Renewable Energy xxx (2014) 1e8

Contents lists available at ScienceDirect

Renewable Energy
journal homepage: www.elsevier.com/locate/renene

Coordinated charging of plug-in hybrid electric vehicles in smart

hybrid AC/DC distribution systems
M.F. Shaaban*, A.A. Eajal 1, E.F. El-Saadany 2
Department of Electrical and Computer Engineering, University of Waterloo, 200 University Avenue, West, Waterloo, ON, Canada N2L3G1

a r t i c l e i n f o a b s t r a c t

Article history: This paper introduces an online coordination approach for plug-in hybrid electric vehicles (PHEVs)
Received 25 May 2014 charging in smart hybrid AC/DC distribution systems. The goal of the proposed method is to optimally
Accepted 5 August 2014 charge the PHEVs in order to maximize the PHEV owners' satisfaction without violating the network
Available online xxx
constraints. The charging costs, which represent the PHEV owners' satisfaction in this work, are based on
real-time pricing. The proposed approach includes consideration of PHEV owners' requirements, PHEV
Keywords:
batteries' characteristics, and hybrid distribution system limitations. Moreover, a sliding window concept
Coordinated charging
is introduced to facilitate managing the PHEV charging and the system interlinking converters in real-
Electric vehicles
Energy management
time. A 38-bus test system has been modified to include DC links and used to validate the developed
Hybrid AC/DC online charging scheme. The test results clearly demonstrate the effectiveness of the proposed method.
Smart distribution systems © 2014 Elsevier Ltd. All rights reserved.

1. Introduction violation, and fuses blowouts [2]. This is due to the fact that the
distribution system was not originally designed to accommodate
Over the last few decades, a growing body of evidence has such extra load and to the expectation that PHEVs will likely be
shown that global temperatures are rising due to greenhouse gases, clustered in specific geographical areas.
which will result in severe climate changes and rises in sea levels. Two categories of solutions have been proposed in the literature
Such predictions have made the reduction of gas emissions a vital as a means of facilitating the PHEV charging. The first focuses on
necessity, and governments around the globe are taking action to upgrading the grid infrastructure or deploying distributed gener-
minimize their emissions. A key component in the reduction of CO2 ation (DG) to meet the excess power demand [3]. The second
emissions is to shift to low- or zero-emission vehicles for trans- category, which is addressed in this paper, focuses on utilizing the
portation. To this end, the electrification of vehicles has become the two-way communication infrastructure of the smart grid to coor-
best option for reducing transportation emissions, given that the dinate the PHEV charging by shifting the excess power demand to
electric power system is best positioned to provide the requisite optimal periods during the day [4]. However, the second solution is
infrastructure for these vehicles. known to be more beneficial for both the PHEV owners and the
Concerns about emissions coupled with developments in elec- utilities [5].
tric vehicle technology have led to the expectation that electric
vehicles numbers will rise rapidly over the next few decades. This
1.1. Background
will result in reductions in emissions and fuel consumption, as well
as an increase in energy security by exploiting diverse energy
The PHEV, as defined by the IEEE, is any hybrid electric vehicle
sources [1]. If the plug-in hybrid electric vehicle (PHEV) charging is
(HEV) that has: (1) a battery storage bank of 4 kWh or more; (2) a
not managed properly, its extra load may lead to negative impacts,
rechargeable battery system that can be charged from an external
such as voltage limits violation, voltage imbalance, feeders' limits
energy source; and (3) the capability of driving at least 10 miles
(16.1 km) with no gasoline used [6]. PHEVs and plug-in pure electric
vehicles (PEVs) with the capability of being charged and discharged
* Corresponding author. Tel.:þ1 519 888 4567x33367. to inject the electricity back into the grid in what is known as a
E-mail addresses: [email protected] (M.F. Shaaban), [email protected]
(A.A. Eajal), [email protected] (E.F. El-Saadany).
vehicle to grid principle (V2G). V2G can bring benefits to the ve-
1
Tel.:þ1 519 888 4567x32872. hicles owners and the society. Examples of these benefits are: less
2
Tel.:þ1 519 888-4567x33035. reliance on gas, reductions in gas prices and emission levels.

http://dx.doi.org/10.1016/j.renene.2014.08.012
0960-1481/© 2014 Elsevier Ltd. All rights reserved.

Please cite this article in press as: Shaaban MF, et al., Coordinated charging of plug-in hybrid electric vehicles in smart hybrid AC/DC distribution
systems, Renewable Energy (2014), http://dx.doi.org/10.1016/j.renene.2014.08.012
2 M.F. Shaaban et al. / Renewable Energy xxx (2014) 1e8

According to Environmental Canada in its report entitled “green- However, the load of the PHEV is lumped and the preference of the
house gas inventory” published in 2007, the conventional vehicles PHEV's owner is not considered.
are responsible for approximately 12% of the total annual emissions The authors of Ref. [20] developed a two-level PHEV coordina-
in Canada [7]. tion scheme that operates on the market level and in real-time. The
Further, due to the recent advances of DC-based technologies work described in Ref. [21] resulted in fuzzy logic controllers for
such as electric vehicles, fuel cells, solar panels, and energy storage managing PHEV charging in real time. Two controllers are proposed
systems, distribution systems will require a tremendous number of at the substation level and at the charging stations level. However,
conversion devices which will in turn lead to operational and the optimality of the fuzzy logic based decisions is not guaranteed.
control challenges. Thus, the best infrastructure that the distribu- A greedy fair charging algorithm for PHEVs is proposed in Ref. [22]
tion system can have is of a hybrid AC/DC type, which can host the based on distributed control. The work assumed the charging de-
AC/DC technologies. Thus, it is envisioned that many local distri- cisions to be based only on local information at the charging loca-
bution companies (LDCs) will adopt the hybrid AC/DC infrastruc- tion. However, centralized controllers offer better performance and
ture. The hybrid AC/DC structure of power grids offers some observability of the system; thus, they are preferred by grid
advantages, which can be summarized in the following [8]: operators.
However, all the aforementioned studies in Refs. [14e22] have
 Reducing the number of conversion devices which are currently failed to account for the dynamics of PHEV arrivals and departures
used to connect the DC loads and DC energy sources to the host in real-time. The work in Ref. [23] presented a new approach for
AC network. PHEV coordination in smart grid based on queuing theory as means
 Reducing the conversion cost. of predicting the incoming vehicles to the parking lots. However,
 Simplifying the control and operation strategies. the work considered AC systems without any DG units.
 Controlling the harmonic injections efficiently by connecting Another aspect related to the PHEV charging, is the potential
the DC loads to the DC side of the hybrid system. consequences of the charging process such as voltage imbalance
 Solving the zero and negative sequence currents caused by and harmonic distortions. DC distribution systems can eliminate
unbalanced AC loads. these negative impacts of the PHEVs charging. As a step towards
adopting the concept of DC power distribution, a DC task team was
1.2. Related work formed to investigate the standards and regulations of DC distri-
bution systems [24]. In Ref. [25], the concept of having DC islands
Previous work in the area of PHEV coordination introduced within the existing AC infrastructure was introduced. The authors
different techniques to coordinate their charging. in Ref. [25] proposed a decentralized management system to con-
In Ref. [9], a methodology is presented to optimally coordinate trol the interlinking converters. However, the PHEVs have not
PHEV charging based on time-of-use energy pricing. An optimal considered as a major component in hybrid AC/DC distribution
charging/discharging approach is presented in Ref. [10] based on systems. In [26], the authors introduced the idea of the hybrid AC/
V2G strategy. An approach for online PHEV charging coordination DC intelligent parking lot having smart charging controllers,
is presented to minimize system losses and improve voltage renewable energy resources. The smart parking lot has the option
profile in Ref. [11]. In Ref. [12], a methodology is proposed for of charging the PHEVs through the DC energy sources and the AC
optimal PHEV charging based on demand side response and utility grid.
vehicle to battery (V2B) strategy. V2G ancillary services are Based on the aforementioned discussions, it is obvious that
managed in a charging/discharging method introduced in there is a need for a proper PHEV charging coordination method-
Ref. [13]. ology. Therefore, this paper proposes to develop a PHEV charging
However, the methods in Refs. [9e13] deal with current instant coordination approach in smart hybrid AC/DC distribution systems.
of the system without considering the future system states. To The paper is organized as follows. The problem description is
rectify this shortcoming, a few studies considered the future PHEV introduced in the next section. Sections 3 and 4 present the
demand in the charging process. The work in Refs. [14,15] takes into modeling and problem formulation, respectively. Finally, Sections 5
consideration the day-ahead system states to coordinate the PHEV and 6 present the simulation result and concluding remarks,
charging. The day-ahead resource and reserve scheduling problem respectively.
with demand response in uncertain environment was tackled in
Ref. [16]. The researchers in Ref. [16] are credited for including some 2. Problem description
critical aspects in the day-ahead energy resource scheduling such
as the stochastic natures of the wind energy and demand, network A centralized control center is assumed to be responsible for
constraints, reserve requirements, and cost/emission reductions. managing the PHEV charging and interlinking converters in a
Nevertheless, they failed to address some key components in the certain territory, as shown in Fig. 1. The following assumptions are
future smart grid, which are the ESS and the PHEV. In their later adopted in this work:
work [17], the authors considered the coordinated PHEV charging
in the day-ahead operational planning such that the operating cost  The real-time electricity price is announced by the utility.
and emissions were minimized. However, the energy resources  PHEV chargers allow only vehicles charging.
considered were only of a dispatchable type, which contradicts  Two-way communication infrastructure is available.
with the fact that the future smart grid will be mostly renewable-  The charging takes place at level 2 (240 V, 30 A).
based.  The AC voltage at the AC bus connected to the converter is
The authors in Ref. [18] proposed a day-ahead approach to purely sinusoidal and balanced.
manage the grid congestion through optimal charging/discharging  The converter and its transformer are lossless.
of PHEVs. The work assumed contract agreements between the  The DC current and voltage are ripple-free.
PHEV fleet operator and the distribution system operator. The work
in [19] developed an economic dispatch approach on the power The control center collects system information and searches for
system level with V2G capability. The work examined different the optimal decisions. Then, these optimal decisions are sent to
PHEV penetration levels on the economic dispatch problem. each system element to be implemented. The cycle of collecting

Please cite this article in press as: Shaaban MF, et al., Coordinated charging of plug-in hybrid electric vehicles in smart hybrid AC/DC distribution
systems, Renewable Energy (2014), http://dx.doi.org/10.1016/j.renene.2014.08.012
 M.F. Shaaban et al. / Renewable Energy xxx (2014) 1e8 3

Fig. 1. A smart hybrid AC/DC distribution system.

information, optimizing the decisions, and sending the decision is follows: first, the EMS central controller request information from
performed during time dt, as shown in Fig. 2. During dt, which is distributed energy resources (DER) local controllers, parking lots
the time step of the operation, the power consumed or delivered by aggregators, and SCADA monitoring points in the system. This in-
different system elements is assumed to be constant. The time step formation includes the real-time measurements and forecasted
dt should be as short as possible to enhance the accuracy of the data if applicable (e.g. forecasted PV-based DG output power).
proposed approach; however, the computational time is a burden Further, the central controller chooses a proper unified sliding
on this time step. window for all system components. Then, the optimization solver is
Moreover, to consider the dynamics of PHEV arrivals and de- used to generate the optimal decisions and settings for various
partures in a parking lot, the moving time window concept is system components. The entire procedure is repeated every time
adopted, as shown in Fig. 2, where the forecasted information, i.e. step dt.
demand and electricity price, during the specified time window
affects the decisions in each time step. While the time step dt can
be in the range of seconds or minutes, the moving time window 3. Modeling
duration T can be in the range of a few hours. The PHEV power
demand is shifted to optimal periods within this time window T. In this section, two types of models are presented, namely load
Thus, longer T leads to more flexibility in the demand coordina- models and grid models. While the normal load models and the AC
tion process. However, two aspects must be considered in grid models have been discussed in several publications, this sec-
choosing the duration of T: the computational time of the opti- tion introduces the PHEV load model and the DC grid model for
mization process, which must be performed within time dt, and load flow application in this study.
the accuracy of the forecast, especially if the system contains
renewable DG.
The procedure of the energy management system (EMS) real- 3.1. PHEV modeling
time operation is illustrated in Fig. 3, and can be explained as
As shown in Fig. 4, the DC grid is connected to the battery pack
of the PHEV via DC/DC converter [27]. The DC/DC converter aims to
maintain either constant current or constant voltage at the termi-
nals of the battery pack depending on the level of state of charge
(SOC), as shown in Fig. 5 [28].
Since, the charging power is based only on the SOC level, and not
affected by the DC grid voltage, due to the buffering via the DC/DC
converter, the PHEV charging load is implemented as constant
power load in the DC distribution network. The power consumed
by each PHEV in the system under study is assumed to follow the
Fig. 2. Sliding time window concept. same function shown in Fig. 5 [28] for simplicity.

Please cite this article in press as: Shaaban MF, et al., Coordinated charging of plug-in hybrid electric vehicles in smart hybrid AC/DC distribution
systems, Renewable Energy (2014), http://dx.doi.org/10.1016/j.renene.2014.08.012
4 M.F. Shaaban et al. / Renewable Energy xxx (2014) 1e8

Fig. 5. Li-ion battery charging characteristics with 7.2 kW charger [28].

steady-state or dynamic study. The steady-state model of the

converter is shown in Fig. 6. Mathematically, the converter can be
represented by a set of equations shown in Eqs. (1)e(5) [30].

fci1 ¼ Vdci  kbi ai jVi jcos fi ¼ 0 (1)

3
fci2 ¼ Vdci  kbi ai jVi jcos ai þ Idci Xci ¼ 0 (2)
p

X
nDC
fci3 ¼ PGi  PDi  Vi Vk Gik  Vdci Idci ¼ 0 (3)
k¼1

spec
fci4 ¼ Vdci  Vdc ¼0 (4)
i

fci5 ¼ ai  aspec
i
¼0 (5)

where Idci ; Vdci are the DC current and voltage of converter i; Vi is

the AC voltage of converter i; fi is the phase angle between the AC
Fig. 3. EMS flow chart.
voltage and current of converter i; ai is the transformer tap ratio of
converter i; Xci is the commutation reactancep offfiffifficonverter i; kbi is the
3.2. Converter modeling parameter of converter i defined by: ½ð3 2=pÞki  and ki is a
parameter above 0.99 for normal operating conditions.
The converters in a hybrid AC/DC system must be coordinated Eqs. (1)e(5) can be expressed in a compact form as shown in Eq.
with the utility grid to ensure an interruptible and a high quality (6).
power supply under variable solar and wind power generations
[29]. Thus, the converter model needs a special care to conduct any fci ðVi ; xi Þ (6)

where xi is the vector of converter variables at bus i and defined by

Eq. (7).
 
xi ¼ Vdci ; Idci ; ai ; ai ; fi (7)

4. Mathematical formulation

The ultimate goal is to minimize the charging expenses for the

PHEV's owners by optimizing the charging decision, u, in corre-
spond to the real-time electricity pricing such that the constrains
imposed by the system and the PHEVs' owners are satisfied. The
PHEV charging problem is formulated as a nonlinear optimization
Fig. 4. PHEV charging system in DC network [27]. problem and mathematically described by (8)e(16).

Please cite this article in press as: Shaaban MF, et al., Coordinated charging of plug-in hybrid electric vehicles in smart hybrid AC/DC distribution
systems, Renewable Energy (2014), http://dx.doi.org/10.1016/j.renene.2014.08.012
 M.F. Shaaban et al. / Renewable Energy xxx (2014) 1e8 5

4.1. Objective function CH

uj;t Pj;t 
dt
SOCj;tþ1 ¼ SOCj;t þ
EjBAT 60
The objective of the proposed charging scheme is to minimize  (13)
the charging expenses as shown in Eq. (3). T
cj2f1; 2; …; nEV g; t2 1; 2; …;
dt
X
nB .
EV
min Costt ¼ r  Pi;t  Dt 60; ct2f1; 2; …; T=dtg (8)
u2½0;1
i¼1 charging cost at time t ($); r is the charging cost
SOC requirements constraints: the battery pack SOC level should
where Costt is the
EV is the
reach to the specified SOCspec during the parking period to satisfy
per kWh ($/kWh); nB is the total number of buses; Pi;t
the customers. The SOCspec is assumed to be set by the vehicle's
aggregated PHEV demand at time t; dt is the time step size; u is the
owner.
charging decision variable.
The objective is subject to a set of equality and inequality con- SOCj;t ¼ SOCspec cj2f1; 2; …; nEV g; t ¼ TjPRK (14)
j
straints, which are explained as follows.
where TjPRK is the parking duration of vehicle j.

4.2. Equality constraints

4.3. Inequality constraints
The equality constraints involve the highly nonlinear AC power
The inequality constraints involve the AC/DC bus voltage and
flow constraints, the DC grid power flow constraints, and the ve-
line thermal limits as expressed by Eqs. (15) and (16), respectively.
hicles SOC update constraints.
The AC power flow equations: the active and reactive power in- 
  T
jections at any AC bus must satisfy Eqs. (9) and (10), respectively. Vmin  Vi;t   Vmax ci; t2 1; 2; …; (15)
dt
 X nAC      
G
Pi;t D
 Pi;t ¼ Vi;t  V Y cos q þ d  di;t
k;t i;k i;k k;t   T
(9) Si;k;t   SMAX ci; k; t2 1; 2; …; (16)
k¼1 i;k dt
ci2f1; 2; …; nAC g; t2f1; 2; …; T=dtg
where Vmin and Vmax are the lower and upper bound on the AC/DC
 X nAC      bus voltage respectively, and SMAX
i;k
is the maximum thermal limit of
G
Qi;t D
 Qi;t ¼ Vi;t  V Y sin q þ d  di;t
k;t i;k i;k k;t line iek.
k¼1 (10)
ci2f1; 2; …; nAC g; t2f1; 2; …; T=dtg 5. Test results and discussion
G and Q G are the active
where nAC is the total number of DC buses; Pi;t i;t The developed charging scheme was implemented within the
D and Q D are the
and reactive power generations at bus i at time t; Pi;t MATLAB® environment interfaced with the General Algebraic
 i;t 
active and reactive power demands at bus i at time t; Vi;t  is the Modeling System (GAMS). A 38-bus radial distribution system, with
voltage magnitude at bus i at time t; di;t is the voltage phase angle at a peak demand of 4.37 MVA and whose single-line diagram is
  depicted in Fig. 7, was used to validate the developed approach [31].
bus i at time t; Y  is the magnitude of the admittance in the bus
i;k
The test system was modified to form a hybrid AC/DC distribu-
admittance matrix; qi;k is the phase angle of the element of the bus tion system. This hybrid system, shown in Fig. 7, has DC segments
admittance matrix. (23e25,38) and (26e33) connected to the AC side through bidi-
The equality constraints also involve the converter equations rectional converters. The hybrid smart grid is connected to the main
(Eqs. (1)e(5)) so that the optimal settings of the converter pa- grid through the substation located at bus 1, which functions as a
rameters are determined. slack bus for the smart grid. The batteries capacities are assumed to
The DC power flow equations: the DC power injection at each DC have the same capacity which is 14 kWh in this study. The chargers
bus must satisfy (11). are of a two-level type with a capacity of 7.2 kW each. Four parking
X
nDC  lots are located at buses 23, 26, 33, and 38. The hourly real-time
G D EV T energy price for 24 h is shown in Fig. 8.
Pi;t Pi;t Pi;t ¼Vi;t Vk;t Gi;k ci2f1;2;…;nDC g; t2 1;2;…;
dt The time step is assumed to be 10 min, i.e., dt ¼ 10 min. On the
k¼1
(11) other hand, the sliding window T duration is assumed to be the
same as the parking duration of the vehicle with the latest depar-
where nDC is the total number of DC buses and Gi,k is the conduc- ture in the system. However, this duration T is limited to a
tance iek in the bus conductance matrix. maximum of 3 h or 18 time segments (3  60 min/10 min) to limit
It should be mentioned that the PHEV power demand depends the computational time.
on the charging decision, u2½0; 1, charger allowable delivered Two case studies are presented, which examine the perfor-
power PCH, and charger efficiency hCH as shown in (12). The mance of the proposed coordination scenario in the hybrid system
j
charging power level varies between 0 and the allowable charging for 20% and 30% PHEV load penetration respectively. In each case
power. study, the coordinated scenario is compared to the uncoordinated
scenario and the no PHEV scenario.
X
nEV . 
EV CH T
Pi;t ¼ Pj;t uj;t hCH
j ci2f1;2;…;n DC g; t2 1;2;…; (12)
j¼1
dt 5.1. 20% PHEV load penetration

SOC update constraints: the SOC level of the battery packs is This case represents 20% PHEV load penetration. Fig. 9(a, b)
updated during the time window based on the energy capacity EjBAT shows the total demand and PHEV demand for three scenarios,
of the battery of vehicle j. namely uncoordinated, coordinated and no PHEV scenarios. It is

Please cite this article in press as: Shaaban MF, et al., Coordinated charging of plug-in hybrid electric vehicles in smart hybrid AC/DC distribution
systems, Renewable Energy (2014), http://dx.doi.org/10.1016/j.renene.2014.08.012
6 M.F. Shaaban et al. / Renewable Energy xxx (2014) 1e8

Fig. 6. Converter model.

Fig. 8. Energy hourly price.

assumed that in the uncoordinated scenario, the PHEVs start

charging once plugged-in. The proposed approach delivers same
amount of charging energy to PHEVs as the uncoordinated Moreover, as a by-product, the system total peak load is reduced
approach by satisfying the customers' desired SOC, as in Eq. (14). by almost 8% in the coordinated scenario compared to the unco-
Thus, both the coordinated and uncoordinated achieved same level ordinated scenario as shown in Fig. 9(b).
of charging success (100%), which is attributable to the low PHEV
penetration level, which enables the system to charge all connected
PHEVs without violating the technical limitations. This can be 5.2. 30% PHEV load penetration
observed from the equivalent area under curve in Fig. 9(a).
However, the charging cost for customers are $1253.5 and In this case study, the total PHEV demand represents 30% of the
$1457.7 for the coordinated and uncoordinated scenarios, respec- regular load. As shown in Fig. 10, the uncoordinated charging
tively. Thus, the coordinated scenario saves almost 14% of the scheme delivered the total required energy to the customers;
charging cost for PHEV owners. This is achieved through allocating
the charging energy in low energy price periods subject to system
technical constraints, desired SOC level, initial SOC levels, and
parking duration. As shown in Fig. 9(a), the PHEV load in the co-
ordinated scenario peaks during low price periods (normal load off-
peak periods) and troughs during high price periods (peak normal
load periods).

Fig. 9. Consumption for 20% PHEV penetration scenario: (a) PHEV load consumption;
Fig. 7. 38-Bus smart hybrid AC/DC distribution system under study. (b) Total system demand.

Please cite this article in press as: Shaaban MF, et al., Coordinated charging of plug-in hybrid electric vehicles in smart hybrid AC/DC distribution
systems, Renewable Energy (2014), http://dx.doi.org/10.1016/j.renene.2014.08.012
 M.F. Shaaban et al. / Renewable Energy xxx (2014) 1e8 7

however, it violates the grid technical constraints (e.g. voltage charging energy in low pricing periods to achieve minimum
limits and/or thermal limits) from 8:40 p.m. to 11:50 p.m. charging cost for PHEV owners, without violating system technical
In contrast, the coordinated charging scheme resulted in constraints.
infeasible solution after 8:40 p.m. due to the constraint in (14), The proposed approach takes into consideration the PHEV
which entails that satisfying the customers' specified SOC levels owners' constraints, PHEV batteries' characteristics, and grid con-
SOCspec is not possible without violating the grid technical con- straints. Moreover, the proposed approach utilizes a sliding time
straints. Thus, the proposed problem formulation in Section 4 is window to include the effect of future system states and future
modified as follows to accommodate high PHEV penetrations. charging decisions on the implemented charging decisions.
During each time step, the maximum SOC that can be attained The proposed coordinated charging scheme is examined on a
by the system without violating the technical constraints is evalu- hybrid AC/DC distribution system and compared with an uncoor-
ated using the following nonlinear optimization problem: dinated charging scenario for low and high PHEV penetrations. The
results prove the effectiveness of the proposed approach, which
nEV 
X 
spec T significantly reduces the charging costs for the PHEV owners and
min SOCj  SOCj;t ct2 1; 2; …; (17)
u2½0;1 dt the peak demand for the electric utility.
j¼1
Moreover, the approach is modified for high PHEV penetration
Subject to (9)e(13), (15) and (16), while the constraints in (14) to include consideration of maximizing the SOC level of the vehicles
are replaced by: batteries limited to the customers' desired SOC. Thus, the proposed
PHEV charging approach aims to satisfy the PHEV owners in terms
SOCj;t  SOCspec
j
cj2f1; 2; …; nEV g; t ¼ TjPRK (18) of their batteries SOC levels and the charging costs.

Thus, the maximum SOC (limited by SOCspec )

is evaluated for
j
References
each time step. Denote this maximum SOC as SOCMAX j . Then, the
problem in Section 4.1 is solved while replacing the value of SOCspec
j [1] Eberle U, von Helmolt R. Sustainable transportation based on electric vehicle
in (14) by SOCMAX
j . This guarantees that the problem is feasible and concepts: a brief overview. Energy Environ Sci 2010;3:689e99.
the SOC levels are maximized without violating the grid technical [2] Liu R, Dow L, Liu E:A survey of PEV impacts on electric utilities. In: Innovative
smart grid technologies (ISGT), 2011 IEEE PES; 2011. pp. 1e8.
constraints. Then, as a second priority, the charging costs are [3] Shaaban MF, Atwa YM, El-Saadany EF. PEVs modeling and impacts mitigation
minimized for the customers. in distribution networks. IEEE Trans Power Syst 2013;28(2):1122e31.
The resulted PHEV consumption is shown in Fig. 10. The coor- [4] Khayyam H, Ranjbarzadeh H, Marano V. Intelligent control of vehicle to grid
power. J Power Sources 2012;201:1e9.
dinated charging scheme is able to deliver 96.18% of the total
[5] Verzijlbergh RA, Grond MO, Lukszo Z, Slootweg JG, Ilic MD. Network impacts
required energy to the vehicles. This value can be assumed as the and cost savings of controlled EV charging. IEEE Trans Smart Grid 2012;3:
success factor of the charging scheme. Although, the uncoordi- 1203e12.
nated charging delivered 100% of the required energy, it is infea- [6] Dyke KJ, Schofield N, Barnes M. The impact of transport electrification on
electrical networks. IEEE Trans Ind Electron 2010;57:3917e26.
sible due to the violation of the grid technical limits. The charging [7] Environment Canada. Canada's 2007 greenhouse gas inventory: a summary of
costs are $1741.24 and $2143.36 for the coordinated and uncoor- trends. Environment Canada; 2007.
dinated charging schemes. The 18.76% reduction in the charging [8] Wang P, Goel L, Liu X, Choo FH. Harmonizing AC and DC: a hybrid AC/DC
future grid solution. IEEE Power Energy Mag 2013;11:76e83.
costs is due to the amount of energy delivered, which is less by [9] Cao Y, Tang S, Li C, Zhang P, Tan Y, Zhang Z, Li J. An optimized EV charging
3.82% in the case of coordinated charging scheme, and the model considering TOU price and SOC curve. IEEE Trans Smart Grid 2012;3:
reshaping of the charging curve similar to the low penetration 388e93.
[10] Ota Y, Taniguchi H, Nakajima T, Liyanage KM, Baba J, Yokoyama A. Autono-
case. mous distributed V2G (vehicle-to-grid) satisfying scheduled charging. IEEE
Trans Smart Grid 2012;3:559e64.
[11] Deilami S, Masoum AS, Moses PS, Masoum MA. Real-time coordination of
6. Conclusions
plug-in electric vehicle charging in smart grids to minimize power losses and
improve voltage profile. IEEE Trans Smart Grid 2011;2:456e67.
This paper introduces a coordinated charging scheme for PHEVs [12] Pang C, Dutta P, Kezunovic M. BEVs/PHEVs as dispersed energy storage for
in smart hybrid AC/DC distribution systems. The coordinated V2B uses in the smart grid. IEEE Trans Smart Grid 2012;3:473e82.
[13] Quinn C, Zimmerle D, Bradley TH. An evaluation of state-of-charge limitations
approach relies on the two-way communication infrastructure and actuation signal energy content on plug-in hybrid electric vehicle, vehicle-
under the paradigm of the smart grids. The proposed PHEV to-grid reliability, and economics. IEEE Trans Smart Grid 2012;3:483e91.
charging coordination approach aims to optimally allocate the [14] Derakhshandeh SY, Masoum AS, Deilami S, Masoum MAS, Golshan MEH.
Coordination of generation scheduling with PEVS charging in industrial
microgrids. IEEE Trans Power Syst Aug. 2013;28(3):3451e61.
[15] Soares J, Morais H, Sousa T, Vale Z, Faria P. Day-ahead resource scheduling
including demand response for electric vehicles. IEEE Trans Smart Grid March
2013;4(1):596e605.
[16] Zakariazadeh A, Jadid S, Siano P. Economic-environmental energy and reserve
scheduling of smart distribution systems: a multiobjective mathematical
programming approach. Energy Convers Manage 2014;78:151e64.
[17] Zakariazadeh A, Jadid S, Siano P. Multi-objective scheduling of electric vehi-
cles in smart distribution system. Energy Convers Manage 2014;79:43e53.
[18] Hu J, You S, Lind M, Ostergaard J. Coordinated charging of electric vehicles for
congestion prevention in the distribution grid. IEEE Trans Smart Grid March
2014;5(2):703e11.
[19] Ortega-Vazquez MA, Bouffard F, Silva V. Electric vehicle aggregator/system
operator coordination for charging scheduling and services procurement. IEEE
Trans Power Syst May 2013;28(2):1806e15.
[20] De Craemer K, Vandael S, Claessens B, Deconinck G. An event-driven dual
coordination mechanism for demand side management of PHEVs. IEEE Trans
Smart Grid March 2014;5(2):751e60.
[21] Singh M, Thirugnanam K, Kumar P, Kar I. Real-time coordination of electric
vehicles to support the grid at the distribution substation level. IEEE Syst J,
in press.
[22] Xia L; Mareels I, Alpcan T, Brazil M, de Hoog J, Julian DAT. A distributed electric
Fig. 10. PHEV load consumption for 30% penetration. vehicle charging management algorithm using only local measurements. In:

Please cite this article in press as: Shaaban MF, et al., Coordinated charging of plug-in hybrid electric vehicles in smart hybrid AC/DC distribution
systems, Renewable Energy (2014), http://dx.doi.org/10.1016/j.renene.2014.08.012
8 M.F. Shaaban et al. / Renewable Energy xxx (2014) 1e8

IEEE PES innovative smart grid technologies conference (ISGT); Feb 2014. [27] Monteiro V, Gonçalves H, Ferreira JC, Afonso JL. Batteries charging systems for
pp.1e5, 19e22. electric and plug-in hybrid electric vehicles; 2012.
[23] Shaaban MF, Ismail M, El-Saadany EF, Zhuang W. Real-time PEV charging/ [28] Marra F, Yang GY, Traholt C, Larsen E, Rasmussen CN, You S. Demand profile
discharging coordination in smart distribution systems. IEEE Trans Smart Grid study of battery electric vehicle under different charging options. Power and
July 2014;5(4):1797e807. energy society general meeting, 2012. IEEE; 2012. p. 1e7.
[24] Hesla E, Brusso B, Downey L, Giese B, Parise G, Valdes M, et al. DC task team [29] Liu X, Wang P, Loh PC. A hybrid AC/DC microgrid and its coordination control.
report. In: IEEE/IAS 49th industrial & commercial power systems technical IEEE Trans Smart Grid 2011;2:278e86.
conference (I&CPS); 2013. pp. 1e9. [30] Arrillaga J, Smith B. ACeDC power systems analysis. London: The Institute of
[25] Piegari L, Tironi E, Musolino V, Grillo S, Tornelli C. DC Islands in AC smart grids. Electrical Engineers (IET); 1998.
IEEE Trans Power Electron Jan. 2014;29(1):89e98. [31] Singh D, Misra R. Effect of load models in distributed generation planning.
[26] Ma T, Mohamed A, Mohammed O. Optimal charging of plug-in electric ve- IEEE Trans Power Syst 2007;22:2204e12.
hicles for a car park infrastructure. Industry applications society annual
meeting (IAS). IEEE; 2012. p. 1e8.

Please cite this article in press as: Shaaban MF, et al., Coordinated charging of plug-in hybrid electric vehicles in smart hybrid AC/DC distribution
systems, Renewable Energy (2014), http://dx.doi.org/10.1016/j.renene.2014.08.012