Electric Power Systems Research 184 (2020) 106307

Contents lists available at ScienceDirect

Electric Power Systems Research

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

Charge coordination and battery lifecycle analysis of electric vehicles with T

V2G implementation
M. Sufyana,b, N.A. Rahima, , M.A. Muhammadc, C.K. Tana, S.R.S. Raihana, A.H.A. Bakara
⁎

a
Higher Institution Centre of Excellence (HICoE), UM Power Energy Dedicated Advanced Centre (UMPEDAC) Wisma R & D, University of Malaya, Kuala Lumpur,
Malaysia
b
Faculty of Engineering and Sciences Technology, Iqra University, Main Campus, Karachi, Pakistan
c
Electrical Engineering Department, Faculty of Engineering University of Malaya, Kuala Lumpur, Malaysia

ARTICLE INFO ABSTRACT

Keywords: With the growing concerns on the energy depletion and the reduction of CO2 emission, electric vehicles (EVs)
Electric vehicle have gained popularity in transport sector due to clean and reliable energy source. However, the charging of EVs
Charge coordination has imposed significant load in the electrical distribution system. The stability of power network is disturbed
Battery degradation with uncoordinated charging. This work aims to investigate the optimal EV coordination with vehicle to grid
Renewable energy sources
(V2G) technology for the cost-benefit analysis. However, the cost of EV battery replacement due to degradation
is of paramount concern for the EV users, which constitute the most important ingredient to ensure active
participation in V2G energy exchanges. Therefore, battery degradation cost is formulated for real time analysis
taking depth of discharge at each time interval. The firefly algorithm has been used to optimize the system cost.
The performance of the proposed system is tested on modified 33 bus distribution system in the presence of
renewable energy sources (RES). The impact of system cost and energy losses are analysed for different RES
penetration, EV capacities and travelling distances. The simulation results show significant reduction in oper-
ating cost when RESs are integrated into distribution network. Moreover, V2G technology is beneficial for EV
consumers with high penetration of RES.

1. Introduction
Escalation of greenhouse gas emissions to alarming levels in the
recent decade is instrumental in pivoting the attention of utility com-
panies towards establishing a clean, reliable and sustainable energy
system. Large-scale integration of electric vehicles (EVs) is being con-
sidered an effective solution to decarbonize the environment [1]. The
ongoing research for new developments in charging infrastructures
along with policy support for suitable investment are resulting in lower
battery costs and higher EV production. International Energy Agency
has predicted a significant increase in EV sales after 2020 reaching up
to 100 million every year by 2050 [2]. The advantage of EV over
conventional vehicles are the ability to run on the locally generated
power, vehicle to grid (V2G) power transfer and reducing CO2 emis-
sions.
Given the intermittent nature of renewable energy sources (RES) in
the grid, battery energy storage is proposed as a viable solution to
suppress the voltage dip and power deficit [3] facilitating power grid to
meet the maximum power demand through peak shaving schemes [4].
The benefits from EV are not limited to the automobile industry, rather
their utilization as battery energy storage reduces the power fluctua-
tions from RES, provide energy arbitrage and earn extra revenue [5,6].
Proper charging coordination of EVs is essential in achieving an
efficient interconnection to the power system, without proper co-
ordination electrical distribution system will experience overloading
due to simultaneous charging of EV. Moreover, high penetration of EV
may result in increase of technical losses, higher peak demands and
reduction in voltage profile [7]. These negative impacts can be miti-
gated by metaheuristic techniques to attain efficient charge coordina-
tion of EV.
EVs are the credible candidate for demand side management (DSM),
providing flexibility for charging techniques depending on the time.
Peak shaving and load shifting can be performed with the smart char-
ging techniques. Moreover, EVs are capable of feeding power back to
grid during high peak demands reducing the electricity cost and earning
profit for consumer. The bidirectional converters connect the EV to the
grid; charging the batteries at low demand period and discharging at
high demand period. This concept is referred as V2G. V2G methods can

⁎
Corresponding author.
E-mail address: nasrudin@um.edu.my (N.A. Rahim).

https://doi.org/10.1016/j.epsr.2020.106307
Received 20 September 2019; Received in revised form 26 February 2020; Accepted 4 March 2020
Available online 20 March 2020
0378-7796/ © 2020 Elsevier B.V. All rights reserved.
M. Sufyan, et al.
be categorized into economic and technical strategies. Economic stra-
tegies focus on EV owners profit maximization considering the impact
of battery degradation, user availability and time of use (TOU) tariff.
Whereas technical strategies support the system in performing voltage
and frequency regulation, power balance, demand response and loss
reduction.
Batteries of EVs supplying power to the gird, in V2G configuration
present a promising substitute for a comparatively expensive battery
farm [8]. As such, the research into the V2G and battery energy man-
agement system (BEMS) had seen rapid progress in the last decade. The
principal objective of these researches is to optimize cost savings, which
can be achieved through peak shaving for the utility companies. The
resultant saving is turned into cash incentives for EV users. However,
the cost of EV battery replacement due to degradation is of paramount
concern for EV users as they are the key players in V2G energy ex-
change market. Furthermore, there is also concerns within the EV
manufacturers pertaining to the state of warranty for batteries that
participate in V2G system. As such, intense researches are being con-
ducted to identify the battery degradation pattern under various dy-
namic scenarios of V2G participation, together with its associated life
span and cost of energy exchanges.
A charge coordination problem of EV batteries has been rectified in
[7] to reduce the operational cost of the system. The impact of charging
EVs under specific time period according to EV owner's preference is
evaluated with three techniques, producing an optimal solution by
using a hybrid optimization algorithm. Different EV penetrations are
examined in the presence of distributed generators (DG) on a 449-node
distribution network. The results show that charging of EV improves the
voltage profile and minimizes the system losses when DG is integrated
into the network. An optimal scheduling algorithm with decentralized
controller is proposed in [9] to schedule the charging and discharging
of EV. The decentralized controller is used to communicate with the
aggregator after EV conducts local computation. The proposed method
aims to flatten the demand curve as per consumer battery requirement.
The study recommends to retain the battery state of charge (SOC) be-
tween 20% to 85% to extend the battery lifetime. The proposed algo-
rithm reduces the computational burden in charging EVs to specific
state-of-charge (SOC) compared to other mixed integer non-linear
programming problems. In [10] the charging and discharging co-
ordination of EV is performed for G2V and V2G operation taking cus-
tomers preference to charge EV up to defined SOC. The proposed
method schedules the future charging request as per customer's desire
for the day ahead demand. The customer participation program im-
proves the performance of the control algorithm in load shifting during
peak periods. However, the authors have not taken the power losses
into consideration during charging and discharging event.
In [11] energy management strategy is proposed for microgrid in-
tegrating EV parking lot to minimize the total operation cost. The effect
of market price fluctuation, parking time, uncertainty of battery capa-
city and EV arrival/departure is considered for the cost saving analysis.
The proposed method maximizes the savings for EV users with free
parking facility and compensation for battery wear cost. The savings
increases when market price fluctuation is high and EV users have
maximum parking time facility. A smart charging strategy to integrate
EVs in the carpark is presented in [12] to reduce the cost and minimize
the peak average ratio (PAR). The proposed method has been tested on
37 bus distribution system with maximum EV penetration by con-
sidering the technical constraints of the system. The study focuses on
EV charging at workplace, hence charging to maximum SOC is avoided
and EV battery is charged as per user's next trip details, which are based
on present SOC, traveling distance and departure time. The analysis
found that fast charging results in higher cost and PAR and reduces the
EV penetration into the distribution system. The cost benefit analysis
for the optimal charging and discharging coordination of plug in
2
Electric Power Systems Research 184 (2020) 106307
electric vehicle (PEV) is proposed in [13]. The study addresses the is-
sues of characterizing the charging pattern and computing the max-
imum charging load and available discharge capacity of EV battery. The
two-stage optimization model minimizes the peak load with maximum
EV load in the first stage and then reduces the fluctuation at the peak
load in the second stage. The proposed method calculates the annual
benefit from the savings earned through peak shaving whereby ne-
glecting the battery degradation cost and taking fixed charging power
of EVs. Zhang in [14] has developed a quantitative evaluation method
for discharging PEV fleets by considering power and energy constraints
of the PEVs. The proposed method develops aggregate queueing model
to evaluate the available V2G capacity during real-time operations
without identifying the charging and discharging durations. The diffi-
culty in forecasting of charging demand due to the stochastic behaviour
of EV traveling is reduced to compute the available capacity for V2G. In
addition, smart charging strategy is designed to perform reliable op-
eration for PEV charging and discharging. The numerical analyses re-
sult in gaining the benefits from the reserves accumulated during en-
ergy scheduling of PEV. The economic dispatch of microgrid was
optimized by developing a multi-agent system with different penetra-
tions of EV [15]. Three charging patterns; uncontrolled charging, rapid
charging and smart charging together with V2G were analysed in the
presence of a gas turbine, chillers and photovoltaic system. The smart
charging strategy shifts the EV charging load to valley period during
peak time to improve the stability of the system. The V2G technology
reduces the electricity supply from gas turbine when high penetration
was considered. Hence, the microgrid operating cost was remarkably
minimized with V2G availability compared with the uncontrolled
charging. However, the proposed method had defined a potential limit
for V2G in a specific time to avoid deep discharge depreciating battery
lifetime.
A novel bidirectional operation of EV in grid connected and islanded
mode is proposed in [16] for three phase unbalanced distribution
system. The proposed method implements grid-to-vehicle (G2V)
strategy based on the power availability of buses ensuring the grid
operative conditions are satisfactory. If any failure leads the network to
islanded mode, the reliability and security of the network are increased
by V2G implementation. A stochastic optimization method having
centralized control unit is proposed in [17] for coordinated charging of
PEV. The economic scheduling by factoring high uncertainties of re-
newable energy and load is formulated to reduce the microgrid oper-
ating cost. The study takes constant battery lifecycle in their analysis.
However, the battery lifecycle is dependent on the rated depth of dis-
charge.
In contrast, Czechowski in [18] questions the V2G capability of
turning a profit under specific circumstances. A comprehensive cost
benefit analysis is required for the economic scheduling considering
battery degradation cost. The model assumes V2G to be beneficial with
the decrease in battery capital cost and more ancillary services like peak
shaving and reserve capacity are performed. Lu in [19] proposed a
multi-objective optimal dispatch of microgrid incorporating wind tur-
bine, photovoltaic, diesel generators and microturbine. The Monte
Carlo simulation is used to solve the uncertainties of EV arrival and
departure. The proposed method performs peak shaving by charging
the EVs during low price period to reduce the operating cost. The re-
sults reveal that coordinated charging with higher penetration of dis-
tributed generators minimizes the cost and load variance. However,
weighting factor behaves as a trade-off between system cost and load
fluctuation. The study finds V2G is not economical for EV owners due to
battery degradation cost, but it improves the stability of the system by
minimizing load variance. The charge coordination problem of EV to-
gether with V2G technology in an unbalanced distribution system is
solved in [20]. The authors have taken random arrival and departure
times and EVs state of charge upon arrival with different battery sizes to
M. Sufyan, et al. Electric Power Systems Research 184 (2020) 106307

reduce the energy cost and optimize the operation of distribution

system. The proposed method schedules the optimal EV charging by
considering the uncertainty of load while satisfying the voltage limits
and minimizing the power losses. The results illustrate that V2G utili-
zation in distribution system in not economically beneficial as the
lifetime of EV battery depreciates and cost of replacing the battery
outstrips the benefits associated with V2G. The above literature corre-
sponds to the EV charging and discharging schedule minimizing the
system operating cost. However, the battery degradation cost for every
charging / discharging state is not modelled in the literature.
EV owners start charging their vehicles after returning to home
during night time to the maximum SOC (ratio of energy available in the
battery to the maximum battery capacity). They are most likely to sell
this power to the grid when the electricity tariff is high during day time.
Commercial workplaces and parking lots are the best places to imple- Fig 1. Cost of energy storage for different DOD and power discharge.
ment V2G and earn maximum profit as the vehicle is parked there for
several hours. Thus, the utility grid has to dispatch less power because 0 vh vc, i orvh vc, o
these workplaces can act as generation sources to fulfill the load re- v vc, i
Phw =
w
Pmax * vh vc, i vh vrt
quirement. The amount of power exchange between the grid and EV rt vc, i

depends on the operation cost and the depth of discharge (DOD) of w

Pmax vrt vh vc, o (1)
battery. The cumulative cost of G2V and V2G depends on the number of
EV charging/discharging and the amount of power drawn/delivered by where (kW) is the maximum power output of WT, vh (m/s) is the
w
Pmax
battery. speed of wind at hour h, vrt (m/s) is its rated value whereas the cut-in
The study is different from prior research works in the following speed and cut-out wind speed are shown by vc, i and vc, o respectively.
aspects that reflect its contributions.
2.2. Solar PV model
• The project optimizes the economic power dispatch in V2G en- The power measured by solar photovoltaic array is dependent on
vironment by incorporating the battery degradation model. The
presence of RES, centralized charge coordination of EVs and dis- the solar irradiation and the ambient temperature at each hour. The PV
tributed V2G batteries power dispatches will be considered in the power is given by [22]
economic dispatch problem. I
•
pv
Customarily, large majority of the decision making in battery energy Pout = Prtpv* *[1 + {(Tamb + (0.0256*I)) Tref }]
Iref (2)
management system is simply based on the availability and deficit of
renewable energy respectively to meet the demand. However, this where (kW) is output power of the solar panel,
pv
Pout (kW) is max- Prtpv
study proposes the optimized solution for battery energy manage- imum power at standard conditions, I is solar irradiation (W/m2), Iref is
ment system by taking into account the battery degradation cost. In the solar radiation at standard temperature (Iref = 1 kW/m2), Tref is
addition, the battery degradation cost is modelled for real time standard temperature taken as 25 °C, τ is 3.7 × 10−3(1/ °C), and Tamb (
analysis considering DOD at each time interval and deep discharges °C) is the ambient temperature of the solar panel.
are avoided to prolong the battery lifespan.
• The operation cost and energy losses are investigated for different 2.3. EV battery cost model
RES penetration levels, EV battery capacities and uncertainties in
the renewable energy and EV availability in distribution network. The battery cost analysis for the different values of discharge power
• In the previous works, the V2G is not economically profitable as the and depth of discharge is shown in Fig 1. The depth of discharge re-
battery lifetime deteriorates and degradation cost is high. However, presents the amount of capacity used by the battery. The figure depicts
the battery scheduling in this study considers profit as key factor for that cost of battery increases when the DOD is high and similarly the
EV users during V2G exchange while taking battery degradation cost cost is increased when the battery discharges more power. The increase
into consideration. in discharge power results in decreasing the capacity of battery and
causes the DOD to be high. However, the continuous discharge of high
This paper is organized into seven sections. The first section dis- power from the battery at maximum DOD increases the battery de-
cusses the background of EV roles in a power and renewable energy gradation cost to the maximum value. The cost of discharging battery at
sources grid. Section 2 describes the modelling of renewable energy any time interval as a function of battery power and DOD is formulated
sources and battery storage. The objective function and constraints are in Eq (3) [23]. The lifecycle of the lithium-ion battery is represented by
formulated in Section 3 while Section 4 illustrates the optimization an exponential function taken from [24]. The degradation cost of the
algorithm. Section 5 presents the simulation setup and description. Fi- battery storage during discharge event is given by
nally, Sections 6 and 7 provide the discussion on the results and the
Cb, cap*P ev
h* h
conclusion of the study respectively. Cbh, d =
Ecap*lc (dodhb)* b2
ev
(3)
2. Renewable energy and battery modelling
lc (dodhb) = 694*(dodhb) 0.795 (4)
2.1. Wind turbine model Eev
h
dodhb = 1
Eev
cap (5)
The relation between the wind output power (WT) and the speed is
given by the relationship below [21]: where Cbh, d is the battery degradation cost at hour h in $, Cb, cap is the

3
M. Sufyan, et al. Electric Power Systems Research 184 (2020) 106307

capital cost of battery ($), P ev

h is the amount of power discharged by
battery at hour h, Δh is the time interval taken as 1 h in this study, Eev
cap
represents the electric vehicle battery capacity and Eev h is the present
battery capacity at hour h, dodhb is the battery depth of discharge status
at hour h, lc (dodhb) is the number of cycles of battery storage and ηb is
efficiency of the energy storage.

3. Problem formulation

The incorporation of electric vehicles in the power grid improves

the stability of the system as they can behave as energy consumer and
energy providers during the low and high load periods respectively. The
electrical system faces issues like grid overloading and voltage devia-
tion under uncoordinated EV charging. On the other hand, smart
charging mechanism reduces the power difference between the peak
and valley periods minimizing the system losses. The proposed method
minimizes the operating cost of distribution system and promotes EV
users for V2G exchange while considering battery degradation cost. The
objective function in this research is the sum of costs associated with
the residential and commercial load, EV charging, battery degradation
and the power losses, formulated as:

Objective = Min. (Cres com

h + Ch + Cev b, d los
h + Ch + Ch )
h t (6)

Cres
h = P res
r , h* res
rt b (7)

Ccom
h = Prcom
,h * com
rt b (8)

Cev
h = P rev, h* ev
rt b (9)

Clos
h = I2sr , h*Rsr*
sr l (10)

In the above formulation, Eq (7) and Eq (8) are the costs of the
residential and commercial loads respectively, Eq (9) shows the char-
ging cost of EV and Eq (10) shows the cost of power losses in the dis-
tribution network. It is to be noted that the cost of EV will be negative
when discharging and positive if charging. The variables δres, δcom and
ɛev are the residential, commercial and EV electricity tariffs rate. Isr is
the bus current and ℜsr is the bus resistance associated for the power
loss in the network, while is the cost parameter for losses.
Fig 2. Flowchart of the proposed charging method.
3.1. Active and reactive power constraints

Psr , h + Prg, h + P ren

r ,h = Prt , h + I2rt , h*Rrt + P res
r ,h + P rcom
,h + P ev
r,h
sr l r b rt l rt l rt b rt b rt b (11)

where Psr, h is the active power of branch sr, P ren r , h is the renewable P 2sr , h + Q2sr , h Vr2, h*I2sr , h (15)
power (WT and PV) connected at bus r, Prg, h is the generated power at
bus r, P res
r , h and P r , h are the active residential and commercial loads
com
where Vr2, h represents bus r voltages, Vmax and Imax are maximum bus
connect at bus r. voltage and maximum branch current respectively.
Q sr , h + Qrg,h = Qrt, h + Qres
rt , h + Qrtcom
,h + Irt2 , h*Xrt
sr l r b rt l rt b rt b rt l (12)
3.2. Voltage limit constraints
where Q sr , h is the reactive power of branch sr, is the generated Qrg, h
power at bus r, Qres r , h and Qr , h are the reactive residential and com-
com

mercial loads connect at bus r. Vmin Vr , h Vmax (16)

Psr , h Vmax *Imax (13)

Q sr , h Vmax *Imax (14)

4
M. Sufyan, et al. Electric Power Systems Research 184 (2020) 106307

Fig 3. Pseudo code of firefly algorithm.

3.3. EV power and its state of charge originally developed by Yang [25] based on three main ideas:

ev ev ev ev 1) The fireflies attract their mating partners.

+ + = (17)
ch, h dch, h i, h h
2) The bright firefly gets attracted towards brighter fireflies.
ev ev
P ev ev ev 3) If the firefly cannot find brighter fireflies, then it will move ran-
ch, h *P ch,min ch, h ch, h *P ch,max (18)
domly in the search space
ev ev
dch, h *P dch,min P ev
dch, h
ev ev
dch, h *P dch,max (19)
Like other optimization techniques such as artificial bee colony
Eev
h = Eev
h 1 +( ev ev
ch, h *P ch, h
ev ev
dch, h *P dch, h) (1 ev ev
h )*Etrv, h (20) (ABC), particle swarm optimization (PSO) and harmony search algo-
rithm (HSA), which are based on the population of search space, FA is
Eev
min Eev
h Eev
max (21) also a population-based optimization algorithm. However, FA is dis-
where tinguished from the other optimization techniques by adjusting the
ch, h ,
and are the binary statuses for the charging,
ev ev ev
dch, h i, h
discharging and idle state of EV respectively. hev represents the EV parameters, which have low dependency on the algorithm and appro-
status when it is connected to the distribution system. P ev ch, h and are the
priately identifying the search space. The above mentioned three ideas
EV charging and discharging power, Eev h represents the present EV of firefly algorithm are explained in the mathematical form below:
battery capacity at specific time and Etrv
ev
, h is the energy associated with
EV travelling. The maximum charging and discharging power are spe- 5.1. Separation between fireflies
cified by constraint (18) and (19) whereas the maximum and minimum
SOC limit is defined by (21). The flowchart of the proposed charging The distance between two mating fireflies in the search space is
schedule for EV is shown in Fig 2. This study executes forward back- calculated as vector operation performed in Cartesian framework be-
ward sweep load flow for the distribution system with network load and tween ith and jth firefly given by the expression:
EV charging load. The power losses and bus voltages are computed for s
the distribution system. The algorithm fitness function is evaluated and rij = Yi Yj = (Yi, D Y j, D ) 2
minimum operating cost is obtained for the objective function in Eq (6). D=1 (23)
The intensity and position of fireflies represents the optimal number of
where ris the distance between two fireflies, s is the dimension of
EVs charging within the defined set of constraints.
control vector, Yi, D/Yj, D are the Dth dimensions of Yi/Yj fireflies re-
spectively.
4. Performance Evaluation parameter
5.2. Attraction between firefly
4.1. Profit for EV user
The attraction of the fireflies decreases when the two mating fire-
The EV users are concerned for the profit from V2G application. In
flies moves in opposite direction and the separation between them in-
this study, the cost for battery degradation is deducted from the EV
creases. The attraction of the flies can be represented by the following
users profit. Therefore, EV users prefer to discharge during high tariff
expression:
periods only to get maximum profit. The profit equation is calculated as
ratio of the amount of power exchanged through V2G (PV2G, h) to the (r ) = 0 × exp( r m); m 1 (24)
total power discharged by EV.
where β(r) and β0 represents the attractiveness when the fireflies are at
(PV2G, h* ev ) Cbh, d the distance r and 0. γ is the coefficient of light absorbed by firefly and
Profit= 24 m is the number of fireflies taken as 2.
P ev
h = 1 dch, h (22)
5.3. Movement of the fireflies
5. Optimization algorithm
The fireflies move towards brighter fireflies. The movement be-
The firefly algorithm (FA) analyses the social behaviour of flies and tween the two fireflies, jth firefly (low intensity) towards the ith firefly
is similar to other meta-heuristic techniques. The algorithm was (high intensity) is given by mathematical expression:

5
M. Sufyan, et al. Electric Power Systems Research 184 (2020) 106307

Fig 4. 33 bus distribution system.

Table 1 tested on modified 33 bus medium voltage distribution network. The

Parameters of EV. test system is connected to the 12.66 kV substation. Moreover, each bus
Component Parameter Value
in the distribution network can accommodate 10 EVs except load bus
(bus 1) as shown in Fig 4. The total load of the system is 3.78 MW
Battery Capacity (kWh) 85 (active load) and 2.35 MVAR (reactive load). Three units of PV and WT
SOC minimum/maximum limit (%) 20/100 with total generation of 530 kW and 275 kW respectively are integrated
ch,max (kW)
P ev 20
in the distribution system. The maximum number of EV charged/dis-
ev
Pdch ,max (kW) 8
charged at each instant in the distribution system is 320. In addition, all
Efficiency (%) 0.9
the EVs are assumed to have the maximum capacity of 85 kWh with
initially completely discharged to model the system effectively. The
average distance EV user travels is 34 km per day which constitutes
Yj (t ) = Yj + 0 × exp( r m) × (Yi Yj ) + vj (25)
6 kWh of battery capacity [26,27]. Therefore, it is assumed that EV
vj = (rand 0.5) (26) battery discharges 3 kWh during travel from home to workplace and
3 kWh to return home. The parameters for EV battery are shown in
The first term of the Eq. (25) shows the present position of jth Table 1. The maximum transformer capacity of the distribution system
firefly. The second term represents intensity of brightness by which the is 4.12 MW and the voltage limit of each bus is set as 0.94 p.u.–1.04 p.u.
jth firefly is attracted towards ith firefly. However, the last term vj re- The charge coordination problem is formulated in MATLAB (R2016b)
presents the movement of jth firefly in the entire search space when it and run on the personal computer 2.6 GHz core i5 processor with 6 GB
cannot find fireflies with more intensity. The randomization parameter RAM.
δ is a constant value in the range of 0–0.5. The pseudo code of the The time horizon in this study is set as one day, with equal intervals
firefly algorithm is shown in Fig 3. of 1 h as dispatch period which divides the overall time into 24 periods.
The solar irradiation (Wh/m2) and wind speed (m/s) profiles are shown
6. Simulation setup in Fig 5. The solar irradiation and wind speed have been recorded at the
Wisma R&D, University of Malaya for a sunny day. It can be seen from
In this study, the proposed charging/discharging schedule of EV is the figure that the output power of PV is high at noon when solar

Fig 5. Solar irradiation and wind speed. Fig 6. Temperature data for the day.

6
M. Sufyan, et al.
Fig 7. Residential and commercial electricity tariff.
Fig 8. Residential and commercial load profile.
irradiation is at peak whereas the output power was zero during night
time. The power output for WT varied with the speed of wind at each
time interval and fluctuations are observed within the certain range.
The hourly temperature for the entire day is shown in Fig 6. The hourly
electricity price for the residential and commercial sector are shown in
Fig 7. The commercial tariff is divided into three periods: peak, mid-
peak and off peak whereas the residential tariff consists of peak and off-
peak periods.
Fig 8 shows the actual commercial and residential load of one day. It
can be observed that commercial load rises significantly from 7:00 in
the morning. The peak load exists from 12:00 noon to 16:00 in the
afternoon. On the other hand, peak load in residential sector is from
19:00 to 22:00. In addition, the residential load rises in the morning
during 8:00–9:00. Therefore, proposed method allows minimum EVs to
charge during the peak hours such that distribution system may not face
any overloading issue utilizing the maximum capacity of the power
system.
The impact of coordinated charging/discharging of EV in the dis-
tribution system is analysed by considering three scenarios. The power
dispatch and the total operating cost is compared for each scenario. The
scheduling scenarios are described as:
Scenario1
In this scenario the load model is established under uncoordinated
charging/discharging schedule of EVs. The electricity cost and
system stability is analysed when all the EVs start charging after
returning home. This scenario serves as a reference case when in-
vestigating the effectiveness of algorithm.
Scenario 2
The coordinated charging/discharging schedule of EVs is modelled
in this scenario. The load model ensures that bus voltage does not
fall below the minimum voltage level. The electricity cost
7
Electric Power Systems Research 184 (2020) 106307
Fig 9. EV charging power for different scenarios.
consumption in this scheduling scenario is analysed and compared
with the scenario 1.
Scenario 3
The impact of renewable energy (photovoltaic and wind turbine) is
considered in this scheduling scenario. The electricity cost of the
system with coordinated charging/discharging incorporating the
renewable energy is analysed and the system stability is compared
with other scenarios. The optimal location for the PV and WT system
is located at buses 31, 32 and 33 based on previous work in [28].
7. Results and discussion
The power profiles of charging and discharging of EV under dif-
ferent scheduling scenarios is shown in Fig 9. The time period in this
study starts from 19:00 after EV returns home in the evening till 18:00
on the next day. During uncoordinated charging, the charging power is
high at the start of the time period. The charging power drops to zero
after the EV is completely charged. Similarly, during the day time when
the electricity tariff is high, all the EVs start discharging, which is re-
presented by negative values in the figure. The battery degradation cost
under this scenario increases due to continuous discharge even at high
DOD. On the other hand, the charging power under scenario 2 and 3 is
quite low when coordinated charging is performed. Unlike scenario 1,
all the EVs are not connected to the charging station at 19:00 main-
taining the voltage profile within limits whilst serving high residential
load demand at this time. Moreover, the charging power is almost
uniformly distributed over the night time ensuring the distribution
system is not overloaded. Similarly, when EVs start discharging during
day time, the discharging power is low in scenario 3 compared to
scenario 2 at some of the instances. This is because the renewable en-
ergy generation is high during day time which reduces the overall load
of the distribution system. EV owners consider incentives while
Fig 10. Voltage magnitude of the weakest bus.
M. Sufyan, et al.
Fig 11. Power generation profile for different scenarios.
Fig 12. EV owner profit at each bus of network.
discharging and if there is no profit in discharging, they are unwilling to
discharge as the battery degradation cost increases with increase in
discharge cycles.
The minimum bus voltage of the distribution system is shown in
Fig 10 for all the scheduling scenarios. It is evident that the voltage
level is violated when uncoordinated charging is performed. The bus
voltage drops below 0.94 p.u. when EV starts charging between 19:00
to 22:00. The bus voltage remains within the voltage limit during co-
ordinated charging in scenario 2 and 3.
The power generation profiles between the distribution system and
the power grid is shown in Fig 11 for coordinated and uncoordinated
scheduling. The peak load in uncoordinated scenario is twice as com-
pared to coordinated scenarios because of high EV charging power in
the evening when EVs return homes. Substantial power difference be-
tween charging and discharging creates stability issues in the dis-
tribution system. In the coordinated scenarios the charging power does
not increase above the maximum transformer capacity improving the
stability of the system. The power exchange in scenario 3 is less com-
pared to scenario 2 due to the addition of RES which decreases the
operating cost of the system.
It has been established that suitable incentives are requisite of en-
couraging EV owners to plug in their vehicles, in this regard the profit
to EV owner under coordinated scheduling for scenario 2 and 3 is
Table 2
System cost and profit for different scenarios.
Cases Cost ($)
Uncoordinated charging discharging 22,512.3
Coordinated charging 25,502.6
Coordinated charging discharging 22,292.5
Coordinated charging with RES 22,138.5
Coordinated charging discharging with RES 18,876
Electric Power Systems Research 184 (2020) 106307
shown in Fig 12. The profit of each EV owner varies depending on the
charging and discharging power. The maximum profit of single EV
owner at each bus of the distribution system is represented in Fig 12.
The profit for the EV user is calculated using Eq (22) when the EV user
discharges power to the grid. It can be observed that EV owners have
high profit in scenario 3 when renewable energy is integrated in the
system. However, Fig 8 shows that EV discharges more power in sce-
nario 2 as compared to scenario 3. The battery degradation cost in-
creases with high power discharge reducing the profit for EV user.
Hence, with integration of renewable energy EV discharges less power
during the times when the tariff is low making it an optimal scenario
where the EV owners earn maximum profit and experience relatively
less battery degradation.
The cost analysis for the V2G implementation under different
scheduling scenarios is presented in Table 2. The total cost of the system
when coordinated charging/discharging is performed with renewable
energy integration is lowest compared to other scenarios. The cumu-
lative profit of all EV owners is also reported in the table. The profit is
relatively low with uncoordinated scenario and increases when co-
ordinated charging/discharging is applied. However, EV users have
maximum profit with the integration of renewable energy to the
system. This is because in scenario 3, load of the system is minimized
through energy generated from RES, due to which EV owners do not
discharge during the low tariff times.
The maximum lifecycle of EV battery in all the above defined sce-
narios is also computed in the Table 2. Battery lifecycle is highest in
scenario 3 and lowest in uncoordinated scenario. With the assumption
that battery performs one complete cycle each day, the lifetime (years)
of the battery is dependent on the number of cycles the battery can
sustain until it reaches its end of life. As reported in the table, the
battery can last longer in scenario 3 due to minimal V2G exchange
compared to scenario 2, saving the battery replacement cost for the
system. To analyse the significance of V2G technology, the system cost
for coordinated charging and coordinated charging with RES is also
shown in the Table 2. The cost of the system increases excessively when
EVs are not allowed to discharge and EV owners cannot earn any rev-
enue. However, due to no V2G the profit to EV user is zero and impact
of battery degradation is not computed for this case.
The electric vehicle charging schedule and system cost is examined
for different cases of renewable energy integration in the distribution
network.
7.1. Impact of renewable energy penetration
To test the resilience of the distribution network, renewable energy
(WT and PV) is increased by 30% for high RES penetration and is de-
creased by 30% for low RES penetration. The renewable energy pene-
tration affects the charging / discharging schedule of EVs and the total
cost of system. Table 3 reports the system cost and the cumulative profit
by EV owners. The maximum lifecycle of single EV battery and its
lifetime is also shown in Table 3. It can be seen from the table the cost
of the system reduces with high penetration of renewable energy,
earning more profit for EV owners. On the other hand, when the re-
newable energy penetration is low, the cost of the system increases,
reducing the profit of EV owner. The lifetime of battery storage also
increases proportionally with the penetration of RES output. However,
Cumulative EV Profit ($/kWh) Lifecycle (cycle) Lifetime (year)
23.85 2213 6
– – –
62.88 2788 7.6
– – –
68.35 3276 9
8
M. Sufyan, et al. Electric Power Systems Research 184 (2020) 106307

Table 3
Cost analysis for different cases of scenario 3.
Cases Cost ($) Cumulative EV Profit ($/kWh) Lifecycle (cycle) Lifetime (year)

Coordinated charging discharging with high RES penetration 17,949 70.63 3614 10
Coordinated charging discharging with low RES penetration 19,825.5 67.99 3093 8.4
Coordinated charging discharging with different EV capacities 19,174.4 20.75 2446 6.7
Coordinated charging discharging with long traveling distance 19,519.2 46.01 3055 8.3
Coordinated charging discharging with different travel time 18,949.7 58.06 3133 8.6
Coordinated charging discharging with different RES location 19,028.7 61.18 3182 8.7
Coordinated charging discharging with uncertainty in renewable power 19,945.8 56.37 2956 8.1
Coordinated charging discharging with uncertainty in EV availability 20,148.3 51.74 3249 8.9

the system cost and EV profit for all cases of RES penetration is better
than scenario 2 as shown in Table 2. The reduction in cost compared to
scenario 2 is due to the fact that RES reduces the load of the distribution
network during the day time and EV owner discharges during the times
having maximum profit.

7.2. Impact of different EV capacities

The cost of the system varies with different EV battery capacities. To

analyse the impact of battery capacity, different models of EV are taken
in to account. The analysis in this subsection considers Nissan Leaf
40 kWh, BMWi3 42 kWh, Volkswagen e-Golf 35 kWh, Citroën C Zero
16 kWh and BYD e6 61 kWh with equal participation of 20% each.
[29,30]. The overall cost and losses are reduced and compared to the
case in which all the EVs have 85 kWh battery because small batteries Fig 13. Solar Irradiation and wind speed for a cloudy day.
require less power and time to fully charge. In addition, the profit to the
EV owner is remarkably low because EV users with small battery ca-
pacity are not willing to sell energy to the grid as the financial gain of
selling energy is low.

7.3. Impact of travelling distance

EV loses its energy when travelling and its battery degrades with
high travelling distance. The driving distance in this subsection is as-
sumed to be twice of the average distance EV travels (68 km/day). As
expected, the system cost increases with the distance, and EV owners
earn less profit because travelling distance increases battery degrada-
tion cost. In addition, EV needs to charge back again to sell energy to
the grid.

7.4. Impact of different travelling time

This study is conducted under the assumption that EVs travels from
home to workplace at 8 am and returns back home at 7 pm (19:00).
However, the travel-to-work time in Malaysia varies from 6 am to 8 am
in the morning and return time varies from 17:00 to 18:00 in the
evening. Therefore, in this subsection the impact of different travelling
time is analysed. It is assumed that 30% of EVs travel at 06:00, 30%
travels at 07:00 and the rest 40% travels at 08:00 from home to
workplace. On the way back from work, 60% of EVs travel at 17:00 and
40% travels at 18:00. The system cost increases with different travel
times because there are only 40% of EVs available to discharge during
peak hours at 17:00. Thus, the total profit for EV owners also reduces.
In the morning, when EVs reach early to workplace, EV owners are
unwilling to sell energy due to low electricity tariff.
7.5. Impact of RES location in the distribution system
To investigate the impact of locations of RES, wind turbine and
photovoltaic plants are located at different nodes of the distribution
network as identified by [31]. The load flow analysis is executed to
calculate the total losses and system cost as shown in Table 3. It can be
depicted from table that EV owners earn less profit compared to the
Fig 14. The uncertainty of EV availability.
base case reported in Table 2 when the DGs are connected at the end
buses 31–33.
7.6. Impact of uncertainty in weather conditions
The uncertainties related to the renewable energy are considered in
this section. The intermittent nature of renewable energy can affect the
system cost if the solar irradiation drops during peak time. Moreover,
change in wind energy can also impact the system cost due to increase
in losses. The uncertainty of solar irradiation is shown in Fig 13, when
PV power drops due to cloudy weather. Fig 13 also shows the change in
wind speed due to weather condition. The uncertainty of weather is
recorded at Wisma R&D, University of Malaya on a cloudy day.
7.7. Impact of uncertainty by EV availability
EV availability is also major concern for the charging and dischar-
ging power. In the above analysis, it was assumed that all the EVs are
available and the algorithm decides the number of EVs charging and
discharging based on the system load and RES. However, in this

9
M. Sufyan, et al.
Table 4
Comparison of different algorithms.
Algorithms Cost ($) Cumulative EV Profit ($/kWh)
Artificial bee colony 19,357 64.35
Harmony search algorithm 20,828.7 60.18
Particle swarm optimization 20,541.2 60.96
Firefly algorithm 18,876 68.35
Fig 15. System losses for different cases.
subsection the EV availability is selected as random by algorithm. The
algorithm allows the EV to charge the battery in duration of four hours
to maximum. However, by random selection there are possibilities that
EV cannot charge to maximum level. Fig. 14 shows the histogram of EV
availability for the entire day. The microgrid cost is affected by the EV
availability during V2G exchange as the lower EVs would put more load
on the grid and increase system losses.
7.8. Comparison of proposed optimization algorithm with other algorithms
The robustness of the firefly algorithm (FA) is analysed by im-
plementing artificial bee colony (ABC), harmony search algorithm
(HSA) and particle swarm optimization (PSO) for the proposed method
and the results are reported in Table 4. The base case with scenario 3 is
executed to compare the feasibility of different algorithms. It is evident
from the table that FA has lowest cost with high EV profit and minimum
losses compared to other optimization techniques. The parameters of
the different algorithms are also reported in Table 4 where the equa-
tions for the optimum solutions of ABC, HSA and PSO are taken from
[32, 33] and [34] respectively.
The system losses for total time horizon in each of the above defined
cases are shown in Fig 15. The system losses are minimum in case 4
when EVs with different battery capacities are modelled. This is be-
cause the overall charging power in this case is low, minimizing the
generation power exchanged between the grid and distribution system.
However, the operating cost is high in this case due to low V2G ex-
change. The system losses are also reduced in case 1 when the RES
penetration is increased with lowest operating cost. However, the losses
escalate when the RES penetration is decreased. Hence, the integration
of RES reduces the system cost and total losses and maximizes the profit
for EV owners.
10
Electric Power Systems Research 184 (2020) 106307
Losses (kW) Computational time (sec) Parameters
3018 284 = 0.7
r = 0.5
l=2
3347 249 w = 0.6
= 0.7
HMCR = 0.9
3546 267 c1 = 2.5
c2 = 1.9
w = 0.95
2831 214 o = 2
= 0.2
=1
8. Conclusion
This paper investigated the impact of charge coordination of EV
with V2G technology to maximize the profit for EV owners and
minimize the system cost. The real time battery degradation cost is
modelled by considering the DOD at each time interval. The proposed
problem was examined on modified 33 bus distribution system with
ten EVs connected at each bus and three wind turbine and photo-
voltaic systems. Three scenarios that involve economic scheduling:
uncoordinated charging, coordinated charging and coordinated char-
ging with RES are simulated with the residential and commercial load
of the system. The results reveal that voltage profile of the test system
has improved by coordinated charging compared to uncoordinated
charging, whereas the system cost is reduced when renewable energy
is integrated with coordinated charging. Moreover, it is also evident
from results that the lifetime of the battery can be prolonged with
renewable energy scenario as the EV owners are not willing to sell
energy during low tariff periods to maximize their profit.
Furthermore, the impact of different cases on the system cost and
losses have been analysed for the third scenario. The results demon-
strate that total cost and losses are reduced with high penetration of
RES. The energy losses are minimized for different EV capacities.
However, due to minimal power sharing between EV and the grid, the
profit for EV owner in this case is low and system cost is high.
Author's contribution
The study is different from prior research works in the following
aspects that reflect its contributions.
• The project optimizes the economic power dispatch in V2G en-
vironment by incorporating the battery degradation model. The
presence of RES, centralized charge coordination of EVs and dis-
tributed V2G batteries power dispatches will be considered in the
economic dispatch problem.
• Customarily, large majority of the decision making in battery energy
management system is simply based on the availability and deficit of
renewable energy respectively to meet the demand. However, this
study proposes the optimized solution for battery energy manage-
ment system by taking into account the battery degradation cost. In
addition, the battery degradation cost is modelled for real time
analysis considering DOD at each time interval and deep discharges
are avoided to prolong the battery lifespan.
• The operation cost and energy losses are investigated for different
RES penetration levels, EV battery capacities and uncertainties in
the renewable energy and EV availability in distribution network.
• In the previous works, the V2G is not economically profitable as the
battery lifetime deteriorates and degradation cost is high. However,
the battery scheduling in this study considers profit as key factor for
M. Sufyan, et al.
EV users during V2G exchange while taking battery degradation cost
into consideration.
Declaration of Competing Interest
None.
Acknowledgement
The authors thank the technical and financial assistance of UM
Power Energy Dedicated Advanced Centre (UMPEDAC), the Higher
Institution Centre of Excellence (HICoE) Program Research Grant,
UMPEDAC - 2018(MOHE HICOE – UMPEDAC), Ministry of Education
Malaysia, TOP100UMPEDAC and RU005-2015.
References
[1] X. Dong, et al., A charging pricing strategy of electric vehicle fast charging stations
for the voltage control of electricity distribution networks, Appl. Energy 225 (2018)
857–868 2018/09/01/.
[2] S. Falahati, S.A. Taher, M. Shahidehpour, A new smart charging method for EVs for
frequency control of smart grid, Int. J. Electr. Power Energy Syst. 83 (2016)
458–469 2016/12/01/.
[3] Q. Fu, A. Hamidi, A. Nasiri, V. Bhavaraju, S.B. Krstic, P. Theisen, The role of energy
storage in a microgrid concept: examining the opportunities and promise of mi-
crogrids, IEEE Electrif. Mag. 1 (2) (2013) 21–29.
[4] M. Shakeri, et al., An intelligent system architecture in home energy management
systems (HEMS) for efficient demand response in smart grid, Energy Build. 138
(2017) 154–164 2017/03/01/.
[5] R.J. Bessa, M.A. Matos, Economic and technical management of an aggregation
agent for electric vehicles: a literature survey, Int. Trans. Electr. Energy Syst. 22 (3)
(2012) 334–350.
[6] X. Hu, Y. Zou, Y. Yang, Greener plug-in hybrid electric vehicles incorporating re-
newable energy and rapid system optimization, Energy 111 (2016) 971–980 2016/
09/15/.
[7] N.B. Arias, J.F. Franco, M. Lavorato, R. Romero, Metaheuristic optimization algo-
rithms for the optimal coordination of plug-in electric vehicle charging in dis-
tribution systems with distributed generation, Electr. Power Syst. Res. 142 (2017)
351–361 2017/01/01/.
[8] M.J.E. Alam, K.M. Muttaqi, D. Sutanto, A controllable local peak-shaving strategy
for effective utilization of PEV battery capacity for distribution network support,
IEEE Trans. Ind. Appl. 51 (3) (2015) 2030–2037.
[9] H. Xing, M. Fu, Z. Lin, Y. Mou, Decentralized optimal scheduling for charging and
discharging of plug-in electric vehicles in smart grids, IEEE Trans. Power Syst. 31
(5) (2016) 4118–4127.
[10] H.N.T. Nguyen, C. Zhang, M.A. Mahmud, Optimal coordination of G2V and V2G to
support power grids with high penetration of renewable energy, IEEE Trans.
Transp. Electrif. 1 (2) (2015) 188–195.
[11] E. Mortaz, J. Valenzuela, Microgrid energy scheduling using storage from electric
vehicles, Electr. Power Syst. Res. 143 (2017) 554–562 2017/02/01/.
[12] R. Mehta, D. Srinivasan, A.M. Khambadkone, J. Yang, A. Trivedi, Smart charging
strategies for optimal integration of plug-in electric vehicles within existing dis-
tribution system infrastructure, IEEE Trans. Smart Grid 9 (1) (2018) 299–312.
11
Electric Power Systems Research 184 (2020) 106307
[13] Z. Luo, Z. Hu, Y. Song, Z. Xu, H. Lu, Optimal coordination of plug-in electric ve-
hicles in power grids with cost-benefit analysis—part I: enabling techniques, IEEE
Trans. Power Syst. 28 (4) (2013) 3546–3555.
[14] H. Zhang, Z. Hu, Z. Xu, Y. Song, Evaluation of achievable vehicle-to-grid capacity
using aggregate PEV model, IEEE Trans. Power Syst. 32 (1) (2017) 784–794.
[15] H. Lin, Y. Liu, Q. Sun, R. Xiong, H. Li, R. Wennersten, The impact of electric vehicle
penetration and charging patterns on the management of energy hub – a multi-
agent system simulation, Appl. Energy 230 (2018) 189–206 2018/11/15/.
[16] Y.R. Rodrigues, A.Z. de Souza, P. Ribeiro, An inclusive methodology for plug-in
electrical vehicle operation with G2V and V2G in smart microgrid environments,
Int. J. Electr. Power Energy Syst. 102 (2018) 312–323.
[17] S. Tabatabaee, S.S. Mortazavi, T. Niknam, Stochastic scheduling of local distribu-
tion systems considering high penetration of plug-in electric vehicles and renewable
energy sources, Energy 121 (2017) 480–490 2017/02/15/.
[18] K. Czechowski, "Assessment of profitability of electric vehicle-to-grid considering
battery degradation," ed, 2015.
[19] X. Lu, K. Zhou, S. Yang, H. Liu, Multi-objective optimal load dispatch of microgrid
with stochastic access of electric vehicles, J. Clean. Prod. 195 (2018) 187–199
2018/09/10/.
[20] C.S. Antúnez, J.F. Franco, M.J. Rider, R. Romero, A new methodology for the op-
timal charging coordination of electric vehicles considering vehicle-to-grid tech-
nology, IEEE Trans. Sustain. Energy 7 (2) (2016) 596–607.
[21] K.K. Mehmood, S.U. Khan, S.-J. Lee, Z.M. Haider, M.K. Rafique, C.-H. Kim, Optimal
sizing and allocation of battery energy storage systems with wind and solar power
DGs in a distribution network for voltage regulation considering the lifespan of
batteries, IET Renew. Power Gen. 11 (10) (2017) 1305–1315.
[22] H. Borhanazad, S. Mekhilef, V.G. Ganapathy, M. Modiri-Delshad, A. Mirtaheri,
Optimization of micro-grid system using MOPSO, Renew. Energy 71 (2014)
295–306.
[23] M. Sufyan, N.A. Rahim, C. Tan, M.A. Muhammad, S.R.S. Raihan, Optimal sizing and
energy scheduling of isolated microgrid considering the battery lifetime degrada-
tion, PLoS ONE 14 (2) (2019) e0211642.
[24] C. Zhou, K. Qian, M. Allan, W. Zhou, Modelling of the cost of EV battery wear due to
V2G application in power systems, IEEE Trans. Energy Convers. 26 (4) (2011)
1041–1050.
[25] X.-S. Yang, Nature-inspired Metaheuristic Algorithms, Luniver press, 2010.
[26] E. Gustafsson and F. Nordström, "Impact of electric vehicle charging on thedis-
tribution grid in Uppsala 2030," ed, 2017.
[27] R. Gough, C. Dickerson, P. Rowley, C. Walsh, Vehicle-to-grid feasibility: a techno-
economic analysis of EV-based energy storage, Appl. Energy 192 (2017) 12–23.
[28] R.S. Rao, K. Ravindra, K. Satish, S.V.L. Narasimham, Power loss minimization in
distribution system using network reconfiguration in the presence of distributed
generation, IEEE Trans. Power Syst. 28 (1) (2013) 317–325.
[29] Electric vehicle database (2019). Available: https://ev-database.org/.
[30] BYD E6 (2019). Available: https://www.carmudi.com.ph/newcars/brands/
byd/e6/.
[31] R. Li, W. Wang, M. Xia, Cooperative planning of active distribution system with
renewable energy sources and energy storage systems, IEEE Access 6 (2018)
5916–5926.
[32] K. Sundareswaran, P. Sankar, P. Nayak, S.P. Simon, S. Palani, Enhanced energy
output from a PV system under partial shaded conditions through artificial bee
colony, IEEE Trans. Sustain. Energy 6 (1) (2015) 198–209.
[33] E. Khorram, M. Jaberipour, Harmony search algorithm for solving combined heat
and power economic dispatch problems, Energy Convers. Manage. 52 (2) (2011)
1550–1554.
[34] T. Kerdphol, K. Fuji, Y. Mitani, M. Watanabe, Y. Qudaih, Optimization of a battery
energy storage system using particle swarm optimization for stand-alone micro-
grids, Int. J. Electr. Power Energy Syst. 81 (2016) 32–39.