,(((&DQDGLDQ&RQIHUHQFHRI(OHFWULFDODQG&RPSXWHU(QJLQHHULQJ&&(&(

The Impact of Charging Battery Electric Vehicles on

the Load Profile in the Presence of Renewable Energy
Danilo Yu Min Prasad Adhikari Aurélien Guiral Alan S. Fung
Departmentt of Electrical and Department of Aerospace Institut Catholique d’Arts et Department of Mechanical and
Computer Enginnering Engineering Métiers Industrial Engineering
Ryerson University Ryerson University Toulouse, France Ryerson University
Toronto, ON, Canada Toronto, ON, Canada [email protected] Toronto, ON, Canada
[email protected] [email protected] [email protected]

Farahnaz Mohammadi Kaamran Raahemifar

Department of Electrical and Department of Electrical and
Computer Engineering Computer Engineering
Ryerson University Ryerson University
Toronto, ON, Canada Toronto, ON, Canada
[email protected] [email protected]

Abstract— The driving and charging patterns of battery in scheduling and controlling the charging of EVs [3-8] with
electric vehicles can have a significant impact on the load profile varying objectives, such as minimizing the operating cost of the
of the distribution grid. Renewable resources and controlled aggregator, charging station or EV owners; maximizing the
charging can mitigate this effect by reducing the energy imported number of EVs charged; minimizing supply and demand
from the grid. The smart real-time electric vehicle charging imbalance; minimizing peak load; minimizing load variance;
scheduling proposed in this paper reduced grid peak demand and meeting technical grid constraints.
through dynamic programming that uses future renewable supply
and electric vehicle demand predictions to optimize the balance of Precise modelling of EV charging requires accurate and
energy in the electric grid. The large-scale driving, charging and detailed EV driving patterns [9,10]. Due to lack of data, the
parking availability patterns used in the simulations were majority of the research found in literature ignored or assumed
synthesized from an actual experiment made up of a commercial the driving and charging dynamics of EV drivers by
fleet of battery electric vehicles. Numerical simulations showed extrapolating the patterns from national travel surveys [11-13]
that for commercial-use vehicles, solar photovoltaic and hybrid which do not capture the changes in driver behavior due to the
photovoltaic+wind technologies can reduce annual grid demand anxieties of using a new technology and the fear of losing battery
by 24%. Smart scheduling of electric vehicles and the integration power before reaching a charging station. Majority of studies
of renewable energy resulted in lower peaks but no significant shift focus on EVs for private use which can be charged overnight at
in the peak hour demand was observed. home and assumed the EVs were always plugged in when
Keywords— Plug-in electric vehicles, control strategy, smart
parked, with fixed travel hours, fixed schedules and routes
grids, net zero-energy community, scheduling optimization, (work, home, shop) and do not leave until fully charged or the
renewable energy. requested charge is fulfilled [14,15].
Global renewable energy resources are expected to grow by
8,300 TWh by 2040 with wind energy contributing 3,568 TWh
I. INTRODUCTION
and solar photovoltaic (PV) energy contributing 1,521 TWh
The transportation sector’s contribution to greenhouse gas [16]. Although integrating renewable energy resources into the
emission (GHG) accounts for almost a quarter (23%) of the distribution system is becoming a popular green energy
global energy-related GHG emissions [1]. Switching from approach to meet the large-scale penetration of EVs, predicting
internal combustion engine (ICE) vehicles to battery electric their behavior brings uncertainties into the equation. The
vehicles (BEV) and plug-in electricity vehicles (PHEV) will random arrival and departure times and the battery capacity of
significantly contribute to meeting the Paris Agreement EVs will increase the demand side uncertainties. The
(COP21) goal of limiting the global temperature increase to 2ºC unpredictability of solar and wind energy imposes operational
necessitates [1]. The Paris Declaration on Electro-Mobility and challenges to the distribution network. Studies that can be found
Climate Change and Call to Action sets a goal of global in literature on solar energy-supplied EV charging focused on
deployment of 100 million EVs and 400 million 2- and 3- on-site generation [17,18] while research on wind technology
wheelers by 2030 [2]. To prepare for the high penetration of EVs focus on large-scale wind farms [19,20]. One of the few works
into the electricity network, various research has been conducted on the integration of large scale-renewable energy into the

k,(((
 ,(((&DQDGLDQ&RQIHUHQFHRI(OHFWULFDODQG&RPSXWHU(QJLQHHULQJ&&(&(

power system to support the increase in EV load demand can be C. State of Charge
found in [21,22]. During the optimization period, each EV j can take several
The main contributions of this paper are as follows: 1) trips and the kilometers (km) travelled in each trip ‫ܦ‬௜ were
simulation of a large-scale commercial fleet of EVs based on “guessed” from the average μ and standard deviation . The
actual EV driving and charging behavior; 2) a 4-hr rolling- energy consumed per unit distance travelled and the power
horizon predictive charging control based on time-of-use transfer rate during charging is assumed to be a linear function
pricing, future forecasts of renewable energy generation and of the battery capacity and the driving range. The battery SOC
historical EV driving and charging patterns; and 3) numerical of EV j at each time interval of the rolling horizon was modeled
results show that solar energy is more effective than wind energy as follows:
in meeting the demands of EVs used for commercial runs. In this
work, the charging scheduling of 1 million EVs and 500,000 SOCi + min(Omax,q , EEVi , Eop,i ) if charging
charging stations located at home, at work and at public stations SOCi+1 = ቐSOCi - EV * Di if travelling (1)
were simulated. Renewable energy resources investigated were
solar PV, wind and hybrid PV+wind technologies. SOCi if parked or idle

The rest of this paper is structured as follows: the where i is the time interval within the rolling horizon, Omax,q is
mathematical models are introduced in Section II, the impact of the maximum power output of charging outlet q (in this work,
integrating PV, wind and hybrid PV+wind, into the power 1.8 or 4.8 kW), EEVi is the energy in kWh required by EV j to
system for controlled charging is discussed in Section III. The fully charge its battery, Eop,i is the optimized energy given to the
conclusions drawn are provided in Section IV. In the succeeding EV, and EV is the efficiency of the EV in kWh per km.
discussions the term EV will denote battery electric vehicle.
D. Optimization Algorithm
II. PROBLEM FORMULATION
A rolling horizon mechanism was used for the central
A. System Model Architecture controller. The algorithm performs a new optimization at every
It is assumed that the charging stations, distribution time step using the next n periods of demand, supply, energy
controllers, central microgrid controller and utility company prices and weather forecasts to calculate the EV charging
have two-way real-time communications capability (Fig. 1). A energy.
charging station can be made up of one or several charging Improved solar and wind power forecasting will result in the
outlets. The smart charging outlet can capture the state of charge reduction of grid reserves that must be carried to accommodate
(SOC) of the EV battery. The central controller has access to the the uncertainty of solar and wind power output. Although the
actual and forecast community load and renewable energy 15-minute ahead (15MA) forecast is the most accurate [23], the
generation in real time. It receives the aggregated EV demand 4-hour ahead (4HA) forecast and hence the 4-hour rolling
and supply from all the charging stations in real time through the horizon was adopted in this paper due to the non-availability of
distribution controllers. Historical EV demand are stored in the higher resolution weather data and load profile and as a
central controller for EV load forecasting. The central controller compromise between accuracy and computational performance.
thus has a picture of the overall supply and demand scenarios of Even though an EV may need to be plugged in for longer than
the community. four hours to get a full charge because of the type of chargers
used, a four-hour limit was imposed on the model to mimic the
4HA forecast. If the controllers were not able to fully charge the
EV in four hours, the EV must be given the maximum power on
the fourth hour if still plugged in, without exceeding the outlet’s
charging rate or the EV’s maximum SOC.
The central controller optimization problem was formulated
as follows:

cG,i PG,i - σl cl,i Pl,i +

min σt+n-1
i=t [ 2 (2)
(DEVj ,i - PEVJ ,i )2 + ൫PG,i - PG,ሺi+1ሻ ൯ ]

subject to:
Fig. 1. Proposed smart charging model made up of three-tier
control 0  PG,i  PGmax,i (3)
B. EV Driving and Charging Pattern Generator
0  Pl,i  Plmax,i (4)
Findings from the WA EV trial [10] were used to synthesize
the dynamic arrival and departure of EVs, and the SOC of the
EV battery. An EV can be in any of the following states: DL,i + FEVmin DEV,i  PG,i + σl Pl,i  DL,i + DEV,i (5)
charging, parking, and travelling. Charging statistics were
separated into home, work, and public stations. σk (DL,t+k + DEV,t+k ) = σk ሺPG,t+k + σl Pl,t+k )
 ,(((&DQDGLDQ&RQIHUHQFHRI(OHFWULFDODQG&RPSXWHU(QJLQHHULQJ&&(&(

for k = 0,1,2, … ,n-1 (6) optimized charging is used as shown in Figs. 3 and 4 and in
Table II. Solar PV came second with 23.8% reduction in grid
σt+n
i=t
-1
(DL,i + DEV,i ) = σt+n
i=t
-1
[PG,i + σl Pl,i ] (7) imports and 3.6% renewable energy wasted. In terms of grid
savings during mid- and peak-hours, solar PV provided the
DEV = σm (8) highest savings.
j=1 min( Rj,max ,EEVJ )
IV. CONCLUSIONS
The central controller optimization problem (2) is made up
of four parts: 1) minimize the total cost of charging the EVs; 2) In this paper, a rolling-horizon, cost-minimization approach to
maximize the use of renewable sources of energy; 3) maximize control the charging of a large community fleet of EVs for
the power given to the EVs; and 4) minimize the variation of the business use was developed and the effectiveness of different
grid load during the rolling horizon window. In the above renewable energy resources in meeting the demand was
formulations, t is the current time step; ܿீǡ௜ is the price of analyzed. The simulated arrival and departure times and the
battery capacity of EVs were based on actual EV driving and
importing power from the grid at interval ݅; ܿ௟ǡ௜ is the price of
charging patterns. Past historical EV demand, renewable energy
renewable technology ݈ at the same interval ݅; ܲୋǡ௜ , ܲ௟ǡ௜ are the forecasts and time-of-use pricing were used to predict the 4-hour
grid and renewable resources energy, respectively, allocated to demand of the EVs. The results show that for business fleets, EV
the EV and non-EV load; ‫ܦ‬ா௏೔ is the EV demand and ߱ and ߙ requirement in the mid-morning cannot be shifted to another
are the penalties for maximizing the charging of EVs and for time period due to the nature of the EV usage. It was further
minimizing the grid load fluctuations, respectively. ‫ܦ‬௅ǡ௜ is the observed that for coordinated charging to be highly effective,
non-EV demand, and m is the total number of EVs plugged in at there has to be a large-scale EV penetration in the community.
time i. ‫ܨ‬ா௏௠௜௡ , the minimum energy given was set to 20%. In
the optimization, it was assumed that the grid is able to supply ACKNOWLEDGMENT
all load requirements, therefore ܲீ௠௔௫ǡ௜ was given a value equal The authors would like to acknowledge the funding support
to the EV and non-EV demands. Table I lists the parameters from the following organizations: Natural Sciences and
used in the simulation. Engineering Research Council (NSERC) of Canada and its
Smart Net-Zero Energy Building Strategic Research Network
TABLE I. PARAMETERS USED IN THE SIMULATION (SNEBRN), Center for Urban Energy-Hydro One Network Inc.
Number of EVs 1,000,000 (CUE-HONI) and Ryerson University.
Number of stations 500,000
Annual non-EV demand (‘000 MWh) 51,200 TABLE II. ANNUAL UNMET EV DEMAND, RENEWABLE
Annual PV energy available (‘000 MWh) 12,823 ENERGY WASTED AND GRID REDUCTION
Annual Wind energy available (‘000 MWh) 12,823 Uncontrolled Charging Optimized Charging
Annual PV+Wind energy available (‘000 MWh) 12,823 PV Wind Hybrid PV Wind Hybrid
Unmet EV 0 0 0 265 187 211
III. RESULTS Demand 0% 0% 0% 28.7% 21.6% 24%
The simulations produced an EV demand contribution of [‘000
MWh]
1.5% of the total electricity consumption. Annual unmet EV Renewable 483 1003 94 465 992 91
demand, wasted renewable energy, and reduction in grid usage Energy 3.8% 7.9% 0.8% 3.6% 7.7% 0.7%
using different technologies and charging schemes are presented Wasted
in Table II. Table II compares the performance of uncontrolled [‘000
and controlled charging using no renewable energy supply as the MWh]
baseline. The Unmet EV Demand in Table II is the result of Grid 9,517 8,501 12,168 12,358 11,831 12,733
Reduction 18.4% 16.5% 23.6% 23.8% 22.8% 24.5%
curtailing the charging energy given to the EV units. In the [‘000
uncontrolled charging scheme, every EV started charging the MWh]
moment they are plugged in and all the demands were satisfied,
increasing the overall peak demand starting at 8 am when EVs
start charging at the workplace and at public charging stations
as shown in the annual demand plots of Fig. 2. The figure also
shows that energy available from renewable resources did not
always coincide with the demand resulting in wasted energy if
no energy storage is employed. With optimized charging, it can
be gathered from Table II that wind energy had the best annual
charging performance with only 21.6% of the total EV demand
not met, compared to PV with 28.7% and hybrid system at 24%,
respectively, but it has the worst timing in meeting EV energy
demand when it is most needed with 7.7% renewable energy
wasted. This pattern is also shown in Fig. 2 and Fig. 3. The
hybrid PV+wind technology follows more closely the new load
curve introduced by the EV demand, with 99.3% renewable Fig. 2. EV and non-EV load profile versus available PV, wind and
utilization and 24.5% savings in grid energy usage when hybrid PV+wind
 ,(((&DQDGLDQ&RQIHUHQFHRI(OHFWULFDODQG&RPSXWHU(QJLQHHULQJ&&(&(

[8] M. Amjad, A. Ahmad, M.H. Rehmani, and T. Umer, "A review of

EVs charging: From the perspective of energy optimization,
optimization approaches, and charging techniques,"
Transportation Research Part D, vol. 62, pp. 386-417, 2018.
[9] N. Daina, A. Sivakumar, and J.W. Polak, "Electric vehicle
charging choices: Modelling and implications for smart charging
services," Transportation Research Part C, vol. 81, pp. 36-56,
2017.
[10] S. Speidel and T. Braunl, "Driving and charging patterns of
electric vehicles for energy usage," Renewable and Sustainable
Energy Reviews, vol. 40, pp. 97-110, 2014.
[11] X. Liu et al., "Optimal Day-ahead Charging Scheduling of
Electric Vehicles through an Aggregative Game Model," IEEE
Transactions on Smart Grid, vol. 9, no. 5, pp. 5173-5184, 2017.
[12] M. Crow and M.L. Crow, "Economic Scheduling of Residential
Fig. 3. Renewable energy available vs energy used Plug-In (Hybrid) Electric Vehicle (PHEV) Charging," Energies,
vol. 7, pp. 1876-1898, 2014.
[13] C.G. Hoehne and M.V. Chester, "Optimizing plug-in electric
vehicle and vehicle-to-grid charge scheduling to minimize carbon
emissions," Energy, vol. 115, pp. 646-657, 2016.
[14] M.H.K. Tushar, A.W. Zeineddine, and C. Assi, "Demand-Side
Management by Regulating Charging and Discharging of the EV,
ESS, and Utilizing Renewable Energy," IEEE Transactions on
Industrial Informatics, vol. 14, no. 1, pp. 117-126, January 2018.
[15] X. Lu and S. Yang, "Multi-objective optimal dispatch of
microgrid containing electric vehicles," Journal of Cleaner
Production, vol. 165, pp. 1572-1581, November 2017.
[16] International Energy Agency, "World Energy Outlook 2015,"
Paris, France, 2015.
[17] R. Wang, P. Wang, and G. Xiao, "Two-Stage Mechanism for
Fig. 4. Comparison of energy source usage under different Massive Electric Vehicle Charging Involving Renewable Energy,"
charging schemes IEEE Transactions on Vehicular Technology, vol. 65, no. 6, pp.
4159 - 4171, Jun 2016.
[18] L. Yao, Z. Damiran, and W.H. Lim, "Optimal Charging and
Discharging Scheduling for Electric Vehicles in a Parking Station
REFERENCES with Photovoltaic System and Energy Storage System," Energies,
vol. 10, no. 550, pp. 1-20, 2017.
[1] International Energy Agency, "Global EV Outlook 2016: Beyond
one million electric cars," Paris, France, 2016. [19] H. Nguyen, C. Zhang, and M.A. Mahmud, "Smart Charging and
Discharging of Electric Vehicles to Support Grid with High
[2] (2015, Nov.) United Nations Climate Change. [Online]. https:
Penetration of Renewable Energy," in 19th World Congress
//unfccc.int/news/the-paris-declaration-on-electro-mobility-and-
International Federation of Automatic Control, Cape Town, South
climate-change-and-call-to-action
Africa, 2014, pp. 8605-8609.
[3] J. Garcia-Villalobos, I. Zamora, J.I. San Martin, F.J. Arsenio, and
[20] C.T. Li, C. Ahn, H. Peng, and J. Sun, "Integration of Plug-In
V. Apperibay, "Plug-in electric vehicles in electric distribution
Electric Vehicle Charging and Wind Energy Scheduling on
networks: A review of smart charging approaches," Renewable
Electricity Grid," in IEEE PES Innovative Smart Grid
and Sustainable Energy Reviews, vol. 38, pp. 717-731, 2014.
Technologies (ISGT), Washington D.C., USA, 2012.
[4] J. Hu, H. Morais, T. Sousa, and M. Lind, "Electric vehicle fleet
[21] M. Zhang and J. Chen, "The Energy Management and Optimized
management in smart grids: A review of services, optimization
Operation of Electric Vehicles Based on Microgrid," IEEE
and control aspects," Renewable and Sustainable Energy Reviews,
Transactions on Power Delivery, vol. 29, no. 3, pp. 1427-1435,
vol. 56, pp. 1207-1226, 2016.
June 2014.
[5] J.Y. Yong, V.K. Ramachandaramurthy, K.M. Tan, and N.
[22] X. Jiang, J. Wang, Y. Han, and Q. Zhao, "Coordination Dispatch
Mithulananthan, "A review on the state-of-the-art technologies of
of Electric VehiclesCharging/Discharging and Renewable Energy
electric vehicle, its impacts and prospects," Renewable and
Resources Power in Microgrid," in International Congress of
Sustainable Energy Reviews, vol. 49, pp. 365–385, 2015.
Information and Communication Technology (ICICT 2017) ,
[6] L. Liu, F. Kong, Y. Peng, and Q. Wang, "A review on electric Sanya, China, 2017, pp. 157-163.
vehicles interacting with renewable energy in smart grid,"
[23] J. Zhang et al., "Baseline and Target Values for PV Forecasts:
Renewable and Sustainable Energy Reviews, vol. 51, pp. 648–
Toward Improved Solar Power Forecasting," in IEEE Power and
661, 2015.
Energy Society General Meeting, Denver, Colorado, 2015.
[7] I. Rahman, P.M. Vasant, B.S.M. Singh, M. Abdullah-Al-Wadud,
and N. Adnan, "Review of recent trends in optimization
techniques for plug-in hybrid, and electric vehicle charging
infrastructures," Renewable and Sustainable Energy Reviews, vol.
58, pp. 1039-1047, 2016.