Electric Power Systems Research 220 (2023) 109240

Contents lists available at ScienceDirect

Electric Power Systems Research

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

Charging strategy in electric vehicle chargers by utilizing demand side

management scheme
Aziah Khamis a, *, M.H. Aiman a, W.M. Faizal b, C.Y. Khor b
a
Fakulti Kejuruteraan Elektrik, Universiti Teknikal Malaysia Melaka, Hang Tuah Jaya, Durian Tunggal, Melaka, 76100, Malaysia
b
Faculty of Mechanical Engineering & Technology, Universiti Malaysia Perlis, Arau, Perlis, 02600. Malaysia

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

Keywords: Electric vehicles (EVs) have received significant attention recently, given their potential technical, environ­
EV charger mental, and economic benefits. However, given the fact that EV charging could happen at peak demand. This
Demand side management situation may deteriorate the distribution network’s overall performance, particularly in terms of voltage sta­
Load shifting
bility. The novelty of this paper is to evaluate the impacts of the integration of EV chargers in the distribution
network and to propose a charging strategy in EV chargers. The proposed charging strategy manages the power
in the distribution network during the EV charger connection by utilizing demand side management. The IEEE 9-
bus radial distribution system was used as the test network. Four scenarios were conducted by relevant assess­
ments using the deterministic and probabilistic approaches. The result showed the influence of charging strategy
on EV charging systems where it can reduce the impact of integration EV chargers by shifting the load of EV
charging from the on-peak period to the off-peak period. As a result of using EV charging strategy, peak active
power losses were reduced by 2.2 to 3.2 percent. The proposed method contributes the idea to the engineer and
researcher in designing the EV charging strategy. However, the charging EV strategy could be further improved
by optimizing the value of the relay setting of both the UVLS relay and the power demand relay in future
research.

1. Introduction more mileage and lower emissions. The worldwide transportation

industry’s participation is also regularly increasing the popularity of
Environmentalists are concerned about the air pollution produced by electric cars, with the ultimate objective of lowering harmful pollution.
fossil fuels in transportation, manufacturing, and energy generation. The substitution of conventional combustion-based engines with
Climate change, rising energy prices, and dependency on fossil fuels are electric cars is an economical solution due to the current electrification
all significant issues today. These challenges are primarily relevant to of large sections in the electricity and transport sectors [3,4]. Electric
the three key businesses that rely heavily on fossil fuels, including Vehicles (EVs) may be narrowly categorized as Hybrid EVs (HEVs) and
transportation, manufacturing, and power. Academics and governments Plug-in EVs (PEVs). PEVs are classified as either Plug-in Hybrid EVs
worldwide are focusing on lowering reliance on fossil fuels and replac­ (PHEVs) or Battery EVs (BEVs). In HEVs, the battery cannot be charged
ing them with more sustainable alternatives [1,2]. The growth of sus­ from an external power source as opposed to BEVs [5,6]. Therefore,
tainable energy is more concerned with the environment, and the need PEVs would be integrated with EVs in the sense of this research. The
for renewable and clean energy has led to the demand for electric ve­ rapid expansion of EVs also offers a number of challenges to the power
hicles as a mode of transport. Several countries are contributed to system in terms of increasing charging capacity [7]. Therefore, it is also
accomplishing some sustainable energy market goals. In mitigating the essential to determine the effect of EVs on the power grid in the worst
consequences of rising gasoline prices and imposing more sustainability case of charging. The impact of the charging load on the power network
rules, electric cars are a viable solution for meeting the need for a may be organized in two ways: the change in setting position and the
greener transportation choice with lower emissions and improved fuel spontaneous charging time caused by driving decision-making. Exten­
economy. Developing and adopting electric vehicle technology is a sive studies were performed in this field on the dependability of the
developing answer to the issues above, with its enticing strategy giving voltage at a particular moment based on the charging load on the power

* Corresponding author.
E-mail address: aziah83@gmail.com (A. Khamis).

https://doi.org/10.1016/j.epsr.2023.109240
Received 8 November 2022; Received in revised form 13 January 2023; Accepted 20 February 2023
Available online 14 March 2023
0378-7796/© 2023 Elsevier B.V. All rights reserved.
A. Khamis et al. Electric Power Systems Research 220 (2023) 109240

Fig. 1. Single-Line diagram of the IEEE 9-bus radial distribution system [26].

grid bus, which is specified by the power grid’s moving behaviour [8]. compared to conventional vehicles, although when trying to charge for
The influence of varied penetration loads on voltage stability has been fossil fuel-generated power [16]. Undoubtedly, the demand for EVs will
discussed [9] based on the broad-scale charging demand analysis. The increase with the maturity of battery and charging techniques. There­
effect of load charging on the voltage and the loading gap in the peak fore, the relatively rapid integration of EV loads into the electricity
valley was analyzed [10], where a special load rate was incorporated in network is expected. Thus, carrying out system studies and taking
some specified power grid buses. However, the disadvantages are in­ remedial action to address any inconsistencies are essential in the early
dependent of the spatial and temporal variation of the charging load. stages of the deployment of EVs to sustain a reliable power system. For
These are due to the transfer properties of mobile energy induced by the instance, Fernandez et al. [16] discussed EV charging on energy losses in
mobility of EVs in the vehicle energy network. the device. The device energy losses can rise to 40% when EVs are not
Long-distance transportation helps improve economic growth, and used. Masoum et al. [17] studied delivery transformer losses due to
increased demand means more energy is needed to fuel car engines. residential EV charging. An increase in delivery transformer losses of up
However, consuming fossil fuels adds more carbon dioxide to the to 300% is possible in areas with a high adoption rate of EVs. It is critical
environment, causing numerous issues, such as air pollution and climate to spot and addresses any voltage drop or harmonics in a circuit as soon
change. Hence, EVs are developed as an alternative to internal com­ as possible to keep the power system running smoothly.
bustion engine vehicles due to their zero CO2 emissions during service. Farkas et al. [18] reported that extra load from chargers may over­
Incorporating EVs into the transportation system entails offering a va­ load grid components and cause voltage quality issues if the voltage
riety of EV chargers to meet the demand for EVs. High EV chargers can drop is greater than allowed. Meanwhile, [19] claims that a slight in­
influence voltage drops, system losses, an increase in power consump­ crease in the load on device components indicates that EV charging
tion, phase imbalance, and stability difficulties [11–13]. While these would not exceed voltage limits. A significant phase imbalance neces­
impacts are negligible in the distribution system for a single EV holder, sitated extra care when planning future EV deployments. Besides, the
multiple EVs can be charged from the distribution grid at the same time. current and voltage mismatch in the distribution network is not signif­
The harmful effects on the total delivery system might be significant. icantly affected by incorporating many EV chargers into the network
Therefore, it is also critical to investigate the EV charger’s effects and set [20]. The phase imbalance remains within reasonable bounds under a
up an adequate monitor to analyze the impact that EV charging may wide range of test conditions. Some studies revealed a significant in­
have on the safety and functioning of the distribution network. Mullan fluence, while others showed a minor impact. Several factors impacted
et al. [14] demonstrated that the grid could handle an additional these results, including the test network’s power, the position of the EV
charging load of 200,000 EVs during on-peak hours. However, 900,000 connection point, EV penetration pace, and EV charging preferences.
EVs may be absorbed without negatively impacting the grid by moving Besides, the lithium-ion battery also plays a crucial role by providing
EV charging to off-peak hours utilizing the Time-of-Use (TOU) price energy storage in EVs. Wang et al. [21] reviewed the Artificial Intelli­
plan. Implementing a TOU tariff system [15] may move EV loads from gence (AI) approach in investigating the lithium-ion battery state pre­
on-peak hours to off-peak periods, so resolving the issue of connecting diction method based on the data-driven method. They also examined
an EV charger during peak hours. Introducing the TOU tariff could solve accurate whole-life-cycle of state-of-charge (SOC) prediction by effec­
the integration of EV chargers into the distribution network. However, tively considering the current, voltage, and temperature variations [22].
the TOU tariff scheme does not have the capability and flexibility to The lithium-ion battery charging capacity decreases noticeably over
solve the issues when the power demand of the distribution network time; after 200 cycles, it has decreased by 21.30% and 22.61%,
unexpectedly changes. respectively.
Various research and development programs are being carried out Recently, power system operators have become increasingly con­
worldwide to find effective and affordable strategies for electrifying the cerned about their systems’ voltage reliability. The stable voltage across
transport system. Replenishing petroleum resources, health and envi­ all system buses after a disturbance generated by a particular starting
ronmental issues correlated with vehicle pollution, and energy protec­ operating condition is meant by voltage stability in the context of power
tion concerns have contributed to a rise in the rate of electrical EVs. systems [23]. At this time, pre-designed or online computed
Recent studies revealed that EVs had enhanced energy savings load-shedding is executed. Load shedding is terrible from a consumer

2
A. Khamis et al. Electric Power Systems Research 220 (2023) 109240

2. System modelling

Fig. 1 shows the IEEE 9-bus radial distribution system used to test the
proposed EV charging strategy with demand side management scheme.
The system consists of three power generators, G1, G2, and G3, six
transmission lines, three transformers, and three loads. The parameters
for the system components have been collected from previous studies
[26]. Meanwhile, Fig. 2 depicts the hourly load profile load with a daily
power consumption where the 24-h load profile relied on historical data
that was extracted from [26]. The total amount of load when the load
profile is 100% is 315 MW and 115 MVar real and reactive load,
respectively, which is depicted in Fig. 2 (a) and (b).
Meanwhile, Fig. 3 can be divided into three main part named A, B,
and C. Part A represent active and reactive power control, where details
modeling is shown in Figs. 3 and 4. The real and reactive power daily
consumption data from each load will be imported to PQ_4 as depicted
in Fig. 4. The signal of the phase voltage and PQ_4 were then determined
in Part B as illustrated in Fig. 3. The Simulink input signals from part B
were converted into an analogous current source using the regulated
current source in part C. This generated current is driven by the block’s
input signal to present the actual value of both an active and a reactive
load consumption at a specific time. In this study, a 24-h daytime period
has been selected in order to simulate daily consumption.

2.1. Electric vehicle modelling

In Malaysia, the EV can be charged with 3.7 kW at 240 V AC and

connected with a 16 A single-phase circuit in the regular low voltage
distribution grid. Note that most Malaysian low-voltage distribution
circuits in the residential area are single-phase circuits. Therefore, this
study used the EV charger infrastructure onboard single-phase with a
rating of 3.7 kW. This rate is primarily applicable in households without
an additional expenses for the charging infrastructure. In order to
demonstrate the effect of EV chargers at the distribution sides, a lump of
EV charger load will be connected in parallel with the load as illustrated
in Fig. 5.
All EVs are assumed to utilize lithium-ion batteries to simplify the
generation process. Thus, the real power demanded by an EV remains
constant during its charging duration [27] according to the EV specs on
the market [28], where the EV was characterized load as follows:
The state of charge (SOC), when the individual trip starts, can be
Fig. 2. Hourly load profile (a) Real and (b) Reactive Power. formulated in Eq. (1)
D
SOC = 1 − , (1)
voltage efficiency standpoint. However, the power grid uses load shed­ R
ding as a safety net to prevent voltage collapse in an emergency. Many
countries, including the United States, Canada, Belgium, Sweden, where SOC is a state of charge EV battery [%], D is the kilometers
Ireland, and Spain, use under-voltage load-shedding systems [24]. For traveled since the last charge [km], and R is the EV’s driving range [km].
example, in the Sammis-Star 345 kV grid breakdown, the blackout may Meanwhile, the time charging EV battery, tch can be represented in Eq.
have been prevented if 1,500 MW of manual or automatic load shedding (2)
had occurred before the critical occurrence [25]. In short, the EV Bc ∗ (1 − SOC)
charging strategies could solve the integration of EV chargers into the tch = , (2)
C
distribution network during an on-peak period, even if there is an un­
expected change in the power demand. where tch represents the time of charging EV battery [hour], Bc repre­
Therefore, the following are the primary contributions of the current sents the battery’s capacity [kWh], and C represents the rate of charging
research study: [kW].
Meanwhile, the limitations of inequality refer to the relay’s design
i This research investigates the impact of EV charging on the distri­ limits in the power system to ensure that the system is maintained in the
bution system in terms of injecting a lump of EV chargers and specified security edge.
determining the location of the weak buses in terms of voltage Limit of voltage: The minimum voltage limits, Vmin
i in bus-k for each
stability. bus must remain within the standard limits where it can be address in
ii The proposed EV charging strategy is designed to contribute stable Eq. (3)
distribution network’s voltage level while connecting EV charging
power at peak periods. Vmin
i < Vi , = 1, 2, 3, …, ni (3)

where Vmin
i is a minimum voltage at bus i, Vi is a voltage at bus i.

3
A. Khamis et al. Electric Power Systems Research 220 (2023) 109240

Fig. 3. EV charging strategy model.

Fig. 4. Active and reactive power control.

Fig. 5. EV load connection at test network system.

4
A. Khamis et al. Electric Power Systems Research 220 (2023) 109240

Fig. 6. Implementation of EV charging strategy using demand side management.

Limit of EV charger power: This constraint identified minimum active

Max voltage deviation from average line voltage
power limits of EV charger power units connected at bus i. The limit of %LVUR = ∗ 100% (5)
average line voltage
EV charger power can be given in Eq. (4)
Network losses: The power losses in the distribution network are
Pd + PEV , i ≤ Pmax
setting,i , i = 1, 2, 3, …ni (4)
mainly driven by the I2 R Joule Effect on both transmission line and
transformer. As discussed before, EV charging will introduce higher
where Pd is power demand at bus i, PEV is the power of EV chargers at
demand for EV charging; thus, higher energy losses are expected.
bus i, Pmax
setting is maximum power demand at bus i.
Therefore, the impact of network losses by EV charging was assessed in
During the operation of EV charging in the distribution networks, the
this study. The network losses assessment was carried out using Eq. (6),
following constraints should be considered:
which can be expressed as
Voltage fluctuation: The statutory tolerance for voltage excursion on
the distribution network in Malaysia [29] should be considered during Network energy losses
% of network losses = ∗ 100% (6)
the EV charger integrations (see Table 3). Peak power demand of base case
Voltage unbalance: EV charger load charging is mainly in a single
Penetration of EV: The impact of the EV on the distribution network’s
phase, likely to increase the distribution system’s voltage unbalance
performance was evaluated through different scenarios, from 0% to
level. Malaysia’s voltage unbalance factor’s statutory limit is 2.0% [29].
100% EV penetration levels. The definition of EV penetration level can
Thus, the mathematical expression of voltage unbalance, also known as
be expressed in Eq (7)
the line voltage unbalance rate (LVUR) at the load according to the
National Electrical Manufacturers Association (NEMA) definition, was Numbers of EV throughout feeder
EV penetration level = ∗ 100% (7)
used [30], which can be formulated in Eq. (5) Total numbers of EV
A case with injecting an EV charger into the distribution network
would influence system operation stability. Therefore, the EV charging

5
A. Khamis et al. Electric Power Systems Research 220 (2023) 109240

Fig. 7. UVLS relay and power demand relay model.

Table 1 Table 2
Steady-state voltage level regulation limits under normal Control setting of UVLS relay and power demand relay.
conditions. Relay Number UVLS Relay (PU) Power Demand Relay (PU)
Voltage Level % Variation
1 0.970 0.90
400V and 230V -6% & +10% 2 0.965 0.85
6.6kV, 11kV, 22kV,33kV +/-5% 3 0.960 0.80
132kV & 275kV +/-5% 4 0.955 0.75
500kV +/-5%

strategy is introduced here to reduce the influence of injecting EV Table 3

Battery charging scenarios.
chargers into the distribution network. The EV charging strategy (Fig. 6)
manages the power of EV charging in the distribution network through Scenarios Battery SOC, % Charging time
demand side management, in which the shifting period of EV charging is Maximum Discharge for Nissan Leaf 40 kw/h 20% ~ 80% 6 h 48 min
used based on the determination of limits towards the inequality con­ Maximum Discharge for BMW i3 42kw/h 20% ~ 80% 7 h 21 min
straints. The charging strategy was designed to maintain the distribution Average daily use for Nissan Leaf 50.8%~ 80% 3 h 23 min
Average daily use for BMW i3 49.9% ~ 80% 3 h 40 min
network’s voltage and total demand within normal operation conditions
(Fig. 7). The innovation of the proposed monitoring system is expected
to be applied in the household EVs charging system to facilitate the charging profiles. The Average Annual Kilometers Travelled (AAKT)
Malaysian EVs driver. Besides, it provides stable voltage and minimizes [31] indicated the average in 2013 was 24,129 km, the highest recorded
power loss during the charging process. The proposed EV charging in Selangor at 28,575 km, and the lowest at 16,342 km per year in Johor,
strategy could be realized by setting two switches. The first switch respectively. This data translates into an average of 78.3 km daily
represent UVLS relay. The output of this relay depends on a voltage commuted distance for Malaysian drivers. Based on the Environmental
magnitude level of the related phase at the local busbar, where the Protection Agency, US (EPA) laboratory test in LA4 mode drive cycle
charger is located between EV charger loads and the local bus. Ac­ (city driving conditions), the travelled distance by Nissan Leaf and BMW
cording to the model shown in Fig. 7, the decision either the voltage i3 on a fully charged battery is approximately 270 km (Nissan Leaf EV,
limit match the inequality were depands on (3), with two outputs signal 2019) and 260km (BMW i3s, 2020). With the assumption of 270 km and
either 0 or 1. Meanwhile, the second switch presents the power demand 260 km travelled distance for a fully charged battery, Nissan Leaf and
relay. This relay’s output is similar to the UVLS relay output, where the BMW i3 consume approximately 29.2% and 30.1% of the battery’s en­
condition stated in (4) with two output signals, 0 or 1. Note that, for both ergy daily commuted travelling distance.
relays, the output with 0 represents the mismatch condition, while 1 Therefore, four charging cases were considered regarding the bat­
represents the match condition. Table 1 summarizes the control setting tery’s state of charge (SOC). Table 2 summarizes the required charging
of the UVLS relay and power demand relay. times of the Nissan Leaf 40 and BMW i3 with a 3.7 kW onboard EV
charger. Considering the assumption that people might start charging
their EVs any time at home after office hours at 5.00 pm [29]. Thus, the
2.2. Probabilistic approach in assessing EV integration
charging required 3.7 kW, where the battery scenario condition is
depicted in Table 2 with a charging rating. The maximum charging time
Integrating EV charging in the distribution network involves many
is around 7.5 h, and the minimum is about 3.5 h.
uncertainties, especially in dealing with different driver behaviours. The
different driving patterns result in different EV energy consumptions
3. Results and discussion
and charging profiles. Uncontrolled single-phase EV charging also in­
creases the uncertainty of the LV network performance.
Four scenarios were considered to assess the impact of EV charging
In this study, the basic Malaysian driving pattern creates different EV

6
A. Khamis et al. Electric Power Systems Research 220 (2023) 109240

Table 4 controlled and uncontrolled charging conditions. Meanwhile, Table 5

Various case scenarios of controlled and uncontrolled EV charging. summarizes the number of EV lump chargers and the charging time in all
Scenario Condition SOC setting scenarios. The results revealed that inserting uncontrolled EV chargers
into the distribution network during on-peak periods significantly
EV 1 Uncontrolled EVs charging 20%
EV 2 Controlled EVs charging 20% reduced voltage and increased active power losses. This situation
EV 3 Unbalanced EVs charging 50% occurred because EV charging demand surpassed the distribution
EV 4 Controlled EVs charging during an unbalanced 50% network demand limit. The suggested method improved system stability
connection by minimizing voltage drops and active power losses (Table 6). This
improvement was achieved by moving the charging time from on-peak
on the distribution network. These scenarios include the impact of un­ to off-peak based on the charging EV strategy. The charging EV strategy
controlled EV charging, unbalanced EV charging, and controlled EV received data from the power system monitor on load demand, voltage
charging as stated in Table 4. The comparison was made between the level, and EV charger demand. The system checks all grid components,
such as the voltage level of each bus, load demand, and EV charging

Table 5
Charger number and the duration time of charging.
Scenario Bus Number Charger Number Load in kW Time Duration
Phase A Phase B Phase C Phase A Phase B Phase C

EV 1 Bus 5 10 Units 10 Units 10 Units 74kW 74kW 74kW 7h

Bus 6
Bus 8
Bus 9
EV 2 Bus 5 10 Units 10 Units 10 Units 74kW 74kW 74kW 7h
Bus 6
Bus 8
Bus 9
EV 3 Bus 5 10 Units 5 Units 10 Units 74kW 37kW 74kW 5h
Bus 6
Bus 8
Bus 9
EV 4 Bus 5 10 Units 5 Units 10 Units 74kW 37kW 74kW 5h
Bus 6
Bus 8
Bus 9

Table 6
Performance of Four scenarios based on peak active power losses.
Scenario Minimum Voltage Drop Peak Active Power Losses Deviation
Bus 5 Bus 6 Bus 8 Bus 9

EV 1 0.877 0.875 0.873 0.869 6.8% 2.8%

EV 2 0.879 0.980 0.966 0.892 4.0%
EV 3 0.873 0.979 0.954 0.889 6.8% 3.2%
EV 4 0.911 0.982 0.965 0.923 3.6%

Fig. 8. Scenario EV 2 daily voltage magnitude at Bus 5, Bus 6, Bus 8, and Bus 9.

7
A. Khamis et al. Electric Power Systems Research 220 (2023) 109240

Fig. 9. Scenario EV 2 daily network active power losses.

Fig. 10. Scenario EV 4 and EV 3 daily voltage magnitude at (a) Bus 5, (b) Bus 6, (c) Bus 8, (d) Bus 9.

demand. This data enables the charging EV strategy to select the ideal 3.1. Scenario EV 2: controlled EV charging
time for injecting EV chargers so that the EV charging demand does not
exceed the distribution network demand restrictions. The proposed This section compares the controlled EV charging time scenario (EV
technique’s effectiveness and a detailed evaluation of the controlled 2) with the uncontrolled EV charging (EV 1). In EV 2 scenario, the EV
charging are discussed in the following sub-sections. charging time was divided into two different charging periods. The
network accommodates a higher level of EV integration by shifting the

8
A. Khamis et al. Electric Power Systems Research 220 (2023) 109240

Fig. 10. (continued).

charging time from the on-peak period to the off-peak period. Fig. 8 scenario EV 1. The shifting of the charging from the on-peak period to
shows the voltage magnitude in scenario EV 2 at Bus 5, Bus 6, Bus 8, and the off-peak period decreases the peak demand. Hence, as the peak de­
Bus 9, respectively, where the distribution network has contacted EV mand decreases, so does the current line, resulting a reduction in active
chargers. Although the EV charging duration in this scenario is similar to power losses.
scenario EV 1, the minimum voltage drop recorded in this scenario was
less than the minimum voltage drop recorded in scenario EV 1. Thus, it is 3.2. Scenario EV 4: controlled EVs charging during an unbalanced
proven that introducing a charging EV strategy reduces the negative connection
impact of EV charger integration in the distribution network. Further­
more, the result shows the optimal period of EV charger connection. In scenario EV 4, the control strategy was implemented to maintain
Meanwhile, Fig. 9 shows the active power losses of the distribution the voltage within the normal operating range and to reduce the per­
network in scenario EV 2. The voltage drop (Fig. 8) corresponds to the centage of unbalanced voltage during EV charging. The results (Fig. 10)
power loss (Fig. 9). Voltage and power drops occur at a similar time (16- reveal that the minimum voltage drop from all buses has improved
17 h). The active power loss from 16 to 17 h is about 50%. From the significantly compared to the minimum voltage drop in scenario EV 3.
result, the peak active power losses could be reduced to 2.8% based on This situation is observed starting from 16 to 21 h in the second charging

Fig. 11. Scenario EV 3 daily unbalanced voltage percentage.

9
A. Khamis et al. Electric Power Systems Research 220 (2023) 109240

Fig. 12. Scenario EV-4 daily network active power losses.

period. The voltage drop of E V3 phase C from 16 to 17 h is about 4.35%, (i) The on-peak period was susceptible to EV charging which can
and 2.17% for EV3 phase B (Fig. 10a). A similar phenomenon was also easily cause the voltage collapse and the location of the EV
observed in Figs 10 (b)-(d), where the voltage distribution during charger connection.
charging for controlled EV 4 is better than the uncontrolled EV3. The (ii) Applying the controlled EV charging reduces the peak active
percentage of unbalance voltage has been reduced. This situation power losses by 2.2%-3.2%.
happened because the demand side management. The chargers become (iii) This proposed charging EV strategy reduces the impact of the
active again when several EVs are fully charged. This situation makes integration of EV chargers by shifting the period of EV charging
space for the chargers to react again. The output of this result could be from the on-peak period to the off-peak period.
improved if the UVLS relay and the power demand relay were set up (iv) The output of the charging EV strategy could be improved by
properly. At the same time, the charge will activate at the lower point of optimizing the value of the relay setting of both the UVLS relay
distribution network demand. As a result, the voltage drop will get and the power demand relay in future research.
better than in scenario EV 3. Further, the optimal duration time will
influence the unbalanced voltage when some load shifts to a different CRediT author statement
time.
Fig. 11 depicts the daily voltage unbalance values for several bus Aziah Khamis: Conceptualization, Methodology, Data curation,
locations on the distribution network. After implementing a charging EV writing-original draft, Writing - Review & Editing, Visualization
strategy in an unbalanced connection, the results confirmed that the M.H. Aiman: Investigation, software, Writing - Original Draft
percentage of unbalanced voltage is directly influenced by the location W.M. Faizal: Writing - Review & Editing, Visualization
of the bus and the period. The output of this scenario shows a decrease in C.Y. Khor: Writing - Review & Editing
the overall percentage of voltage unbalances. Compared to an uncon­
trolled EV charger connection (scenario EV 3), scenario EV 4 shows that
Bus 5 improved by 0.8%. While, Bus 9 improved by 0.4%. Over limit Declaration of Competing Interest
unbalance voltage at Bus 5 did not exceed the limit of unbalance voltage.
In contrast, the uncontrolled EV charger connection in EV 3 revealed The authors declare that they have no known competing financial
that the unbalanced voltage at Bus 5 exceeded the limit of unbalance interests or personal relationships that could have appeared to influence
voltage by 2.4%. However, the limit of unbalance voltage should be the work reported in this paper.
below 2.0%, as stated in the Tenaga Nasional Berhad (TNB) handbook.
As planned, implementing a charging EV strategy reduces the peak Data availability
active power losses caused by the connection of EV chargers to the
distribution network during peak hours. Fig. 12 shows the improvement Data will be made available on request.
of implementing a control strategy in an unbalanced connection. The
shifting of the charging from the on-peak period to the off-peak period
decreases the peak demand. The decline in peak demand lowers the References
current line, which lowers operational power losses. Thus, as mentioned
in the previous section, it will prevent any harm to network components. [1] Zahra Darabi, Mehdi Ferdowsi, Impact of plug-in hybrid electric vehicles on
electricity demand profile, in: Smart Power Grids 2011, Berlin, Heidelberg,
Springer, 2012, pp. 319–349.
4. Conclusion [2] Saeed Rezaee, Ebrahim Farjah, Benyamin Khorramdel, Probabilistic analysis of
plug-in electric vehicles impact on electrical grid through homes and parking lots,
This research examined the impact of EV charging system integration IEEE Trans. Sustain.Energy 4 (4) (2013) 1024–1033.
[3] Larminie, J., and J. Lowry. "Electric vehicle technology explained, electric vehicle
and in steady-state behaviour of the distribution network. Four scenarios technology explained." (2003).
were considered in this research to reflect the probability impact of EV [4] M.F.M. Sabri, Kumeresan A. Danapalasingam, Mohd Fuaad Rahmat, A review on
charging on the distribution network and the influence of an introducing hybrid electric vehicles architecture and energy management strategies, Renew.
Sustain. Energy Rev. 53 (2016) 1433–1442.
control strategy on the EV charging system. For EV charging, the depth
[5] Andrew J. Nathan, Andrew Scobell, How China sees America: the sum of Beijing’s
of impact was high depending on the time and location of the EV charger fears, Foreign Aff 91 (2012) 32.
connection. The research revealed, as mentioned: [6] Jiuyu Du, Danhua Ouyang, Progress of Chinese electric vehicles industrialization in
2015: a review, Appl. Energy 188 (2017) 529–546.
[7] Petr Kadurek, Christos Ioakimidis, Paulo Ferrao, Electric vehicles and their impact
to the electric grid in isolated systems, in: 2009 International Conference on Power
Engineering, Energy and Electrical Drives, IEEE, 2009, pp. 49–54.

10
A. Khamis et al. Electric Power Systems Research 220 (2023) 109240

[8] Noraidah Binti Mohd Shariff, Mohammed Al Essa, Liana Cipcigan, Probabilistic [20] Jian Liu, Electric vehicle charging infrastructure assignment and power grid
analysis of electric vehicles charging load impact on residential Distributions impacts assessment in Beijing, Energy policy 51 (2012) 544–557.
Networks, in: 2016 IEEE international energy conference (ENERGYCON), IEEE, [21] Shunli Wang, Pu Ren, Paul Takyi-Aninakwa, Siyu Jin, Carlos Fernandez, A critical
2016, pp. 1–6. review of improved deep convolutional neural network for multi-timescale state
[9] Ying Zheng, Jinwen Sun, Chong Zhang, X. Lin, Study of voltage stability margin for prediction of lithium-ion batteries, Energies 15 (14) (2022) 5053.
the distribution network with electric vehicle integration, Trans. China Electrotech. [22] Shunli Wang, Paul Takyi-Aninakwa, Siyu Jin, Chunmei Yu, Carlos Fernandez,
Soc. 29 (8) (2014) 20–26. Daniel-Ioan Stroe, An improved feedforward-long short-term memory modeling
[10] X.C. Du, Y.M. Liu, Z.C. Xu, T.T. Li, J.H. Huang, P. Jia, Modeling of random load of method for the whole-life-cycle state of charge prediction of lithium-ion batteries
electric vehicle and its influence on node voltage distribution of distribution considering current-voltage-temperature variation, Energy (2022), 124224.
network, J. Electr. Power Autom. Equip. 38 (2018) 124–130. [23] Prabha Kundur, John Paserba, Venkat Ajjarapu, Göran Andersson, Anjan Bose,
[11] C.H. Dharmakeerthi, N. Mithulananthan, Tapan Kumar Saha, Overview of the Claudio Canizares, Nikos Hatziargyriou, et al., Definition and classification of
impacts of plug-in electric vehicles on the power grid, in: 2011 IEEE PES power system stability IEEE/CIGRE joint task force on stability terms and
Innovative Smart Grid Technologies, IEEE, 2011, pp. 1–8. definitions, IEEE Trans. Power Syst. 19 (3) (2004) 1387–1401.
[12] II Green, C. Robert, Lingfeng Wang, Mansoor Alam, The impact of plug-in hybrid [24] O.P. Rahi, Amit Kr Yadav, Hasmat Malik, Abdul Azeem, Bhupesh Kr, Power system
electric vehicles on distribution networks: a review and outlook, Renew. Sustain. voltage stability assessment through artificial neural network, Procedia Eng. 30
Energy Rev. 15 (1) (2011) 544–553. (2012) 53–60.
[13] James Dixon, Keith Bell, Electric vehicles: Battery capacity, charger power, access [25] Muir, A., and J. Lopatto. "Final report on the August 14, 2003 blackout in the
to charging and the impacts on distribution networks, ETransportation 4 (2020), United States and Canada: causes and recommendations." (2004).
100059. [26] Aziah Khamis, Hussain Shareef, Azah Mohamed, Zhao Yang Dong, A load shedding
[14] Jonathan Mullan, David Harries, Thomas Bräunl, Stephen Whitely, Modelling the scheme for DG integrated islanded power system utilizing backtracking search
impacts of electric vehicle recharging on the Western Australian electricity supply algorithm, Ain Shams Eng. J. 9 (1) (2018) 161–172.
system, Energy policy 39 (7) (2011) 4349–4359. [27] Francesco Marra, Guang Ya Yang, Chresten Træholt, Esben Larsen, Claus
[15] Woo-Jae Park, Kyung-Bin Song, Jung-Wook Park, Impact of electric vehicle Nygaard Rasmussen, Shi You, Demand profile study of battery electric vehicle
penetration-based charging demand on load profile, J. Electric. Eng. Technol. 8 (2) under different charging options, in: 2012 IEEE power and energy society general
(2013) 244–251. meeting, IEEE, 2012, pp. 1–7.
[16] Luis Pieltain Fernandez, Tomás Gómez San Román, Rafael Cossent, Carlos [28] Alireza Hatefi Einaddin, Ahmad Sadeghi Yazdankhah, A novel approach for multi-
Mateo Domingo, Pablo Frias, Assessment of the impact of plug-in electric vehicles objective optimal scheduling of large-scale EV fleets in a smart distribution grid
on distribution networks, IEEE Trans. Power Syst. 26 (1) (2010) 206–213. considering realistic and stochastic modeling framework, Int. J. Electric. Power &
[17] Mohammad AS Masoum, Paul S. Moses, Keyue M. Smedley, Distribution Energy Syst. 117 (2020), 105617.
transformer losses and performance in smart grids with residential plug-in electric [29] Tenaga Nasional Berhad, Electricity supply application handbook, Distribution
vehicles, in: ISGT 2011, IEEE, 2011, pp. 1–7. Division TNB, Kuala Lumpur, Malaysia, 2011, pp. 7–130.
[18] Csaba Farkas, Gergely Szűcs, Laszlo Prikler, Grid impacts of twin EV fast charging [30] M.H.J. Bollen, Definitions of voltage unbalance, IEEE Power Eng. Rev. 22 (11)
stations placed alongside a motorway, in: 2013 4th International Youth Conference (2002) 49–50.
on Energy (IYCE), IEEE, 2013, pp. 1–6. [31] Akmalia Shabadin, N.Megat Johari, H.Mohamed Jamil, Car annual vehicle
[19] A. Rautiainen, et al., Case studies on impacts of plug-in vehicle charging load on kilometer travelled estimated from car manufacturer data–an improved method,
the planning of urban electricity distribution networks, in: 2013 Eighth Pertanika 25 (1) (2014) 171–180.
International Conference and Exhibition on Ecological Vehicles and Renewable
Energies (EVER), 2013, pp. 1–7, https://doi.org/10.1109/EVER.2013.6521542.

11