Microsoft Word - 42-3517_s_ETASR_V10_N3_pp5837-5844
Engineering, Technology & Applied Science Research Vol. 10, No. 3, 2020, 5837-5844 5837
www.etasr.com Ponnam & Swarnasri: Multi-Objective Optimal Allocation of Electric Vehicle Charging Stations …
Multi-Objective Optimal Allocation of Electric
Vehicle Charging Stations and Distributed Generators
in Radial Distribution Systems using Metaheuristic
Optimization Algorithms
Venkata K. Babu Ponnam
Department of EEE
Acharya Nagarjuna University
Guntur, India
kishore.ponnam@gmail.com
K. Swarnasri
Department of EEE
RVR & JC College of Engineering
Guntur, India
swarnasrik@gmail.com
Abstract—The acceptance rate of Electric Vehicles (EVs) in the
transport industry has increased substantially due to the
augmented interest towards sustainable transportation
initiatives. However, their impact in terms of increased power
demand on the electric power market can increase real power
losses, decrease voltage profile, and consequently decrease
voltage stability margins. It is necessary to install Electric Vehicle
Charging Stations (EVCSs) and Distributed Generators (DGs) at
optimal locations to decrease the EV load effect in the Radial
Distribution System (RDS). This paper addresses a multi-
objective optimization technique to obtain simultaneous EVCS &
DG placement and sizing. The problem is formulated to optimize
real power losses, Average Voltage Deviation Index (AVDI), and
Voltage Stability Index (VSI) of the electrical distribution system.
Simulation studies were performed on the standard IEEE 33-bus
and 69-bus test systems. Harries Hawk Optimization (HHO) and
Teaching-Learning Based Optimization (TLBO) algorithms were
selected to minimize the system objectives. The simulation
outcomes reveal that the proposed approach improved system
performance in all aspects. Among HHO and TLBO, HHO is
reasonably successful in accomplishing the desired goals.
Keywords-Electrical Vehicles (EVs); Electric Vehicle Charging
Stations (EVCSs); renewable energy sources; Distributed
Generators (DGs); Average Voltage Deviation Index (AVDI);
Voltage Stability Index (VSI); Harris Hawks Optimization (HHO);
Teaching-Learning Based Optimization (TLBO)
I. INTRODUCTION
Ever-increasing demands for energy, the limited availability
of fossil fuels, and the global climate change are some major
concerns of the 21
st
century. The CO2 emissions from the
transportation sector are one of the major causes of global
climate change. Researchers highlighted the significant impact
of shifting from conventional vehicles to EVs to reduce the
transport sector's greenhouse gas contribution. Developing
countries are also attracted to EVs due to the low electricity
costs as compared to fossil fuel. One of the disadvantages of
EVs is the driving range limitation. After running a certain
distance, EVs need to recharge the battery at Charging Stations
(CSs). Thus, focus must be put on the construction of charging
infrastructure for large-scale installation of EVs. Coordinated
EVCS planning is of prevailing importance for the sustainable
growth of the EV industry.
In [1], a multi-objective function is built that takes into
account the performance indexes of the traffic and electrical
distribution network, including traffic conditions, charging wait
time, power loss, and node voltage, to plan the charging
behavior of EVs and achieve optimum efficiency of the entire
system. Nevertheless, for optimum location, the effect on the
voltage profile and branch power losses needs to be examined.
The low sensitivity voltage bus [2] could be the best location
for the charging station. Nevertheless, this approach fails to
provide the optimum EVCS on the higher radial bus systems.
Heuristic algorithms have been developed to optimize the
placement of EVCS not only for servicing EVs but also to
minimize losses and thus to increase voltage profile. To
examine the effect of charging EVs on the electrical
distribution network the Fast Charging Station (FCS) load is
designed in [3] by taking into account the relationship between
the power consumption of EVs and grid voltage. The suggested
model is also compared to the constant ZIP model and constant
power load model. In [4], the Genetic Algorithm is used mainly
to optimally place the EVCSs without considering losses and
conductor thermal limits. This method of positioning EVCSs
often impacts the reliability of the grid-connected distribution
system. By reconfiguring the given RDS [5], the effect of the
charging stations could be minimized. In [6], the location of the
EVCSs is identified in two stages. The EVCSs’ operation
spectrum is calculated in the 1
st
stage using the trip success
ratio with the account of the volatility of trip distance and
variability in the remaining electrical charge of the EVs. In the
2
nd
stage, the CS spectrum of operation is calculated for
optimal CS location. In [7], a tri-objective optimization
function that minimizes the charging station’s levelized cost of
energy, emissions, and losses on the electrical grid by
proposing the use of photovoltaic cells and of a battery storage
system. Most studies concentrate on the optimum location and
Corresponding author: Venkata K Babu Ponnam
Engineering, Technology & Applied Science Research Vol. 10, No. 3, 2020, 5837-5844 5838
www.etasr.com Ponnam & Swarnasri: Multi-Objective Optimal Allocation of Electric Vehicle Charging Stations …
size of EVCSs [8–12]. Improper positioning and sizing of
EVCSs hurt the distribution system by increasing the network
power losses and cause more depreciation in voltage profile. A
detailed review of EV technologies and the impact of electrical
demand from Plug-in Electric Vehicles (PEVs) on load profiles
have been provided in [13]. Authors in [14, 15] addressed the
problem of dynamic economic dispatch by incorporating PEVs
that charge patterns into a 24-hour load demand in economic
and environmental dispatch. Authors in [16] suggested a
control mechanism to increase the efficiency of the energy
shared between a photovoltaic generator and an unbalanced
system. In [17], the detrimental impact on the distribution
network of resulting EVCS loads is analyzed for voltage
stability, reliability indices, power losses, and economic loss of
the IEEE 33 bus system. In [18], the simultaneous placement of
EVCSs and DGs in the distribution system is proposed in order
to reduce EV user charge and Network Power Losses (NPL)
cost for the same station development cost and DG investment.
Non-dominated sorting Genetic Algorithm II was utilized for
achieving the optimal placement of EVCS and DGs. The
accuracy of the proposed methodology is checked on the 118
bus system.
Authors in [19] proposed a multi-objective function by
considering the daily real power losses and voltage deviations
under a 24-hour load pattern, including residential, industrial,
and commercial loads. Particle Swarm Optimization (PSO) and
Butterfly Optimization (BO) algorithms were applied to
minimize the desired objectives. A repeated forward-backward
load flow was used to determine losses and voltages.
Compared to PSO, BO does well in achieving the desired
goals. In [20], a multi-objective hybrid Shuffled Frog Leap
Teaching Learning Based Optimization (SFL-TLBO)
algorithm was proposed for better planning of EVCSs and DGs
in the combined electrical distribution and transportation
network, considering the objectives of voltage deviation, power
loss, DG power cost, and energy consumption. The proposed
model has been assessed on the IEEE 118 bus system and was
confirmed to be accurate and stable for different EV population
levels. When considering a large number of PEVs charging and
discharging during peak and off-peak grid hours, the optimum
positioning of EVs on the IEEE-33 radial distribution system
using the PSO-CFA optimization algorithm was suggested in
[21] to reduce power losses and enhance voltage profile. In
[22], the Artificial Bee Colony (ABC) algorithm was proposed
for optimal placement of Battery Swapping Stations (BSS) and
DG in the RDS to mitigate the power loss increment due to the
EV load. If both the DG and BSS are configured concurrently,
the system's power losses and reliability are significantly
improved. Accordingly, the authors concluded that the
simultaneous placement of DG and BSS systems is the best
choice to get the maximum benefit from the deployment of DG
and BSS in the network. PSO algorithm [23] was used to
reduce real power losses and enhance the voltage profile of the
system with optimal placement of charging stations.
This paper's main contributions are:
• Identifying the optimal locations of EVCSs by considering
the distribution system losses, AVDI, and VSI of RDS.
• Finding optimum DG size and location by considering
optimum EVCS load to boost system bus voltage profile
and reducing RDS power losses.
• Placing different categories of DGs and EVCSs
simultaneously while taking into account system constraints
for minimizing system power losses and better system bus
voltage regulation.
• An efficient multi-objective function formed using TLBO
[24] and HHO [25] algorithms is proposed for finding the
optimal solution of the optimal number of EVCSs, DGs,
and their simultaneous positioning and sizing for
minimizing distribution system losses, AVDI and enhance
VSI of RDS.
II. PROBLEM FORMULATION
A. Objective Function
The size and placement of EVCSs and DGs should be
designed to minimize the impact of the increased EV load on
distribution system’s performance. Radial Distribution System
(RDS) has a large R/X ratio and because of this basic load flow
tools such as Newton Raphson or fast decoupled approaches do
not provide accurate results. An efficient load flow method is
presented in [26] based on the forward-backward method for
solving the power flows of RDS. The objectives of this study
are minimizing power losses, AVDI, and enhancing VSI. The
objective function for power losses of the system is given by:
����� = min ∑ �� ∗ �������� (1)
where Ri is the resistance of the i
th
branch, Ii is the current
flowing in the i
th
branch, i is the branch number and br is the
total number of branches. AVDI is defined in terms of all bus
voltage magnitudes are deviation w.r.t. reference voltage
magnitude 1.0p.u. and is given in:
����� = �� ∑ |1 − ��|����� (2)
where Vk is the voltage at the k
th
bus, k is the bus number, and b
is the total number of buses. The VSI of a receiving end bus of
a branch can be determined by [27]:
����� = �|��|� − 4���� � + "�# �$� − 4���# � + � �$|��|�% (3)
where Vk is the voltage at k
th
bus, Pk is the total real power
demand at the k
th
bus, Qk is the total reactive power demand at
the k
th
bus, rjk is the resistance of the branch ‘jk’, and xjk is the
reactance of the branch ‘jk’.
VSIk should be greater than 0 for stable RDS operation. The
lowest value of VSI can be considered as the overall stability of
the given system. The ultimate objective function is developed
to minimize power losses, average voltage deviation, and
optimize the index of voltage stability. Mathematically:
&��� = min '(������ + (������ + (� ) �*+���,- (4)
B. Constraints
Equality constraints: The constraints for power balance
requirements:
Engineering, Technology & Applied Science Research Vol. 10, No. 3, 2020, 5837-5844 5839
www.etasr.com Ponnam & Swarnasri: Multi-Objective Optimal Allocation of Electric Vehicle Charging Stations …
∑ �.�/0��� − �1 = �2 (5)
∑ ".�/03�� − "1 = "2 (6)
where PGk and QGk are the active and reactive powers injected
by DG at the k
th
bus, PL and QL are the active and reactive
power losses at the k
th
bus, Pd and Qd are the active and reactive
power demands at the k
th
bus, and NG is the total number of
buses.
Inequality Constraints: The acceptable limitation of
distribution systems should not extend the highest possible
allowable power generated from DGs, voltage deviation should
be within 5%, the number of charging points (nCP) and the
number of Charging Stations (nCSs) should be between
minimum and maximum allowable charging points and
charging stations respectively, i.e.:
�.�4�5 ≤ �.� ≤ �.�478 (7)
".�4�5 ≤ ".� ≤ ".�478 (8)
0.95 ≤ �� ≤ 1.05, k = 1,2,3 … … ., B (9)
BC�4�5 ≤ BC� ≤ BC�478 (10)
BCD4�5 ≤ BCD ≤ BCD478 (11)
III. HHO FOR EVCS AND DG SIMULTANEOUS OPTIMAL
PLANNING
Harris Hawks Optimization (HHO) was proposed in 2019.
The harris hawk is a predator bird that lives at steady
communities found in the southern side of Arizona, United
States. Harris hawk’s main tactic to catch prey is “surprise
pounce” and it can also be recognized as the strategy of “seven
kills”. In this strategy, many hawks try to strike together from
different directions and concentrate on one rabbit. The assault
can be easily accomplished by catching the surprised victim in
a couple of seconds, although sometimes, with regards to the
prey's escape capability and habits, the “seven kills” can
include several, short length, rapid dives near the victim for
several minutes. Harris hawks can display a diverse range of
hunting styles depending on complex nature and prey's escape
patterns. When the best hawk (leader) bows down a prey and
departs, a switching tactic occurs, and one of the party
members can continue the chase. In different circumstances,
such switching activities could be examined as they are helpful
for annoying the escaping rabbit. The main benefit of such
collaborative tactics is that the harris hawks can pursue to
exhaustion the detected rabbit which cannot restore its
defensive capabilities by confusing the predators and usually
one of the hawks, which is often the most effective and skilled,
catches the exhausted rabbit quickly and shares it with the
others. Generally, HHO is modeled into exploration and
exploitation stages inspired by harris hawks exploration of a
prey, surprise pounce, and various attacking strategies.
A. Exploration Phase
The position of the hawk is changed in the exploration
phase by random location and the other hawks and is given by
(12) [25]:
1 2
3 4 lim lim
( ) | ( ) 2 ( ) |, 0,5
( 1)
( ( ) ( )) ( ( )), 0.5
k k
r m
L t d L t d L t r
L t
L t L t d lb d u l r
− − ≥
+ =
− − + − <
(12)
where L is the hawk's position, Lk is the position of a randomly
chosen hawk, Lr is the location of prey, t is the present
iteration, ulim and llim are the threshold limits of the search area,
d1, d2, d3, d4, and r are five discrete random numbers in [0, 1],
and Lm is the average location of the present hawks’
community, which can be calculated by:
E4�F� = �G ∑ EH�F�GH�� (13)
where Lp is the p
th
hawk in the community, and H is the
number of hawks.
B. Transition from Exploration to Exploitation
The HHO algorithm can be transposed from exploration to
exploitation. The behavior of exploration is modified
depending on the prey's escaping strength. Mathematically,
prey's strength to escape can be measured as
D = 2DI�1 − JK� (14)
DI = 2L − 1 (15)
where T is the max number of iterations, S0 is the initial
strength produced arbitrary in [-1, 1], and d is a random
number in [0, 1]. When the prey’s escaping strength S is greater
than 1, HHO grants the hawks the possibility to hunt on various
regions globally. On the other hand, HHO boosts the local
search of the best alternatives around the neighborhood when
the prey's escaping strength is less than 1.
C. Exploitation Phase
During this phase, the position of the hawk is changed
according to the scenarios described below. Its behavior is
controlled based on the prey's escape strength (S) and the
prey’s chance to escape successfully (q<0.5) or not (q>0.5)
before a surprise bounce.
1) Soft Besiege
This phase appears when q≥0.5 and |S|≥0.5, and by using
(16), the hawk adjusts its position:
E�F + 1� = ∆E�F� − D|NEO�F� − E�F�| (16)
where S is the prey's escaping strength, Lq is the position of
prey, ∆L is the difference between the prey’s position and
present hawk position, and j is the dive strength. The j and ∆L
are calculated by [25]:
∆E�F� = EO�F� − E�F� (17)
N = 2�1 − LP� (18)
where d5 is a random number in [0, 1] that is modified in each
iteration.
2) Hard Besiege
This phase appears when q≥0.5 and |S|<0.5. In this scenario,
the hawk updates its position using [25]:
E�F + 1� = EO�F� − D|∆E�F�| (19)
Engineering, Technology & Applied Science Research Vol. 10, No. 3, 2020, 5837-5844 5840
www.etasr.com Ponnam & Swarnasri: Multi-Objective Optimal Allocation of Electric Vehicle Charging Stations …
3) Soft Besiege with Progressive Rapid Dives
This phase occurs when q<0.5 and |S|≥0.5. The hawk is
slowly selecting the best possible dive for capturing the prey.
In this situation the hawk's new position is created as [25]:
Q = EO�F� − D|NEO�F� − E�F�| (20)
" = Q + R ∗ ESTU�V� (21)
where Y and Q are two recently produced hawks, j is the
leaping power, N is the number of dimensions, a is an N
dimension random vector, and Levy is the Levy flight function
that can be computed as:
ESTU�W� = 0.01 ∗ X∗Y
|Z|[/]
(22)
where µ, ϑ are two separate random numbers created from the
normal distribution, and σ is described as:
^ = _ `��ab�∗cde)fgh ,
`)[ijh ,∗b∗�)
gk[h ,
l
[]
(23)
where γ is a constant set to 1.5, and the location of the hawk is
modified in this phase by:
E�F + 1� = mY if &�Q� < &�E�F��" if &�"� < &�L�F�� (24)
where F is the fitness function and Y and Q are two solutions
derived from (20) and (21).
4) Hard Besiege with Progressive Rapid Dives
This is the last scenario and appears when q<0.5 and
|S|<0.5. In this condition, the following two new solutions are
created [25]:
Q = EO�F� − D|NEO�F� − E4�F�| (25)
" = Q + R ∗ ESTU�V� (26)
where Lm is the average location of the hawks in the current
community. The position of the hawk is subsequently modified
as:
E�F + 1� = mY if &�Y� < &�E�F��Q if &�"� < &�E�F�� (27)
IV. HHO APPLICATION FOR SOLVING THE OBJECTIVE
FUNCTION
In this section, the sequential steps involved in solving the
problem of optimal allocation of EVCSs and DGs using the
HHO algorithm are given.
Step 1) Define the objective function subjected to equality
and inequality constraints of the design variables and read the
line and load data.
Step 2) Initialize the number of hawk searches within the
upper and lower limits of DG sizes in kW, locations of EVCSs
and DGs, HHO parameters and maximum number of iterations
T.
Step 3) Set iteration count k=1.
Step 4) Using forward-backward load flow, determine the
objective function value using (4) for each search hawk.
Step 5) Identify the best hawk (Lr), which has given the best
objective function value.
Step 6) For each hawk Lp, update the values of S, S0, and j
by using (14), (15), and (18), respectively.
Step 7) If S≥1, update the hawk’s position by using (12),
otherwise go to Step 8.
Step 8) If S≥0.5, go to Step 9 otherwise go to Step 10.
Step 9) If q≥0.5, update the hawk’s position by using (16)
otherwise update by using (24) and go to Step 11.
Step 10) If q≥0.5, update the hawk’s position by using (20)
otherwise update by using (25) and go to Step 11.
Step 11) Increment the iteration count k by 1 and if (number
of iterations<maximum number of iterations) then go to Step 4
otherwise go to Step 12.
Step 12) Return the final best solution stored (DG sizes and
EVCSs-DGs locations).
Step 13) Run the load flow and determine the power losses,
average voltage deviation index, and voltage stability index.
V. RESULTS AND OBSERVATIONS
This section summarizes the simulation results of the 33-
bus, 69-bus [28] IEEE test systems using the prescribed
methods for optimizing the location and capacity of DGs along
with the best placement of EVCS. In this paper, EVCSs and
DGs locations are categorically studied with three EVCSs and
Type I, II, III, and IV DGs [29]. In addition to the EV models
(Chevrolet Volt, Chang An Yidong, Tesla Model X and BMW
i3) given in Table I, AC/DC level-2 type charging ports (CPs)
are also considered in the CS design. According to SAE J1772
standard, this type of CPs can suit both Battery Electric
Vehicles (BEVs) and Plug-in Hybrid Electric Vehicles
(PHEVs), and has a maximum power rating of 7kW [30]. Some
design features of CPs are the EVs types which can charge at a
time in a particular CS and their power ratings in kW, the
minimum and maximum number of CPs for different types of
EVs and correspondingly the minimum and maximum power
rating of CS (Table I). For instance, if all the CSs are designed
for only minimum CPs as given in Table I, the power rating of
1 CS is 975kW. Similarly, for maximum number CPs, the
demand may increase to 1675.5kW.
TABLE I. EVCS FEATURES FOR SIMULATION
EV Type
EV power
rating (kW)
No. of CPs CS rating (kW)
Min Max Min Max
Chevrolet Volt 2.2 25 35 55 77
Chang An Yidong 3.75 20 30 75 112.5
Tesla Model X 13 15 25 195 325
BMW i3 44 10 20 440 880
SAE J1772 Standard 7 30 40 210 280
Total power rating of CS (kW) 100 150 975 1674.5
HHO and TLBO algorithms are utilized to decide the
optimal position of EVCSs and the optimal position and size of
Engineering, Technology & Applied Science Research Vol. 10, No. 3, 2020, 5837-5844 5841
www.etasr.com Ponnam & Swarnasri: Multi-Objective Optimal Allocation of Electric Vehicle Charging Stations …
the DGs for minimizing distribution system losses, AVDI, and
enhance VSI of RDS. In this article, the following summaries
are presented to address the effect of type-I, II, III & IV DGs
on the system.
• Scenario 1: RDS without integrating DGs and EVCSs.
• Scenario 2: RDS with minimum and maximum EVCS load,
without integrating DGs.
• Scenario 3: RDS with three optimal locations of EVCSs
with minimum EVCS load without integrating DGs.
• Scenario 4: RDS with three optimal locations of EVCSs
with minimum EVCS load and Type-I, II, III, and IV DG
integration.
• Scenario 5: RDS with three optimal locations of EVCSs
with maximum EVCS load without integrating DGs.
• Scenario 6: RDS with three optimal locations of EVCSs
with maximum EVCS load and Type-I, II, III, and IV DG
integration.
A. Scenario 1
1) 33 Bus System
The cumulative real and reactive demand on the system is
3715kW and 2300kVar. The minimum voltage of 0.9038p.u is
observed at bus number 18, which is outside of the prescribed
limit of ±5% after executing the distribution power flow
algorithm by using the forward-backward method [26].
Moreover, the cumulative active power loss of the system is
210.9897kW, and the cumulative reactive loss of the system is
143.1284kVar. The AVDI of the system is 0.0041 and VSImin is
0.6661. It is recognized that 5.68% of the power is absorbed in
the form of active power loss in the system. The regulation of
voltage in the system is 9.63%.
2) 69 Bus System
The cumulative real and reactive demand on the system is
3801.4kW & 2693.6kVAr. The maximum voltage deviation
occurred at bus 65, and it is the worst among all other busses.
The voltage at the 65
th
bus is 0.9092p.u., the cumulative active
power loss of the system is 224.8807 kW, and the cumulative
reactive loss of the system is 102.1647 kVAr. The AVDI of the
system is 0.0014, and VSImin is 0.6823. It can be noticed that
5.92% of the power is absorbed in the form of active power
loss in the system. The regulation of voltage in the system is
9.08%.
B. Scenario 2
1) 33 Bus System
Incorporation of a total of 3 EVCSs load in the system is
assumed. By imposing a total power demand of 2925kW
(975kW×3) for 3 CSs with the minimum number of CPs at all
EVCSs as given in Table I, the new loading condition of the
test system is determined. According to this, the real power
load is increased from 3715kW to 6640kW. The corresponding
test system performance is given in Table II. Due to the
increased EVCS load, the real losses are increased to
576.1705kW from 210.9897kW, the average voltage deviation
index is increased to 0.0108 from 0.0041, the stability index is
decreased to 0.4984 from 0.6661, and the minimum voltage is
decreased to 0.8408p.u. from 0.9038p.u. Similarly, for the
maximum number of CPs in each EVCS, the load can increase
on the network by 5023.5kW (1674.5×3) for the 3 EVCSs. The
corresponding system performance is given in Table II. The
losses are increased to 1024.3908kW, the voltage deviation
index is raised to 0.0187, the stability index is decreased to
0.3854, and the minimum voltage reaches 0.7888p.u.
2) 69 Bus System
As in the case of the 33-bus test system, 3 EVCSs were
considered with different CPs. By imposing a total power
demand of 2925kW for 3 CSs with minimum CPs, the test
loading condition increased from 3801.4kW to 6726.4kW. Due
to the increased EV load, the real losses increased to
613.4994kW from 224.8807kW, the average voltage deviation
index increased to 0.004 from 0.0014, the stability index
decreased to 0.5114 from 0.6823, and the minimum voltage
was reduced to 0.8462p.u. from 0.9092p.u. Similarly, for the
maximum number of CPs in each EVCS, the load on the
network increased by 5023.5kW (1674.5×3) for the 3 EVCSs.
The corresponding system performance is given in Table II.
The losses increased to 1108.6266kW, the voltage deviation
index increased to 0.0072, the stability index decreased to
0.3949, and the minimum voltage reached 0.7935p.u.
TABLE II. SYSTEM PERFORMANCE WITH MINIMUM AND MAXIMUM
EVCS LOAD WITHOUT PLACING-INTEGRATING DGS.
System Loading condition Ploss (kW)
AVDI
(p.u.)
VSImin
(p.u.)
Vmin
(p.u.)
33
Without EVCS load 210.9897 0.0041 0.6661 0.9038
With minimum EVCS load
(without optimal
placement)
576.1705 0.0108 0.4984 0.8408
With maximum EVCS load
(without optimal
placement)
1024.3908 0.0187 0.3854 0.7888
69
Without EVCS load 224.8807 0.0014 0.6823 0.9092
With minimum EVCS load
(without optimal
placement)
613.4994 0.004 0.5114 0.8462
With maximum EVCS load
(without optimal
placement)
1108.6266 0.0072 0.3949 0.7935
C. Scenario 3
It is assumed that 3 EVCSs are integrated at optimal
locations in the test systems with a minimum number of CPs at
best locations (bus-2, 19, and 25), see Table III. The real losses
are decreased to 295.6474kW from 576.1705kW, the average
voltage deviation index decreased to 0.0047 from 0.0108, the
voltage stability index improved to 0.6499 from 0.4984, and
the minimum voltage increased to 0.8982p.u from 0.8408p.u
for the 33 bus system. For the 69 bus system, the best locations
are bus-2, 28, and 47. The losses decreased to 225.2186kW
from 613.4994kW, AVDI decreased to 0.0014 from 0.004, and
VSI increased to 0.6822 from 0.51114. Also, the minimum
voltage increased to 0.9092p.u from 0.8462p.u.
Engineering, Technology & Applied Science Research Vol. 10, No. 3, 2020, 5837-5844 5842
www.etasr.com Ponnam & Swarnasri: Multi-Objective Optimal Allocation of Electric Vehicle Charging Stations …
TABLE III. SYSTEM PERFORMANCE WITH EVCS AT OPTIMAL
LOCATIONS WITHOUT INTEGRATING DGS
System
Algo
rithm
EV
locations
Ploss
(kW)
AVDI
(p.u.)
VSImin
(p.u.)
Vmin
(p.u)
33
HHO 2,19,25 295.6474 0.0047 0.6499 0.8982
TLBO 2,19,25 295.6474 0.0047 0.6499 0.8982
69
HHO 2,28,47 225.2186 0.0014 0.6822 0.9092
TLBO 2,28,47 225.2186 0.0014 0.6822 0.9092
D. Scenario 4
The line diagrams of the 33 and 69 bus systems with
integration of EVCS and DGs are shown in Figures 1-2.
Placing different types of DGs results to the reduction of power
loss and to the improvement of the voltage profile as shown in
Tables IV-V. Total power losses are reduced by 68.07%,
25.85%,89.05%, and 44.79% for Type-I, II, III, and IV DGs
placed respectively for HHO and 68.07%, 25.85%,89.02%, and
45.16% for TLBO. Besides, with multiple DGs placed, the
voltage profile improved significantly. The minimum voltage
was 0.9685, 0.9267, 0.9845, and 0.9412p.u. for HHO and
0.9685, 0.9268, 0.9845, and 0.9409p.u. for TLBO for Type-I,
II, III, and IV DGs respectively.
Fig. 1. Single line diagram of the 33 bus system with EVCS and DG
integration
Fig. 2. Single line diagram of the 69 bus system with EVCS and DG
integration
TABLE IV. SYSTEM PERFORMANCE OF THE 33 BUS SYSTEM WITH EVCS AND DG INTEGRATION AT OPTIMAL LOCATIONS
Algorithm /
Parameter
EV and DG locations
DG size
(kVA)
Total DG size
(kVA)
Ploss
(kW)
% Reduction
in Ploss
AVDI
(p.u.)
VSImin
(p.u.)
Vmin
(p.u.)
Type-I DG
HHO 2,19,25 and 13,24,30 837.01,1500 and 1137 3474.01 94.3844 68.07 0.0047 0.8799 0.9685
TLBO 2,19,25 and 13,24,30 834.61,1500 and 1139.39 3474 94.3847 68.07 0.0047 0.8799 0.9685
Type-II DG
HHO 2,19,25 and 13,24,30 382.84,598.3 and 1035 2016.14 219.2107 25.85 0.0047 0.7378 0.9267
TLBO 2,19,25 and 13,24,30 388.77,584.61 and 1039.01 2021.39 219.2106 25.85 0.0047 0.7378 0.9268
Type-III DG
HHO 2,19,25 and 13,24,30
787.76+j436.64,
1500+j427.41 and 1155.62+j989
3443.38+
j1853.05
32.3824 89.05 0.0047 0.9392 0.9845
TLBO 2,19,25 and 13,24,30
852.64+j362.65,
1500+j579.53 and
1090.52+j1013.99
3443.16+
j1956.17
32.4586 89.02 0.0047 0.9391 0.9845
Type-IV DG
HHO 2,19,25 and 13,24,30
663.07-j217.94,
1400-j460.16 and 717.08-j235.69
2780.15-
j913.79
163.2147 44.79 0.0047 0.7846 0.9412
TLBO 2/19/25 and 13/24/30
645.41-j212.13,
1500-j493.02 and 727.71-j239.18
2873.12-
j944.33
162.1358 45.16 0.0047 0.7837 0.9409
TABLE V. SYSTEM PERFORMANCE OF THE 69 BUS SYSTEM WITH EVCS AND DG INTEGRATION AT OPTIMAL LOCATIONS
Algorithm/
Parameter
EV and DG
locations
DG size
(kVA)
Total DG size
(kVA)
Ploss
(kW)
% Reduction
in Ploss
AVDI
(p.u.)
VSImin
(p.u.)
Vmin
(p.u.)
Type-I DG
HHO 2,28,47 and 11,17,61 516.98, 387.08 and 1716.7 2620.76 69.6232 69.09 0.0014 0.9185 0.9789
TLBO 2,28,47 and 11,17,61 523.68,382.42 and 1715.97 2622.07 69.6231 69.09 0.0014 0.918 0.9788
Type-II DG
HHO 2,28,47 and 11,17,61 408.34,233.12 and 1231.97 1873.43 145.1159 35.57 0.0014 0.7526 0.9314
TLBO 2,28,47 and 11,17,61 413.19,230.61 and 1232.42 1876.22 145.1166 35.57 0.0014 0.7526 0.9314
Type-III DG
HHO 2,28,47 and 11,17,61
388.15+j438.23,
440.43+j179.53 and 1692.93+j1214.89
2521.51+
j1832.65
4.7502 97.89 0.0014 0.9773 0.9942
TLBO 2,28,47 and 11,17,61
622.64+j277.77,
300+j300.08 and 1629.48+j1275.96
2552.12+
j1853.81
4.7817 97.88 0.0014 0.9773 0.9943
Type-IV DG
HHO 2,28,47 and 11,17,61
377.71-j124.12,
294.12-j98.58 and 1255.9-j412.68
1927.73-
j633.4
133.0426 40.93 0.0014 0.8283 0.954
TLBO 2,28,47 and 11,17,61
400.03-j131.45,
276.05-j90.71 and 1255.55-j412.57
1931.63-
j634.73
133.091 40.91 0.0014 0.8283 0.9539
Engineering, Technology & Applied Science Research Vol. 10, No. 3, 2020, 5837-5844 5843
www.etasr.com Ponnam & Swarnasri: Multi-Objective Optimal Allocation of Electric Vehicle Charging Stations …
E. Scenario 5
As in Scenario 3, the integration of 3 EVCSs at optimal
locations in the test systems with a maximum number of CPs at
best locations (buses 2, 19, and 25) is assumed (Table VI). The
real losses are decreased to 390.6266kW from 1024.3908kW,
the average voltage deviation index is decreased to 0.0053 from
0.0187, the voltage stability index is increased to 0.6381 from
0.3854 and the minimum voltage at bus-18 is increased to
0.8941p.u from 0.7888p.u. for the 33 bus system. For the 69
bus system, the best locations are buses 2, 28, and 47. The
losses are decreased to 225.5766kW from 1108.6266, AVDI is
decreased to 0.0014 from 0.0072 and VSI increases to 0.6821
from 0.3949. Also, the minimum voltage at bus-65 increases to
0.9092p.u from 0.7935p.u.
F. Scenario 6
Placing different types of DGs by HHO and TLBO
technique results in the reduction of power losses and better
voltage profile, (Tables VII-VIII). The total power losses are
reduced by 64.81%, 20.36%, 81.05%, and 45.93% for type-I,
II, III, and IV DGs placed respectively for HHO and 64.81%,
20.36%, 81.05%, and 45.93% for TLBO. Besides, with
multiple DGs placed, the voltage profile has improved
significantly. The minimum identified voltage was 0.9633,
0.923, 0.9706, and 0.9395p.u. for HHO and 0.9633, 0.923,
0.9706, and 0.9395p.u. for TLBO for type-I, II, III, and IV DGs
respectively.
TABLE VI. SYSTEM PERFORMANCE WITH EVCS AND DG
INTEGRATION AT OPTIMAL LOCATIONS
System Algorithm
EV
locations
Ploss
(kW)
AVDI
(p.u.)
VSImin
(p.u.)
Vmin
(p.u.)
33
HHO 2,19,25 390.6266 0.0053 0.6381 0.8941
TLBO 2,19,25 390.6266 0.0053 0.6381 0.8941
69
HHO 2,28,47 225.5766 0.0014 0.6821 0.9092
TLBO 2,28,47 225.5766 0.0014 0.6821 0.9092
TABLE VII. SYSTEM PERFORMANCE OF THE 33 BUS SYSTEM WITH EVCS AND DG INTEGRATION AT OPTIMAL LOCATIONS
Algorithm/
Parameter
EV and DG
locations
DG Size
(kVA)
Total
DG size
(kVA)
Ploss
(kW)
% Reduction
in Ploss
AVDI
(p.u.)
VSImin
(p.u.)
Vmin
(p.u)
Type-I DG
HHO 2,19,25 and 13,24,30 880.91,1500 and 1227.81 3608.72 137.4507 64.81 0.0053 0.8606 0.9633
TLBO 2,19,25 and 13,24,30 881.1, 1500 and1229.58 3610.68 137.4506 64.81 0.0053 0.8606 0.9633
Type-II DG
HHO 2,19,25 and 13,24,30 389.72,638.12 and 1040.4 2068.24 311.0932 20.36 0.0053 0.7258 0.923
TLBO 2,19,25 and 13,24,30 389.82,638.11 and 1041.18 2069.11 311.0911 20.36 0.0053 0.7258 0.923
Type-III DG
HHO 2,19,25 and 13,24,30
878.82+j381.86,
1500+j520.82 and 1204.88+j1003.6
3583.7+
j1906.28
74.0126 81.05 0.0053 0.8872 0.9706
TLBO 2,19,25 and 13,24,30
856.93+j378.37,
1500+j525.62 and 1230.96+j1004.27
3587.89+
j1908.26
74.0212 81.05 0.0053 0.8872 0.9706
Type-IV DG
HHO 2,19,25 and 13,24,30
682.16-j224.21,
1500-j493.03 and 804.86-j264.54
2987.02-
j981.78
211.221 45.93 0.0053 0.7792 0.9395
TLBO 2,19,25 and 13,24,30
682.44-j224.31,
1500-j493.03 and 805.87-j264.88
2988.31-
j982.22
211.2217 45.93 0.0053 0.7791 0.9395
TABLE VIII. SYSTEM PERFORMANCE OF THE 69 BUS SYSTEM WITH EVCS AND DG INTEGRATION AT OPTIMAL LOCATIONS
Algorithm/
Parameter
EV and DG
locations
DG Size
(kW)
Total DG
size (kW)
Ploss
(kW)
% Reduction
in Ploss
AVDI
(p.u.)
VSImin
(p.u.)
Vmin
(p.u.)
Type-I DG
HHO 2,28,47 and 11,17,61 490.3,396.56 and 1724.6 2611.46 69.876 69.02 0.0014 0.9186 0.979
TLBO 2,28,47 and 11,17,61 529.49, 380.46 and 1719.77 2629.72 69.8763 69.02 0.0014 0.9185 0.979
Type-II DG
HHO 2,28,47 and 11,17,61 412.84,230.23 and 1231.34 1874.41 145.7912 35.37 0.0014 0.7525 0.9314
TLBO 2,28,47 and 11,17,61 413.21,230.62 and 1232.41 1876.24 145.797 35.37 0.0014 0.7525 0.9314
Type-III DG
HHO 2,28,47 and 11,17,61
416.13+j336.94,
416.13+j249.35 and 1681.09+j1205.83
2513.35+
j1792.12
4.7654 97.89 0.0014 0.977 0.9942
TLBO 2,28,47 and 11,17,61
438.25+j336.86,
412.29+j262.19 and 1683.56+j1173.63
2534.1+
j1772.68
4.7698 97.89 0.0014 0.977 0.9942
Type-IV DG
HHO 2,28,47 and 11,17,61
319.64-j105.03,
308.58-j101.4 and 1262.96-j415
1891.18-
j621.43
133.6038 40.77 0.0014 0.8282 0.9539
TLBO 2,28,47 and 11,17,61
402.82-j132.37,
276.06-j90.71 and 1256.14-j412.77
1935.02-
j635.85
133.6074 40.77 0.0014 0.8282 0.954
VI. CONCLUSIONS
Massive EV deployment has adverse impacts on the power
grid. A significant amount of EVCSs connected to the grid
increases the system power losses and significant voltage
variation at remote buses from sources. In this paper, a novel
approach for the simultaneous allocation of DGs and EVCSs to
Engineering, Technology & Applied Science Research Vol. 10, No. 3, 2020, 5837-5844 5844
www.etasr.com Ponnam & Swarnasri: Multi-Objective Optimal Allocation of Electric Vehicle Charging Stations …
increase bus voltage profile and decrease power losses in the
distribution network. In addition to the AC/DC Level-2 EV-
CSs suitable for BEVs and PHEVs, different EV models
(Chevrolet Volt, Chang An Yidong, Tesla Model X and BMW
i3) are taken into account while designing the EVCS with
multiple CPs and all four types of DGs. The multi-objective
function is formulated and optimized using HHO and TLBO
algorithms. The simulation results presented on standard IEEE
33-bus and 69-bus test systems have highlighted the technical
benefits which can be achieved through the optimal allocation
of EVCSs and DGs even with increased EV loading conditions.
The main findings of the simulation are summarized as
follows:
• HHO algorithm has superior performance when compared
to TLBO in assessing optimal size and location of EVCSs
and DGs.
• The optimum placement and sizing of the three type-III
DGs led to significant reduction in power losses and to
voltage enhancement.
REFERENCES
[1] S. Wan, T. Zhu, Y. G. Luo, S. Zhang, “Large scale EVs charging
scheduling ensuring secure and efficient operation of traffic and
distribution”, World Electric Vehicle Journal, Vol. 7, No. 4, pp. 605-
612, 2015
[2] K. Schneider, C. Gerkensmeyer, M. K. Meyer, R. Fletcher, “Impact
assessment of plug-in hybrid vehicles on pacific northwest distribution
systems”, IEEE Power and Energy Society General Meeting-Conversion
and Delivery of Electrical Energy in the 21st Century, Pittsburgh, USA,
July 20-24, 2008
[3] A. Shukla, K. Verma, R. Kumar, “Voltage-dependent modelling of fast
charging electric vehicle load considering battery characteristics”, IET
Electrical Systems in Transportation, Vol. 8, No. 4, pp. 221-230, 2018
[4] M. E. Amoli, K. Choma, J. Stefani, “Rapid-charge electric-vehicle
stations”, IEEE Transactions on Power Delivery, Vol. 25, No. 3, pp.
1883-1887, 2010
[5] A. Hajimiragha, C. A. Canizares, M. W. Fowler, A. Elkamel, “Optimal
transition to plug-in hybrid electric vehicles in Ontario, Canada,
considering the electricity-grid limitations”, IEEE Transactions on
Industrial Electronics, Vol. 57, No. 2, pp. 690-701, 2010
[6] Y. A. Alhazmi, H. A. Mostafa, M. M. A. Salama, “Optimal allocation
for electric vehicle charging stations using trip success ratio”,
International Journal of Electrical Power & Energy Systems, Vol. 91, pp.
101-116, 2017
[7] A. Wanitschke, O. Arnhold, “Multi-objective optimization of an
autobahn BEV charging station supplied by renewable energy”, World
Electric Vehicle Journal, Vol. 8, No. 4, pp. 911-922, 2016
[8] G. H. Fox, “Electric vehicle charging stations: Are we prepared?”, IEEE
Industry Applications Magazine, Vol. 19, No. 4, pp. 32-38, 2013
[9] D. Garas, G. O. Collantes, M. A. Nicholas, City of vancouver EV
infrastructure strategy report, Working Paper–UCD-ITSWP-16-04,
University of California at Davis, 2016
[10] P. S. Barzani, A. R. Ghahnavieh, H. K. Karegar, “Optimal fast charging
station placing and sizing”, Applied Energy, Vol. 125, pp. 289-299,
2014
[11] Z. Liu, F. Wen, G. Ledwich, “Optimal planning of electric-vehicle
charging stations in distribution systems”, IEEE Transactions on Power
Delivery, Vol. 28, No. 1, pp. 102-110, 2013
[12] M. Miralinaghi, B. B. Keskin, Y. Lou, A. M. Roshandeh, “Capacitated
refueling station location problem with traffic deviations over multiple
time periods”, Networks and Spatial Economics, Vol. 17, pp. 129–151,
2017
[13] J. Y. Yong, V. K. Ramachandaramurthy, K. M. Tan, N. Mithulananthan,
“A review on the state-of-the-art technologies of electric vehicle, its
impacts and prospects”, Renewable and Sustainable Energy Reviews,
Vol. 49, pp. 365-385, 2015
[14] H. Ma, Z. Yang, P. You, 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
[15] Z. Yang, K. Li, Q. Niu, Y. Xue, A. Foley, “A self-learning TLBO based
dynamic economic/environmental dispatch considering multiple plug-in
electric vehicle loads”, Journal of Modern Power Systems and Clean
Energy, Vol. 2, pp. 298–307, 2014
[16] R. Abbassi, “SOGI-based flexible grid connection of PV power three
phase converters under non-ideal grid conditions”, Engineering
Technology & Applied Science Research, Vol. 10, No. 1, pp. 5195-
5200, 2020
[17] S. Deb, K. Tammi, K. Kalita, P. Mahanta, “Impact of electric vehicle
charging station load on distribution network”, Energies, Vol. 11, No. 1,
Article ID 178, 2018
[18] G. Battapothula, C. Yammani, S. Maheswarapu, “Multi-objective
simultaneous optimal planning of electrical vehicle fast charging stations
and DGs in distribution system”, Journal of Modern Power Systems and
Clean Energy, Vol. 7, pp. 923–934, 2019
[19] S. K. Injeti, 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 bioinspired optimization algorithms”, Protection and
Control of Modern Power Systems, Vol. 5, Article ID 3, 2020
[20] G. Battapothula, C. Yammani, S. Maheswarapu, “Multi-objective
optimal planning of FCSs and DGs in distribution system with future EV
load enhancement”, IET Electrical Systems in Transportation, Vol. 9,
No. 3, pp. 128-139, 2019
[21] M. Dixit, R. Roy, “PSO-CFA based optimal placement of EVs in radial
distribution network for loss minimization”, IEEE International
Conference on Electrical, Computer and Communication Technologies,
Coimbatore, India, March 5-7, 2015
[22] J. J. Jamian, M. W. Mustafa, H. Mokhlis, M. A. Baharudin, “Simulation
study on optimal placement and sizing of battery switching station units
using artificial bee colony algorithm”, International Journal of Electrical
Power & Energy Systems, Vol. 55, pp. 592-601, 2014
[23] M. S. K. Reddy, K. Selvajyothi, “Optimal placement of electric vehicle
charging stations in radial distribution system along with
reconfiguration”, 1st International Conference on Energy, Systems and
Information Processing, Chennai, India, July 4-6, 2019
[24] R. V. Rao, “Review of applications of TLBO algorithm and a tutorial for
beginners to solve the unconstrained and constrained optimization
problems”, Decision Science Letters, Vol. 5, No. 1, pp. 1-30, 2016
[25] A. A. Heidari, S. Mirjalili, H. Faris, I. Aljarah, M. Mafarja, H. Chen,
“Harris hawks optimization: Algorithm and applications”, Future
Generation Computer Systems, Vol. 97, pp. 849-872, 2019
[26] P. V. K. Babu, K. Swarnasri, P. Vijetha, “A three phase unbalanced
power flow method for secondary distribution system”, Advances in
Modelling and Analysis B, Vol. 61, No. 3, pp. 139-144, 2018
[27] M. S. S. Danish, T. Senjyu, S. M. S. Danish, N. R. Sabory,Narayanan K,
P. Mandal, “A recap of voltage stability indices in the past three
decades”, Energies, Vol. 12, No. 8, Article ID 1544, 2019
[28] M. E. Baran, F. F. Wu, “Optimal capacitor placement on radial
distribution systems”, IEEE Transactions on Power Delivery, Vol. 4, No.
1, pp. 725-734, 1989
[29] S. Kansal, V. Kumar, B. Tyagi, “Optimal placement of different type of
DG sources in distribution networks”, International Journal of Electrical
Power & Energy Systems, Vol. 53, No. 1, pp. 752–760, 2013
[30] F. Mwasilu, J. J. Justo, E. K. Kim, T. D. Do, J. W. Jung, “Electric
vehicles and smart grid interaction: A review on vehicle to grid and
renewable energy sources integration”, Renewable and Sustainable
Energy Reviews, Vol. 34, pp. 501–516, 2014