Journal of Energy Storage 50 (2022) 104655

Contents lists available at ScienceDirect

Journal of Energy Storage

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

Research papers

A novel consumer-friendly electric vehicle charging scheme with vehicle to

grid provision supported by genetic algorithm based optimization
Syed Abdullah-Al-Nahid, Tafsir Ahmed Khan, Md. Abu Taseen, Taskin Jamal, Tareq Aziz *
Department of Electrical and Electronic Engineering, Ahsanullah University of Science & Technology, Dhaka, Bangladesh

A R T I C L E I N F O A B S T R A C T

Keywords: Due to escalating demand for electric vehicles (EVs) in the worldwide transportation sector, the charging fa­
Electric vehicle cilities supported by an efficient charging scheme across the power network have a significant impact on the
Residential consumer operation of the power system. The unsupervised and decentralized charging of the EVs have several adverse
Optimization
effects on various operational features of the power system. On the other hand, the centralized charging system
Genetic algorithm
with proper charging schemes offers less complexity in power system operation to avoid unnecessary power
Vehicle to grid
system network stress. This article proposes a valley-filling technique-based EV charging scheme for residential
consumers facilitated by a centralized charging system. The charging scheme comprises a genetic algorithm-
based optimization for EV acceptance to utilize the available energy of the supply network in the best way.
The charging scheme also includes vehicle to grid (V2G) facilitation and the reallocation technique for shifting
EVs to resolve any overloading of charging slots. Simulation results of Genetic Algorithm (GA) based optimi­
zation show the attainment of near-optimum EV data set with less than 50 iteration numbers and usage of more
than 99.5% of available kWh for charging. Other result segments indicate the efficient use of available supply
energy by incorporating EVs in the valley time slab of the demand curve. Also, the findings point out that the
charging scheme successfully encounters unwanted network stress issues during EV integration by V2G provision
and reallocation technique, prioritizing consumer satisfaction at the overloaded charging time slots. The test case
simulation results indicate that the ‘average to peak’ demand ratio in the demand curve is increased to 90% from
68% by incorporating the proposed charging scheme for EV charging. Moreover, the outcomes indicate the
novelty of the proposed methodology in allocating charging slots considering complex driving parameters like
network stress and customer prioritization.

for the development of Smart Grid [3]. They approve various policies
such as subsidies, duty exemption, tax reduction, and more [4]. Auto­
1. Introduction
mobile manufacturing companies such as Nissan [5], Ford Motor [6],
BMW [7] etc. are expanding their ability to get a commanding position
This section discusses the background and motivation of the work,
in this exponentially growing market of EVs. Because of these new
the relevant literature survey and the consequent gap analysis. The core
policies and the advancements in technology, the electric vehicle in­
contributions and novelty of the work are also discussed elaborately in
dustry is going to experience an overwhelmed situation. According to
this section.
the projection of IHS Markit's analysts, around 41% of total vehicles on
the road will run by electricity in 2050, provided that electric vehicle
1.1. Background and motivation
sales will progressively increase to 60% in the next 30 years [8]. If EV
sales jumped up to 100% in half the time, nearly all vehicles on the road
One of the many steps taken by several countries to achieve net zero-
would be electric by 2050 [8]. Roughly 28% of total energy in the world
emission is to revolutionize the transportation sector by reducing the
is currently being utilized for transportation [9]. As a large portion of
emission caused by vehicles [1]. It is observed that most future vehicles
market shares in transportation sectors will be taken by electric vehicles,
will be run by clean and pollution-free electricity as electric vehicles
developing wide-scale and effective charging schemes is necessary.
(EVs) are becoming more popular in every instance [2]. The adminis­
With the growth of electric vehicles, a complementary challenge will
trations in many countries are encouraging their citizens to shift to EVs

* Corresponding author.
E-mail address: [email protected] (T. Aziz).

https://doi.org/10.1016/j.est.2022.104655
Received 28 August 2021; Received in revised form 5 April 2022; Accepted 10 April 2022
2352-152X/© 2022 Elsevier Ltd. All rights reserved.
S. Abdullah-Al-Nahid et al. Journal of Energy Storage 50 (2022) 104655

Nomenclature tkWhreq Total required energy (in kWh) by the EVs for the day
tkWhsys Total available energy (in kWh) from the supply for the day
EV Electric vehicle d Difference in energy between supply energy and required
V2G Vehicle to grid energy for EV charging
G2V Grid to vehicle tkWhp1 Total energy (in kWh) by the EVs in first parent
CO2 Carbon di oxide tkWhp2 Total energy (in kWh) by the EVs in second parent
GA Genetic algorithm p1(k) Parent-1 with k number of genes
SOC State of charge p2(k) Parent-2 with k number of genes
kWhEach slot
(before EV integration) Available energy in each charging slot (unit in k Number of genes in each parent
kWh) before the integration of EV tkWhreq1 Total required energy (in kWh) by the parent-1
kWhEach slot
(after EV integration) Available energy in each charging slot (unit in tkWhreq2 Total required energy (in kWh) by the parent-2
kWh) after the integration of EV data1 First set of EV data
Each slot
kWh(required by EVs) Required energy in each charging slot (unit in data2 Second set of EV data
kWh) for the EVs fregen. Parent regeneration fitness function
kWhDisposable V2G Amount of energy (unit in kWh) supplied by EVs in fterm. Program termination fitness function
V2G mode Day_Alloc.n. group Availability of EVs for charging in a given time
SOCpresent Current SOC of EVs during V2G mode period
kWhratting of EV Ratting of EVs in kWh Dayint.char. Interval of repeated charging
kWhavailable
(after V2G)
in each slot
Available energy in each charging slot (unit in tperiod Time period
kWh) after the V2G mode of the EVs SOCthr Lower threshold limit of SOC of EVs
kWheach slot
(supply) Calculated energy in each charging slot (unit in kWh) of EVmileage Millage of EVs
the supply system Disavg. tr Average distance travelled per-day
kWhoverloaded slot
(Disposable V2G) Amount of energy (in kWh) supplied by EVs in n. group Cluster of EVs in group
V2G mode at the overloaded charging slots supply_kWhmax. mon Maximum supply in kWh per month
Pnsurplus Surplus power at ‘n’ th slot kWhavg. mon(per house) Average kWh consumption per house per
Pndemand Demand without EV integration of the ‘n’ th slot month
Pmax Maximum power capacity of the distribution line EVnum. thr Considered EV number percentage in the area
tna. slot Availability of ‘n’ th slot for EV charging kWhEV. av. charg. Available kWh for EV charging in a day
thr Risk threshold (20%) of maximum demand kWhEV. ratting Average kWh ratting of EVs in the residential area

also rise [10]. Accommodating these EVs in the distribution network for the load shifting, utilizes the excess power during the valley periods.
charging will have serious impacts. The EV charging infrastructures can These hours have low demand and the stress on the distribution network
be of two fundamental types: centralized and decentralized. In a is consequently lower. However, charging a large amount of EVs without
centralized charging method, a master controller connects with each any supervision may erupt a new peak in valley hours and cause network
electric vehicle and exchanges essential information about the timing, overloading [20]. Due to this reason, centralized coordination in EV
state of charge (SOC), tariff, etc. [11]. After gathering information, the charging is essential [21]. The central controller sets the EVs' charging
controller finds a solution to certain optimization problems and regu­ pattern and prevents any new peak in the valley. If any overloading
lates the charging patterns of individual EVs [12]. Although, the situation arises while charging, the controller shifts or removes that EV
centralized charging infrastructure, without a proper strategy, can be from charging. This might be undesirable for some EV owners as their
much more complex. It may suffer from dimensionality issues while convenience is neglected.
solving optimization problems [13]. Another issue with the centralized To guarantee the most possible charging service to EVs in valley
approach is that it generally ignores or gives less importance to EV users. hours, an additional energy source can be integrated. This involves a
On the other hand, decentralized methods are simpler to establish. The substantial amount of investment and infrastructure modification [22].
EV owners control their own charge patterns according to the electricity Another viable, cost-effective alternative to this strategy can be sharing
pricing structure or their time slot preferences [11]. This convenience to charges from vehicles that are sitting idle. This method is often referred
users can be a matter of grave problem for the power distribution to as Vehicle to Grid, Vehicle to Vehicle or Vehicle to Home mode [23].
network. The uncoordinated decentralized EV charging may introduce The Vehicle to Grid (V2G) is becoming popular since it can be utilized
unwanted peaks in the load curve of the distribution system [14]. For for many ancillary services at different stages of the grid [24]. V2G mode
instance, transformers and distribution lines may face severe over­ of operation is employed to bring back some power from the vehicles to
loading and severe voltage sagif EVs are charged at peak hours [15]. the distribution network when the system requires extra power during
This potentially can widen the span between peak and valley levels of a peak hours [25]. By introducing this mode of operation, the EVs with
load curve [16]. A nearly flat load curve is desirable to make the elec­ urgency can be charged and possible overloading can be avoided.
tricity infrastructure more efficient [17]. A common approach to deal Moreover, V2G integration for solving overloading in valley hours does
with this problem is adjusting the electricity price, usually by a dual- not require a huge change in current distribution infrastructures.
tariff structure [18]. This dual pricing structure, to some extent, moti­
vates the consumers to charge their EVs in valley periods. However, this 1.2. Literature survey
does not guarantee the protection from reoccurrence of unwanted peaks.
In uncoordinated charging, there is a possibility of getting a new peak Electric vehicles have turned into the centre of research recently.
even in valley hours. It may happen if a good number of EVs are tempted Numerous studies have been performed to enhance the user experience
to use charging facilities simultaneously when the tariff is comparatively and valley filling performance during the charging process of residential
low [19]. EVs. In the field of EV charge scheduling, a remarkably large amount of
For avoiding the peaks getting bigger, most charging schemes follow literature can be found. Some authors have tried to utilize the valley
the valley filling strategy. The valley filling strategy, often referred to as hours, while others have tried to alleviate the difference between peak

2
S. Abdullah-Al-Nahid et al. Journal of Energy Storage 50 (2022) 104655

Table 1
Comparative summary of this study and other similar studies.
Key parameters Defining factors status addressed or not

This paper Ref. [11] Ref. [12] Ref. [26] Ref. [29] Ref. [37] Ref. [38] Ref. [39] Ref. [40] Ref. [41]

Valley filling strategy Yes Yes Yes Yes Yes Yes Yes No No Yes
Optimization Yes No Yes Yes No Yes No Yes Yes Yes
Network stress Yes No No No No No Yes No Yes No
Customer preference Yes Yes No No No No No No Yes No
V2G operation Yes No No No Yes No No Yes Yes No
V2G day ahead scheduling Yes No No No No No No No No No
V2G discharging threshold Yes No No No No No No Yes No No
Emergency threshold addressing Yes Yes No No No No No No No No
Reallocation using tariff Yes Yes No No No No No No No No
Reallocation using time preference Yes Yes No No No No No No No No

and valley of the load curve. In some studies, authors have emphasized consumption by up to 20%. Ref. [31] offers a charging scheme for grid-
the user's convenience to some extent. connected EVs based on V2G behaviour management. This method de­
In terms of valley filling electric vehicle charging, there are some livers optimized peak shaving and rapid power balancing facilities for
notable articles. Ref [12] has offered a novel decentralized valley filling the microgrid. The proposed V2G controller schedules EVs' charging and
charging method for a large number of EVs, where a day ahead time- discharging based on frequency state information to achieve rapid
power-varying scheme for the price is suggested by addressing the power balancing benefits for the grid. Although this study hasn't
least-cost optimization issue. In this method, vehicle owners' charging addressed users' convenience. In addition, some studies have used V2G
behaviours are coordinated indirectly after the price scheme is broad­ for environmental purposes. For example, a V2G-based study in
cast. EV users react to the tariff scheme by independently optimizing Ref. [32] investigates daily charging strategies aiming to curtail CO2
their charge models. The simulated output of this study illustrates that emissions. The results from the study reflect that the CO2 emissions can
valley-filling by coordinated charging reduces the power generation cost be reduced by one-third if the EVs are charged optimally. The emissions
of utility significantly in comparison with the uncoordinated charging can be reduced to 59% if V2G is integrated. Some studies have utilized
method. In this study, distribution network stress or overloading situa­ V2G for analyzing the possibility of incorporating this technology to
tions are not addressed. The authors have claimed that their charging their grid system. For example, the study in Ref. [33] examines the
strategy can achieve a flawless valley-filling result under various sce­ capability of V2G technology for ancillary services in the Moroccan grid
narios at three-fourths of the EV load generation cost. Another unique system. They predicted the growth of the market for electric vehicles and
and efficient valley filling approach for centralized EV charging is pre­ charging infrastructure. A power scenario is also given according to the
sented in [23]. In this study, a massive amount of EVs is considered for prediction. The study has shown that the scheduled EV charging and
simulation. The capacity margin and charging priority of EVs are taken V2G discharging provide control over reactive power injection as well as
into account while investigating. Surplus power is calculated based on balancing the magnitudes of voltage and frequency. Some studies have
capacity margin to estimate the available time slots for EV charging. also worked on battery longevity management alongside developing
This scheduling scheme prevents the charging load demand from going centralized charging and discharging scheduling methods. The study in
above the capacity margin. Although the study does not give enough Ref. [34] has presented a V2G scheduling approach for achieving grid
preference to user convenience. In [26], a decentralized charge sched­ frequency control as well as active anti-aging control for batteries. This
uling strategy is investigated to fill the valley of the load curve. The method protects the installed batteries from being overused, which
strategy utilizes the non-cooperative games theory developed by John guarantees a long battery life.
Nash to minimize individual charging costs. Even though the study has On the subject of the optimization method used in charge scheduling,
effectively filled the valley by load shifting yet failed to address user numerous studies have been carried out. In Ref. [35], a bi-level, two-
convenience. This method has not used any strategy to overcome stage, and multi-objective hybrid robust-stochastic framework is
overloading in any particular charging slot. In [27], a coordinated valley created. The higher-level problem includes two objectives: minimizing
filling scheduling strategy is suggested for charging EVs connected in a total cost and reducing operational emission, whereas the lower-level
microgrid. The strategy takes both the minimum and maximum state of works on maximizing social welfare. In Ref. [36], a structure for day-
charge (SOC) level of the batteries into account. The charging mode is ahead optimal vehicle charging is offered where the optimal exchange
chosen depending on the urgency indicator. An optimization model of active and reactive power among EVs, parking lots, and the grid is
controls the charging patterns to minimize the difference between the taken into account. Additionally, the optimization framework works on
peak and valley load. the benefit of EV users and system operators simultaneously. To make
Regarding vehicle to grid operation, there have been some inspiring the framework more realistic, a route mapping is executed for partici­
developments. Ref. [28] offers a fresh day-ahead peak shaving and pating EVs. A metaheuristic optimization-based charging strategy for
valley filling scheduling strategy for parking lot microgrids. The study big numbers of EVs is presented in Ref. [37]. A genetic algorithm (GA) is
has considered a realistic electrical demand scenario for a parking lot used to obtain a flat load curve by compressing the level of power
that has its own generation facilities. A V2G technology-based peak fluctuation. The algorithm optimally varies the charging power to fill the
shaving and valley filling method is elaborately discussed in [29]. The valley.
authors have analyzed the effects of the total number of connected ve­
hicles along with the mean of the expected load curve. The proposed 1.3. Contributions
scheduling strategy utilized V2G technology only to trim the peak of the
load curve. Another peak shaving and valley filling method is suggested Table 1 shows a summary of similar publications and this research.
in [30], illustrating a charging-enabled EV parking facility for non- Despite their significant contributions, the examined studies and a few
residential housing, where a mathematical model for peak shaving others contain a few shortcomings. These limitations can be summarized
and valley filling the load profile of a university campus was simulated as follows:
by planning the charging and discharging phases. By discharging elec­
tric vehicles, the proposed method was able to shrink the peak power

3
S. Abdullah-Al-Nahid et al. Journal of Energy Storage 50 (2022) 104655

• The studies in [12,26,27] fill the valley by charging EVs during N=1 N=2 N=3 N=4 N=24
valley hours. To the best of the authors' knowledge, none of these
studies has addressed distribution network stress, overloading situ­
ations, and users' convenience. 12 AM 01 AM 02 AM 03 AM 04 AM 11 PM 12 AM

• The study in [38] fills the valley successfully and takes care of dis­ Time period in 24 hours
tribution network stress. But this study does not give enough pref­
erence to users' convenience. Fig. 1. Time slots of a day.
• The studies in [28–31] have used V2G technology solely for peak
shaving and power balancing. To the best of the authors' knowledge, 2. Methodology
none of these studies has utilized V2G technology to resolve network
overloading and satisfy EV users. 2.1. Proposed charging scheme
• The study in [11] has used the valley filling strategy to schedule
charging but has not used any optimization. This study has not used In this study, a day ahead scheduling is considered based on fore­
V2G or other methods to solve an overloading situation while casting demand data. The forecasted demand data is achieved from
charging. historical demand and weather data through the short-term load fore­
• The studies in [33,34] have incorporated V2G technology in the EVs casting method. The length of the day has 24 slots. Each time slot has a
charge-discharge process but have not clearly stated users' duration of 1 h which is shown in Fig. 1. The valley filling approach is
convenience. used to create each identical time slot for EV charging. In this method­
ology, each charging slot is assumed to have constant power. Fig. 2
To address these research gaps, a day ahead consumer-friendly EV depicts the suggested methodology's flow chart. The methodology works
charging scheme is proposed in this paper. The proposed methodology in three sequential steps to complete the charging scheduling for vehicle
meets the complex network stress and user satisfaction issues. The customers.
problem of determining charging slots is finished by the valley filling Step-1: At the beginning, the methodology generates the surplus
technique for using the supply energy efficiently based on forecasted power curve from the day ahead demand curve and the maximum
demand data. A safety threshold of 20% is considered to meet urgent allocated power of the area using equation no. (1). The maximum rating
forthcomings of the power system network in each charging slot. Next, of the feeder transformer is considered as maximum allocated power.
the problem of near-optimum use of supply energy by the EV charging is After that, the algorithm determines the available charging slots using
met by the GA-based EV acceptance optimization in the proposed Eq. (2) with the threshold mark at 20%.
method. This algorithm produces a near-optimum set of EVs from many
inputs for allocating charging slots. It is achieved by eliminating some of Pnsurplus = Pmax − Pndemand (1)
the EVs from the input data set by the GA to use the supply energy near- ( )
optimally. The outcome of this optimization supports avoiding network n
ta.slot = f Pnsurplus , thr (2)
stress partially even before the scheduling task. Furthermore, additional
support to network stress and customer convenience problems are Where,
encountered by the V2G provision and reallocation technique. The V2G Psurplusn: Surplus power at ‘N’ th slot.
provision supports both the user convenience and network stress by Pmax: Maximum power capacity of the distribution line.
providing energy to the charging facility to resolve the overloading Pndemand: Demand without EV integration of the ‘N’ th slot.
while keeping the EVs in the desired slots. The maximum discharge of tna. slot: Availability of ‘N’ th slot for EV charging.
EVs is up to 30% SOC in V2G provision. Moreover, the problem of loss of thr: Risk threshold (20%) of maximum demand.
energy storage or even being incapable of resolving overloading in V2G The threshold value, which is 20%, is considered to avoid strain on
mode is supported by the reallocation technique in the proposed the distribution network. The timeslots having the surplus power larger
scheme. In this technique, some EVs are shifted to resolve the over­ than the threshold value are the charging slots for EVs. The proposed
loading. During the allocation of new time slots to the shifted EVs, this method furthermore calculates the available kWh of each time slot.
technique prioritizes user satisfaction by providing choices near the Then, a block meter tariff rate is planned against each time slot based on
preferred charging slots of the EV customers. Finally, the proposed the availability of the kWh. The tariff is planned such that the vehicles
methodology becomes capable of completing the day-ahead scheduling are encouraged to charge in the time slots having lower demand by
task by prioritizing network stress and user satisfaction. Therefore, the setting lower prices. All of this information is sent to individual EV
core novelty and contributions presented in this paper can be abridged customers. Thus, vehicle users select the desired time slot.
as follows: The data from the EV consumers are received in the form of power
rating, kWh rating, and SOC level. The methodology then calculates the
• Development of a day ahead EV charging scheme for a centralized total required kWh for the vehicles for charging. If the summed required
residential charging station. kWh by the vehicles is greater than the total available kWh from the
• The network stress and customer comfort complexity are taken into supply system for charging, the acceptance of EVs is optimized by the
consideration in allocating charging slots using valley filling tech­ Genetic Algorithm (GA). However, if the total required kWh is less than
nique to the EVs. the available kWh of the day, the GA accepts all the EVs for preparing
• Deployment of GA incorporation with V2G provision and realloca­ the charging schedule. The GA identifies which vehicles can be elimi­
tion technique in the methodology to address the user convenience nated from the charging before even scheduling. The allowable vehicles
and network stress near-optimally. are optimized in such a way by the GA that the total surplus kWh from
the supply system can be used at the maximum possible value. After
that, the methodology calculates the required kWh in each slot by the
1.4. Organization of paper vehicles for charging. Then, the requisite kWh data are compared with
the available kWh of each time slot for finding any overloading situa­
The remainder of the article contains Section 2, which describes the tion. If no overloading slots are found, the scheduling process is
methodology of the proposed scheme; Section 3, which presents simu­ completed. For any overloading slots, two separate sets of measures in
lated results along with the necessary analysis; Section 4, which repre­ the algorithm resolve the overloading situation. Thus, the distribution
sents the key concluding remarks with a few suggestions for future work. network has the chance to experience less stretch due to EV charging.

4
S. Abdullah-Al-Nahid et al. Journal of Energy Storage 50 (2022) 104655

Fig. 2. Flow chart of the proposed methodology.

Fig. 3 shows the EV acceptance optimization based on GA. The al­ (ii) Next, the algorithm checks the fitness functions of the parents.
gorithm identifies the eliminated vehicles to use the available kWh of Two fitness functions are used in this algorithm. The first one
the supply system near-optimally. The objective function of this algo­ (fterm.) is the program termination fitness function that is the total
rithm is to categorize a set of EVs to eradicate before the scheduling task. required kWh by the EVs must be less than the available kWh. The
The corresponding variables are the EV kWh ratings, SOC values, and second fitness function (fregen.) in this study, parent's regeneration
the total available kWh of the day. The constraint in this GA is that the fitness, is defined as a range of the difference value of required
total required kWh by the EVs must be less than the available kWh from kWh with available kWh. The lower limit of the range is greater
the system. The proposed algorithm can be described in the following than the difference between required kWh and available kWh.
four stages-. The upper limit is not greater than the difference value plus the
10 kWh threshold value. If any of the parent sets fulfil the fregen.,
(i) At first, the total required kWh for the EVs is calculated by the then stage (iii) is executed. If not, then stage (iv) is executed.
algorithm. Then, the algorithm calculates the difference between (iii) In this sub-step, the algorithm generates two new EV data set (EV-
the total required kWh by the vehicles and the total kWh avail­ data1 and EV-data2) by eliminating the parent sets (p1(k) and
able from the supply for the vehicle charging. The algorithm p2(k)) from the acquired EV data set. After that, the algorithm
creates the population set that is the possible eliminated vehicles. calculates the total required kWh of the two new EV data sets.
The selection is based on the SOC level. The vehicles with two Then, the algorithm checks the termination fitness function for
maximum SOC levels are considered as the population for the both EV-data1 and EV-data2. If any of the two data sets satisfy the
algorithm. The algorithm then generates two-parent sets, p1(k) fitness value, the iteration stops. After that, the corresponding
and p2(k), with ‘k’ number of genes randomly from the population parent's information is stored. Then the algorithm performs the
set. The value of ‘k’ is 3 in this paper. After that, the algorithm greedy selection between these two parents. The position of the
sums the total kWh of each parent and calculates the difference best fit parent set is determined, and the EVs are eliminated
from the system available kWh. before the scheduling task. Thus, the program of this algorithm

5
S. Abdullah-Al-Nahid et al. Journal of Energy Storage 50 (2022) 104655

Fig. 3. Flow chart of the used genetic optimization algorithm.

ends. However, if none of the parents satisfies the fitness func­ randomly from p1(k) and p2(k) is the mutation function in this
tion, the algorithm executes stage (iv). paper. After that, the algorithm again calculates the distance
(iv) In this step, a single-point crossover is performed among the p1(k) between available kWh and the total kWh of the parents. The
and p2(k). After that, the total kWh of the two parents set is algorithm checks the fitness fregen.. If the fitness is satisfied, the
determined. Then, the distance from the available kWh is calcu­ algorithm executes sub-step (iii). If not, the algorithm then
lated for the parents. Next, the algorithm checks the fitness fregen.. randomly updates the zero-valued genes from the EV data. Thus,
If the fitness is satisfied, the algorithm executes sub-step (iii) for the parents p1(k) and p2(k) are updated. After that, the sequential
terminating the program by identifying the eliminated EV posi­ steps from sub-step (ii) are executed again. This iterative process
tion. If the fitness (fregen.) is not satisfied, the algorithm performs continues until near-optimum EV data sets are achieved. Finally,
mutation among p1(k) and p2(k). The ‘deletion’ of a single number the possible accepted vehicles for charging during the day are

6
S. Abdullah-Al-Nahid et al. Journal of Energy Storage 50 (2022) 104655

selected. The processed vehicle data (kWh rating, SOC, and time
slot) are utilized by the methodology of this study for scheduling.

Step-2: In this measure, the overloading slot's available kWh is

increased by the V2G operation from the EVs with higher SOC levels.
The methodology schedules for V2G operation. The time slots infor­
mation is sent to vehicle users. After acquiring the data from vehicle
users for the V2G operation, the algorithm checks for the highest SOC
valued EV. Then the disposable kWh is calculated. The threshold mark
for discharging of EV during V2G operation is considered 30% in this
article. The calculated kWh is added with supply kWh for EV charging at
the overloaded slot. Then the overloading is again checked. If over­
loading is resolved, the scheduling task becomes completed. If not, then
the methodology adds other vehicles in the same sequential process, if
available for V2G, with a higher SOC level at the overloading slots to
resolve. If still overloading continues, the reallocation technique comes
into action for shifting the vehicles as a second measure for resolving to
overload.
Step-3: In this reallocation process, the methodology finds the
possible shifted EVs from the overloading slot based on SOC level and Fig. 4. Day ahead demand curve of a day in 24-hour time slot.
kWh rating. The vehicles with the lowest SOC level get the priority for
charging at the overloaded time slot. If shifting of the vehicle with the
highest SOC does not settle the overloading situation, the method then
shifts the vehicles with the highest kWh. In this case, a vehicle with a
lower kWh rating is given priority for charging at the overloaded slot. In
both cases, the feasible vehicles are reallocated to the new time slot.
These time-slots are grouped based on the tariff plan and time of the day
selected by the possible shifted vehicles. The method generates two
closest individual sets of timeslots based on the tariff plan and time
choice of the day picked by the EVs. The method also determines the
overloading condition of the time slots where the EVs are reallocated
from the overloaded time slots. Thus, the iteration lasts until an
appropriate solution is achieved. Finally, the new time-slots information
is referred to as the shifted EVs for rescheduling. After following all these
sequential steps, the scheduling task for the EV charging station is
completed. The charging station becomes ready to accommodate vehicle
consumers the next day.

2.2. Modeling for EV's availability analysis

The availability of EVs for charging depends on the per-day distance Fig. 5. Surplus power curve of a day in the 24-hour time slot.
travelled and the millage of the EV. The travelled distance per day highly
depends on the two factors of the EV consumer, one is behaviour, and
another is necessity. These two factors make the forecasting task of EVs
availability exceptionally difficult. However, mathematical modeling
comprising average travelled distance, EV mileage, and minimum SOC
level is developed in this paper to analyze the daily availability of EVs in
charging stations.
The availability of EVs for charging in a given time period can be
calculated from the following-
( )
Day Alloc.n.group = f Dayint.char. , tperiod (3)

where, tperiod is the time period; Dayint.char. is the interval of repeated

charging; n. group is cluster of EVs in group.
The Dayint.char. can be calculated as-
(1 − SOCthr ) × EV mileage
Dayint.char. = (4)
Disavg.tr

Now, the number of EV cluster (n. group) can be formulated as-

supply kWhmax.mon
× EV num.thr
n.group =
kWhavg.mon (per house)
(5) Fig. 6. Available time slots for EV charging.
kWhEV.av.charg.
(1− SOCthr )×kWhEV.ratting

7
S. Abdullah-Al-Nahid et al. Journal of Energy Storage 50 (2022) 104655

3. Results and analysis

The simulation results of the proposed methodology are shown from

Fig. 4 to Fig. 17. The number of EVs considered in this article is 24. The
minimum SOC values of the EVs are considered 30% in this paper. Fig. 4
shows the predicted demand curve of a single feeder in the residential
complex of ‘Japan Garden City’ located in Dhaka city for G2V operation
[42]. The maximum demand is also assumed to be the equal rating of the
distribution transformer for the complex. This paper considers a
centralized charging station at an 11 kV voltage level for vehicle
charging. The charging station is comprised of three fast chargers. Each
fast charger is assumed with an equal rating of 50 kW. The reason for
considering three chargers is that the limit supposed of the distribution
transformer is 152 kW for the case residential area.
Fig. 5 shows the surplus power curve that is determined from Eq. (1).
Fig. 6 shows the possible time slots for EV charging of the day. The
timeslots are identified such that the available power corresponding to
these slots is at least more than 20% of the maximum demand. The time-
slots with the value “1” indicate the availability for charging and vice-
Fig. 7. Tariff chart based on the availability of surplus power. versa for those with values “0”. The block meter tariff rate for the
available charging slots is shown in Fig. 7. The unavailable time-slots do

Fig. 8. Outcomes of GA based EV acceptance optimization during 1st run time considering the sample size of (a) 24 number of EVs (b) 25 number of EVs (c) 27
number of EVs.

8
S. Abdullah-Al-Nahid et al. Journal of Energy Storage 50 (2022) 104655

Table 2
Summary of GA based EV acceptance optimization before allocating time slots.
Run No. of EV before No. of EV after Total required kWh Total required kWh after Converged Considered Total
time optimization optimization before optimization optimization iteration number iteration number available kWh

Data set-1
1 24 23 1135 1105 07 100 1111
2 24 23 1135 1105 06
3 24 23 1135 1105 05
4 24 23 1135 1105 08
5 24 23 1135 1105 06

Data set-2
1 25 23 1185 1105 07 100 1111
2 25 23 1185 1105 21
3 25 23 1185 1110 11
4 25 23 1185 1105 12
5 25 23 1185 1110 13

Data set-3
1 27 25 1205 1110 31 100 1111
2 27 25 1205 1110 11
3 27 25 1205 1110 22
4 27 25 1205 1110 25
5 27 25 1205 1110 18

Fig. 9. EVs demand in each preferred slot. Fig. 10. Demand curve after EV integration.

not have any tariff plan. Hence, it is excluded and is represented as a “0” the number of vehicles before the optimization is 25 having a required
tariff in the figure. The tariff plan is such that the time-slots with the kWh of 1185. After optimization, two vehicles are eliminated with a
highest available power have the lower unit price and vice versa for the total required kWh of 1105. In Fig. 8(c), the required kWh by the ve­
time-slots with the least available power. hicles is found at 1110 by eliminating two vehicles from 27 vehicles that
Fig. 8 shows the simulation results for EV acceptance optimization are processed for optimization. In this article, EV information from Fig. 8
during run time-1. The genetic algorithm is used in the optimization (a) is used as data for the simulation of the rest of the methodology.
technique. The fitness function is defined such that the total kWh of the Fig. 8(b) & (c) are the test case for validation of the genetic algorithm
accepted EVs does not exceed the total number of kWh available for EV acceptance optimization.
charging, which is 1111 kWh, supplied from the system. The population The summarized information on the GA-based EV acceptance opti­
selected for the possible elimination of vehicles is based on the SOC mization is listed in Table 2. The algorithm is tested for three different
level. The algorithm identifies the population set based on the two data set for five run times with a total number of iterations is 100 in this
highest values of SOC. The optimization technique eliminates one or paper. The first data set contains 24 EVs with a required 1135 kWh.
more EVs to use the maximum possible kWh from the supply. The figures Similarly, the second and third data set possess 25 and 27 EVs. The first
with ‘0’ SOC & ‘0’ kWh against EV position indicate that these vehicles data set in Table 2 has optimized the required kWh by the EVs is 1105
are eliminated by the optimization algorithm. Three different numbers kWh. The maximum iteration number required to converge near the
of EVs are considered and tested in the optimization technique. Fig. 8(a) optimum value is 8. For the second data set, the algorithm generates two
shows the simulation results for the total number of EVs considered is near-optimum data sets. Each solution has the value of required kWh are
24. The optimization algorithm eliminated one vehicle to accomplish 1110 and 1105. The maximum iteration number required to converge
the fitness value. After optimization, the total required kWh is 1105 less near-optimum value in this data set is 21 among five different run times.
than the total kWh value of the supply for vehicle charging. In Fig. 8(b), Similarly, the maximum iteration number required to reach the near-

9
S. Abdullah-Al-Nahid et al. Journal of Energy Storage 50 (2022) 104655

Fig. 11. Graphical demonstration for locating overloaded slots steps in methodology (a) surplus kWh curve against each charging slots before EV addition (b) surplus
kWh curve against each charging slot after EV addition (c) overloaded slots.

optimum solution in the third data set is 31 among five consecutive run Case-4: SOC based EV shifting does not resolve the overloading.
times. Case-1: In this case, the V2G operation comes into action to resolve
The demands by the EVs against their preferred time slots are plotted the overloading situation. Simulation results for this case are shown in
in Fig. 9. The day ahead demand curve for the residential complex after Fig. 12. In Fig. 12(a), the available kWh curve after EV addition is
EV integration is presented in Fig. 10. In contrast to Fig. 4, Fig. 10 is plotted. Fig. 12(b) shows the timeslots for the V2G operation. Some of
flatter with the increase in maximum demand, minimum demand, and the EVs with higher SOC are adding power at these time slots. Fig. 12(c)
average demand value. shows the available kWh curve after the EV integration and V2G oper­
Fig. 11 shows the simulation results for the identification of the ation at overloaded timeslots. The overloading timeslots before the V2G
overloaded slots. Fig. 11(a) shows the available kWh curve for each operation is displayed in Fig. 12(d). The overloading is resolved after the
charging time slot. Fig. 11(b) shows the available kWh curve after EV V2G operation in this case. Fig. 12(e) indicates that with plotting ‘0’. In
addition. This curve is achieved by subtracting available kWh and addition, two previously overloaded timeslots (from 7 AM to 8 AM &
required kWh by the EVs. The detected overloaded time slots are indi­ from 7 PM to 8 PM) now have sufficient kWh to accommodate more EVs.
cated in Fig. 11(c). In this case, three overloaded time-slots are found. Case-2: The addition V2G operation does not resolve the overloading
The first overloaded time slot is from 7 AM to 8 AM. The second overload situation in this case. Fig. 13(a) shows the available kWh after the
is identified from 10 AM to 11 AM, with the last being found from 7 PM integration of the EVs. Some of the time-slots have negative available
to 8 PM. kWh that confirms the overloading at those time-slots. The proposed
The overloading situation is resolved in this article by V2G operation methodology then schedules for V2G operation shown in Fig. 13(b). The
at the overloaded time-slots and shifting the EVs prioritizing customer available kWh after V2G operation at the overloaded time slot is plotted
preference from the overloaded time-slots. Four different cases are in Fig. 13(c). In this case, the V2G operation does not resolve over­
investigated to establish the effectiveness of the proposed methodology. loading. The indication of the continuation of overloading is shown in
Those cases are-. Fig. 13(e). If the V2G operation does not resolve the overloading (Case-
Case-1: The addition of V2G operation resolves the overloading. 2), the reallocation technique happens to solute overloading. The
Case-2: V2G operation does not resolve the overloading. simulation results for this reallocation method are shown in Fig. 14 to
Case-3: SOC based EV shifting resolves the overloading. Fig. 16. The algorithm highly prioritizes customer satisfaction during

10
S. Abdullah-Al-Nahid et al. Journal of Energy Storage 50 (2022) 104655

Fig. 12. The visual representation of case-1 consideration (a) available kWh after EV integration (b) time-slots for V2G operation (c) available kWh after V2G
operation (d) original overloading situation (e) identified overloading slot after V2G operation.

the reallocation process. The simulation result for finding new time slots the overloaded time slot. In contrast with Fig. 14(a), Fig. 14(b) shows
for the EVs at the overloaded time slots is plotted in Fig. 14. Fig. 14(a) new time slots based on time slot preference. Here, two nearby time slots
shows the likely new time slots for reallocation based on tariff selection. are being discovered contrasted with overloaded timeslots. In the real­
The two closest tariff-based time slots are categorized for the EV users in location process, the lowest SOC EV is prioritized for charging at the

11
S. Abdullah-Al-Nahid et al. Journal of Energy Storage 50 (2022) 104655

Fig. 13. The pictorial presentation of case-2 consideration (a) available kWh after EV integration (b) time-slots for V2G operation (c) available kWh after V2G
operation (d) original overloading situation (e) identified overloading slot after V2G operation.

overloaded time slots. If shifting of EV with the highest SOC does not level among all the considered EVs. Two more considerations, Case-3
resolve the overloading, the algorithm shifts the EV from the overloaded and Case-4, are examined during the reallocation process mentioned
time slot based on the highest rating of required kWh. Fig. 14(c) shows earlier.
the possible shifted EV position based on the highest SOC level. In this Case-3: In this case, the relocation process shifts the EV based on the
case, EV at the 11th position is identified for shifting based on the SOC SOC level. The possible shifted EV position based on SOC from Fig. 14(c)

12
S. Abdullah-Al-Nahid et al. Journal of Energy Storage 50 (2022) 104655

Fig. 14. Graphical presentation of finding possible new time-slots from overloaded slots segment in the methodology (a) possible new time slots based on tariff (b)
possible new time slots based on time slot preference (c) eliminated EV position based on SOC level.

Fig. 15. The reallocation segment's outcomes during case-3 (a) EV kWh data based on consumer time slot choice (b) EV kWh data after reallocation.

13
S. Abdullah-Al-Nahid et al. Journal of Energy Storage 50 (2022) 104655

Fig. 16. The reallocation segment's outcomes during case-4 (a) EV kWh with position data based on customer time slot preference (b) EV kWh data with position
after reallocation.

Fig. 17. Day ahead demand curve after the completion of EV scheduling (a) when case-1 is validated (b) when case-3 is validated.

is 11th. Fig. 15(b) shows that the corresponding EV from these positions demand time slots of the day are appropriately used by fixing minimal
is shifted and resolves the overloading. tariffs. A 20% threshold is considered while estimating the available
Case-4: In case-4, the EVs from the overloaded time slots are resolved energy of the charging slots to address the urgent state partially. Next,
based on shifting either or both of the highest and next highest requisite the genetic algorithm (GA)-based optimization produces a set of
kWh ratings of the EVs. If case-3 does not solve the overloading, the accepted EV data before the scheduling. The algorithm accepts vehicles
algorithm identifies the EVs with the maximum required kWh ratings to allocate scheduling by using the total available kWh of the day for
from the overloaded slot. The algorithm shifts the EVs, shown in Fig. 16 charging. In achieving this, the algorithm identifies which EVs are
(b), and notifies EV customers to select a new time slot preference. eliminated before scheduling for near-optimum use of available kWh.
Fig. 17 shows the day ahead demand curve with the EV integration Vehicles having higher SOC are the considered deciding variables in the
after allocating charging slots to EV users' through the execution of the optimization algorithm. Moreover, the V2G facilitation and reallocation
proposed charging scheme. Fig. 17(a) & (b) is the final outcome in the technique supports network stress issues and prioritizes consumer
demand curve after executing previously discussed case-1 and case-3. It satisfaction. The V2G highlights the allocation of charging slots ac­
is observed that the average to peak demand ratio is enhanced signifi­ cording to customer preference with support to avoid overloading sit­
cantly to 90% from 68% (from Fig. 4) by incorporating EV charging in uations. In V2G operation, a 30% threshold is considered to calculate the
the lower demand slots in a day. disposable kWh from the EVs for discharging. In the scenario of unre­
solved overloading, the reallocation technique recognizes and shifts the
4. Conclusions EVs to a new time slot so that the supply system remains stable due to
excessive EV charging. The new time slots are identified based on the
This article proposes an electric vehicle (EV) charging scheme for tariff and time slot preference from the possible shifted EVs by the
residential consumers. At first, the proposed methodology focuses on the reallocation method. The shifted vehicles from the overloading slots are
valley filling method in determining charging slots. In doing so, low- chosen based on SOC level and required kWh ratting. Different case

14
S. Abdullah-Al-Nahid et al. Journal of Energy Storage 50 (2022) 104655

studies have been investigated during the simulation of the proposed [17] V.K. Mehta, R. Mehta, Variable load on power stations, in: Principles of Power
System, 3rd ed, S. Chand & Co. Ltd., 2006, pp. 41–68.
methodology. The outcomes display that the proposed method suc­
[18] P. Finn, C. Fitzpatrick, D. Connolly, Demand side management of electric car
cessfully handles the complex driving parameters such as overloading charging: benefits for consumer and grid, Energy 42 (2012) 358–363.
situations after EV integration and consumer prioritization in allocating [19] A. Mohsenian-Rad, V. Wong, J. Jatskevich, R. Schober, A. Leon-Garcia,
day-ahead charging schedules. The significant outcomes attained from Autonomous demand-side management based on game-theoretic energy
consumption scheduling for the future smart grid, IEEE Trans. Smart Grid 1 (2010)
this paper can be listed as follows: 320–331.
[20] K. Qian, C. Zhou, M. Allan, Y. Yuan, Modeling of load demand due to EV battery
• A day ahead charging scheme prioritizing customer satisfaction for charging in distribution systems, IEEE Trans. Power Syst. 26 (2011) 802–810.
[21] M.T. Hussain, D.N. Sulaiman, M.S. Hussain, M. Jabir, Optimal management
the EVs is introduced. strategies to solve issues of grid having electric vehicles (EV): a review, J. Energy
• The GA-based optimization deployed in the methodology ensures Storage 33 (2021), 102114, https://doi.org/10.1016/j.est.2020.102114.
near-optimum use of supply energy in less than 50 iterations. [22] M. Shaaban, A. Eajal, E. El-Saadany, Coordinated charging of plug-in hybrid
electric vehicles in smart hybrid AC/DC distribution systems, Renew. Energy 82
• Both the network stress issues and consumer satisfaction parameters (2015) 92–99.
are confronted successfully by the V2G provision and reallocation [23] T.U. Solanke, V.K. Ramachandaramurthy, J.Y. Yong, J. Pasupuleti, P. Kasina-than,
technique. A. Rajagopalan, A review of strategic charging–discharging control of grid-
connected electric vehicles, J. Energy Storage 28 (2020), 101193, https://doi.org/
10.1016/j.est.2020.101193.
In future, a more significant number of EVs will be considered, using [24] S.S. Ravi, M. Aziz, Utilization of electric vehicles for vehicle-to-grid services:
an intelligent algorithm to address overcrowding scenarios during progress and perspectives, Energies 15 (2022) 589, https://doi.org/10.3390/
en15020589.
vehicle charging slot scheduling. Also, the number of possible solutions
[25] M.A. Islam, J.G. Singh, I. Jahan, M.S. Lipu, T. Jamal, R.M. Elavarasan, L. Mihet-
that may come from a thorough search of all possible combinations of Popa, Modeling and performance evaluation of ANFIS controller-based
the problem variables will be computed to demonstrate the superiority bidirectional power management scheme in plug-in electric vehicles integrated
of employing GA in the methodology. In addition, a concise study of EV with electric grid, IEEE Access. 9 (2021) 166762–166780, https://doi.org/
10.1109/access.2021.3135190.
charging on a monthly and annual basis will be conducted to integrate [26] Z. Ma, D.S. Callaway, I.A. Hiskens, Decentralized charging control of large
EV demand in generation expansion planning. populations of plug-in electric vehicles, IEEE Trans. Control Syst. Technol. 21
(2013) 67–78, https://doi.org/10.1109/tcst.2011.2174059.
[27] K. Zhou, L. Cheng, L. Wen, X. Lu, T. Ding, A coordinated charging scheduling
Declaration of competing interest method for electric vehicles considering different charging demands, Energy 213
(2020), 118882, https://doi.org/10.1016/j.energy.2020.118882.
The authors declare that they have no known competing financial [28] M.K. Daryabari, R. Keypour, H. Golmohamadi, Robust self-scheduling of parking
lot microgrids leveraging responsive electric vehicles, Appl. Energy 290 (2021),
interests or personal relationships that could have appeared to influence 116802, https://doi.org/10.1016/j.apenergy.2021.116802.
the work reported in this paper. [29] Z. Wang, S. Wang, Grid power peak shaving and valley filling using vehicle-to-grid
systems, IEEE Trans. Power Deliv. 28 (2013) 1822–1829, https://doi.org/10.1109/
TPWRD.2013.2264497.
References [30] C.S. Ioakimidis, D. Thomas, P. Rycerski, K.N. Genikomsakis, Peak shaving and
valley filling of power consumption profile in non-residential buildings using an
[1] I.S. Bayram, S. Galloway, G. Burt, A probabilistic capacity planning methodology electric vehicle parking lot, Energy 148 (2018) 148–158, https://doi.org/10.1016/
for plug-in electric vehicle charging lots with on-site energy storage systems, j.energy.2018.01.128.
J. Energy Storage 32 (2020), https://doi.org/10.1016/j.est.2020.101730. [31] S. Li, C. Gu, X. Zeng, P. Zhao, X. Pei, S. Cheng, Vehicle-to-grid management for
[2] X. Zeng, M.S. Nazir, M. Khaksar, K. Nishihara, H. Tao, A day-ahead economic multi-time scale grid power balancing, Energy 234 (2021), 121201, https://doi.
scheduling of microgrids equipped with plug-in hybrid electric vehicles using org/10.1016/j.energy.2021.121201.
modified shuffled frog leaping algorithm, J. Energy Storage 33 (2021), https://doi. [32] C.G. Hoehne, M.V. Chester, Optimizing plug-in electric vehicle and vehicle-to-grid
org/10.1016/j.est.2020.102021. charge scheduling to minimize carbon emissions, Energy 115 (2016) 646–657,
[3] M. Tuballa, M. Abundo, L. Kazmerski, A review of the development of smart grid https://doi.org/10.1016/j.energy.2016.09.057.
technologies, Renew. Sustain. Energy Rev. 59 (2016) 710–725. [33] H. Ben Sassi, C. Alaoui, F. Errahimi, N. Es-Sbai, Vehicle-to-grid technology and its
[4] M. Wikström, L. Eriksson, L. Hansson, Introducing plug-in electric vehicles in suitability for the Moroccan National Grid, J. Energy Storage 33 (2021), 102023,
public authorities, Res. Transp. Bus. Manag. 18 (2016) 29–37. https://doi.org/10.1016/j.est.2020.102023.
[5] Nissan North America Inc., 2014 Nissan Leaf Brochure. Dealer E-Process, 2014. [34] Q. Yang, J. Li, W. Cao, S. Li, J. Lin, D. Huo, et al., An improved vehicle to the grid
[6] Ford Motor Company, 2014 Ford Focus Electric Brochure, Ford Motor Company, method with battery longevity management in a microgrid application, Energy 198
2013. (2020), 117374, https://doi.org/10.1016/j.energy.2020.117374.
[7] BayerischeMotoren Werke, The New BMW i3-Launches November 2013, BMW, [35] S. Zeynali, N. Nasiri, M. Marzband, S.N. Ravadanegh, A hybrid robust-stochastic
UK, 2013. Printed in the UK. framework for strategic scheduling of integrated wind farm and plug-in hybrid
[8] B. Plumer, N. Popovich, B. Migliozzi, Electric Cars Are Coming. How Long Until electric vehicle fleets, Appl. Energy 300 (2021), 117432, https://doi.org/10.1016/
They Rule the Road? The New York Times, 2021 (Mar. 10). j.apenergy.2021.117432.
[9] U.S. Energy Information Administration (EIA), Energy Consumption by Sector, [36] A. Shahkamrani, H. Askarian-abyaneh, H. Nafisi, M. Marzband, A framework for
2020. day-ahead optimal charging scheduling of electric vehicles providing route
[10] R.S. Gupta, A. Tyagi, S. Anand, Optimal allocation of electric vehicles charging mapping: Kowloon case study, J. Clean. Production. 307 (2021), 127297, https://
infra-structure, policies and future trends, J. Energy Storage 43 (2021), 103291, doi.org/10.1016/j.jclepro.2021.127297.
https://doi.org/10.1016/j.est.2021.103291. [37] J. Zheng, X. Wang, K. Men, C. Zhu, S. Zhu, Aggregation model-based optimization
[11] S. Abdullah-Al-Nahid, T.A. Khan, M.A. Taseen, T. Aziz, in: A Consumer-Friendly for electric vehicle charging strategy, IEEE Trans. Smart Grid 4 (2013) 1058–1066,
Electric Vehicle Charging Scheme for Residential Consumers, 2020 International https://doi.org/10.1109/TSG.2013.2242207.
Conference on Smart Grids and Energy Systems (SGES), 2020, pp. 893–897, [38] L. Jian, Y. Zheng, Z. Shao, High efficient valley-filling strategy for centralized
https://doi.org/10.1109/SGES51519.2020.00164. coordinated charging of large-scale electric vehicles, Appl. Energy 186 (2017)
[12] K. Zhang, L. Xu, M. Ouyang, H. Wang, L. Lu, J. Li, Z. Li, Optimal decentralized val- 46–55, https://doi.org/10.1016/j.apenergy.2016.10.117.
ley-filling charging strategy for electric vehicles, Energy Convers. Manag. 78 [39] N. Erdogan, F. Erden, M. Kisacikoglu, A fast and efficient coordinated vehicle-to-
(2014) 537–550, https://doi.org/10.1016/j.enconman.2013.11.011. grid discharging control scheme for peak shaving in power distribution system,
[13] M. Peng, L. Lian, C. Jiang, A review on the economic dispatch and risk J. Mod. Power Syst. Clean Energy 6 (2018) 555–566, https://doi.org/10.1007/
management of the large-scale plug-in electric vehicles (PHEVs)-penetrated power s40565-017-0375-z.
systems, Renew. Sustain. Energy Rev. 16 (2012) 508–515. [40] B. Yagcitekin, M. Uzunoglu, A double-layer smart charging strategy of electric
[14] Y. Noorollahi, A. Golshanfard, A. Aligholian, B. Mohammadi-ivatloo, S. Nielsen, vehicles taking routing and charge scheduling into account, Appl. Energy 167
A. Hajinezhad, Sustainable energy system planning for an industrial zone by (2016) 407–419, https://doi.org/10.1016/j.apenergy.2015.09.040.
integrating electric vehicles as energy storage, J. Energy Storage. 30 (2020), [41] Z. Hu, K. Zhan, H. Zhang, Y. Song, Pricing mechanisms design for guiding electric
101553, https://doi.org/10.1016/j.est.2020.101553. vehicle charging to fill load valley, Appl. Energy 178 (2016) 155–163, https://doi.
[15] M. Khan, A. Hossain, A. Ullah, M.H. Lipu, S. Siddiquee, M. Alam, T. Jamal, org/10.1016/j.apenergy.2016.06.025.
H. Ahmed, Integration of large-scale electric vehicles into utility grid: an efficient [42] A. Rahman, S.I.Chowdhury Arnob, Developing load profile for domestic customers
approach for impact analysis and power quality assessment, Sustainability 13 of Dhaka city through statistical prediction, in: 2016 3rd International Conference
(2021) 10943, https://doi.org/10.3390/su131910943. Electrical Engineering Inf. Communication Technology (ICEEICT), 2016, pp. 1–6,
[16] Y. Mu, J. Wu, N. Jenkins, H. Jia, C. Wang, A spatial-temporal model for grid impact https://doi.org/10.1109/CEEICT.2016.7873121.
analysis of plug-in electric vehicles, Appl. Energy 114 (2014), https://doi.org/
10.1016/j.apenergy.2013.10.006.

15