Journal of Intelligent & Fuzzy Systems 43 (2022) 6651–6662 6651

DOI:10.3233/JIFS-220726
IOS Press

Economic analysis of a hybrid intelligent

optimization-based renewable energy
system using smart grids
K. Suresha,∗ and P. Jagatheeswarib
a Department of EEE, Ponjesly College of Engineering, Nagercoil, Tamil Nadu, India
b Department of ECE, Ponjesly College of Engineering, Nagercoil, Tamil Nadu, India

Abstract. Renewable energy has seen a substantial increase in deployment as an alternative to traditional power sources.
However, two fundamental constraints exist that preclude widespread adoption: the availability of the generated power and
the expense of the equipment. One of the most critical difficulties with this sort of hybrid system is to appropriately design
the Hybrid Renewable Energy System (HRES) elements so that they fulfill all load requirements while requiring the least
amount of investment and running expenditures. This research proposes a novel technique for evaluating the optimal smart
grid linking Hybrid Renewable Energy (Solar photovoltaic and wind) with battery, to increase profitability, dependability,
and feasibility. A multiobjective function is suggested and constructed to be optimized utilizing two optimization algorithms:
Enhanced Particle Swarm Optimization (EPSO) and Harris Hawks Optimization (HHO) algorithm with Fuzzy-Extreme
Learning Machine (ELM). The primary goal for the HRES is to operate optimally to reduce the cost of energy generat ion
through hourly day-ahead. Here, the Fuzzy-ELM is utilized to predict the required load of the smart grid-connected system
and hybrid EPSO-HHO, which are introduced to solve the problem of HRES economic analysis. Finally, the suggested
EPSO-EHO method is implemented in the MATLAB software, and its performance comparison is made with other existing
methods such as PSO, WOA, and HHO. The simulation result shows that the cost of the newly suggested EPSO-HHO
technique-based Hybrid Renewable Energy System is less than PSO, WOA, and HHO by 4.89 %, 4.51 %, and 4.05 %,
respectively.

Keywords: Harris Hawks’ Optimization, economic analysis, renewable energy sources, Extreme Learning Machine, smart
grid

1. Introduction cost, the scattered character of the region, and low

load factor [1]. The economic and environmental ben-
The enlarged use of fossil fuel, along with the efits of the reduction of CO2 emissions and power
degradation of the environment, has resulted in the losses. Because of the increased worldwide percep-
growth of environmentally friendly renewable energy tion of the necessity for ecological guarantee and the
sources. However, after more than five decades, pro- necessity for decreased reliance on fuels for power
gression in remote rural regions could not take place generation, power generation techniques will be cru-
because the grid could not be enlarged due to its high cial [2] in future power supply. For power generation,
a substantial percentage of power generation meth-
∗ Corresponding author. K. Suresh, Assistant Professor, Depart- ods rely on non-renewable energy sources such as
ment of EEE, Ponjesly College of Engineering, Nagercoil, coal, gas & oil. Because of these traditional energy
Tamil Nadu, 629003, India. E-mail: [email protected]. supplies, greenhouse gas emissions and other envi-

ISSN 1064-1246/$35.00 © 2022 – IOS Press. All rights reserved.

6652 K. Suresh and P. Jagatheeswari / Economic analysis of a hybrid intelligent optimization

ronmental damage are on the rise [3]. The energy Eltamaly et al. [8] presents the HOMER software
produced by Renewable Energy Sources (RES) is implemented hourly wind speed, radiation, and load
clean and does not pollute the environment, but it power with many different types of wind turbines and
is difficult to control and stochastic [4]. Due to this PV module types in ten Saudi sites [9] to achieve the
drawback, a large penetration of RES can cause best component size and the lowest price of kWh
power quality issues in the electrical grid, stability, generated while maintaining the greatest level of
and reliability [5]. dependability. It modifies the wind/ PV share ratio
Various conventional and non-conventional energy with certain increments, and it is implemented in the
sources are utilized to create electricity. One of the new proposed computer program (NPCP) [10].
most prominent energy sources is solar and wind Zahra Pooranian et al. [11] create smart Microgrids
energy systems. Because of their modularity and (MGs) networked based on the NSGA II technique,
environmental friendliness, wind and solar energy which has a low running cost & emits the minimum
systems are becoming increasingly popular. For the contaminants. To compare, the issue is reduced to a
past two decades, the area of wind and solar has seen single objective function and then addressed by two
amazing progress in the widespread usage of stan- investigative methods, PSO and Imperialist Compet-
dalone to utility-interactive wind and solar systems itive Algorithm (ICA). Mohamed et al. [12] offer
[6, 7]. an optimal utilization of HRES based on available
Nowadays, various types of renewable energy sys- generation utilizing a smart grid load management
tems are the primary energy sources that are used application and particle swarm optimization. Roy et
to meet the load demand. The growing number of al. [13] presents a fusion method for minimizing gen-
RES and distributed generators necessitates innova- eration costs while also making greater use of RES.
tive techniques for management and management This approach combines the performance of the Bac-
of the electrical grid, with the ultimate objective of terial Foraging Optimization Algorithm (BFOA) with
maintaining or even improving steadfast dependabil- the artificial neural network (ANN) technique for the
ity and quality of power supply [8]. As a result, a energy flow structure of energy sources as well as the
power supply backup, as well as control and reg- grid.
ulation methods, are necessary for optimum power Amirtharaj et al. [14] developed an effective
management in a microgrid setting to compensate for inverter for the use of HRES using a fusion method.
the constraints of individual resources [28]. This approach was a mix of the ANN and the Adap-
This work employs a hybrid Enhanced Parti- tive Grasshopper Optimization Algorithm (AGOA).
cle Swarm Optimization (EPSO) and Harris Hawks Based on power variations, the AGOA generates pre-
Optimization (HHO) with Fuzzy-Extreme Learning cise control signals in the load and source sides in this
Machine (ELM) for SG maximum power production presented system. Venkatesan et al. [15] suggested a
at a minimum cost. The major objective of the pro- hybrid approach for OPF regulation of HRES with
posed approach is to give optimum power to the load battery. The created approach was a combination of
at a minimal cost. Fuzzy-ELM is used to predict the ANN and a Whale Optimization Algorithm (WOA).
load demand, while EPSO-HHO is used for economic The merger of two approaches was thereafter gen-
analysis. The remainder of the work is organized as erally referred to as the WOANN technique. WOA
below. Section 2 describes the connected research improved the learning process of ANN by using the
efforts as well as the research work’s basis. 3. Mod- least error objective function.
eling of HRES with smart grid-connected is depicted Islam et al. [16] introduce the Harris Hawks Opti-
in Section 3. The multiobjective problem formulation mization (HHO) strategy for fossil fuel emission
is explained in section 4. The proposed technique is management by resolving multi and single-objective
explained in section 5. Section 6 clearly illustrates the Optimal Power Flow (OPF) concerns. Mathemati-
results and debates. Section 7 depicts the conclusion. cal analysis of hawks’ synchronized behavior and
dynamic chase patterns to pounce on escaping tar-
gets is used to solve the optimization problem. Suresh
2. Related works – A review et al. [17] presents a cost-effective hybrid technique
for Power Flow Management (PFM) in SG sys-
There are several study papers in the literature that tems using HRES. TFW-BRO refers to the suggested
are founded on the smart grid energy management hybrid approach, which combines Turbulent Flow
structure. Few of them are covered in this section. of Water-Based Optimization (TFWO) with Battle
 K. Suresh and P. Jagatheeswari / Economic analysis of a hybrid intelligent optimization 6653

Fig. 1. Schematic of proposed hybrid renewable energy system.

Royale Optimization. TFWO generates the Voltage ing the efficacy of the grid-connected energy system.
Source Inverter (VSI) control signal grounded on the Power management for the suggested system is dif-
fluctuation of power delivery among the load and ficult to work that must be completed to deliver an
source sides. inexpensive and continuous supply.
An ideal hybrid strategy is presented to ensure
power fluctuation. The proposed hybrid approach
combines the Adaptive Neuro-Fuzzy Interference 3. Modeling of HRES with smart
System (ANFIS) with the Advanced Salp Swarm grid-connected
Optimization Algorithm (ASOA) by Kallol Roy et al.
[18]. ASOA finds the best MG combination in terms The proposed smart grid-integrated hybrid renew-
of the expected needed load. ANFIS is used to con- able energy system is revealed schematically in
tinually track the needed load of the MG-connected Fig. 1. The suggested system comprises wind energy
system. sources, battery storage, Solar Photovoltaic (PV),
To solve challenges, certain control mechanisms bi-directional converter, load, and utility grid [19].
are presented in the section; the constraints identi- The PV system and storage system are connected to
fied in the earlier studies prompted us to perform this the Direct Current (DC) side of the bus bar, while
investigation. According to the research, solar and the wind turbine is linked to the Alternating current
wind energy-based technologies might be combined (AC) side. Hybrid Renewable Energy Sources (PV
with grid-connected systems to provide cost-effective and wind) are the major source of power generating
electrical power. The grid integration with these grid- [20–22]. When solar wind turbines and PV panels
connected PV and wind energy systems improves fail to fulfill the load requirements, grid as a major
system dependability; nonetheless, power manage- source for pofr can be used to solve the load demand
ment and economic analysis is a key and difficult [23]. If neither source can meet the load require-
issues that must be addressed. As a result, the offered ment, the bi-directional converter functions as both
study proposes a complete mathematical model of a an inverter converting Dc to AC to fulfil AC load
grid integrated hybrid renewable energy-grid linked requirements and a rectifier to convert AC from grid to
energy system to meet the load requirement of a DC. Each component’s precise mathematical models
small distant location. The economic study of a grid- were assessed further below.
connected HRES system to provide an optimal size
solution while taking into account competing objec- 3.1. Wind turbine (WT)
tives, such as LPSP and LCOE, has been examined.
Furthermore, the cost of power exchange with the grid The speed of the wind is the most significant ele-
becomes a helpful and efficient method for improv- ment for a WT. At time t, the generated power of a
6654 K. Suresh and P. Jagatheeswari / Economic analysis of a hybrid intelligent optimization

wind turbine (PWT ) is calculated as follows: may be calculated as follows:

⎧
⎪ 0 V Vci orV  Vco SOC (t) = SOC (t − 1) . (1 − σ)
⎪
⎪
⎨ V −Vci (4)
PWT (t) = P, Vr −Vci Vci < V < Vr Pload (t)
⎪
(1) + − (PPV (t) + PWT (t)) /ηBD
⎪
⎪ ηInv
⎩ Pr Vr < V < Vco
Where, ηBD denotes the battery bank’s discharging
Where, Pr is the wind turbine’s rated power, Vr , efficiency.
Vco and Vci are rated, cut-out and cut-in wind speeds,
respectively, and v (t) is the wind speed at time t.
4. Multi-objective problem formulation
3.2. Solar PV The primary purpose of this research is to develop
an optimization approach based on SG applicants to
First, it receives irradiance and temperature as tackle the problem of optimizing renewable energy
inputs and outputs the quantity of PV energy pro- sources to fulfill the demand of load while paying the
duced (Ppv ). The solar radiation may be used to lowest costs and meeting a predefined trustworthiness
compute the output power of a PV panel (Ppv ) at index.
time t using the following expression:
 
I (t)  4.1. Objective function
PPV = Rpv . . 1 + λp (TC − 25) (2)
1000
The primary goal of the reliability factors is to opti-
Where, TC denotes the temperature of the PV mize the HRES such that the power supply to the load
panel (◦ C), λp denotes temperature coefficient (◦ C), is maintained. The economic constraint, on the other
I denote the solar radiation (W/m2 ), and Rpv denotes hand, tends to identify the most cost-effective system
the PV panel’s rated power (W). arrangement. The optimum design problem’s objec-
tive purpose is to minimize the Annualized Cost of
the System (ACS) and energy balance.
3.3. Battery The number of PV systems, wind turbines, and bat-
tery banks is treated as decision variables, while the
The battery’s state of charge (SOC), together with loss of power supply probability (LPSP) is considered
its time consumption and productivity, is calculated the problem constraint. The constraints and objective
as below. As a result, the charge amount of the battery function are represented as.
at time t may be calculated as follows: OF=Min (COE)
subject to
SOC (t) = SOC (t − 1) . (1 − σ)
(3) Nxmin ≤ Nx ≤ Nxmax , x = {PV, WT, battery}
Pload (t)
+ (PPV (t) + PWT (t)) − × ηBC
ηInv
LPSP ≤ LPSP max
Where, ηBC denotes the battery bank’s charge
efficiency, Pload (t) denotes the load demand for a The maximal and minimal limits for decision vari-
specific hour, ηInv denotes the inverter efficiency, σ ables are significantly depending on the task.(number
denotes the hourly self-discharge rate, and SOC (t) of variables and search space complexity). The com-
and SOC (t − 1) are the battery bank’s charge ponent limits are determined by several factors, as
amounts at time t and (t – 1). Whenever the total power wind turbines are installed only in allowed areas and
production of wind turbines and PV panels exceeds the available space within the working area for estab-
the load demand, the battery bank is now charging. lishing the solar system. The estimation was carried
When the combined total power production of out, the initial limitations of the system components
wind turbines and PV panels is fewer than the load were recognized, and the optimization strategies were
requirement, the battery bank is discharged [24]. As then tested numerous times using the trial-and-error
a result, the charge amount of the battery at time t method to discover the optimal answer.
 K. Suresh and P. Jagatheeswari / Economic analysis of a hybrid intelligent optimization 6655

4.1.1. Annualized cost of system (ACS) 4.2. Constraints

The proposed HRES’s primary goal is to minimize
costs. The annualized cost of the complete system is Power Reliability Constraint Based on LPSP:
minimized by minimizing replacement cost, overall Because of the presence of several factors, such
capital cost, and operational and maintenance cost. as grid disruption, wind speed, and intermittent solar
The following Equation (5) expresses the ACS; radiation characteristics, which influence the capacity
and reliability of the proposed system
ACS = COM + Cinv + Cr + Cpc + Cgp − Cgs The LPSP is the likelihood that the hybrid model
(5) will be unable to satisfy the load demand due to inad-
Where Cpc and Cr denotes penalty cost, replace- equate power supply.
ment cost, Cgp and Cgs denotes costs of power The dependability of the system is characterized
drawn from the grid and power supplied to the grid, in this study in assessing the probability of power
respectively, COM and Cinv denotes the operation & supply failure (LPSP). It is expressed as below;
maintenance cost and investment cost, which also
N
comprises the cost of several elements in the system PL (ti ) − Psupplied (ti )
such as batteries, PV panels, wind turbines, and so LPSP= N (11)
i =1 i =1 PL (ti )
on. The penalty cost is supplied by;
Where, Psupplied (ti ) denotes the entire quantity of
N
energy provided by the framework at a time ti and N
Cpc = Cpc1 (LPSP − βL ) PL (ti )
denotes the overall number of hours.
i =1 (6)
Dgs − βg
+Cpc2 × 100
βg 5. Proposed methodology

 This study proposes an effective control approach

if (LPSP > βL ) and Dgs > βg (7)
for smart grid power management using distribution
Where, Cpc1 represents the penalty cost for a sup- system resources. The suggested hybrid approach is
ply deficiency (Rs/kWh) & Cpc2 denotes the cost of a combination of Fuzzy-Extreme Learning Machine
supply variation (Rs/%). and Enhanced Particle Swarm Optimization (EPSO)
and Harris Hawks Optimization (HHO) (EPSO-
HHO). The Fuzzy-Extreme Learning Machine is
4.1.2. levelized cost of energy (LCE)
used to forecast system load demand, and the optimal
cost functions are derived using a hybrid EPSO-HHO.
ACS
LCE = (8)
Etotal 5.1. Load demand prediction using Fuzzy-ELM
Where, Etotal denotes the total amount of energy In this section, the proposed Fuzzy ELM is used
delivered by all sources combined. That LCE may be to predict the load demand of the grid-linked hybrid
utilized to economically optimize the system. renewable energy system. The ELM trains feed-
forward networks with a single hidden layer. The
4.1.3. Energy balance extreme learning machine can predict loads with
Energy balancing is achieved in grid-linked mode more accuracy while consuming low time. In this
as follows. If the need for load exceeds the supply, proposed model, the load forecast is generated in two
the remaining power will be acquired from the grid. stages. The first phase is the prediction phase, dur-
ing which the input data is fed into the ELM. The
PL (t)=PPV (t) +PWT (t) + PBD (t) + Pgp (t) (9) power demand at the output node is then forecasted
by a back-propagation ELM, which is subsequently
Pgs (t) = PPV (t) +PBC (t) − PL (t) (10) controlled by a fuzzy expert system to get the final
prediction. The FELM is trained by comparing input
Where Pgp (t) represents the grid-purchased elec- time intervals of a day using the goal power demand,
tricity and Pgs (t) represents the power provided to that is, because of the available RES daily demand
the grid at a given point in time. dataset. ELM has 3 layers: an input, output, and a
6656 K. Suresh and P. Jagatheeswari / Economic analysis of a hybrid intelligent optimization

Fig. 2. Structure of Fuzzy-ELM based load prediction.

single hidden layer [25]. decreases by wkI +1 during iterations. This concept
l proposes changes to the fundamental PSO method

ti = βj g j .xi + bj (12) that result in enhanced particle swarm optimization
j =1 (EPSO). And, while the HHO has a strong capac-
ity to utilize the search space, its exploration ability
Where, i ∈ 1; 2; . . . .; N, j ∈ 1; 2; . . . .; 1, g (x) might be improved [27]. As a result, we employed
denotes an activation function and j .xi represents the HHO to accomplish this improvement based on
the inner product of j andxi . The output matrix T EPSO’s capacity to rapidly explore the given search
can be expressed as below; space. As a result, the primary goal of EPSO-HHO is
 
g (ω1 .x1 + b1 ) · · · g (ωl .xl + bl ) to solve the issue optimally while improving the con-
H= (13) vergence of the standard HHO utilizing EPSO. The
g (ω1 .x1 + bn ) · · · g (ωl .xl + bn )
N×1 inertia weight wI of EPSO can be expressed as;
The output weight
wkI +1 = α.wk0 (15)
+
β=H T (14)
Each particle’s velocity is modified in the follow-
The fuzzy expert system knows load and weather ing way;
characteristics. Time series data is separated into  
training and testing data during the load prediction. vki +1 = wkI .vki + C1 · r1 Pbest
k
− xik
Further training data is set as much as feasible to   (16)
ensure the fitting impact of the prediction results as +C2 · r2 gbest
k
− xik
well as the test data. The overall structure of load pre-
diction using the Fuzzy-ELM method is depicted in Where, w0 is the initial positive value and α is
Fig. 2. the inertia weight damping ratio (here 0.99, which
Once the system has been fully trained, the network is set by error and trial). HRES optimal utilization
training procedure is completed based on the input can be mathematically represented and expressed as
time interval for identifying the load requirement. In follows;
this study, the EPSO-HHO approach improves the Min W (A, U)
hybrid RE management of the smart grid based on Where W is the weighted objective function to be
the Fuzzy-ELM output network. minimized as given in Equation (5).
The proposed hybrid method begins by finding
5.2. Economic analysis using EPSO-HHO the initial value for each potential solution and then
computes the fitness value for each solution using
In this part, the Enhanced PSO and HHO, known Equation (5). The next stage is to find the optimal
as EPSO-HHO, are introduced to answer the prob- solution (Xb ) with the lowest fitness value (Fb ) and
lem of HRES economic analysis. An inertial constant then update the other solutions that use it as well as the
w governs the degree to which the particles recalls EPSO and HHO operators, which is shown in Fig. 3.
its previous velocity [26]. which starts at w0 and This updating procedure is carried out in a series of
 K. Suresh and P. Jagatheeswari / Economic analysis of a hybrid intelligent optimization 6657

Fig. 3. Proposed economic analysis flow diagram.

Fig. 3. (a) Load demand for study (b) Actual and predicted load demand.

phases that begin by determining the probability of The process of updating the solutions is repeated until
each candidate solution (Xi ) based on its fitness value the halting criteria are met, which are either a max-
(FFi ) as follows; imal number of populations or a suitable value for
the fitness function for the generation. Furthermore,
N
the proposed algorithm was used to find the global
Pi = FFi / FFi (17)
minimum.
i =1
The EPSO algorithm is faster than the PSO algo-
In the event of Pi 0.5, the HHO operators will rithm and provides better outcomes at a lower cost.
be employed; otherwise, the EPSO operators will be HHO has a strong capacity to utilize the search space,
used to update the current population of solutions. its exploration ability might be improved. As a result,
6658 K. Suresh and P. Jagatheeswari / Economic analysis of a hybrid intelligent optimization

Table 1 ied from their average annual value to assess the

Energy resources parameters influence of resource variation on the system econ-
Power sources Details Ranges omy. Second, the capital expenses of a PV panel and
Battery SoC max (%) 100 a wind turbine differ. These results show that wind
Capacity (h) 6.5 provides the majority of the energy, as seen in Fig. 4.
System Voltage (V) 200
Photovoltaic (PV) Temperature (◦ C) 25 Thus, the sensitivity analysis also forecasts that low-
Number of cells 36 ering the installation cost of a PV or wind energy
Irradiance (W/m2 ) 1000 system leads to per-unit power costs close to grid
Cell temperature (◦ C) 55
electricity prices.
Wind Turbine Rated speed (kW) 17.5
(WT) Cut out & cut in speed (m/s) 18 & 3.5 When assessing battery power, it was discovered
Wind speed (m/s) 5.8 that battery power was consumed from 1 to 11 hours
throughout the mode of charging. The Maximum bat-
tery power is calculated in the discharge operating
we employed the HHO to accomplish this improve- mode to be 12–16 hours. Figure 4(a) depicts the sug-
ment based on EPSO’s capacity to rapidly explore gested method’s battery power. Peak hours 8 to 19
the given search space. As a result, the primary goal hours provide the most PV power. PV achieved max-
of EPSO-HHO is to solve the issue optimally while imum output at 3.6 kW and grew to 6 kW during peak
improving the convergence of the standard HHO uti- hours of 15 hours. Power is gradually reduced till it
lizing EPSO. reaches 0.2 kW after peak hours, which is shown in
Fig. 4(b). When wind turbine electricity is generated,
the maximum power is achieved at the wind speed.
6. Results and discussion Figure 4(c) depicts the greatest power produced by
the suggested approach at times 14 to 19 hours. Wind
The results of the simulated assessment of current power is determined by reducing the produced power
techniques of the established algorithm are reported from 4.9 to 1.9 kW in time 5 to 9 hours. Power gradu-
in this part. The effectiveness of the established strat- ally increases to around 7.8 kW at time 13 to 16 hours.
egy and existing methods such as the HHO, PSO, At time 17 to 24 hours after peak hours, the output
and WOA, are studied. The recommended technique approaches 3 kW.
employed throughout the 24 hours is examined in this Figure 5 demonstrates the examination of the sys-
section. On MATLAB R2019a, a simulation with a tem’s SOC using the presented EPSO-HHO and
period of hours is conducted. Table 1 lists the sim- existing HHO, PSO, and WOA methodologies. The
ulation parameters of the suggested approach. The SOC comparative study of the presented and exist-
suggested approach, which operates in terms of cost ing approaches is shown in Table 2. The proposed
minimization, selects the smart grid source allocation EPSO-HHO approach achieves a high level of SOC
for each load demand value. The load demand curve is (80 %). Despite the proposed algorithms, the system
depicted in the following Fig. 3. The proposed system can meet all of the power demands of load in normal
is suitable for small remote locations. operation, as shown in this Fig. 5. The battery runs
From Fig. 3(a), the electrical energy demand is in discharging mode at the time instant time of 17
25 kW/day is peak demand. A single time slot equals to 24 hrs, achieving SOC. Despite the higher power
one hour, which means that input and goal data are from the utility grid offer relative to the battery, the
gathered from the consumer end every hour to achieve suggested approach recognizes that if it can utilize the
high efficacy in neural network training. Figure 3(b) utility grid to compensate for power shortages while
depicts the derived findings from energy demand pre- using the remaining generated power to charge the
diction and actual energy use. As a consequence, battery, the overall generating cost will be reduced.
the acquired results for the Fuzzy-ELM approach are To save the cost, the battery conserved extra energy
rather near to the real consumption, indicating that the for the next day’s load supply.
Fuzzy-ELM provided the maximum accuracy. The The comparative investigation of produced PV
following Fig. 4(a)–(c) depicts the generated power of power is shown in Fig. 6. Using the existing approach,
the battery, PV, and wind turbine using the proposed the maximum PV power generation first climbed
EPSO-HHO method. by 22.8 KW and then gradually fell. The current
Two types of sensitivity variables are used in this approach generates maximum power with greater dis-
research. To begin, wind and solar resources are var- ruptions and is incapable of providing power to the
 K. Suresh and P. Jagatheeswari / Economic analysis of a hybrid intelligent optimization 6659

Fig. 4. Proposed generated power analysis of (a) battery (b) PV (c) WT.

Fig. 6. Comparative analysis of generated PV power.

Fig. 5. Comparative analysis of battery SOC.

Figure 7 depicts the wind turbine power of the pro-
posed and current approaches. At first, the maximal
Table 2 wind power production improved by 28 KW utiliz-
SOC comparison analysis
ing the existing approach, but the power production
SOC (%) Techniques gradually declined. With increasing interruptions, the
PSO WOA HHO Proposed
72 75 77 80
existing approach obtains maximum power and can-
not deliver stable energy to the SG and distorts the
power produced by wind turbines, preventing the
system from achieving extreme power. Using our sug-
SG system. The maximum generation of PV power gested method, optimum wind electricity production
is more ideal and less distortion is discovered uti- is more ideal, lower distortion is discovered, and the
lizing our proposed technique, and the system gains system obtains peak energy with optimum SG func-
extreme power with optimum SG functioning after tioning after immediate distortion. The battery power
the immediate distortion. of the proposed EPSO-HHO and PSO, HHO, and
6660 K. Suresh and P. Jagatheeswari / Economic analysis of a hybrid intelligent optimization

Fig. 7. Comparative analysis of generated WT power.

Fig. 9. Cost comparative analysis.

Table 3
Optimized results of the HRES for proposed and existing
methods
HRES HRES elements Cost of Annualized LPSP
(PV/WT/ energy Cost (%)
battery) (Rs/ (Rs/kWh)
NWT NPV Nbat kWh)
PSO 1 2 4 0.216 491,958 0.92
WOA 1 5 5 0.225 490,137 0.81
HHO 1 3 4 0.223 488,004 0.90
Proposed 1 2 4 0.140 468,982 0
Fig. 8. Comparative analysis of generated battery power. EPSOHHO

WOA approaches is depicted in Fig. 8. The battery is

charging from 1 to 11 hours, and it moves into dis-
charging state from 12 to 14 hours. In this case, the
suggested approach generates less distorted power
than previous techniques. By employing the pro-
posed strategy, the overall cost is drastically reduced
as compared to HHO, PSO, and WOA procedures.
Figure 9 depicts the total cost function comparison
graph. The proposed EPSO-HHO algorithm-based
hybrid system produces the lowest LCOE of 0.140
Rs/kWh than the existing methods, as shown in
Table 3. Based on their load demand, the annualized
cost of a hybrid PV/WT/battery system is 468,982 Fig. 10. Comparison of convergence.
Rs/kWh. The proposed EPSO-HHO technique-based
hybrid renewable energy system annual costs less Table 4 shows a comparative study of the suggested
than PSO, WOA, and HHO by 4.89 %, 4.51 %, and with several current methodologies for elapsed time.
4.05 %, respectively. The suggested method outper- When compared to the various current strategies,
forms the other strategies in terms of outcomes. Table 3 indicates that the suggested method rapidly
The proposed approach converges after 20 iter- provides the ideal result. The suggested EPSO-HHO
ations and yields fitness values ranging from 3 to method’s standard deviation (SD), median, and mean
0.3. The fitness comparison proposed with existing are 0.5423, 0.3574, and 0.5408, respectively. Table 5
approaches is depicted in Fig. 10. HHO converges at demonstrates that the suggested method has the
iteration count 24 and produces fitness values ranging best geometric parameters among the known strate-
from 3.4 to 0.4. WOA converges after 36 iterations gies. The approach can address the grid-connected
and returns the value of fitness ranging from 3.6 to HRESs power management scheme. Furthermore,
0.6. PSO meets at 38 iterations and produces fitness the suggested technique provides cost-effective grid
values ranging from 4.2 to 0.6. power generation and effective RES usage without
 K. Suresh and P. Jagatheeswari / Economic analysis of a hybrid intelligent optimization 6661

Table 4 Acknowledgments
Statistical evaluation of proposed and existing methods
Methods Proposed HHO WOA PSO The authors express their heartfelt gratitude to the
EPSO-HHO
supervisor for his direction and unwavering support
Mean 0.5408 0.7653 0.9702 1.2455 in this study.
Median 0.3574 0.4952 0.5415 0.6118
SD 0.5423 0.7312 0.9597 1.1012

Statement on fund providing

Table 5
The elapsed time of proposed and existing techniques This study was not funded in any way by the
Methods Proposed HHO WOA PSO researchers.
EPSO-HHO
Time (Sec) 34.901 35.332 36.026 36.571
Conflicts of interest

squandering energy available. In conclusion that the The authors said that they do not have competing
proposed approach has the following benefits over interests to disclose in this work.
previous techniques: little dependency on particle
initialization positions, quick and high convergence
rate, ease of use, simplicity, and low storage demand. References
Furthermore, the overall proposed HRES system is
economically viable, technically practical, and pro- [1] B. Sivasankari and A. Ahilan, Smart Energy Harvesting for
vides social and environmental benefits. Intelligent Railway Condition Monitoring System, Journal
of Electrical Engineering 19(2) (2019), 7–7.
[2] A.K. Akella, R.P. Saini and M.P. Sharma, Social, economi-
cal and environmental impacts of renewable energy systems,
7. Conclusion Renewable Energy 34(2) (2009), 390–396.
[3] D.P. Pathak and D. Khatod, Optimum utilization of alterna-
tive sources of energy for an un-electrified remote area, in
An efficient hybrid EPSO-HHO with fuzzy-ELM Proc. 2017 14th IEEE India Council International Confer-
technique in a smart grid linked hybrid RES system is ence INDICON, Roorkee, India, pp. 1–6, 2017.
[4] M.A. Mohamed, T. Jin and W. Su, An effective stochastic
presented in this research. The optimization aim is to
framework for smart coordinated operation of wind park and
reduce system costs while maintaining the demand of energy storage unit, Applied Energy 272 (2020), 115228.
load and meeting a set of optimization criteria. On a [5] X. Gong, F. Dong, M.A. Mohamed, E.M. Awwad, H.M.
MATLAB work site, the proposed EPSO-HHO with Abdullah and Z.M. Ali, Towards distributed based energy
transaction in a clean smart island, Journal of Cleaner Pro-
fuzzy-ELM technique is implemented and evaluated. duction 273 (2020), 122768.
The performance of EPSO-HHO with the fuzzy-ELM [6] H. Lund and B.V. Mathiesen, Energy system analysis of
approach is compared with recent techniques such 100% renewable energy systems—The case of Denmark in
as PSO, WOA, and HHO. The simulation findings years 2030 and 2050, Energy 34(5) (2009), 524–531.
[7] K. Rahbar, J. Xu and R. Zhang, Real-time energy storage
show that the proposed EPSO-HHO technique-based management for renewable integration in microgrid: An off-
hybrid renewable energy system costs less than PSO, line optimization approach, IEEE Transactions on Smart
WOA, and HHO by 4.89 %, 4.51 %, and 4.05 %, Grid 6(1) (2014), 124–134.
respectively, and takes less time, 34.901 secs. Also, [8] A.M. Eltamaly, K.E. Addoweesh, U. Bawah and M.A.
Mohamed, New software for hybrid renewable energy
the total power consumption of 25KWh/day was met assessment for ten locations in Saudi Arabia, Journal of
by a hybrid PV, wind, and battery system and the Renewable and Sustainable Energy 5(3) (2013), 033126.
power exchange has been efficiently maintained. The [9] A.M. Eltamaly, K.E. Addoweesh, U. Bawa and M.A.
proposed EPSO-HHO algorithm produces the lowest Mohamed, Economic modeling of hybrid renewable energy
system: a case study in Saudi Arabia, Arabian Journal for
LCOE of 0.140 Rs/kWh. According to the research Science and Engineering 39(5) (2014), 3827–3839.
made, the EPSO-HHO exceeds existing approaches [10] A.M. Eltamaly and M.A. Mohamed, A novel design and
in terms of producing better optimal sizing outcomes. optimization software for autonomous PV/wind/battery
hybrid power systems, Mathematical Problems in Engineer-
This proposed system may be expanded in the future
ing, 2014.
to accommodate various resources such as biogas, [11] Z. Pooranian, N. Nikmehr, S. Najafi-Ravadanegh, H.
hydro energy, fuel cells, diesel generators, and so on. Mahdin and J. Abawajy, Economical and environmental
6662 K. Suresh and P. Jagatheeswari / Economic analysis of a hybrid intelligent optimization

operation of smart networked microgrids under uncer- [20] M. Sarwar and B. Asad, A review on future power systems;
tainties using NSGA-II, in Proc. 2016 24th International technologies and research for smart grids, in Proc. 2016
Conference on Software, Telecommunications and Com- International Conference on Emerging Technologies ICET,
puter Networks SoftCOM, Split, Croatia, pp. 1–6, 2016. Islamabad, Pakistan, pp. 1–6, 2016.
[12] M.A. Mohamed, A.M. Eltamaly and A.I. Alolah, PSO- [21] N. Phuangpornpitak and S. Tia, Opportunities and chal-
based smart grid application for sizing and optimization of lenges of integrating renewable energy in renewable energy,
hybrid renewable energy systems, PloS One 11(8) (2016), Energy Procedia 34 (2013), 282–290.
e0159702. [22] O. Palizban, K. Kauhaniemi and J.M. Guerrero, Microgrids
[13] K. Roy, K.K. Mandal, A.C. Mandal and S.N. Patra, Anal- in active network management—Part I: hierarchical control,
ysis of energy management in micro grid–A hybrid BFOA energy storage, virtual power plants, and market participa-
and ANN approach, Renewable and Sustainable Energy tion, Renewable and Sustainable Energy Reviews 36 (2014),
Reviews 82 (2018), 4296–4308. 428–439.
[14] S. Amirtharaj, L. Premalatha and D. Gopinath, Optimal [23] V. Khare, S. Nema and P. Baredar, Solar–wind hybrid renew-
utilization of renewable energy sources in MG connected able energy system: A review, Renewable and Sustainable
system with integrated converters: an AGONN approach, Energy Reviews 58 (2016), 23–33.
Analog Integrated Circuits and Signal Processing 101(3), [24] V.M. Dhivya, A. Sridevi and A. Ahilan, A high speed area
513–532. efficient FIR filter using floating point dadda algorithm,
[15] K. Venkatesan and U. Govindarajan, Optimal power flow In 2014 International Conference on Communication and
control of hybrid renewable energy system with energy Signal Processing, pp. 1640–1644, 2014.
storage: A WOANN strategy, Journal of Renewable and [25] G.B. Huang, Q.Y. Zhu and C.K. Siew, Extreme learning
Sustainable Energy 11(1) (2019), 015501. machine: theory and applications, Neurocomputing 70(1-3)
[16] M.Z. Islam, N.I.A. Wahab, V. Veerasamy, H. Hizam, N.F. (2006), 489–501.
Mailah et al., A Harris Hawks optimization based single-and [26] E.H. Houssein, A.G. Gad, K. Hussain and P.N. Suganthan,
multi-objective optimal power flow considering environ- Major advances in particle swarm optimization: theory,
mental emission, Sustainability 12(13) (2020), 5248. analysis, and application, Swarm and Evolutionary Com-
[17] G. Suresh, D. Prasad and M. Gopila, An efficient approach- putation 63 (2021), 100868.
based power flow management in smart grid system with [27] A.A. Heidari, S. Mirjalili, H. Faris, I. Aljarah, M. Mafarja
hybrid renewable energy sources, Renewable Energy Focus and H. Chen, Harris hawks optimization: Algorithm and
39 (2021), 110–122. applications, Future Generation Computer Systems 97
[18] K. Roy, K.K. Mandal and A.C. Mandal, Application (2019), 849–872.
of ANFASO for optimal power flow management of [28] M.A. Mohamed, T. Jin and W. Su, Multi-agent energy
MG-connected system with energy storage, International management of smart islands using primal-dual method of
Journal of Energy Research 44(7) (2020), 5272–5286. multipliers, Energy 208 (2020), 118306.
[19] M. Dali, J. Belhadj and X. Roboam, Hybrid solar–wind
system with battery storage operating in grid-connected
and standalone mode: control and energy management–
experimental investigation, Energy 35(6) (2010),
2587–2595.
Copyright of Journal of Intelligent & Fuzzy Systems is the property of IOS Press and its
content may not be copied or emailed to multiple sites or posted to a listserv without the
copyright holder's express written permission. However, users may print, download, or email
articles for individual use.