International Journal of Applied Power Engineering (IJAPE)

Vol. 14, No. 2, June 2025, pp. 282~290

ISSN: 2252-8792, DOI: 10.11591/ijape.v14.i2.pp282-290  282

Optimal allocation of PV units using metaheuristic optimization

considering PEVs charging demand

A. Manjula, G. Yesuratnam
Department of Electrical Engineering, University College of Engineering, Osmania University, Hyderabad, India

Article Info ABSTRACT

Article history: The distribution system is seeing a dramatic shift as a result of the increasing
use of distributed generators (DGs) and plug-in electric vehicles (PEVs), or
Received Mar 25, 2024 plug-in hybrid electric cars. The research endeavors to optimize the
Revised Aug 18, 2024 allocation of photovoltaic (PV) based DGs within radial distribution systems
Accepted Oct 23, 2024 (RDS) while accommodating the load demand stemming from PEVs. A
weighted-sum based multiobjective (WMO) technique is employed in this
study to optimize three fundamental technical metrics of the distribution
Keywords: network: achieving the best possible voltage stability index (VSI) while
reducing real power loss and total voltage variation to a minimum. Initially,
Distributed generators the study investigates the impact of both conventional and PEVs load
Photovoltaic demand, considering PEVs load demand on distribution system performance
Plug-in electric vehicles under three charging scenarios: a situation involving peak charging, scenario
Radial distribution system involving off-peak charging, and scene of random charging. Subsequently,
WMODE PV units are strategically planned, taking into account the PEVs demand
WMOEEFO within the distribution system utilizing an innovative weighted
WMOGWO multiobjective electric eel foraging optimization (WMOEEFO) algorithm, its
effectuality is validated with weighted multiobjective differential
evolutionary (WMODE) and weighted multiobjective grey wolf
optimization (WMOGWO) algorithms on standard test system IEEE 33-bus.
This is an open access article under the CC BY-SA license.

Corresponding Author:
A. Manjula
Department of Electrical Engineering, University College of Engineering, Osmania University
Hyderabad-500007, Telangana, India
Email: [email protected]

1. INTRODUCTION
The demand for electricity is significant and is on the rise across the world. According to recent
reports [1], traditional energy sources still maintain a dominant share of 62% in the global energy mix.
However, the finite nature of fossil fuel resources coupled with their detrimental environmental impacts, such
as pollution and climate-altering emissions leading to ecological imbalance, render them increasingly
untenable as primary energy sources. As a result, the rapid incorporation of renewable energy sources (RESs)
into distribution networks has intensified the need of decarbonizing the electrical energy industry. Solar
photovoltaic (PV) units and wind turbines are the most well-known and fully functional examples of RESs
among the many technologies [2] that may generate electricity from low-carbon sources. Among these
renewable energy sources, solar-powered RESs have exploded in popularity during the last 18 years,
outpacing all others. A notable jump of 24% in the global energy mix was seen in 2022 alone for solar-based
energy systems.
Extensive literature documents the potential benefits of optimally allocating (sizing and siting)
distributed generators (DGs) within the distribution system [3], including minimizing of energy loss [4],
enhancement of the voltage profile [5], maximization of load ability [6], and augmentation of the voltage

Journal homepage: http://ijape.iaescore.com

Int J Appl Power Eng ISSN: 2252-8792  283

stability limit [7]. However, the optimal allocation of DGs presents a large-scale, nonlinear, and multi-
objective optimization problem, often posing significant challenges in finding near-optimal solutions [8].
Consequently, nature-inspired metaheuristic algorithms have gained prominence as effective approaches for
addressing this intricate optimization problem [9]. Electric vehicles (EVs) are becoming more common and
are expected to play a big part in reducing carbon emissions from road travel, in addition to PV system
integration [10]. Nonetheless, the inclusion of plug-in electric vehicle (PEV) charging loads remains largely
unaddressed in the literature concerning DG allocation [11]. Various studies have incorporated different PEV
charging profiles to evaluate the impact of PEV loads [12]. In [13], the application of a novel lightning search
calculation is proposed to address the DG assignment issue. Nevertheless, the study [14] overlooks a crucial
objective, namely, the voltage stability index (VSI), during DG allocation and confines its investigation to
dispatchable DGs exclusively. In contrast, Sankar and Chatterjee [15] determine the placement and
dimensions of DGs by use of the gorilla troops optimization technique.
The literature review underscores the necessity of considering plug-in electric vehicles (PEVs)
charging demand when allocating PV units within the context of contemporary research. In this vein, we
employ the weighted multiobjective electric eel foraging optimization (EEFO) algorithm to solve the PV unit
allocation problem. The EEFO algorithm as detailed by [16], emulates the foraging behavior of electric eels
and has been rigorously tested and compared with various renowned algorithms. This study contributes to the
existing state-of-the-art in the following aspects:
- Incorporating PEVs charging demand in the allocation of PV units. Considering the stochastic modeling
of the uncertain nature of PV generation.
- Comprehensive assessment of PEVs demand comprising off-peak charging scenario (OPCS), peak
charging scenario (PCS), and stochastic charging scenario (SCS) scenarios on the distribution network.
Distributing PV units while keeping in mind a number of important objectives such as power loss, voltage
deviation, and stability index.
- Introducing a novel application of the weighted multiobjective electric eel foraging optimization
(WMOEEFO) algorithm to address the complex PV allocation problem and comparing WMOEEFO with
the weighted multiobjective grey wolf optimization (WMOGWO) [17] and weighted multiobjective
differential evolutionary (WMODE) [18].
- Simulating numerous study scenarios to assess the impact of the number of PV units installed, and
considering test cases to quantitatively evaluate the objectives in each scenario.
The following is an overview of the paper: i) Section 2 describes the PV modeling; ii) The multiobjective
problem formulation is made in section 3; iii) In section 4, the EEFO algorithm is described in great length;
iv) Section 5 delves the results and discussions; and v) In section 6 the final conclusion is summarized.

2. SOLAR PV UNCERTAINTY MODELLING

A beta probability density function (PDF) was utilized to depict the arbitrary behaving of sun-based
illumination [19]. Within a designated time frame t, the beta PDF 𝑓(𝑧) supported by historical data used for
assessing the probability of solar irradiation is expressed as (1) [20].

𝛤(𝑥+𝑦)
𝑧 (𝑎−1) , 0 ≤ 𝑧 ≤ 1, 𝑥 ≥ 0, 𝑦 ≥ 0
𝑓(𝑧) = {𝛤(𝑥)𝛤(𝑦) (1)
0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒

Where 𝑧 signifies the solar irradiance, 𝑥 and 𝑦 define the parameters that delineate the configuration of the
PDF. Potential values of the solar irradiance state (z) at any given hour may be expressed as (2) [21].
𝑧
𝑃𝑧 (𝐻) = ∫𝑧 2 𝑓(𝑧). 𝑑𝑧 (2)
1

PV module output power may be expressed as (3).

𝑃𝑝𝑣0 (𝑧) = 𝑁𝑚 ∗ 𝐹𝐹𝑚 ∗ 𝑉𝑚 ∗ 𝐼𝑚 (3)

Where 𝐹𝐹𝑀 is the fill factor of PV module, 𝑁𝑀 is no of modules, 𝑉𝑀 is the voltage of the PV module, and 𝐼𝑀
is the current of the PV module. Under varying solar irradiance conditions, the specific performance
characteristics of PV panels output power are calculated as (4) [22].

𝑃𝑝𝑣 (𝑡) = 𝑃𝑧 (𝐻) ∗ 𝑃𝑝𝑣0 (𝑧) (4)

Optimal allocation of PV units using multiobjective metaheuristic optimization approach … (A. Manjula)
284  ISSN: 2252-8792

3. MULTIOBJECTIVE FUNCTION FORMULATION

In this investigation, three pivotal boundaries of the distribution system have been meticulously
examined to formulate the objective function. These parameters encompass energy loss (𝐸𝑙𝑜𝑠𝑠 ), total voltage
deviation (TVD), and VSI. The objective function (OF) is calculated as (5) [15].
(𝑓1 )𝐷𝐺 (𝑓2 )𝐷𝐺 1
𝑂𝐹 = 𝛾1 ∗ (𝑓 ) + 𝛾2 ∗ (𝑓 ) + 𝛾3 ∗ (𝑓3 )𝐷𝐺 (5)
1 𝑊𝑖𝑡ℎ𝑜𝑢𝑡 𝐷𝐺 2 𝑊𝑖𝑡ℎ𝑜𝑢𝑡 𝐷𝐺
(𝑓3)𝑊𝑖𝑡ℎ𝑜𝑢𝑡 𝐷𝐺

Where 𝛾1 , 𝛾2 , and 𝛾3 denote the preference weights given to the objectives following ∑3𝑖=1 𝛾𝑖 = 1 and 𝛾𝑖 for a
given objective is decided based on the preference given to that objective and 𝛾𝑖 ∈ [0,1]. (𝑓)𝐷𝐺 and
(𝑓)𝑊𝑖𝑡ℎ𝑜𝑢𝑡 𝐷𝐺 denote the value of the parameter before and after the installation of the DG. Here 𝑂𝐹 is the
overall objective function to be minimized. In this work, 𝛾1 , 𝛾2 , and 𝛾3 are assigned to 0.4, 0.3, and 0.3,
respectively. The individual objectives are calculated as given in (6)-(9).
𝑛𝑏𝑢𝑠−1 2
𝑓1 = 𝐸𝑙𝑜𝑠𝑠 = ∑24
𝑡=1 ∑𝑗=1 𝐼𝑡,𝑗 𝑅𝑗 (6)

𝑛𝑏𝑢𝑠 2
𝑓2 = 𝑇𝑉𝐷 = ∑24
𝑡=1 ∑𝑚=1 (|1 − 𝑉𝑡,𝑚 |) (7)

𝑓3 = 𝑉𝑆𝐼 = ∑24
𝑡=1 min (𝑆𝐼𝑡,𝑛 ) 𝑛 = 2 … … 𝑛 𝑏𝑢𝑠 (8)
4 2 2
𝑆𝐼𝑡,𝑛 = |𝑉𝑡,𝑚 | − 4[𝑃𝑡,𝑛 𝑋𝑚𝑛 − 𝑄𝑡,𝑛 𝑅𝑚𝑛 ] − 4[𝑃𝑡,𝑛 𝑅𝑚𝑛 + 𝑄𝑡,𝑛 𝑋𝑚𝑛 ]|𝑉𝑡,𝑚 | (9)

Where 𝐼𝑡,𝑗 , 𝑅𝑗 , and 𝑛𝑏𝑢𝑠 respectively indicate the 𝑗th branch current, resistance of the branch 𝑗, and the total
buses in the network. For a given bus 𝑚, 𝑉𝑡,𝑚 , 𝑃𝑡,𝑛 , 𝑋𝑚𝑛 , 𝑄𝑡,𝑛 , and 𝑅𝑚𝑛 represent the bus voltage, injected
real power, the reactance of the line between m and n buses, the injected reactive power injected,
and the resistance of the line between m and n buses. The objective function framed in (5) is bound to the
below constraints:

|𝑉𝑚𝑖𝑛 | ≤ |𝑉𝑡,𝑚 | ≤ |𝑉𝑚𝑎𝑥 | (10)

𝑃𝑡,𝑠𝑠 + 𝑃𝑡,𝐷𝐺 = 𝑃𝑡,𝐷 + 𝑃𝑡,𝑙𝑜𝑠𝑠 + 𝑃𝑡,𝑃𝐸𝑉 (11)

𝑃 𝑚𝑖𝑛,𝐷𝐺 ≤ 𝑃𝐷𝐺 ≤ 𝑃𝐷𝐺𝑚𝑎𝑥,𝐷𝐺 (12)

Where 𝑉𝑚𝑖𝑛 and 𝑉𝑚𝑎𝑥 respectively define the minimum and maximum values of the bus voltage. 𝑃𝑡,𝑠𝑠 , 𝑃𝑡,𝐷𝐺 ,
𝑃𝑡,𝐷 , 𝑃𝑡,𝑙𝑜𝑠𝑠 , and 𝑃𝑡,𝑃𝐸𝑉 respectively denote substation power, power injected by the DG, power demand of the
network, power losses in the network, and demand due to PEVs. 𝑃𝑚𝑖𝑛,𝐷𝐺 and 𝑃𝑚𝑎𝑥,𝐷𝐺 indicate the minimum
and maximum sizes of the DG rating.

4. ELECTRIC EEL FORAGING OPTIMIZATION ALGORITHM

A metaheuristic approach based on natural processes, the EEFO [16] takes its cues from the foraging
tactics used by electric eels. The algorithm aims to emulate the complex foraging behaviors displayed by
electric eels within their ecological environment. Notably, electric eels demonstrate four key foraging
behaviors: interaction, idling, resettlement, and hunting. In Figure 1, we can see the whole process flowchart.

4.1. Interaction
This behavior, also termed churning, occurs among eels as they engage in hunting fish. Throughout
this activity, eels exchange information by maneuvering randomly in various directions. In the framework of
EEFO, each eel represents a potential solution, with the most optimal solution identified thus far serving as
the target prey. The interacting phase can be modeled as (13) [16].

𝑣𝑖 (𝑡 + 1) = 𝑥𝑗 (𝑡) + 𝐶 × (𝑥̅ (𝑡) − 𝑥𝑖 (𝑡)) 𝑞1 > 0.5

{ } 𝑓𝑖𝑡 (𝑥𝑗 (𝑡)) < 𝑓𝑖𝑡(𝑥𝑖 (𝑡))
𝑣𝑖 (𝑡 + 1) = 𝑥𝑗 (𝑡) + 𝐶 × (𝑥𝑟 (𝑡) − 𝑥𝑖 (𝑡)) 𝑞1 ≤ 0.5
(13)
𝑣𝑖 (𝑡 + 1) = 𝑥𝑖 (𝑡) + 𝐶 × (𝑥̅ (𝑡) − 𝑥𝑗 (𝑡)) 𝑞2 > 0.5
{ } 𝑓𝑖𝑡 (𝑥𝑗 (𝑡)) ≥ 𝑓𝑖𝑡(𝑥𝑖 (𝑡))
[ 𝑣𝑖 (𝑡 + 1) = 𝑥𝑖 (𝑡) + 𝐶 × (𝑥 𝑟 (𝑡) − 𝑥𝑗 (𝑡)) 𝑞2 ≤ 0.5 ]

Int J Appl Power Eng, Vol. 14, No. 2, June 2025: 282-290
Int J Appl Power Eng ISSN: 2252-8792  285

Where 𝑞1 and 𝑞2 represent random numbers between (0, 1), 𝑓𝑖𝑡 (𝑥𝑗 (𝑡)), 𝑥𝑗 (𝑡), and 𝑛 respectively represent
candidate fitness, eel position, population size, and a random number between (0, 1).

4.2. Idling
Idling characterizes the behavior of an eel as it rests within an idling zone distinct from the
interaction zone. The idling zone is defined in EEFO by projecting an arbitrary eel position dimension onto
the search region's diagonal and then normalizing it within the range of 0 to 1, which enhances exploring
capabilities. The eels will modify their position for idling, which is modeled as (14) [16].

𝑣𝑖 (𝑡 + 1) = 𝑅𝑖 (𝑡 + 1) + 𝑛 × (𝑅𝑖 (𝑡 + 1)𝑟𝑜𝑢𝑛𝑑(𝑟𝑎𝑛𝑑) × 𝑥𝑖 (𝑡)); 𝑛~𝑁(0.1) (14)

start

Read test system data

and solar irradiance data

Define Objective function

and initialize EEFO
parameters

Generate initial
population(PV sizes and
locations)

Stopping YES Return the best

criteria met ? solution
NO

Interaction +
Idling+ End
Hunting+
Resettlement

Evaluate the fitness of

the population by load
flow analysis

Figure 1. EEFO algorithm flowchart

4.3. Hunting
During prey hunting, electric eels create an electric communication circle around the target. They
encircle the prey and communicate with each other through organ electric discharges, thereby forming an
electric circle that delineates the hunting zone. This hunting behavior in electric eels entails a curling
movement, which is represented as (15) [16], where 𝜂 denote curling parameter.

𝑣𝑖 (𝑡 + 1) = 𝐻𝑝𝑟𝑒𝑦 (𝑡 + 1) + 𝜂 × (𝐻𝑝𝑟𝑒𝑦 (𝑡 + 1) − 𝑟𝑜𝑢𝑛𝑑(𝑟𝑎𝑛𝑑) × 𝑥𝑖 (𝑡)) (15)

4.4. Resettlement
Resettlement is a migratory behavior observed in electric eels, wherein they transition from the
idling zone to the hunting zone. The (16) delineate the resettlement trait in EEFO [13].

𝑣𝑖 (𝑡 + 1) = −𝑟1 × 𝑅𝑖 (𝑡 + 1) + 𝑟2 × 𝐻𝑟 (𝑡 + 1) − 𝐿 × (𝐻𝑟 (𝑡 + 1) − 𝑥𝑖 (𝑡)) (16)

Where 𝐻𝑟 denotes any position within the hunting zone, 𝑟1 and 𝑟2 represent randomly selected values within
the interval (0.1).
Optimal allocation of PV units using multiobjective metaheuristic optimization approach … (A. Manjula)
286  ISSN: 2252-8792

5. RESULTS AND DISCUSSION

Using the IEEE 33-bus radial distribution systems (RDS), a standard test system, this research
validates the appropriate deployment of PV units in a distribution system that supports PEVs. Reference [23]
is used to get data of bus and line information for the 33-bus test system. Figure 2 shows a 33-bus RDS with
a base voltage of 12.66 KV and a base of 100 MVA; the peak values of real power demand of 3.715 MW,
and reactive power demand are 2.300 MVAR. The test systems' hourly power needs for each bus are derived
from the normal daily load pattern [24] shown in Figure 3. A PV unit with a maximum capacity of 3200 kW
and a minimum capacity of 100 kW is considered [25]. A total of 100 iterations were taken into account for
all algorithms in this research, with a population size of 200. The parameter-free optimization methods
WMOGWO and WMOEEFO are used in WMODE with mutation rates and crossover rates set at 0.7. The
optimal values are selected after each algorithm undergoes 30 separate runs. The MATLAB simulations were
performed on a computer with 8 GB of RAM and an Intel(R) Core (TM) i5-7200U 2.50 GHz CPU. This
research analyses the distribution system's performance in each of the following scenarios:
- Scenario 0: without PV units and without PEVs load demand, only conventional load demand in RDS.
- Scenario 1: without PV units and with PEVs, load demand and conventional load in RDS.
- Scenario 2: optimal deployment of one PV unit in RDS hosting PEVs' load demand and conventional load.
- Scenario 3: optimal deployment of two PV units in RDS hosting PEVs' load demand and conventional load.
- Scenario 4: optimal deployment of three PV units in RDS hosting PEVs' load demand and conventional load.
Scenario 0 involves a distribution system subjected to a load flow algorithm without PV units to get
a high-level picture of the system's technical parameters. In scenario 1, the load flow algorithm is utilized to
dissect the influence on system technical metrics caused by the addition of PEVs to the traditional load
demand. For scenarios 2, 3, and 4, the best way to meet the load demand of PEVs is to use one, two, or three
PV units in an RDS. This will reduce the network 𝐸𝑙𝑜𝑠𝑠 , TVD, and improve the system VSI.

Figure 2. IEEE 33-bus system single line diagram Figure 3. Load curve in p.u.

The results produced by WMOEEFO for scenarios 2-4, are summarized in Table 1, are as follows.
A total of 2677 kW of 𝑃𝑙𝑜𝑠𝑠 , 1.6931 p.u. of TVD, and 0.757 p.u. of VSI were recorded in scenario 0 load
flow results. Scenario 1 considers a 33-bus system with a total load of 288 PEVs, PEVs of 9 per bus as
shown in Figure 2, to concentrate on the interest on the electric circulation framework brought about by
PEVs. It is assumed that the state of charge (SOC) of PEVs is 50%, and all PEVs use 25 kWh batteries [15].
The daily charging of 288 PEVs requires a total of 3600 kW of electrical power, calculated as 288*25*0.5.
Three different scenarios for charging PEVs are shown in Figure 4: PCS, OPCS, and SCS. This research
considers that PEVs charge equally under PCS, OPCS, and SCS. Scenarios PCS, OPCS, and SCS are used to
calculate the electric power needed to charge PEVs in a day. Scenario 2 involves the execution of the load
flow algorithm. Figure 5 displays the hourly variation of substation power in scenarios 0 and 1, indicating
that the system's load demand from PEVs causes an increase in substation power. Three technical measures
have deteriorated subsequently: 𝑃𝑙𝑜𝑠𝑠 of system has deteriorated to 2913 kW, which accounts for an 8.1%
improvement; TVD has deteriorated to 1.8581 p.u., and VSI has further aggravated to 0.745 p.u. In scenario 2,
a single 3194 kW PV unit is optimally connected to the 7th bus, reducing the system's 𝑃𝑙𝑜𝑠𝑠 to 2106 kW
(a 27.70% decrease), improving TVD to 1.0837 p.u., and maximizing VSI to 0.824 p.u. Scenario 3's efficient
linking of two 932 kW and 1424 kW PV units at the 13th and 30th buses reduces the system's 𝑃𝑙𝑜𝑠𝑠 to 1845 kW,
which is a 36.36% improvement; it also improves TVD to 1.0481 p.u. and maximizes VSI to 0.831 p.u. As a
result of connecting three PV units at the 14th, 24th, and 30th buses, with a capacity of 844 kW, 992 kW, and
1313 kW, respectively, in scenario 4, the system's 𝑃𝑙𝑜𝑠𝑠 is reduced to 1742 kW, accounting for 40.19%, the

Int J Appl Power Eng, Vol. 14, No. 2, June 2025: 282-290
Int J Appl Power Eng ISSN: 2252-8792  287

TVD is enhanced to 1.0437 p.u, and the VSI is maximized to 0.842 p.u. Figure 6 shows the power production
curves for the hourly PV units produced for scenario 4 of the 33-bus system.

Table 1. Summary of outcomes generated by WMOEEFO for scenarios 0-4 of 33-bus system for 24 hours
S.L. Technical metrics Scenario 0 Scenario 1 Scenario 2 Scenario 3 Scenario 4
1 PV loc’s/PV sizes (kW) - - 7/3194 13/0932 14/0844
30/1424 24/0992
30/1313
2 Substation power (kVA) 78351 81813 53365 60201 53250
3 Objective function (OF) - - 0.7356 0.6916 0.6733
4 Real power loss (𝑃𝑙𝑜𝑠𝑠 ) in kW 2677 2913 2106 1845 1742
5 Total voltage deviation (TVD) in p.u. 1.6931 1.8581 1.0837 1.0481 1.0437
6 Voltage stability index (VSI) in p.u. 0.757 0.745 0.824 0.831 0.842
7 % 𝑃𝑙𝑜𝑠𝑠 reduction - - 27.70 36.36 40.19

Figure 4. Distribution of probabilities for the PEVs in Figure 5. Hourly substation power of 33-bus
PCS, OPCS, and SCS scenarios system without and with PEVs

Figure 6. PV unit output curves for scenario 4 of the 33-bus system

In Figure 7, we can see the 33-bus system average voltage profile for scenarios 0–4. Figure 7 further
demonstrates that, in scenario 1, the load demand of PEVs worsens the system voltage, while the optimal
possible deployment of PV units further develops the voltage profile of the system. Figure 8 shows the hourly
loss of the system for scenarios 0–4. In order to determine the WMOEEFO algorithm's effectiveness using
the WMODE and WMOGWO algorithms to run in scenarios involving the fourth scenario of the 33-bus test
system. The summary of outcomes by comparing different methods is illustrated in Table 2. Figure 9 also
shows how the WMOEEFO, WMODE, and WMOGWO algorithms converge for the test system of scenario 4.
The WMOEEFO algorithm outperforms WMODE and WMOGWO in achieving optimal results.

Optimal allocation of PV units using multiobjective metaheuristic optimization approach … (A. Manjula)
288  ISSN: 2252-8792

Figure 7. 33-bus system mean voltage profile for scenarios 0–4

Figure 8. Power loss in the 33-bus system for scenarios 0-4 on an hourly basis

Figure 9. Convergence of WMOEEFO, WMODE, and WMOGWO algorithms for 33-bus system

Table 2. Summary of outcomes generated by WMOEEFO, WMOGWO, and WMODE for scenario 4 of
the 33-bus system
S.No System Optimization technique PV loc’s/PV sizes (kW) OF value 𝑃𝑙𝑜𝑠𝑠 (kW) TVD (p.u.) VSI (p.u.)
WMOEEFO 14/0844, 24/0992, 30/1313 0.6733 1742 1.0437 0.842
1 33-bus WMOGWO 10/1074, 25/0649, 30/1209 0.6772 1758 1.0463 0.839
WMODE 10/1272, 25/0667, 30/1010 0.6799 1788 1.0481 0.834

6. CONCLUSION
This research focused on optimizing the placement of PV units in an RDS that caters to the load
demand of PEVs. An evaluation of the suggested approach was carried out utilizing the IEEE 33-bus RDS.
The study aimed at optimizing three key technical metrics of the system: maximizing VSI, minimizing TVD,
and minimizing 𝑃𝑙𝑜𝑠𝑠 . In order to achieve these objectives, a weighted-based multiobjective approach was

Int J Appl Power Eng, Vol. 14, No. 2, June 2025: 282-290
Int J Appl Power Eng ISSN: 2252-8792  289

developed and used the EEFO algorithm to minimize its corresponding function. The study was considered
into two parts. The first part looked at how PEVs' load demand affected the operation of the distribution
system. The second part showed how to best allocate PV units in the system to accommodate PEVs' load
demand. Three different charging scenarios (PCS, OPCS, and SCS) were used to model the electrical power
requirement of PEVs on an hourly basis. Research showed that the load demand from PEVs degraded the test
system's performance. Maximal gains in all three technical parameters were achieved by strategically placing
three PV units in the 33-bus distribution system. Real power loss in 33-bus system was reduced by
approximately 40-42%. Nevertheless, the minimum voltage of the system did not improve due to the absence
of PV unit power during peak load times. Further enhancements in loss reduction and minimum voltage were
observed through dispatchable DGs, suggesting a potential future avenue for this research. Comparing
optimization algorithms, the WMOEEFO algorithm demonstrated superior performance in achieving the
optimal solution when contrasted with WMODE and WMOGWO.

FUNDING INFORMATION
Authors state no funding involved.

AUTHOR CONTRIBUTIONS STATEMENT

This journal uses the Contributor Roles Taxonomy (CRediT) to recognize individual author
contributions, reduce authorship disputes, and facilitate collaboration.

Name of Author C M So Va Fo I R D O E Vi Su P Fu
A. Manjula ✓ ✓ ✓ ✓ ✓ ✓
G. Yesuratnam ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓

C : Conceptualization I : Investigation Vi : Visualization

M : Methodology R : Resources Su : Supervision
So : Software D : Data Curation P : Project administration
Va : Validation O : Writing - Original Draft Fu : Funding acquisition
Fo : Formal analysis E : Writing - Review & Editing

CONFLICT OF INTEREST STATEMENT

Authors state no conflict of interest.

DATA AVAILABILITY
Derived data supporting the findings of this study are available from the corresponding author
[AM], on request.

REFERENCES
[1] A. Lebsir and T. Benamimour, “Renewable energies in the twenty-first century: a global-view,” 2023 Second International
Conference on Energy Transition and Security (ICETS), 2023, pp. 1-6, doi: 10.1109/ICETS60996.2023.10410757.
[2] P. Singh, N. K. Meena, J. Yang, E. Vega-Fuentes, and S. K. Bishnoi, “Multi-criteria decision making monarch butterfly
optimization for optimal distributed energy resources mix in distribution networks,” Applied Energy, vol. 278, p. 115723, Nov.
2020, doi: 10.1016/j.apenergy.2020.115723.
[3] M. M. Sankar and K. Chatterjee, “A posteriori multiobjective approach for techno-economic allocation of PV and BES units in a
distribution system hosting PHEVs,” Applied Energy, vol. 351, p. 121851, Dec. 2023, doi: 10.1016/j.apenergy.2023.121851.
[4] B. Ahmadi, O. Ceylan, A. Ozdemir, and M. Fotuhi-Firuzabad, “A multi-objective framework for distributed energy resources
planning and storage management,” Applied Energy, vol. 314, p. 118887, May 2022, doi: 10.1016/j.apenergy.2022.118887.
[5] S. K. Injeti and V. K. Thunuguntla, “Optimal integration of DGs into radial distribution network in the presence of plug-in electric
vehicles to minimize daily active power losses and to improve the voltage profile of the system using bio-inspired optimization
algorithms,” Protection and Control of Modern Power Systems, vol. 5, no. 1, p. 3, Dec. 2020, doi: 10.1186/s41601-019-0149-x.
[6] V. K. Thunuguntla and S. K. Injeti, “Butterfly optimizer assisted max–min based multi-objective approach for optimal connection
of DGs and optimal network reconfiguration of distribution networks,” Journal of Electrical Systems and Information
Technology, vol. 9, no. 1, p. 8, Dec. 2022, doi: 10.1186/s43067-022-00049-y.
[7] A. Fathy, “A novel artificial hummingbird algorithm for integrating renewable based biomass distributed generators in radial
distribution systems,” Applied Energy, vol. 323, p. 119605, Oct. 2022, doi: 10.1016/j.apenergy.2022.119605.
[8] P. Singh, N. K. Meena, S. K. Bishnoi, B. Singh, and M. Bhadu, “Hybrid elephant herding and particle swarm optimizations for
optimal DG integration in distribution networks,” Electric Power Components and Systems, vol. 48, no. 6–7, pp. 727–741, Apr.
2020, doi: 10.1080/15325008.2020.1797931.
[9] R. O. Bawazir and N. S. Cetin, “Comprehensive overview of optimizing PV-DG allocation in power system and solar energy
resource potential assessments,” Energy Reports, vol. 6, pp. 173–208, Nov. 2020, doi: 10.1016/j.egyr.2019.12.010.

Optimal allocation of PV units using multiobjective metaheuristic optimization approach … (A. Manjula)
290  ISSN: 2252-8792

[10] N. Jabalameli and A. Ghosh, “Online centralized coordination of charging and phase switching of PEVs in unbalanced LV networks
with high PV penetrations,” IEEE Systems Journal, vol. 15, no. 1, pp. 1015–1025, 2021, doi: 10.1109/JSYST.2020.3000504.
[11] A. Selim, S. Kamel, A. A. Mohamed, and E. E. Elattar, “Optimal allocation of multiple types of distributed generations in radial
distribution systems using a hybrid technique,” Sustainability, vol. 13, no. 12, p. 6644, Jun. 2021, doi: 10.3390/su13126644.
[12] H. Ma, Z. Yang, P. You, and M. Fei, “Multi-objective biogeography-based optimization for dynamic economic emission load
dispatch considering plug-in electric vehicles charging,” Energy, vol. 135, pp. 101–111, 2017, doi: 10.1016/j.energy.2017.06.102.
[13] Y. Thangaraj and R. Kuppan, “Multi-objective simultaneous placement of DG and DSTATCOM using novel lightning search
algorithm,” Journal of Applied Research and Technology, vol. 15, no. 5, pp. 477–491, Oct. 2017, doi: 10.1016/j.jart.2017.05.008.
[14] J. N. Nweke, A. O. Salau, and C. U. Eya, “Headroom-based optimization for placement of distributed generation in a distribution
substation,” Engineering review, vol. 42, no. 1, pp. 109–120, 2022, doi: 10.30765/er.1748.
[15] M. M. Sankar and K. Chatterjee, “Optimal accommodation of renewable DGs in distribution system considering plug-in electric
vehicles using gorilla troops optimizer,” in 2023 International Conference on Recent Advances in Electrical, Electronics & Digital
Healthcare Technologies (REEDCON), May 2023, pp. 368–373, doi: 10.1109/REEDCON57544.2023.10151205.
[16] W. Zhao et al., “Electric eel foraging optimization: A new bio-inspired optimizer for engineering applications,” Expert Systems
with Applications, vol. 238, p. 122200, Mar. 2024, doi: 10.1016/j.eswa.2023.122200.
[17] A. B. Alyu, A. O. Salau, B. Khan, and J. N. Eneh, “Hybrid GWO-PSO based optimal placement and sizing of multiple PV-DG
units for power loss reduction and voltage profile improvement,” Scientific Reports, vol. 13, no. 1, p. 6903, Apr. 2023, doi:
10.1038/s41598-023-34057-3.
[18] N. Karuppiah, “Optimal siting and sizing of multiple type DGs for the performance enhancement of distribution system using
Differential Evolution Algorithm,” Turkish Journal of Computer and Mathematics Education (TURCOMAT), vol. 12, no. 2, pp.
1140–1146, 2021.
[19] J. Radosavljevic, N. Arsic, M. Milovanovic, and A. Ktena, “Optimal placement and sizing of renewable distributed generation
using hybrid metaheuristic algorithm,” Journal of Modern Power Systems and Clean Energy, vol. 8, no. 3, pp. 499–510, 2020,
doi: 10.35833/MPCE.2019.000259.
[20] K. Fettah, T. Guia, A. Salhi, S. Mouassa, A. Bosisio, and R. Shirvani, “Optimal allocation of capacitor banks and distributed
Generation: A comparison of recently developed metaheuristic optimization techniques on the real distribution networks of ALG-
AB-Hassi Sida, Algeria,” Sustainability, vol. 16, no. 11, p. 4419, May 2024, doi: 10.3390/su16114419.
[21] K. N. Maya and E. A. Jasmin, “Optimal integration of distributed generation (DG) resources in unbalanced distribution system
considering uncertainty modelling,” International Transactions on Electrical Energy Systems, vol. 27, no. 1, p. e2248, Jan. 2017,
doi: 10.1002/etep.2248.
[22] A. Soroudi, M. Aien, and M. Ehsan, “A probabilistic modeling of photo voltaic modules and wind power generation impact on
distribution networks,” IEEE Systems Journal, vol. 6, no. 2, pp. 254–259, Jun. 2012, doi: 10.1109/JSYST.2011.2162994.
[23] A. Hamouda and K. Zehar, “Efficient load flow method for radial distribution feeders,” Journal of Applied Sciences, vol. 6, no.
13, pp. 2741–2748, Jun. 2006, doi: 10.3923/jas.2006.2741.2748.
[24] T. Gu et al., “Placement and capacity selection of battery energy storage system in the distributed generation integrated
distribution network based on improved NSGA-II optimization,” Journal of Energy Storage, vol. 52, p. 104716, Aug. 2022, doi:
10.1016/j.est.2022.104716.
[25] V. K. Thunuguntla and S. K. Injeti, “Ɛ-constraint multiobjective approach for optimal network reconfiguration and optimal
allocation of DGs in radial distribution systems using the butterfly optimizer,” International Transactions on Electrical Energy
Systems, vol. 30, no. 11, p. e12613, Nov. 2020, doi: 10.1002/2050-7038.12613.

BIOGRAPHIES OF AUTHORS

A. Manjula received her B.Tech. in Electrical and Electronics Engineering from

the JNTU, Hyderabad in 2010 and M.Tech. from JNTU, Hyderabad in Electrical Power
Engineering in 2012. She is currently a Ph.D. student at Osmania University, Hyderabad. Her
research interests include renewable energy sources and power system analysis. She can be
contacted at email: [email protected].

G. Yesuratnam completed his B.Tech. from JNTU, Hyderabad in 1995 and

M.Tech. in Power Systems Engineering, Regional Engineering College, Warangal in 1997. He
was awarded a Ph.D. in Electrical Engineering from the Indian Institute of Science, Bangalore in
2007. Currently, he is working as a Sr. professor in the Department of Electrical Engineering at
Osmania University, Hyderabad. His research interests include computer-aided power system
analysis, reactive power optimization, power system security, voltage stability, AI applications
in power systems, and gas-insulated substations. He is serving as a peer reviewer of
international journals. Throughout his career, he has supervised Ph.D. students as well as
published many papers in IEEE as well as international conference proceedings and journals. He
can be contacted at email: [email protected].

Int J Appl Power Eng, Vol. 14, No. 2, June 2025: 282-290