar
Fig. 4. Logical block diagram for management the battery energy. radiation and energy demand of the load for T = 288 h (Fig. 15).
916 H. Zahboune et al. / Energy Conversion and Management 126 (2016) 909–922

Table 1 The MESCA method has for each month a table showing the
The parameters of the PV system, the 1 kW Wind Turbine and the batteries storage energy generated by sources, energy charged and discharged by
system.
the battery and the variation of energy storage units. The two
Characteristics of the time analysis tables (Table 7 and Table 8) show that the Pinch Point takes place
Period of analysis 1y at 8 AM/February, and the maximum of energy stored in the bat-
Time step 1h
teries is located at 6 pm/August. In Fig. 7; the excess energy at
Characteristics of the PV Panel the final point represents a small increase of 1.5 kWh. This FEE
Type of solar PV Mono-crystalline silicon
A 1.07 m2
(1.5 kWh) is equivalent to 4.11% of the total battery capacity
gPV 12% (36.43 kWh), less the MFEE fixed (10%) in the beginning of the
Power of a single solar PV module 120 Wp algorithm. According to the battery characteristic and Eacc(max)
CC.PV(each PV panel) 110 € = 25.5 kWh, the optimal configuration is 25 PVPs, 5 wind turbines
Cm,PV (each PV panel) 10 €
and 8 batteries. Fig. 6 shows the accumulation of electrical energy
CI,PV (each PV panel) 15 €
CBos 500 € in the storage unit (NW = 5 and NPV = 25 with and without the ini-
LPV 25 y tial value of the battery).
Characteristics of the wind turbine Any negative value in the energy accumulation presents that
Pr 1 kW LPSP equal to 0%. Consequently, the energy supplies from sources
vin 2.5 m/s (PV / Wind) are meeting the energy demands 100%.
vout 13 m/s
vr 11 m/s
Cc.w 1300 € 6.1.2. Second algorithm result: T = 8760 h
Cm,w (each wind turbine) 80 € The MESCA result (T = 288 h, NW = 5, NPV = 25, NBat = 8) for the
Ci,W (each wind turbine) 50 €
i 5%
first algorithm will be verified during all the year (T = 8760 h).
LW 25 y The user fixed the desired value of the LPSP in 2%. The obtained
Characteristics of the battery storage
LPSP with NW = 5, NPV = 25, NBat = 8 and T = 8760 h is 2.57%. The
CBat 210 Ah difference between the two LPSP values is equivalent to hourly def-
VBat 24 V icit energy 18.38 Wh. The sum lacking in energy throughout all
DOD 70% year (T = 8760 h) is shown in Fig. 8.
gch 90%
According to the difference between ED,T and ED,min, the energy
gdch 90%
Cc,bat 580 € luck is less than the produced energy of a single 1 kW wind tur-
Cr,bat 580 € bine. For that, only the PVPs number will be resized (scenario 3).
LBat 5y The Table 3. presented the results of the first algorithm (NW = 5,
NPV = 25 and NBat = 8) and the second algorithm (NW = 5, NPV = 26
and NBat = 8).
1200 Table 3 shows the LPSP for the two algorithms result 2.57% and
1000
1.97%, respectively. The optimal solution (NPV = 26, NW = 5,
Power output (W)

NBat = 8) with LPSP = 1.97% and CTACS = 2391 €/y, meets the energy
800 demand and respect the constraints. Hourly expected amount of
600 stored energy during a year in the eight batteries is presents in
Fig. 9.
400
Fig. 10 presents the total power consumption and the energy
200 deficiency for each month. According to Fig. 10, December is the
0 month with highest deficiency of the energy. The power demand
1 3 5 7 9 11 13 15 17 19
has been met 100%, during the months of March, April, June, July
Wind speed (m/s) and August, the LPSP for these months are 0%.
Fig. 5. Wind power curve of 1 kW wind turbine.
6.2. Homer result

The second step is the sizing correction with the second algorithm The search space alternatives used in the simulation presents in
shown in Fig. 3 for T = 8760 h (Fig. 16). the Table 4. A total of 360 a small isolated power system (SIPS),

6.1.1. First algorithm result: T = 288 h 30

The PVPs, wind turbines and batteries numbers are calculated
25
according to the FEE and the fractions energy generated by the
20
devices (fw and fpv). The iterations results for the first algorithm
Energy (kWh)

are present in Table 2. 15

10
5
Table 2
Various iterations (P) for MESCA method. 0
Eacc without initial value
P Sp NW NPV Eacc(t = 0) LPSP FEE NBat CTACS -5 Eacc with initial value

1 S1 16 65 0 0.01 444 2 3961 -10

1
25
49
73
97
121
145
169
193
217
241
265

2 S2 4 24 0 9.3 22 4 1659

3 S3 5 26 0 1.7 8.6 5 1989
Time (h)
3 S30 5 26 3,35 0 11.94 9 2525
4 S4 5 25 0 1.6 1.6 8 2361
Fig. 6. The accumulation of electrical energy in the storage unit without and with
4 S40 5 25 3,73 0 1.5 8 2361
initial value.
 H. Zahboune et al. / Energy Conversion and Management 126 (2016) 909–922 917

30 700 ED EL
FEE = Eacc(T) - Eacc(t=0) 600
25
Eacc(max) 500

Energy (kWh)
20
Eacc (Wh)

400
Eacc(t=0) Eacc(T)
15 300

200
10
100
5
0

0
1 24 47 70 93 116 139 162 185 208 231 254 277
Month
Time (h)
Fig. 10. Different between deficit energy and load demand with MESCA method.
Fig. 7. The accumulation of electrical energy in the storage unit with Eacc(t = 0)
= 3.7 kWh.
Table 4
Search space alternatives used in the simulation.
200
Sensitivity variables Value
175
ED,T=161kWh LPSP=2.57% Wind turbine 3;4;5;6;7;8;9;10
150 PV panels 22;23;24;25;26;27;29;30
ED,min=125kWh LPSP=2% Battery storage 6; 7; 8; 9; 10
ED,T (kWh)

125
LPSPD 2%
100 ED,T=123kWh LPSP=1.97%
75
50 system comprises of 26 PV modules (3.12 kW), 5 wind turbines
(1 kW each), 1.5 kW converter and eight batteries (5 kWh each).
25
The optimisation results for an annual average wind speed of
0 3.48 m/s, an annual average global solar radiation of 5.26 kWh/
1
517
1033
1549
2065
2581
3097
3613
4129
4645
5161
5677
6193
6709
7225
7741
8257

m2/day and the primary load of the case is 18.7 kW/day. The PV
array system will produce more compared to the wind turbine sys-
Time (h)
tem. Fig. 11 presents the Annual of Excess Electricity Produced
Fig. 8. The accumulation of lacking energy during the whole year (T = 8760 h). (AEEP) and shows the calculated Annual Energy Production (AEP).
The monthly average electricity production (for wind turbine
and PVP system) is shown in Fig. 12.
Table 3 The total annual electricity production from the optimal config-
LPSP variation depending to batteries, Wind turbine and PVP numbers.
uration of the hybrid system is 10,131 kWh/y. From that 62.44%
MESCA algorithms NPV NW NBat LPSP CTACS (6326 kWh/y) electricity is generated by the PVPs system and
First algorithm 25 5 8 2.57 2361 37.56% (3805 kWh/y) from the wind source. The optimisation
Second algorithm 26 5 8 1.97 2391 result has annual capacity shortage of 1.8% of the total power
demand during one year operation. The model results show that
the battery input and output are 3103 kWh/y and 2479 kWh/y,
45 respectively. The difference between the battery input and output
40 is mainly due to the charging and discharging losses (624 kWh/y)
and storage depletion (15 kWh/y). The total power consumption
battery capacity (kW)

35
amounts by the load and deficiency of energy for each month are
30
indicated in Fig. 13.
25
The total Energy deficit and surplus of the optimized system are
20 112.5 kWh/y (1.79% of the total power demand) and 2210 kWh/y
15 (21.8% of the total electricity production), respectively. In general,
10 the hybrid system will be able to utilise excess electricity if a dump
5 load is available as heating or cooling for the household.
0 The HOMER Pro result shows that the TAC is 2351 €/y. For the
1
501
1001
1501
2001
2501
3001
3501
4001
4501
5001
5501
6001
6501
7001
7501
8001
8501

annual cost, the wind turbine system is 534 €, the PVP system is
803 € and battery storage is 1014 €. The total net present cost
Time(h)
(NPC) is 37,878 € and a levelized cost of energy (LCOE) of
Fig. 9. The stored energy variant in the batteries Banks. 0.3747 €/kWh.

6.3. Discussion
HOMER grades the feasible system according to increasing net
present cost and eliminates all infeasible solutions. The choice of the optimum configuration using Homer Pro soft-
In Homer Pro, LCE and LPSP are used to choose the optimum ware is made according to the proposals provided by the user (Sen-
size of hybrid power system. Table 5 presents the minimum LCE sitivity Analysis). So the optimal configuration is selected by the
values and their corresponding optimal configurations. According calculated solutions. Therefore, Homer Pro is not used to find the
to the results of the optimization process, the optimal hybrid global optimum but the best possible configuration in the SIPS
918 H. Zahboune et al. / Energy Conversion and Management 126 (2016) 909–922

Table 5
Optimization results of PV/WT hybrid configuration using Homer Pro.

System type NPV NW Nbat Einv (kW) COE (€/kWh) NPC (€) OC (€/y) IC (€)
26 5 8 1.5 0.3757 37,818 774.38 25,360

27 5 8 1.5 0.378 38,314 798.42 25,470

24 6 8 1.5 0.383 38,542 752.30 26,440

25 6 8 1.5 0.386 39,039 776.34 26,550

26 6 8 1.5 0.39 39,536 800.38 26,660

23 7 8 1.5 0.393 39,764 754.27 27,630

27 4 9 1.5 0.396 39,858 776.30 27,370

27 6 8 1.5 0.393 40,033 824.42 26,770

27 6 8 1.5 0.393 40,033 824.42 26,770

24 7 8 1.5 0.397 40,261 778.30 27,740

24 7 8 1.5 0.3972 40,261 778.30 27,740

25 5 9 1.5 0.403 40,583 754.22 28,450

25 7 8 1.5 0.404 40,757 802.34 27,850

26 5 9 1.5 0.406 41,080 778.26 28,560

26 7 8 1.5 0.404 41,254 826.38 27,960

23 6 9 1.5 0.411 41,308 732.15 29,530

27 5 9 1.5 0.409 41,577 802.30 28,670

27 7 8 1.5 0.408 41,751 850.42 28,070

27 7 8 1.5 0.408 41,751 850.42 28,070

24 6 9 1.5 0.413 41,805 756.19 29,640

24 6 9 1.5 0.413 41,805 756.18 29,640

25 6 9 1.5 0.416 42,301 780.22 29,750

25 6 9 1.5 0.416 42,301 780.22 29,750

26 6 9 1.5 0.420 42,798 804.26 29,860

PVPs Power Output

10 1000 WTs Power Output
Electrical power production
Annual Energy (MWh)

8 800
(kWh)

6 600

400
4

200
2

0
0 1 2 3 4 5 6 7 8 9 10 11 12
Er,A Epv,A Ew,A Eex,A Months
Type of Energy
Fig. 12. Electrical power production of hybrid energy system.
Fig. 11. The AEP for each of sources and all system with the potential of the AEEP.

In another way, Homer Pro software found >360 possible com-

given by the sensitivity analysis is selected. In the MESCA method, binations between the proposed variation sensitivity. However,
the algorithm searches the optimal configuration respecting all the MESCA made only 6 iterations to find the optimal configuration.
constraints fixed by the user. The components number is Even if the difference in time to find the optimal solution is not
calculated with sets of iterations and obeying the conditions significantly pronounced due to our chosen site, it would be
(fPV, fW and LPSP). noticeable in complex sites.
 H. Zahboune et al. / Energy Conversion and Management 126 (2016) 909–922 919

Electrical power (KWh) 700 Table 6

Edeficit EL The comparison between the found results by HOMER Pro and MESCA.
600
Unit Homer MESCA Error
500
AEP kWh/y 10,131 10,127 0.03
400 APVEP kWh/y 6326 6421 1.50
300 AWTEP kWh/y 3805 3,745.2 1.57
AEE kWh/y 2210 2090 5.42
200 RF % 100% 100% 0.00
100 LPSP % 1.79 1.97 9,13
ACPVS € 803 776,4 3.31
0 ACWTS € 534 542,8 1.64
1 2 3 4 5 6 7 8 9 10 11 12 ACBU € 1014 1072 5.72
Month TAC € 2351 2391 1.71
EL,A kWh/y 6285 6391 1.67
Fig. 13. Different between deficit energy and load demand with Homer Pro LCE €/kWh 0.379 0.374 0.07
software.

the annual energy produced (10,127 kWh). The TAC values of the
12
optimal solutions found by Homer Pro and MESCA are 2,350.8 €/y
Electrical power production

Homer Pro MESCA

10 and TAC = 2391 €/y, with the error is 1.67%. Using the MESCA
method, the LCE of generated electricity by the hybrid energy sys-
8 tem is calculated to be 0.374 €/kWh, which is about 0.07% cheaper
(MWh)

than the obtained LCE by Homer Pro (0.379 €/kWh).

6
According to this comparison, the MESCA method can be used
4 for sizing of hybrid systems, whether with one type of energy
source with a storage system, or with two types of power sources
2 with storage system. Finally, Table 6 validates the result obtained
by the MESCA method, because the difference is negligible when
0
Er,A(t) EPV,A EW,A EEx,A we compare it with the result obtained by Homer Pro. More than
that, the method includes other advantages: the flexibility to
Fig. 14. The energy produced by the renewable sources and the excess electricity. choose the type of energy sources and its period of analysis is very
short to find the optimal configuration of the system. The valida-
tion of the method MESCA provides a main line to expand, to trans-
In Fig. 14, comparing the annual energy produced by renewable
form and include a set of the tools graphical and numerical, highly
sources with Homer Pro and MESCA method, the error between the
developing and using when sizing the total integration site based
two results is less than 0.04%. It can be concluded that two anal-
on Pinch Analysis.
ysed models provide the same annual energy produced.
After satisfying energy demand and the constraints, the renew-
able fractions for both methods are 100%. The LPSPs for Homer Pro 7. Conclusions
and MESCA method are 1.79% and1.97%, respectively. The two val-
ues of the LPSP respect the desired value of the LPSP by the user. This paper has discussed the performance of the MESCA
The Potential annual excess electricity calculated by MESCA and method, for finding the optimal design of a standalone hybrid
Homer Pro is 2210 kWh/y and 2090 kWh/y, respectively. The error PV/WT power system with battery storage, the Power Pinch Anal-
between the two AEE (120 kWh) is less than 1.2% compared with ysis is used as a guideline. The optimal design is chosen based on

1.0

0.8
Rs (kW/m2)

0.6

0.4

0.2

0.0
2 4 6 8 10 12 14 16 18 20 22 24
Time ( h)
1250
6
1000
Wind speed (m/s)

5
EL(Wh)

750
4

3 500

2 250

1 0
2 4 6 8 10 12 14 16 18 20 22 24 2 4 6 8 10 12 14 16 18 20 22 24
Time ( h) Time ( h)

Fig. 15. Hourly radiation solar, wind speed and load profile to Oujda city T = 288 h.
920 H. Zahboune et al. / Energy Conversion and Management 126 (2016) 909–922

1200 14.00

Solar radiation (W/m2)

1000 12.00

Wind speed (m/s)

10.00
800
8.00
600
6.00
400
4.00
200 2.00
0 0.00
1

8001

1
501
1001
1501
2001
2501
3001
3501
4001
4501
5001
5501
6001
6501
7001
7501

8501

501
1001
1501
2001
2501
3001
3501
4001
4501
5001
5501
6001
6501
7001
7501
8001
8501
Time (h) Time (h)

1400
1200
load demand (wh)

1000
800
600
400
200
0
1
501
1001
1501
2001
2501
3001
3501
4001
4501
5001
5501
6001
6501
7001
7501
8001
8501
Time (h)

Fig. 16. Hourly radiation solar, wind speed and load profile to Oujda city T = 8760 h.

Table 7
HCT for NW = 5, NPV = 25, NBat = 8 and Eacc (t = 0) = 3733 Wh month: February.

t v Ew Ew.T Rs EPV EPV,T EL Er Chbat Dbat EN Eacc

(h) (m/s) (Wh) (Wh) (W/m2) (Wh) (Wh) (Wh) (Wh) (Wh) (Wh) (Wh) (Wh)
1 3.86 75.4 377 0 0 0 240 144 130 0 130 2171
2 3.82 72.7 364 0 0 0 240 131 118 0 118 2288
3 3.74 67.4 337 0 0 0 240 105 94 0 94 2382
4 3.63 60.4 302 0 0 0 240 70 63 0 63 2446
5 3.51 52.9 264 0 0 0 358 72 0 91 91 2355
6 3.40 46.3 231 0 0 0 1038 724 0 905 905 1451
7 3.30 40.4 202 7 1 22 985 686 0 857 857 593
8 3.24 37.0 185 108 13.9 347 1069 475 0 593 593 0
9 3.31 41.0 205 289 37.1 928 1015 142 128 0 128 128
10 3.57 56.6 283 457 58.7 1467 1038 703 633 0 633 761
11 3.82 72.7 364 571 73.3 1833 626 1505 1355 0 1355 2116
12 4.08 90.6 453 602 77.3 1932 724 1592 1433 0 1433 3548
13 4.35 110.4 552 604 77.6 1939 806 1612 1451 0 1451 4999
14 4.58 128.3 642 567 72.8 1820 1038 1363 1227 0 1227 6226
15 4.70 138.0 690 434 55.7 1393 985 1050 945 0 945 7171
16 4.67 135.6 678 273 35.1 876 1038 493 444 0 444 7614
17 4.46 118.9 594 100 12.8 321 985 68 0 85 85 7530
18 4.33 108.9 545 7 0.9 22 724 153 0 191 191 7339
19 4.16 96.3 482 0 0 0 895 402 0 502 502 6837
20 4.01 85.7 428 0 0 0 1069 624 0 780 780 6057
21 3.88 76.7 384 0 0 0 1015 618 0 773 773 5284
22 3.81 72.0 360 0 0 0 1069 697 0 872 872 4412
23 3.81 72.0 360 0 0 0 1015 649 0 811 811 3601
24 3.82 72.7 364 0 0 0 368 16 0 20 20 3581

the hourly meteorological data, the desired value of the loss of standalone hybrid PV/wind system in order to power supply
power supply probability and techno-economic constraints. The residential, located in the city Oujda Morocco (latitude : 34° 410 ,
MESCA method can be used to determine the optimal design longitude 1° 540 ).
according two algorithms: the first algorithm to size the According to the simulation results for the case study used, the
standalone hybrid solar-wind system for T = 288 h and the second optimal design (26 PV panels, a 8 wind turbines and a 8 battery
algorithm to correct the sizing for T = 8760 h if the obtained LPSP is bank) has been found to be the optimum hybrid configuration,
superior to the desired value of the LPSP by the user. The cost of the energy is 0374 €/kWh, the proposed
The validation of the MESCA method has been performed in wind–PV-battery hybrid system is reasonable with the TAC equal
comparison with the solution provided by HOMER software, for to 2391 €.
 H. Zahboune et al. / Energy Conversion and Management 126 (2016) 909–922 921

Table 8
HCT for NW = 5, NPV = 25, NBat = 8 and Eacc (max) = 25.5 kWh month: August.

t v Ew Ew. T Rs EPV EPV,T EL Er Chbat Dbat EN Eacc

(h) (m/s) (Wh) (Wh) (W/m2) (Wh) (Wh) (Wh) (Wh) (Wh) (Wh) (Wh) (Wh)
1 2.13 0 0 0 0 0 296 330 0 412 412 16,104
2 2.20 0 0 0 0 0 302 336 0 420 420 15,684
3 2.24 0 0 0 0 0 296 329 0 411 411 15,273
4 2.23 0 0 0 0 0 302 335 0 419 419 14,854
5 2.26 0 0 0 0 0 296 329 0 411 411 14,443
6 2.23 0 0 14 1.80 45 673 703 0 879 879 13,565
7 1.97 0 0 134 17.20 430 660 322 0 402 402 13,162
8 1.81 0 0 340 43.70 1091 704 258 232 0 232 13,395
9 1.85 0 0 560 71.90 1798 690 945 851 0 851 14,245
10 2.02 0 0 739 94.90 2372 686 1496 1346 0 1346 15,592
11 2.26 0 0 849 109.00 2725 457 2085 1876 0 1876 17,468
12 2.51 0 2 902 115.80 2895 504 2197 1977 0 1977 19,445
13 2.75 11.4 57 885 113.60 2841 605 2087 1878 0 1878 21,323
14 3.00 24.0 120 816 104.80 2619 740 1787 1609 0 1609 22,932
15 3.27 38.7 194 656 84.20 2106 724 1389 1250 0 1250 24,182
16 3.56 56.0 280 481 61.80 1544 740 919 827 0 827 25,009
17 3.78 70.1 350 288 37.00 924 755 382 344 0 344 25,353
18 3.89 77.4 387 114 14.60 366 538 125 113 0 113 25,466
19 3.61 59.1 296 13 1.70 42 664 407 0 509 509 24,957
20 3.06 27.1 136 0 0 0 761 705 0 881 881 24,076
21 2.63 6 29 0 0 0 743 786 0 982 982 23,094
22 2.36 0 0 0 0 0 761 832 0 1040 1040 22,054
23 2.15 0 0 0 0 0 743 812 0 1015 1015 21,039
24 2.06 0 0 0 0 0 314 343 0 429 429 20,609

The MESCA result is validated and compared with the optimum [9] Shiroudi A, Rashidi R, Gharehpetian GB, Mousavifar SA, Akbari Foroud A. Case
study: Simulation and optimization of photovoltaic-wind-battery hybrid
hybrid configuration determined by Homer Pro, based on the con-
energy system in Taleghan-Iran using homer software. J Renew Sustain
straints and sensitivity variables. Firstly, the comparison shows Energy 2012;4:053111.
that the total annual electricity productions of these two methods [10] Prasetyaningsari I, Setiawan A, Setiawan AA. Design optimization of solar
are substantially equal. After meet the power demand and charg- powered aeration system for fish pond in Sleman Regency, Yogyakarta by
HOMER software. Energy Procedia 2013;32:90–8.
ing the storage unit fully, the difference between the excess elec- [11] Sen R, Bhattacharyya SC. Off-grid electricity generation with renewable energy
tricity of the two results hybrid configuration determined by technologies in India: an application of HOMER. Renew Energy
MESCA and Homer Pro, is 1.2%. Secondly, the methods validation 2014;62:388–98.
[12] Bekele G, Tadesse G. Feasibility study of small Hydro/PV/Wind hybrid system
is made by comparing the costs. The difference between the two for off-grid rural electrification in Ethiopia. Appl Energy 2012;97:5–15.
results for TAC and LCE are -1.719% and -0.07%, respectively. [13] Gan LK, Shek JKH, Mueller MA. Hybrid wind – photovoltaic – diesel – battery
The MESCA method can be used for sizing the hybrid system, to system sizing tool development using empirical approach, life-cycle cost and
performance analysis : a case study in Scotland. Energy Convers Manage
supply whatever the isolated systems. The extension of this work 2015;106:479–94.
includes the use of MESCA method with other sources of green [14] Sigarchian SG, Malmquist A, Fransson T. Modeling and control strategy of a
energy (e.g. Geothermal) and storage systems (Hydrogen storage). hybrid PV/Wind/Engine/Battery system to provide electricity and drinkable
water for remote applications. Energy Procedia 2014;57:1401–10.
[15] Zahboune H, Kadda FZ, Zouggar S, Ziani E, Klemes JJ, Varbanov PS, et al. The
Acknowledgment new electricity system cascade analysis method for optimal sizing of an
autonomous hybrid PV/wind energy system with battery storage. IREC 2014-
5th Int Renew Energy Congr. p. 6.
The authors would like acknowledge the EC supported project [16] Krajacic G, Martins R, Busuttil A, Duic N, Carvalho M da G. Hydrogen as an
‘‘Distributed Knowledge-Based Energy Saving Networks” DISKNET, energy vector in the islands’ energy supply. Int J Hydrogen Energy 2008;33
Grant Agreement No: PIRSES-GA-2011-294933. (4):1091–103.
[17] Segurado R, Krajacic G, Duic N, Alves L. Increasing the penetration of
renewable energy resources in S. Vicente, Cape Verde. Appl Energy 2011;88
References (2):466–72.
[18] Krajacic G, Duic N, Carvalho M da G. H2RES, Energy planning tool for island
[1] International Energy Agency. Energy and climate change (2015) World Energy energy systems – The case of the Island of Mljet. Int J Hydrogen Energy
Outlook Special Report OECD/IEA. International Energy Agency (IEA). 2009;34(16):7015–26.
[2] Petruschke P, Gasparovic G, Voll P, Krajačić G, Duić N, Bardow A. A hybrid [19] Østergaard PA. Reviewing EnergyPLAN simulations and performance indicator
approach for the efficient synthesis of renewable energy systems. Appl Energy applications in EnergyPLAN simulations. Appl Energy 2015;154:921–33.
2014;135:625–33. http://dx.doi.org/10.1016/j.apenergy.2015.05.086.
[3] Panayiotou G, Kalogirou S, Tassou S. Design and simulation of a PV and a [20] Wierzbowski M, Lyzwa W, Musial I. MILP model for long-term energy mix
PVeWind standalone energy system to power a household application. Renew planning with consideration of power system reserves. Appl Energy
Energy 2012;37:355–63. 2016;169:93–111. http://dx.doi.org/10.1016/j.apenergy.2016.02.003.
[4] Enevoldsen P, Sovacool BK. Integrating power systems for remote island [21] Shah KK, Mundada AS, Pearce JM. Performance of U.S. hybrid distributed
energy supply: lessons from Mykines, Faroe Islands. Renew Energy energy systems: Solar photovoltaic, battery and combined heat and power.
2016;85:642–8. Energy Convers Manage 2015;105:71–80.
[5] Zakeri B, Syri S. Electrical energy storage systems: a comparative life cycle cost [22] Jirı Klemes, Friedler F, Bulatov I, Varbanov P. Sustainability in the Process
analysis. Renew Sustain Energy Rev 2015;42:569–96. Industry: Integration and Optimization; 2010.
[6] Mahesh A, Sandhu KS. Hybrid wind/photovoltaic energy system [23] Nemet A, Klemeš JJ, Varbanov PS, Mantelli V. Heat Integration retrofit
developments: Critical review and findings. Renew Sustain Energy Rev analysis—an oil refinery case study by Retrofit Tracing Grid Diagram. Front
2015;52:1135–47. Chem Sci Eng 2015;9(2):163–82.
[7] Chua KH, Lim YS, Morris S. Cost-benefit assessment of energy storage for utility [24] Klemeš JJ, Varbanov PS, Liew PY, Čuček L, Kravanja Z, Wan Alwi SR, Abdul
and customers: a case study in Malaysia. Energy Convers Manage Manan Z. Recent developments in advanced process integration: learning the
2015;106:1071–81. lessons from industrial implementations. Appl Mech Mater 2014;625:454–7.
[8] Adaramola MS, Agelin-Chaab M, Paul SS. Analysis of hybrid energy systems for [25] Klemeš JJ, Kravanja Z. Forty years of heat integration: Pinch Analysis (PA) and
application in southern Ghana. Energy Convers Manage 2014;88:284–95. Mathematical Programming (MP). Curr Opin Chem Eng 2013;2:461–74.
922 H. Zahboune et al. / Energy Conversion and Management 126 (2016) 909–922

[26] Wan Alwi SR, Mohammad Rozali NE, Abdul-Manan Z, Klemeš JJ. A process [32] Ho WS, Hashim H, Hassim MH, Muis ZA, Shamsuddin NLM. Design of
integration targeting method for hybrid power systems. Energy 2012;44:6–10. distributed energy system through Electric System Cascade Analysis (ESCA).
[27] Mohammad Rozali NE, Alwi SRW, Manan ZA, Klemeš JJ, Hassan MY. Process Appl Energy 2012;99:309–15. http://dx.doi.org/10.1016/j.apenergy.2012.
integration techniques for optimal design of hybrid power systems. Appl 04.016.
Therm Eng 2013;61:26–35. [33] Bhattacharjee S, Acharya S. PV–wind hybrid power option for a low wind
[28] Mohammad Rozali NE, Wan Alwi SR, Abdul Manan Z, Klemeš JJ, Hassan MY. topography. Energy Convers Manage 2015;89:942–54. http://dx.doi.org/
Process integration of hybrid power systems with energy losses considerations. 10.1016/j.enconman.2014.10.065.
Energy 2013;55:38–45. http://dx.doi.org/10.1016/j.energy.2013.02.053. [34] Baghdadi F, Mohammedi K, Diaf S, Behar O. Feasibility study and energy
[29] Mohammad Rozali NE, Wan Alwi SR, Abdul Manan Z, Klemeš JJ, Hassan MY. conversion analysis of stand-alone hybrid renewable energy system. Energy
Optimal sizing of hybrid power systems using power pinch analysis. J Clean Convers Manage 2015;105:471–9. http://dx.doi.org/10.1016/j.enconman.
Prod 2014;71:158–67. http://dx.doi.org/10.1016/j.jclepro.2013.12.028. 2015.07.051.
[30] Mohammad Rozali NE, Wan Alwi SR, Manan ZA, Klemeš JJ. Peak-off-peak load [35] Li C, Ge X, Zheng Y, Xu C, Ren Y, Song C, et al. Techno-economic feasibility
shifting for hybrid power systems based on Power Pinch Analysis. Energy study of autonomous hybrid wind/PV/battery power system for a household in
2015;90:128–36. http://dx.doi.org/10.1016/j.energy.2015.05.010. Urumqi. China Energy 2013;55:263–72. http://dx.doi.org/10.1016/j.energy.
[31] Ho WS, Tohid MZWM, Hashim H, Muis ZA. Electric System Cascade Analysis 2013.03.084.
(ESCA): solar PV system. Int J Electr Power Energy Syst 2014;54:481–6. http://
dx.doi.org/10.1016/j.ijepes.2013.07.007.