Received November 30, 2020, accepted December 21, 2020, date of publication December 31, 2020,

date of current version January 26, 2021.

Digital Object Identifier 10.1109/ACCESS.2020.3048487

Reliability Support of Undependable Grid Using

Green Energy Systems: Economic Study
MOHAMED MAHMOUD SAMY 1, MOHAMED I. MOSAAD 2, MOHAMED F. EL-NAGGAR3,4 ,
AND SHIMAA BARAKAT1
1 Facultyof Technology and Education, Beni-Suef University, Beni Suef 62521, Egypt
2 Yanbu Industrial College (YIC), Yanbu 30436, Saudi Arabia
3 Department of Electrical Engineering, College of Engineering, Prince Sattam Bin Abdulaziz University, Al-Kharj 11942, Saudi Arabia
4 Department of Electrical Power and Machines Engineering, Faculty of Engineering, Helwan University, Helwan 11795, Egypt

Corresponding author: Mohamed I. Mosaad ([email protected])

ABSTRACT Developing countries’ energy sector faces a multitude of challenges, ranging from inadequate
generation to unstable grids. Power outages are among the most common issues, particularly in remote
areas. Utilizing grid-tied green energy resources to address this issue and to cover for power outages from
local grids. This article presents a cost-effective design of a grid-tied, hybrid green energy system (GES)
consisting of wind, PV, and batteries considering the influence of the grid availability. A multi-objective,
optimal techno-economic design, optimized by multi-objective particle swarm optimization technique is
presented for the grid-tied GES linked to a small hamlet in the north of Egypt. The multi-objective function
introduced in this work includes three objective functions which are, the Loss of Power Supply Probability
(LPSP), the Cost of Energy (COE), and the System Surplus Energy Rates (SSER) considering the grid
availability. The grid availability (GA) of 100% was considered as a base case and it was reduced to 70%
with a step of 5%. The simulation consequences had cleared that the lowest and largest percentage values of
SSER were obtained at GA of 85 %, and 70 %, respectively. When the value of SSER equal to 0.33%, the
system design for solving the grid unavailability consists of 12 PVs, one WT, and 1420 batteries with COE
of 0.145$/kWh and TNPC of 3,699,800 ($).

INDEX TERMS Grid availability, multi-objective optimization, loss of power supply probability, system
surplus energy rates, sizing, and optimization.

ACRONYMS AND NOMENCLATURE G Global utility fallen on the titled

AASI Average availability service index plane (kW/m2)
Cgrid Cost of energy purchase Gref Solar radiation at base conditions (1 kW/m2)
CI Initial capital cost of system devices ($) I Interest rate
CO&M Operation and maintenance cost If Inflation rate
CR Replacement cost Ir Real interest rate
Df PV derating factor (0.85) KT Peak power temperature coefficient
Eb Amount of energy in the battery Nbat Number of batteries
EBat,cap Battery capacity Np Number of pareto front points
Ebat,max Maximum allowable capacity of the battery NPV Number of PV modules
Ebat,min Minimum allowable capacity of the battery NWT Number of the wind turbines
Egrid_buy,t Energy purchased from the utility at assured Pch,b Charging power of the battery
period, t. Pdis,b Discharging power of the battery
Egrid_sell,t Energy sold to the utility at assured period, t. PPV PV produced power (kW)
Eserved Primary load served (kWh/year) Pr Rated wind turbine power (kW)
Prated Module output power (kW)
Pre Power produced by the renewable system
The associate editor coordinating the review of this manuscript and at hour t
approving it for publication was Ton Duc Do . Pw Produced power of the wind turbine (kW)

This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
14528 VOLUME 9, 2021
M. M. Samy et al.: Reliability Support of Undependable Grid Using GESs: Economic Study

Rgrid Sales of electricity to the public utility Mondal and Islam analyzed the frugal viability of
ratefeed−in Feed-in tariff rate utility-connected PV systems and concluded that this system
S Salvage value must be utilized to overcome power shortages [2]. Robert and
Tamb Ambient temperature (◦ C) Goplan had made twenty scenarios to investigate the integra-
Tc PV cell temperature (◦ C) tion of GES into the grid and concluded that the utility-linked
Tcomp Component lifespan (year) GES is the most expense-efficient energy system compared to
Tproj Project lifespan (year) the island systems [3]. The optimal sizing of grid-connected
Tref Temperature at base conditions (◦ C) three renewable energy sources (PV, wind, and fuel cell) was
Vci Cut-in wind speed (m/s) investigated to supply some domestic loads [4].
Vco Cut-out wind speed (m/s) A grid-linked grouped heat and energy project for sup-
Vr Rated wind speed (m/s) plying household loads was investigated with several tar-
iffs for buying and selling electricity from the utility [5].
A model was developed for estimating the energy produc-
ABBREVIATIONS tion, selling electricity to the local grid using wind turbine
COE Cost of electricity (WT)/Hydrogen system, thermal recovery from fuel cells
CRF Capital recovery factor (FCs) [6]. The potential of both off-grid and on-grid PV, WT,
DOD Depth of discharge FC, and biomass hybrid systems for the electrical energy
GA Grid Availability generation with a focus on the optimal size of the GES has
GES Green Energy System been inspected [7]–[14].
LPSP Loss of Power Supply Probability A detailed study about how to overcome the power outages
MOPSO Multi-Objective Particle Swarm Optimization in a rural area in Egypt by integrating a hybrid PV/WT/FC to
NWCT Normal working cell temperature the electrical utility was presented in [14]. This study focused
PPA Power Purchase Agreements on maximizing the profit from integrating these three GESs
SSER System Surplus Energy Rates to the Egyptian electric utility considering the cost of these
Pch,b Charging power of the battery GES sources. The objective function in this study was a single
Pdis,b Discharging power of the battery objective function with one objective, which is to minimize
PPV PV produced power (kW) the cost of electricity (COE). It is worth noting that this study
Pr Rated wind turbine power (kW) had presented a proposal to use FCs as storage units instead
Prated Module output power (kW) of batteries.
Pre Power produced by the renewable system Another investigation for the optimal size of an island
at hour t hybrid GES supplies a new society in Egypt was presented
TNPC Total net present cost in [15], [16]. In the preceding studies, the optimal size with
WT Wind turbine minimum capital and running costs for this proposed feeding
system was introduced. Wang et al. investigated the common
GREEK SYMBOL consequences of combining renewable energy with the elec-
ηb Efficiency of each battery trical network on provincial energy efficiency. The authors
ηinv Inverter efficiency have concluded that the effect of renewable energy genera-
ηw Wind turbine efficiency tors integrated into extra-high voltage and ultra-high voltage
power grids are either not noteworthy or even negative [17].
Kamal, et al proposed a grid-connected GES including WT,
I. INTRODUCTION solid oxide FC, Electrolyzer, and batteries to serve the electric
Instability, power block-outs, significant distribution network demand of twenty-five homes in Islamabad, the metropolis
outages, and frequently scheduled load shedding impact the of Pakistan [18]. A comprehensive literature exploration was
centralized electricity grids, especially the developing coun- conducted to assemble accessible information about unreli-
tries. The reliability of the grid is determined based on the able grid modeling but went over a couple of such exami-
ratio of the total number of customer hours that service was nations [19]. Murphy et al completed an analytical survey
available during a given period to the total customer hours for Uganda dealing with the uncertain utility by devising a
requested, which is known as the average availability service method in which a producer is produced with one hundred
index (AASI). The AASI is expected to be 0.9988 for a efficacy percentage. The survey notified that uncertain utility
reliable grid whereas an unreliable grid is characterized by consideration raised the estimate of energy contrasted to 24h
random power outages [1]. The best way to address these availability of the utility [20]. Harish et al. [21] had made a
grid issues and consequently supporting the grid reliability comparison among utility expansion and distributed genera-
is widely incorporate the production of clean green energy tion (DG) alternates to solve the problem of power outages in
into the power distribution networks. Grid-connected green rural India. Various utility reliability tactics were simulated.
energy systems (GESs) could be an applicable option to The authors concluded that off-grid DG does not seem to
provide sustainable electric outfits to developing countries. be competitive with utility expansion at distances less than

VOLUME 9, 2021 14529

 M. M. Samy et al.: Reliability Support of Undependable Grid Using GESs: Economic Study

seventeen kilometers. This is because whenever the location stable and cost-effective method. While Singh et al. [27]
of the new society or the village is far from the network, on- recommended the reformed electric system cascade analy-
grid GES is the best choice from the frugal viewpoint. While sis (RESCA) approach to design the hybrid renewable energy
an island GES is a better choice for closer locations from the system (HRES), using wind conversion systems, photovoltaic
network. systems, the battery storage system, the non-intermittent
The GES integration to the electric utility is a rich area of source, and the grid as components. Four distinct HRES struc-
research in recent times to solve both the energy supply and tures are considered for study in off-grid and grid-connected
environmental problems. This research area has many chal- mode. The four HRES configurations are configured with
lenges. These challenges are summarized around the optimal restrictions from RESCA, such as final surplus capacity,
utilization of the available resources through sound economic renewable energy rate, LPSP, and annual system cost (ASC).
studies especially for the new communities and through track- The topic of integrating renewable energy sources in the
ing the maximum power point for optimal exploitation of the unreliable grid is a major technical and economic challenge,
available GES [22], [23]. especially for small or rural villages.
As discussed before, the economic study of such GES inte- The issue of unreliable grids is addressed in this article
grated into the electrical utility is a difficult issue and almost since these grids commonly suffer from the problem of fre-
to the stage of complexity. This difficulty of such issues quent power outages a grid-connected green energy system
is related and not limited to, a large number of variables, composed of solar PV, wind turbines, and battery banks have
the non-linear characteristics of GES, and the dependence of been suggested as a backup system used to compensate for
these GES on some environmental factors which in turn affect blackouts.
the power generated from these GES. Evolutionary comput- In this work, the situation where the grid is completely
ing techniques are used for such mixed-complex optimization available was first considered, and then the availability of the
problems. grid was decreased by 5 percent per time from 95 percent to
Genetic and particle swarm optimization (PSO) mecha- 70 percent to show the influence of network availability on
nisms were applied to economic studies of off-grid GES [24]. the design of the proposed hybrid system.
This study is presented also a comparison between these The optimal number of components of the grid-connected
two optimization mechanisms and the results showed the PV/WT/battery system has been identified by utilizing the
superiority of PSO. Multi-Objective Particle Swarm Optimization (MOPSO)
Harmony search (HS), enhanced flower pollination algo- which was used for reducing three objective functions of Loss
rithm (FPA), an electromagnetic field optimization is utilized of Power Supply Probability (LPSP), Cost of Energy (COE),
for enhancing the performance of grid-connected GES in the and System Surplus Energy Rates (SSER) under varying
form of FC [25]. conditions in the weather; by using actual on-site hourly
Rezvani et al. utilized lexicographic optimization and power demand data for the studied village. The Pareto front
hybrid augmented weighted epsilon-constraint algorithms for and decision-making processes are used to obtain the final
the optimal tabulating of a micro-grid to mitigate both the optimal system.
overall working estimation and the emission that was resulted This article is structured as follows: Section 1 addresses the
from the generating stations. The results showed that the background and literature review of the study. Section 2 lays
proposed hybrid optimization technique leads to good out- out the methodology. Details of the Fitness Function and
standing solutions [6]. Improvement of off-grid hybrid sys- Energy Management Scenarios are explained in Section 3.
tem performance was improved by optimal PI controllers Section 4 discusses the findings of the simulation and the
optimized by the cuckoo search mechanism [23]. This opti- optimization of the study case. Conclusions are outlined in
mization technique is applied to achieve maximum power section 5.
point tracking of three GESs (PV/WT/FC) and consequently
achieving the maximum economic benefit and reducing the II. METHODOLOGY
cost of generation. The extreme learning machine (ELM) The objective of this research is to manifest the techno-
mechanism was used to develop a model to forecast the economic feasibility through a case study of a village in
amount of the produced power of utility-tied PV system Northern Egypt. This hybrid GES comprises solar PV genera-
installed at a roof-top of PEARL laboratory at the Univer- tors, wind turbines, and battery banks. The system is attached
sity of Malaya, Malaysia [19]. The consequences from the to an unreliable grid for remote rural electrification.
introduced model were contrasted with those obtained from MOPSO has been used to design a hybrid power system to
the support vector regression and the artificial neural network assess the optimum size of its components by performing a
(ANN). The comparison illustrated that the ELM model has techno-economic analysis. Figure 1 shows the approach used
higher exactness with less predestined time. for the effective configuration and the techno-economic study
The risk theory into the allocation and placement of dis- of the proposed scheme.
tributed generation (DG) units, [26] suggested a risk-based, Information concerning energy demand and environmen-
multi-objective optimum allocation model to maximize the tal data, such as solar radiation, ambient temperature,
placement and configuration of DG units and provide a and wind speed were collected for one year between the

14530 VOLUME 9, 2021

M. M. Samy et al.: Reliability Support of Undependable Grid Using GESs: Economic Study

FIGURE 3. The selected village’s annual solar radiation.

FIGURE 1. Framework for optimal design and techno-economic

assessment of the proposed system.

FIGURE 4. Average monthly wind velocity during the year.

FIGURE 2. The daily load profile of the selected village.

1st January 2019 and the 31st December for Qesm Rem-
anah, North Sinai, Egypt (latitude 31.03641 ◦ N, longitude
32.62705 ◦ E). Renewable energy supplies with an estimated
annual solar radiation of 2027kWh/m2 and an average wind
velocity of 4.22 m/s, are the main feature of this area. This
analysis uses actual hourly data in-site for the village energy
demand, with an average of 457 kWh hourly, an average
demand of 897 kW, and a load factor of 0,509. The average
daily load for the selected area is seen in Figure 2. The solar
radiation and the wind speed of the field under study were
shown in Figures 3 and 4.

A. DESCRIPTION OF THE GRID-TIED SYSTEM

The renewable energy system presented in this work includes
two renewable sources which are PV and WT systems. These FIGURE 5. General schematic of the suggested system.
renewable energy sources are connected to an unreliable grid.
The system components include an inverter, converter, charge The grid blackout intervals are randomly modeled.
controller, and battery banks. Figure 5 presents a straightfor-
ward graphic of the given system. Generally, the grid avail- 1) PV MODEL
ability for a particularized period (e.g. a day, week, month, The produced energy by the PV arrays was estimated as [28],
year) can be defined as [30]:
 

G
PPV = Prated Npv Df
Grid availability GA (%)  
Gref
  
Grid available in period NWCT −20
= × 100 (1) × 1+KT Tamb +G −Tref (2)
Hoursi theperiod 0.8
VOLUME 9, 2021 14531
 M. M. Samy et al.: Reliability Support of Undependable Grid Using GESs: Economic Study

The solar system consists of 80 solar panels, 375W each, 4) INVERTER

and three inverters each of 10kW to connect the PV system to The output of both the WT and PV system is DC. This DC
the AC grid side with a 25-years old lifetime assumed [29]. output of the two GREs is connected to a DC link. To attach
The PV system along with the three inverter‘s capital cost is this DC link to the electrical grid at the AC bus, an inverter is
42,000 $ with annual maintenance of 40$ [30]. needed. As the proposed on-grid hybrid system is embedded
with a battery (as a storage system), the inverter used is a
2) WIND TURBINE bi-direction type to allow charging the batteries from the AC
The generated power by the wind turbine was approximately side, if the output power of the REGs is not sufficient as
computed as [7], [31]. shown in Figure 5.

 NWT × ηw × Pr The size of the inverter depends primarily on the demand

 ! for peak load
V (t)3 − Vci3

 XT
, V ≤ Vr

 × max
Pw = t=1 Vr3 − Vci3 Pinv = PL ηinv (9)
NWT × ηw × Pr ,

Vr ≤ V ≤ Vco




 5) GRID MODELLING
0, Vco ≤ V or V ≤ Vci

The grid is represented as an energy supply that can
(3) absorb/generate electric power. Because of the restrictions
where: V(t) is the adjusted wind speed at the hub high in (m/s) imposed on power transformers, transmission lines, and other
that can be defined as: heavy power equipment of the grid, there is a border on how
HWT α much energy in a specific time frame can be absorbed or
 
V (t) = Vr (4) provided. The following equations are used to model a grid
Hr
energy exchange [39]:
where HWT is the wind turbine hub height, Hr is the base
height, and α is the friction coefficient, and equal 0.1428 for Egrid,t ≤ Egrid_max (10)
low roughness, surface, and well-exposed site [32], [33]. Egrid_buy,t < 0 (11)
In this analysis, a 50-kW wind turbine is used. A wind Egrid_sell,t > 0 (12)
turbine’s capital is $50,000, with $25,000 and $ 75 a year as
a replacement and O&M expenses [34]. The wind turbine is The uncertain utility contains accidental or random power
20 years with a hub height of 20 m. outages. The frequency, the time interval, and the diversity
in blackouts have an opposite influence on the estimate of
3) BATTERY BANK utility-tied GES.
In the hybrid system, the batteries are used as storage media.
During the low power demand, the battery bank saves the III. FITNESS FUNCTION
surplus electricity and releases the stored energy at higher A. ENERGY MANAGEMENT SCENARIOS
loads or when the clean energy sources are inadequate. For the suggested system, three scenarios are proposed:
Because of its numerous benefits, lithium-ion batteries • Scenario 1, the generated electricity from the GES
have been used in this research. These batteries have a depth equals the required demand: in this case, neither the
of discharge of up to 100% (DOD), with a self-discharge battery bank nor the grid will participate in the load
of 1 % per month, and operating efficiency of 95% [35], [36]. demand.
A dynamic model is used to describe the operating of the • Scenario 2, the produced electricity from the GES’s is
battery, in which the energy stored in the battery is calculated greater than the demand and the excess energy will be
at any given interval as: stored in the batteries. The surplus energy could be used
Eb (t) = Eb (t −1) (1−σ )+ Pb,ch (t)×ηb −Pb,dis (t) ×dt for charging the batteries, the energy will be sold to


the grid if the grid is available else this energy will be
(5)
directed to a dump load.
Noting that during the discharging process, the battery • Scenario 3, the energy generated from the GES’s is not
discharging efficiency is set equal to 1. enough to meet the demand. In this case, the deficit
The energy storage of the batteries is subjected to the will be drawn from the battery bank, but if the shortage
following constraints: hits the minimum level, the remaining demand will be
EBat,min (t) ≤ Eb (t) ≤ EBat,max (t) (6) purchased from the grid.
EBat,max = EBat,cap (7) A flowchart of the energy administrative scenarios of the
submitted system is cleared in figure 6.
EBat,min = EBat,max (1 − DOD) (8)
A Li-ion battery (LiFePO4 type) is used in this study, with B. PROBLEM FORMULARIZATION
capital and replacement costs of $630 / kWh, $2 per year as The main objective of this work is to optimize the generat-
O&M costs, and 15 years of a lifetime [37], [38]. ing cost of electrical energy while satisfying the stability of

14532 VOLUME 9, 2021

M. M. Samy et al.: Reliability Support of Undependable Grid Using GESs: Economic Study

the system and assessing the hybrid system’s usability and contented, and if its value is one then the load demand will
performance. eternally be served [42], [43].
The cost of electricity (COE) refers to the costs of produc- The LPSP can be described as:
ing electricity over the lifespan of the project. PT
Pdeficit (t) .1t
It is a very valuable indicator because, while the rating LPSP = PTt=1 , T = 8760 (23)
power and investment costs are different, it permits a distinc- t=1 Pdemand (t) .1t
tion between diverse energy technologies [39], [40]. The SSER is the percentage of the surplus energy in the
The COE of the grid-connected PV/WT/battery hybrid hybrid system compared with the total renewable energy
system is defined as: generated by GES [44]. The SSER is subject to the constraint
(CRF × TNPC) + Cgrid − Rgrid 0 ≤ SSER ≤ 1.
COE = (13) P
Eserved + Egridselling surplus energy
SSER = P (24)
where: Capital recovery factor (CRF) is a ratio used over EPV + EWT
the life of a project to determine the total value of a set of The fitness function for this multi-objective optimization
equivalent cash flows. The estimation of this factor is given can be expressed as follows.
by: 
 F1 = COE
Ir −If N
  
Ir −If minimize : F2 = LPSP (25)
1+If 1 + 1+If
F3 = SSER

CRF =  (14)
Ir −If N
  min
1 + 1+If −1 NPV ≤ N PV ≤ NPV max

 min max
 NWT ≤ N WT ≤ NWT


TNPC is the total net present cost of the system which is min max
subjct to : Nbat ≤ N bat ≤ Nbat (26)
composed of the NPC of the PV, WT, and battery banks.
0 ≤ LPSP ≤ LPSP , LPSP = 2%

max max


0 ≤ SSER ≤ SSERmax , SSERmax = 15%

Nproj 
X
TNPC = (CI (n) + CR (n) + CO&M (n) − S (n))
n=0 C. MULTI-OBJECTIVE PARTICLE SWARM OPTIMIZATION
1 ALGORITHM (MOPSO)
× (15)
(1 + Ir )n MOPSO technique had been used in this article to tackle the
energy management dilemma.
X
CI = NPV CIPV + NWT CIWT + Nbat CIbat (16)
X Therefore, the MOPSO algorithm is utilized to achieve
CR = NPV CRPV + NWT CRWT + Nbat CRbat (17) the Pareto front of the optimization problem. The MOPSO
X procedure (Figure 7) in this research work was adapted
CO&M = NPV CO&M PV + NWT CO&M WT + Nbat CO&M bat
from [45], [46].
(18)
X
S = NPV SPV + NWT SWT + Nbat Sbat (19) D. DATA PREPARATION FOR THE SIZING OPTIMIZATION
PROCESS
The salvage value (S) was estimated by:
    The hourly load demand, the solar irradiance (kWh per m2
Nproj per day), the ambient temperature (◦ C), and the wind speed

Ncomp − Nproj − Ncomp × INT Ncomp
S = CR   (20) (m/s) for the hamlet used in this study are recorded in 2019
Ncomp (8760 points). A code for the MOPSO technique is progressed
in MATLABTM software package is introduced to minimize
Sales of electricity to the public utility and cost of energy
the multi-objective function proposed in this study based on
purchase can be estimated as:
the recorded data all over 2019 for this village.
8760
X
Rgrid = ratefeed−in .Egridselling (21) IV. RESULTS AND DISCUSSIONS
t=1 This article suggests a complementary PV / WT / Battery
8760
X system linked to an unreliable grid to support the grid per-
Cgrid = Cp × Egridpurchased (22) formance through compensating the power outage and con-
t=1 sequently increase the reliability of the system in addition to
where: ratefeed−in is the feed-in tariff rate (0.0617 $/kWh) and minimizing the generating cost for a village in Egypt.
Cp is equivalent to $0.0425 / kWh baased on the Egyptian grid The grid unavailability was expressed by the grid avail-
tariff [41]. ability factor (GA), in which the whole situation of the grid
The loss of power supply probability is described as the availability (GA = 100%) was considered as a base case, and
prospect that an inadequate power supply occurs when the then GA is reduced to 70% with an increment of 5%.
GES is incapable of meeting the load needs. The loss of power The optimized objectives presented in this article are to
supply probability of zero value means the load demand is minimize the COE, the LPSP, and the SSER where COE is

VOLUME 9, 2021 14533

 M. M. Samy et al.: Reliability Support of Undependable Grid Using GESs: Economic Study

FIGURE 6. Flowchart of the energy administration scenarios of the submitted system.

used to assess hybrid system economies; LPSP is for relia- A. THE EFFECT OF GA ON COE
bility assessments, and SSER is to measure hybrid system The effect of the GA percentage change on the proposed
usability and efficiency. The optimization variables include system’s COE can be seen in Figure 8, where it can be
the total number of WT, PV, and battery banks. observed that in the base GA case when GA = 100 percent,
The Pareto front estimation of the MOPSO algorithm for the MOPSO has achieved an optimal solution where the
GA from 100% to 70% with a step of 5% is shown in Figure 8. COE was $0.098/kWh, but when this percentage decreased
The Pareto front allows the decision-maker to choose an to 95 percent, the COE value increased to $0.1263/kWh.
optimum solution based on expectations for lower COE, It can be noticed that COE increases with decreasing GA
LPSP, or SSER from the optimal solutions collection. percentage.
The compromise between the three objective functions is
clearly shown in Figure 8. From this figure, the effect of B. THE EFFECT OF GA ON LPSP
GA on the three objective functions can be concluded as The maximum LPSP was set as 2% in the fitness function
follows: constraint, Figure 8 indicates that the MOPSO obtained an

14534 VOLUME 9, 2021

M. M. Samy et al.: Reliability Support of Undependable Grid Using GESs: Economic Study

TABLE 1. MOPSO results of the sizing problem for different GA.

optimum solution, the point that shows the minimum dis-

tance from other points on the front of Pareto is selected
and described as [39], (27), as shown at the bottom of the
page, where, [COE (i) − COE (k)] , [LPSP (i) − LPSP (k)],
FIGURE 7. The MOPSO established procedure. and [SSER (i) − SSER (k)] denotes three random points on
the Pareto front and Np is the total number of points of the
Pareto front. The final optimal solution for each GA is given
optimal solution with a low LPSP of 0.01 percent in the base in Table 1.
GA scenario, but when this percentage dropped to 95 percent, Table 1 explicitly illustrates the impact of increasing the
the LPSP value increased to 0.62 percent. It can be observed GA ratio on optimum system configuration and shows the
that LPSP rises with declining GA percentage, but not lin- impact of the selected three objective functions. The greatest
early because the amount of battery banks has risen, which influence is on the number of batteries used in the pro-
affects the LPSP percentage. posed system is indicated. The less availability of the grid,
the greater the need to store and use energy through batteries.
C. THE EFFECT OF GA ON SSER The table also illustrates that the rising number of batteries
The maximum SSER was set at 15% in the objective function used in the system retained a low LPSP level, despite the
constraint, Figure 8 reveals that the MOPSO obtained an absence of a completely available grid, which confirms that
optimal solution with zero SSER in the base GA scenario, the proposed system was extremely reliable.
however, when this percentage drops to 95%, the SSER value It is also observed that a reasonable SSER was maintained
increased to 2,88%. It can be noted that SSER increases with as the lowest percentage (0.33 %) in the case of GA was
decreasing GA percentage, but not linearly as the number of 85 %, while the largest percentage (9.78 %) in the case of
battery banks has increased, impacting the SSER percentage. GA was 70 %.
The minimum SEER for GA was 85 percent, while the mean Through analyzing the impact of the grid availability rate
SSER for GA was 70%. SEER hits its lowest value when on the cost of energy production of the proposed scheme,
GA was 85% whereas it achieves its highest value when GA it was observed that the lower the grid availability rate,
was 70%. the higher the cost of energy. A drop of 30% in grid supply
In this article, a decision-making approach is implemented contributed to a rise of 178,657 % in kWh costs.
to thoroughly illustrate the availability of the optimization Figure 9 demonstrates the influence of decreasing GA on
mechanism and enable the decision-maker to properly assess the three objective functions.
the capability of the hybrid system [47]. In this analysis, each While Figure 10 indicates the impact of reducing GA on
point on the Pareto front is viewed as a three-dimensional the rising share of renewable energy sources in the proposed
variable [COE(i), LPSP(i), SSER(i)], and then measured scheme, where the percentage increases from 0% for full grid
the Euclidean distance for each three-point. As the final availability to 52.1 % for GA = 70 %.

Np q
X
Optimal Sol. = min (COE(i) − COE(k))2 + (LPSP(i) − LPSP(k))2 + (SSER(i) − SSER(k))2 i ∈ [1, Np ] (27)
k=1

VOLUME 9, 2021 14535

 M. M. Samy et al.: Reliability Support of Undependable Grid Using GESs: Economic Study

FIGURE 8. Comparing the base GA case, the Pareto front of the three objective functions for various GA states.

14536 VOLUME 9, 2021

M. M. Samy et al.: Reliability Support of Undependable Grid Using GESs: Economic Study

evaluating hybrid system economies; the loss of probability

of power supply (LPSP) for reliability assessments, and the
System Surplus Energy Rates (SSER) for evaluating hybrid
system compatibility and efficiency.
Concerning the design variables, three variables, such as
the number of solar photovoltaic systems, wind turbines, and
battery banks were considered. In this work, the condition
where the grid is entirely accessible was considered as a base
case, and then the availability of the grid has been decreased
by 5 % per time from 95 % to 70 % to illustrate the effect of
grid availability on the configuration of the proposed GES.
The study findings revealed that the major influence of
GA was on increasing the usage of the hybrid system battery.
FIGURE 9. The impact of decreasing GA on the three objective functions. It also indicates that in the absence of a full grid the scheme
proposed was extremely efficient, the rising number of bat-
teries used retained a low LPSP level throughout the system.
A fair SSER has also been preserved. Regarding the impact
of GA on the COE, the lower the grid availability levels, the
higher the energy cost. A 30% decrease in grid availability
led to a 178,657% surge in kilowatt-hour price.
Sizing an unreliable grid-connected hybrid system is a
multi-objective optimization challenge, including numerous,
overlapping objectives with simple trade-offs. Such chal-
lenges are hard to address and time-consuming in modeling
and engineering systems.
The suggested technique, therefore, resulted in a practical
and systemic evaluation of such processes which permits the
FIGURE 10. The effect of GA on the renewable energy fraction. simultaneous treatment of several factors, including decision
variables and objective functions. The Pareto front offers
helpful mechanisms to evaluate the interconnections between
The energy costs estimated in this analysis are consistent the objectives and to provide decision-making processes to
with previous research. T Azerefegn et al. have explored support design.
the feasibility of incorporating a hybrid PV/WT system into
existing unreliable grid/diesel generator networks [48]. The REFERENCES
focus of the analysis is to provide vital industrial park loads [1] F. Cadini, G. L. Agliardi, and E. Zio, ‘‘A modeling and simulation frame-
in three distinct Ethiopian territories. For the three loca- work for the reliability/availability assessment of a power transmission
tions, the authors estimated the energy costs of the suggested grid subject to cascading failures under extreme weather conditions,’’ Appl.
Energy, vol. 185, pp. 267–279, Jan. 2017.
approach from 0,044 to 0,049 $/kWh. The cost of energy of [2] M. A. H. Mondal and A. K. M. S. Islam, ‘‘Potential and viability of grid-
a self-sufficient wind / PV / diesel system that electrifies a connected solar PV system in Bangladesh,’’ Renew. Energy, vol. 36, no. 6,
broad resort facility in the Malaysian South China Sea was pp. 1869–1874, Jun. 2011.
stated to be $0.279 / kWh [49]. The energy cost was stated to [3] F. C. Robert and S. Gopalan, ‘‘Low cost, highly reliable rural elec-
trification through a combination of grid extension and local renew-
be $0.307 / kWh for a grid-connected system for PV-fuel cell able energy generation,’’ Sustain. Cities Soc., vol. 42, pp. 344–354,
feeding the Ayazaga campus loads from the Faculty of Elec- Oct. 2018.
tric & Electronics of Istanbul University of Technology [50]. [4] A. Maleki, H. Hafeznia, M. A. Rosen, and F. Pourfayaz, ‘‘Optimiza-
tion of a grid-connected hybrid solar-wind-hydrogen CHP system for
Likewise, the PV / Wind hybrid was planned for a farm residential applications by efficient metaheuristic approaches,’’ Appl.
in Egypt’s Toshka region at $0.597 per kWh for generating Thermal Eng., vol. 123, pp. 1263–1277, Aug. 2017, doi: 10.1016/j.
electricity [51]. applthermaleng.2017.05.100.
[5] A. Maleki and M. A. Rosen, ‘‘Design of a cost-effective on-grid hybrid
wind–hydrogen based CHP system using a modified heuristic approach,’’
V. CONCLUSION Int. J. Hydrogen Energy, vol. 42, no. 25, pp. 15973–15989, Jun. 2017, doi:
This article proposed a MOPSO-based optimization model 10.1016/j.ijhydene.2017.01.169.
[6] A. Rezvani, M. Gandomkar, M. Izadbakhsh, and A. Ahmadi,
for the sizing of a green energy system connected to a ran-
‘‘Environmental/economic scheduling of a micro-grid with renewable
domly disrupted grid. The paper suggests a PV / WT / battery energy resources,’’ J. Cleaner Prod., vol. 87, pp. 216–226,
hybrid power system to mitigate for frequent outages that Jan. 2015.
adversely influence the standard of service delivered to a [7] M. M. Samy, S. Barakat, and H. S. Ramadan, ‘‘Techno-economic analysis
for rustic electrification in egypt using multi-source renewable energy
small hamlet in northern Egypt. In this respect, three objective based on PV/ wind/ FC,’’ Int. J. Hydrogen Energy, vol. 45, no. 20,
functions are proposed, namely: the energy cost (COE) for pp. 11471–11483, Apr. 2020.

VOLUME 9, 2021 14537

 M. M. Samy et al.: Reliability Support of Undependable Grid Using GESs: Economic Study

[8] S. Barakat, M. M. Samy, M. B. Eteiba, and W. I. Wahba, ‘‘Feasibility study [27] R. Singh, R. C. Bansal, A. R. Singh, and R. Naidoo, ‘‘Multi-objective
of grid connected PV-biomass integrated energy system in egypt,’’ Int. J. optimization of hybrid renewable energy system using reformed elec-
Emerg. Electr. Power Syst., vol. 17, no. 5, pp. 519–528, Oct. 2016, doi: tric system cascade analysis for islanding and grid connected modes
10.1515/ijeeps-2016-0056. of operation,’’ IEEE Access, vol. 6, pp. 47332–47354, 2018, doi:
[9] S. Barakat, M. M. Samy, M. B. Eteiba, and W. I. Wahba, ‘‘Viability study of 10.1109/ACCESS.2018.2867276.
grid connected PV/Wind/Biomass hybrid energy system for a small village [28] V. P. Khambalkar, D. S. Karale, and S. R. Gadge, ‘‘Performance evaluation
in egypt,’’ in Proc. 18th Int. Middle East Power Syst. Conf. (MEPCON), of a 2 MW wind power project,’’ J. Energy Southern Afr., vol. 17, no. 4,
Dec. 2016, pp. 46–51. pp. 70–75, Nov. 2006.
[10] M. M. Samy and S. Barakat, ‘‘Hybrid invasive weed optimization— [29] M. M. Samy, S. Barakat, and H. S. Ramadan, ‘‘A flower pollination opti-
Particle swarm optimization algorithm for Biomass/PV micro-grid power mization algorithm for an off-grid PV-fuel cell hybrid renewable system,’’
system,’’ in Proc. 1st Int. Middle East Power Syst. Conf. (MEPCON), Int. J. Hydrogen Energy, vol. 44, no. 4, pp. 2141–2152, Jan. 2019.
Dec. 2019, pp. 377–382. [30] T. J. Stehly, P. C. Beiter, D. M. Heimiller, G. N. Scott, and A. Walker,
[11] S. Barakat, H. Ibrahim, and A. A. Elbaset, ‘‘Multi-objective optimization ‘‘PV O&M cost model and cost reduction,’’ Nat. Renew. Energy Lab.
of grid-connected PV-wind hybrid system considering reliability, cost, (NREL), Denver, CO, USA, Tech. Rep. NREL/TP-5C00-74840, Jun. 2020.
and environmental aspects,’’ Sustain. Cities Soc., vol. 60, Sep. 2020, [Online]. Available: https://www.nrel.gov/docs/fy20osti/74840.pdf
Art. no. 102178. [31] S. K. Ramoji and B. J. Kumar, ‘‘Optimal economical sizing of a PV-wind
[12] A. F. Tazay, M. M. Samy, and S. Barakat, ‘‘A techno-economic feasibility hybrid energy system using genetic algorithm and teaching learning based
analysis of an autonomous hybrid renewable energy sources for univer- optimization,’’ Int. J. Adv. Res. Electr. Electron. Instrum. Eng., vol. 3, no. 2,
sity building at Saudi Arabia,’’ J. Electr. Eng. Technol., vol. 15, no. 6, pp. 7352–7367, 2014.
pp. 2519–2527, Nov. 2020. [32] A. Malheiro, P. M. Castro, R. M. Lima, and A. Estanqueiro, ‘‘Integrated
[13] M. M. Samy, H. I. Elkhouly, and S. Barakat, ‘‘Multi-objective opti- sizing and scheduling of wind/PV/diesel/battery isolated systems,’’ Renew.
mization of hybrid renewable energy system based on biomass and Energy, vol. 83, pp. 646–657, Nov. 2015.
fuel cells,’’ Int. J. Energy Res., to be published. [Online]. Available: [33] K. Wu, H. Zhou, S. An, and T. Huang, ‘‘Optimal coordinate operation
https://onlinelibrary.wiley.com/doi/abs/10.1002/er.5815 control for wind–photovoltaic–battery storage power-generation units,’’
[14] M. M. Samy, M. I. Mosaad, and S. Barakat, ‘‘Optimal economic study Energy Convers. Manage., vol. 90, pp. 466–475, Jan. 2015.
of hybrid PV-wind-fuel cell system integrated to unreliable electric [34] T. Stehly, P. Maples, D. Heimiller, and G. Scott, 2017 Cost of Wind
utility using hybrid search optimization technique,’’ Int. J. Hydrogen Energy Review. Golden, CO, USA: National Renewable Energy Labora-
Energy, to be published. [Online]. Available: https://www.sciencedirect. tory, Feb. 2018, pp. 23–40, doi: NREL/TP-6A20-64281.
com/science/article/pii/S0360319920329335 [35] Simpliphi Phi 2.6 KWH Smart-Tech Hi Power LFP Battery. Accessed:
[15] M. I. Mosaad, M. O. A. El-Raouf, M. A. Al-Ahmar, and F. M. Bendary, 2016. [Online]. Available: SimpliPhiPower
‘‘Optimal PI controller of DVR to enhance the performance of hybrid [36] A. Jaiswal, ‘‘Lithium-ion battery based renewable energy solution for off-
power system feeding a remote area in Egypt,’’ Sustain. Cities Soc., vol. 47, grid electricity: A techno-economic analysis,’’ Renew. Sustain. Energy
May 2019, Art. no. 101469, doi: 10.1016/j.scs.20189.101469. Rev., vol. 72, pp. 922–934, May 2017, doi: 10.1016/j.rser.2017.01.049.
[16] W. S. Hassanein, M. M. Ahmed, M. O. abed el-Raouf, M. G. Ashmawy, and [37] S. Dhundhara, Y. P. Verma, and A. Williams, ‘‘Techno-economic anal-
M. I. Mosaad, ‘‘Performance improvement of off-grid hybrid renewable ysis of the lithium-ion and lead-acid battery in microgrid systems,’’
energy system using dynamic voltage restorer,’’ Alexandria Eng. J., vol. 59, Energy Convers. Manage., vol. 177, pp. 122–142, Dec. 2018, doi: 10.1016/
no. 3, pp. 1567–1581, Jun. 2020. j.enconman.2018.09.030.
[17] Y. Wang, W. Yan, S. Zhuang, and J. Li, ‘‘Does grid-connected clean power [38] S. M. Hakimi, S. M. M. Tafreshi, and A. Kashefi, ‘‘Unit sizing of a stand-
promote regional energy efficiency? An empirical analysis based on the alone hybrid power system using particle swarm optimization (PSO),’’ in
upgrading grid infrastructure across China,’’ J. Cleaner Prod., vol. 186, Proc. IEEE Int. Conf. Autom. Logistics, Aug. 2007, pp. 3107–3112.
pp. 736–747, Jun. 2018. [39] S. Guo, Y. He, H. Pei, and S. Wu, ‘‘The multi-objective capacity optimiza-
[18] T. Kamal, S. Z. Hassan, H. Li, S. Mumtaz, and L. Khan, ‘‘Energy manage- tion of wind-photovoltaic-thermal energy storage hybrid power system
ment and control of grid-connected wind/fuel cell/battery hybrid renewable with electric heater,’’ Sol. Energy, vol. 195, pp. 138–149, Jan. 2020, doi:
energy system,’’ in Proc. Int. Conf. Intell. Syst. Eng. (ICISE), Jan. 2016, 10.1016/j.solener.2019.11.063.
pp. 161–166. [40] M. Petrollese and D. Cocco, ‘‘Optimal design of a hybrid CSP-PV plant
[19] M. Hossain, S. Mekhilef, M. Danesh, L. Olatomiwa, and S. Shamshir- for achieving the full dispatchability of solar energy power plants,’’ Sol.
band, ‘‘Application of extreme learning machine for short term output Energy, vol. 137, pp. 477–489, Nov. 2016.
power forecasting of three grid-connected PV systems,’’ J. Cleaner Prod., [41] Cabinet Decision no. 2532, of 2016 on the Issuance of the Second
vol. 167, pp. 395–405, Nov. 2017. Phase of the Tariff System for Electric Recharge, Egyptian Electr. Util-
[20] P. M. Murphy, S. Twaha, and I. S. Murphy, ‘‘Analysis of the cost of ity Consum. Protection Regulatory Agency (EgyptERA), Cairo, Egypt,
reliable electricity: A new method for analyzing grid connected solar, 2015. Accessed: Feb. 2018. [Online]. Available: http://egyptera.org/
diesel and hybrid distributed electricity systems considering an unreliable Downloads/Laws/the%20Electricity%20Law.pdf
electric grid, with examples in uganda,’’ Energy, vol. 66, pp. 523–534, [42] M. K. Deshmukh and S. S. Deshmukh, ‘‘Modeling of hybrid renewable
Mar. 2014. energy systems,’’ Renew. Sustain. Energy Rev., vol. 12, no. 1, pp. 235–249,
[21] S. M. Harish, G. M. Morgan, and E. Subrahmanian, ‘‘When does unreliable Jan. 2008.
grid supply become unacceptable policy? Costs of power supply and [43] H. Yang, W. Zhou, L. Lu, and Z. Fang, ‘‘Optimal sizing method for stand-
outages in rural india,’’ Energy Policy, vol. 68, pp. 158–169, May 2014. alone hybrid solar–wind system with LPSP technology by using genetic
[22] M. O. A. El-Raouf, M. I. Mosaad, M. A. Al-Ahmar, and F. M. El Bendary, algorithm,’’ Sol. Energy, vol. 82, no. 4, pp. 354–367, Apr. 2008.
‘‘MPPT of hybrid solar-wind-grid power generation system,’’ Int. J. Ind. [44] S. Diaf, D. Diaf, M. Belhamel, M. Haddadi, and A. Louche, ‘‘A method-
Electron. Drives, vol. 2, no. 4, pp. 234–241, 2015. ology for optimal sizing of autonomous hybrid PV/wind system,’’ Energy
[23] N. Alshammari, M. M. Samy, and J. Asumadu, ‘‘Optimal economic anal- Policy, vol. 35, no. 11, pp. 5708–5718, Nov. 2007.
ysis study for renewable energy systems to electrify remote region in [45] S. Mostaghim and J. Teich, ‘‘Strategies for finding good local guides in
kingdom of saudi arabia,’’ in Proc. 20th Int. Middle East Power Syst. Conf. multi-objective particle swarm optimization (MOPSO),’’ in Proc. IEEE
(MEPCON), Dec. 2018, pp. 1040–1045. Swarm Intell. Symp. SIS, Apr. 2003, pp. 26–33.
[24] B. Tudu, S. Majumder, K. K. Mandal, and N. Chakraborty, ‘‘Comparative [46] N. Padhye, J. Branke, and S. Mostaghim, ‘‘Comprehensive comparison of
performance study of genetic algorithm and particle swarm optimization MOPSO methods: Study of convergence and diversity—Survey of state
applied on off-grid renewable hybrid energy system,’’ in Proc. Int. Conf. of the art,’’ in Proc. 11th Conf. Congr. Evol. Comput. (CEC), May 2009,
Swarm, Evol., Memetic Comput., Dec. 2011, pp. 151–158. pp. 2516–2523.
[25] M. I. Mosaad and H. S. Ramadan, ‘‘Power quality enhancement of grid- [47] A. R. Starke, J. M. Cardemil, R. Escobar, and S. Colle, ‘‘Multi-objective
connected fuel cell using evolutionary computing techniques,’’ Int. J. optimization of hybrid CSP+PV system using genetic algorithm,’’ Energy,
Hydrogen Energy, vol. 43, no. 25, pp. 11568–11582, Jun. 2018. vol. 147, pp. 490–503, Mar. 2018.
[26] X. Xiong, W. Wu, N. Li, L. Yang, J. Zhang, and Z. Wei, ‘‘Risk-based multi- [48] T. M. Azerefegn, R. Bhandari, and A. V. Ramayya, ‘‘Techno-economic
objective optimization of distributed generation based on GPSO-BFA analysis of grid-integrated PV/wind systems for electricity reliability
algorithm,’’ IEEE Access, vol. 7, pp. 30563–30572, 2019, doi: 10.1109/ enhancement in Ethiopian industrial park,’’ Sustain. Cities Soc., vol. 53,
ACCESS.2019.2902886. Feb. 2020, Art. no. 101915, doi: 10.1016/j.scs.2019.101915.

14538 VOLUME 9, 2021

M. M. Samy et al.: Reliability Support of Undependable Grid Using GESs: Economic Study

[49] M. Hossain, S. Mekhilef, and L. Olatomiwa, ‘‘Performance evaluation of MOHAMED I. MOSAAD received the B.Sc. and
a stand-alone PV-wind-diesel-battery hybrid system feasible for a large M.Sc. degrees from Zagazig University, Egypt,
resort center in south China sea, Malaysia,’’ Sustain. Cities Soc., vol. 28, and the Ph.D. degree from Cairo University,
pp. 358–366, Jan. 2017, doi: 10.1016/j.scs.2016.10.008. Egypt, all in electrical engineering. He is currently
[50] B. E. Türkay and A. Y. Telli, ‘‘Economic analysis of standalone and an Associate Professor with the Department of
grid connected hybrid energy systems,’’ Renew. Energy, vol. 36, no. 7, Electrical and Electronic Engineering Technology,
pp. 1931–1943, Jul. 2011, doi: 10.1016/j.renene.2010.12.007.
YIC, Saudi Arabia. His research interests include
[51] M. Abdel-Salam, A. Ahmed, H. Ziedan, K. Sayed, M. Amery, and
power system stability, control, and renewable
M. Swify, ‘‘A solar-wind hybrid power system for irrigation in toshka
area,’’ in Proc. IEEE Jordan Conf. Appl. Electr. Eng. Comput. Technol. energy. He is a regular Reviewer of many IEEE
(AEECT), Dec. 2011, pp. 38–43, doi: 10.1109/AEECT.2011.6132491. TRANSACTIONS, IET Electric Power Applications
journal, IET Generation, Transmission and Distribution journal, Interna-
tional Journal of Industrial Electronics and Drives (IJIED), and Interna-
tional Journal of Energy Engineering (IJEE). He is also the Editor-in-Chief
of YJES.

MOHAMED F. EL-NAGGAR was born in

Helwan, Egypt, in September 1972. He received
the B.Sc. and M.Sc. degrees in electrical engi-
neering and the Ph.D. degree from Helwan Uni-
MOHAMED MAHMOUD SAMY received the versity, Cairo, Egypt, in 1995, 2002, and 2009,
B.Sc. degree in electrical engineering from the respectively. He is currently a Teacher of power
Faculty of Engineering and Technology, Banha system protection with the Power and Machines
University, Banha, Egypt, in 1995, the M.Sc. Engineering Department, Faculty of Engineering,
degree in electrical power and machines from Helwan University. His research interest includes
Zagazig University, Zagazig, Egypt, in 2007, and power system relaying.
the Ph.D. degree from Cairo University, Giza,
Egypt, in 2011. SHIMAA BARAKAT was born in Fayoum, Egypt,
He is currently an Associate Professor with in 1980. She received the B.S., M.S., and Ph.D.
the Electrical Engineering Department, Faculty of degrees in electrical engineering from Fayoum
Engineering, Beni-Suef University, Beni-Suef, Egypt, and also on leave University, in 2002, 2014, and 2018, respectively.
as a Visiting Professor at the Faculty of Engineering, Albaha University, Her research interests include renewable energy,
Albaha, Saudi Arabia. He was elected as a Technical Program Committee which includes solar, wind and biomass, as well as
(TPC) for the Third International Conference on Energy Engineering and fuel cells. She is also interested in the application
Environmental Protection (EEEP2018). In 2012, he was awarded the Prize of of optimization techniques in renewable energy.
Best Ph.D. thesis in electrical engineering from the Faculty of Engineering,
Cairo University.

VOLUME 9, 2021 14539