Académique Documents
Professionnel Documents
Culture Documents
Energy
journal homepage: www.elsevier.com/locate/energy
a r t i c l e i n f o a b s t r a c t
Article history: The direct coupling of solar electricity with electric vehicles is very topical due to their many co-benefits,
Received 27 April 2017 such as reducing the electricity flow into and from the grid. This work illustrates this solution using a
Received in revised form park-and-ride lot outside Lisbon to study its solar energy potential to charge electric vehicles. The solar
28 August 2017
resource, its exploitation and coordination with the vehicles under different charging approaches are
Accepted 6 September 2017
discussed in detail, using methods extendible to other cases. A genetic-algorithm optimisation was
Available online 7 September 2017
performed to maximise the solar energy generation. To ascertain the techno-economic feasibility of the
concept, several sets of assumptions are analysed, including the use of energy storage and smart
Keywords:
Solar parking lots
charging. For current market conditions, the payback time is found to be 14 years; a modest financial
Electric vehicles public incentive would significantly improve the project economics e making the payback drop to 7
EV charging years e enabling conditions to put this inescapable and green approach at the forefront.
Photovoltaic energy © 2017 Elsevier Ltd. All rights reserved.
Sustainable energy systems
Feasibility study
1. Introduction PV generation is its variability and uncertainty [5]. As for the EVs,
they could cause a boost in electricity demand and overload the
The growing global electricity consumption e the Interna- grid, degrade power quality and destabilise the electric system
tional Energy Agency estimates a 30% increase by 2040 [1] e [6]. One way to avoid these issues is by charging EVs directly from
creates concern on the sources of electricity production for the PV. This integration should be carefully planned, using, for
future. In order to not compromise the objectives of the COP21 example, scheduled charging plans to coordinate both technolo-
agreement, renewable energy is seen as the main solution. The gies [7].
large amount of solar energy reaching the earth e 101 PWh per By doing this, on the one hand EVs could be used as storage
hour [2], which is about the same as the world annual final en- devices and dispatchable loads, helping the grid maintaining the
ergy demand [3] e combined with the price competitiveness of supply-demand balance while allowing a larger renewable
photovoltaics (PV) is enhancing the role of this technology as part penetration [8]. On the other hand, PV production could also
of that solution. enable a larger penetration of EVs when they charge from it. This
Alongside, the transport sector is also of concern, since it con- way there is no significant net-load increase in demand. There-
tributes significantly to air pollution and greenhouse gases (GHG) fore, the appropriate charging strategy can allow not only a larger
emissions. The replacement of internal combustion engine (ICE) amount of solar energy into the power generation system, but
vehicles for the much more efficient battery electric vehicles (EVs) also a higher EV penetration, improving at the same time the grid
is one of the possible measures to overcome this. However, EV flexibility.
charging must be based on a clean, fossil-free electricity [4]. Decentralised photovoltaic power plants to satisfy the EV
In both PV and EV cases, the grid integration is an issue: their charging demand could improve the grid operation, while
emerging penetration may jeopardise the grid reliability, causing decreasing GHG emissions not only by themselves but also by
stabilisation and flexibility problems. The main disadvantage of supporting EVs adoption through infrastructures creation. This can
be done by means of electric vehicles solar parking lots (EVSPLs),
i.e., solar parking lots to charge EVs. This way the grid does not have
to integrate a large photovoltaic capacity nor does it need a big
* Corresponding author. FCUL, DEGGE, Ed. C8, 1749-016 Campo Grande, Portugal. reinforcement to satisfy the EVs' demand. The drivers' preferences
E-mail address: rvfigueiredo@fc.ul.pt (R. Figueiredo). align with this, since they regard EVs and solar panels as
https://doi.org/10.1016/j.energy.2017.09.024
0360-5442/© 2017 Elsevier Ltd. All rights reserved.
R. Figueiredo et al. / Energy 140 (2017) 1182e1197 1183
complementary [9]. to CO2 emission reduction costs; it is concluded that the cheapest
The interest for EVs solar charging stations has been emerging in alternative is the smaller off-grid solution, but the grid-connected
the literature, which was thoroughly reviewed in Ref. [10]; other option, bigger, reduces the overall emissions the most. A similar
reviews of technological aspects are provided in Refs. [11] and [12]. comparison was performed in Ref. [22], but optimising the system
Below prior recent research on the topic is summarised. lifecycle cost; the results showed that a grid-connected station with
storage using a supply mix of on-site PV and diesel generated
electricity was the most economical option, while the renewable
1.1. Prior recent work off-grid system was found to be the most expensive.
An on-grid EVSPL may help the grid by behaving as dispatchable
One of the most important objects of the (mostly simulation) loads, feeding energy into it (vehicle-to-grid, abbreviated V2G) and
studies concerns charging strategies, which may be of two types: providing ancillary services. From this viewpoint, Ref. [23] presents
uncontrolled, where the EVs start charging immediately when a fuzzy control method to optimize the schedule of EVs charging.
connected to a charger; or controlled, where the charging is Ref. [24] presents an EV charging scheduling aiming to comply with
scheduled according to certain criteria, such as energy price or the grid capacity constraints.
available renewable energy. The latter implies the use of an opti- A microgrid framework is also used to address EV charging
misation algorithm with a certain objective function. Some studies plans, often considering several energy sources, as in Ref. [25]. The
address both strategies from a price perspective, as in Ref. [13], operational optimisation of microgrids is studied cost-wise in
using a multi-agent approach; it was found that controlled Ref. [26], for minimising the fuel consumption on a remote island;
charging adherents were responsive to charge with renewables in Ref. [27], considering the impact of using different EV charging
when that choice was given. Ref. [14] shows that, to reach the same strategies; and in Ref. [28], investigating at what extent the V2G
environmental target, uncontrolled charging requires a higher concept can minimise its cost.
storage power and energy capacity than controlled smart charging. The location and sizing of charging stations have also been
A controlled charging plan works according to what is defined as addressed, since they can be determinant for the projects feasibility
objective in the analytical model. The most common models use a and its grid impact. Refs. [29] and [30] address this issue from the
stochastic approach to model the EV charging (see Table 1). Often grid stability viewpoint, using an evolutionary-based optimisation
the optimisation function is the minimisation of costs (for the technique with positive results.
parking operator or EV owners) or maximisation of the use of Besides the energy viewpoint, the environmental and economic
renewable electricity. angles are also addressed occasionally. The environmental
Costs can be minimised in several ways, such as taking advan- perspective is taken in a limited number of EVSPL studies, such as in
tage of time-of-use tariffs, as in Ref. [15]. Ref. [16] compares de- Refs. [21] and [22]. The economic angle has been explored under
mand response approaches in what regards grid stability, different approaches, emphasising the standpoint of the operator
concluding that an approach oriented to time-of-use prices leads to [13], the EV user [27] or the overall grid operation [25]. Generally,
reduced costs to the aggregator. these studies show a win-win economic situation for the parties
As to the maximisation of renewable energy use, Refs. [17] and involved on V2G services. The literature shows that EVSPLs are not
[18] conclude that storage greatly reduces grid dependency. In yet financially attractive, presenting payback times between 17 and
Ref. [19] the authors use an optimised PV power extraction tech- 26.5 years, mostly due to low revenues and high investment costs
nique to get the most advantage of the solar charging. Power [31].
sources other than photovoltaics are also considered, for example Table 1 makes a summary of the literature referenced above,
wind in Ref. [20]. categorising the studies according to their distinctive
The solar charging stations can be either off- or on-grid. In characteristics.
Ref. [21] these two types of solar systems are optimised according
Table 1
Summary of the literature.
1.2. Scope of this work The annual PV energy generation (Eannual in Wh) is the sum of all
the hourly energy production throughout the year:
The focus of the literature regarding solar powered parking lots
for electric vehicles has been the energy analysis and EV charging X
8760
strategies, mainly using simulated case studies. The use of actual Eannual ¼ Eprod ðiÞ (2)
i¼1
occupation data for an extensive period e which provides a
detailed realistic input of occupancy patterns often missed with The calculation of Gtilted and hPV system are explained below and in
synthesised datasets e may contribute to a better insight into the the next subsection.
potential benefits of this type of solution, enabling a careful eco- To calculate Gtilted, a typical meteorological year (TMY) weather
nomic and financial analysis. data file for the site can be used [32]. These files, which are freely
The present study fills this gap by analysing different business available, contain hourly statistical data of several meteorological
models of solar PV charging to apply to an existing non-solar pri- parameters for an entire year, such as solar global irradiance, hor-
vate-owned park-and-ride facility, i.e., a parking lot with public izontal diffuse irradiance (Dh, in W/m2), direct normal irradiance
transport connections to serve commuters. Its core focus is not only (DNI, in W/m2), temperature, etc. [33]. The following expressions
the understanding of the correspondence of EV charging and solar make use of DNI and Dh to calculate the different irradiance com-
electricity production, but also the comparison and discussion of ponents (in W/m2) reaching the modules plan: direct (Btilted),
possible business models in order to draw conclusions about their diffuse (Dtilted) and ground-reflected (Rtilted) [34,35].
viability. It proposes a not yet existing comprehensive and
straightforward energy-economic approach to solar parking lots for Btilted ðiÞ ¼ DNI ðiÞ cosðAOI ðiÞÞ (3)
electric vehicles.
The paper is structured as follows: Section 2 presents the Dtilted ðiÞ ¼ Dh ðiÞ ðð1 þ cosðbÞÞ=2Þ (4)
methods and its application to a case study. In Section 3 the results
are shown and discussed, and the conclusions are summed up in Rtilted ðiÞ ¼ ðDNI ðiÞ senðaðiÞÞ þ Dh ðiÞÞ r ðð1
Section 4.
cosðbÞÞ=2Þ (5)
2. Approach Here, AOI is the angle of incidence of the solar radiation on the
module surface [ ]; b is the tilt angle of the module [ ], to be
This section presents the methods used in this work (Section determined e see Sections 2.2.2.1.1 and 2.2.2.1.2; a is the solar
2.1), followed by their application to a real park-and-ride facility altitude [ ]; and r is the ground-reflectivity, which for asphalt may
(Section 2.2). Several assumptions were made (as explained in be considered as 0.2 [36]. The parameters AOI and a depend on the
Sections 2.2.1 and 2.2.2), which are combined in different scenarios geographical coordinates and are calculated using established solar
(Section 2.2.3). The assumptions are specific to the case under geometry methods [35,37,38].
study, but any others may be adopted. The sum of the previous components gives Gtilted.
2.1. Methods Gtilted ðiÞ ¼ Btilted ðiÞ þ Dtilted ðiÞ þ Rtilted ðiÞ (6)
DC/AC Inverters = ~
Control Center
Adjoining facilities
AC/AC unidirectional chargers ~ ~ ~
~
~ ~
surplus.
The overall system efficiency, hPV system , is based on a set of in- Eprod ðiÞ CEV ði 1Þ
NEV ðiÞ ¼ (12)
dividual technical specifications of these components and is given Pcharger
by the following expression:
Here, Eprod ðiÞ is the PV energy production at a certain hour and
hPV system ðiÞ ¼ hmodules ðiÞ hinverters hcables hchargers (7) CEV ði 1Þ is the power demand in the hour i correspondent to the
EVs that started charging previously and are still doing it.
The efficiencies of the inverters (hinverters ), cables (hcables ) and The energy management of the park is hereafter briefly
chargers (hchargers ) are considered constant, and the modules' effi- described. The photovoltaic generation is primarily used for sup-
ciency (hmodules ðiÞ) depends on the ambient temperature. When plying the EV charging, and in the case of surplus it is stored in the
the EV chargers' efficiency is already included in the EV con- central battery of the EVSPL. Finally, if there is no EV demand to
sumption, as often is (see Section 2.1.3), hchargers should be assumed supply and if the storage battery is completely charged, the
as 1 in Eq. (7), to avoid double counting [39]. generated electricity is sold to the adjoining facilities. On the other
The temperature is given by Eq. (8), and it is a product of the hand, if neither the PV generation nor the storage battery are
modules nominal efficiency (hnominal ) and the temperature related enough to supply the EV load, grid purchases are required. The
efficiency (hTamb ðiÞ), which decreases as the ambient temperature energy management strategy is summarised in Fig. 2 for both
rises according to Eq. (9) and Eq. (10) [40]. controlled and uncontrolled charging plans. This method is applied
to seasonal, working and weekend mean days of the parking lot, as
hmodules ðiÞ ¼ hnominal hTamb ðiÞ (8) Section 3.1 shows.
Tmodule ðiÞ ¼ ðNOCT TambNOCT Þ=GNOCT Gtilted ðiÞ þ Tamb ðiÞ 2.1.4. Economic feasibility
The economic balance is based on the revenues and costs during
(9)
the project, which may be assumed to last 25 years, which is
generally the lifetime warranty of PV modules [43]. This section
hTamb ðiÞ ¼ ð100% TC PMPP Þ DTamb ðiÞ (10) presents the analytics applied to the economic component, and also
the common assumptions to all the scenarios tested. The assump-
Here, Tmodule ðiÞ is the module temperature at hour i [K]; NOCT is the tions that are scenario dependant (PV investment costs, etc.) are
Normal Operating Cell Temperature (NOCT) [K]; TambNOCT ¼ 20 C is explained in Section 2.2.2.4. It is only considered the PV-EV
the ambient temperature for NOCT conditions, [40]; component of the parking lot, and it is treated as if it is indepen-
GNOCT ¼ 800 W/m2 is the solar irradiance for the same conditions; dent of the other components (parking space rent, etc.), i.e., the cost
Tamb ðiÞ is the ambient temperature [K]; DTamb ðiÞ is the difference of the operation already in place is considered to be sunk.
between the temperatures of the environment and the module, and The nominal value, i.e. cash flow, of energy traded between the
TC PMPP is the temperature coefficient at the maximum power park and the grid/facilities on the year n of the project (the in-
point, i.e., the efficiency degradation factor with temperature of the vestment year corresponds to n ¼ 0), V(n), is calculated according to
module, in %/K, given in the technical datasheet. the following expression:
2.1.3. Energy management where EðnÞ is the amount of energy traded during that year; c is the
The average daily distance driven by a typical car owner varies energy unit price; i is the inflation rate. Equation (13) is used to
between countries, and even between regions, but generally is calculate both the grid sales and the facilities purchases to the park.
under 40 km [41]; in Portugal, for example, it is 35 km [42]. This The EV demand does not change from year to year of the project,
study assumes that the batteries of the cars are topped up ac- but both the sales and purchases are dependent on the annual
cording to the average energy needs. Using the same example, in energy production, Eprod , which varies with the modules degrada-
Portugal each vehicle needs to charge daily 6 kWh on average, tion, as in Eq. (14).
assuming an EV energy consumption of 0.171 kWh/km [39]. Given
the large number of clients, it is assumed that fluctuations in energy Eprod ðnÞ ¼ Eprod ð1Þ ð1 aPV Þn1 (14)
requirements between individual vehicles offset each other.
Two types of charging are simulated: uncontrolled and where aPV is the degradation rate of the modules energy yield per
controlled. In both approaches, after achieving the period necessary year.
to complete the daily charge of each vehicle ( tcharge , Eq. (11)), the Four different well established financial indicators are used to
vehicle stops charging and its demand becomes null. The energy determine the viability of the project: net present value (NPV),
supply options are the PV system and the grid, and in some cases discounted payback time, internal rate of return (IRR) and dis-
also a stationary storage battery. In the uncontrolled charging, the counted return on investment (ROI) [44]. The NPV represents the
EVs charging starts as soon as they arrive at the park. cumulative discounted balance of revenues and expenses for every
year n of the project.
Edaily demand Eq. (15) is used to calculate the NPV (in V), where t is the project
tcharge ¼ (11)
Pcharger duration and r is the nominal discount rate.
where Pcharger is the charging rate, and Edaily demand is the daily X
n¼t
net cash flow ðnÞ
NPV ¼ Investment (15)
energy demand [kWh].
n¼1
ð1 þ rÞn
The controlled charging assumes that the number of vehicles
charging at every hour, NEV ðiÞ, considers the number of vehicles still The discounted payback time indicator (abbreviated below as
charging from the previous hour and is defined by the following ‘payback’) shows the elapsed years required until the NPV achieves
expression: a positive value, i.e., the required time until the revenues equal the
1186 R. Figueiredo et al. / Energy 140 (2017) 1182e1197
costs (both discounted) and the initial investment. It is calculated accounts for about 12,700 m2 of available area for car shading
through Eq. (16). (about half of the total parking lot area) [49]. The layout is mainly of
perpendicular parking spots oriented towards the southeast
cumulative cash flowdisc: ðYÞ (north-based azimuth: 150 ).
discounted payback ¼ Y þ
(16)
cash flow ðY þ 1Þ
disc: The occupation pattern of a park-and-ride facility is dependent
on the users' working schedule, and so it follows a bell curve on
where Y is the elapsed years since the beginning of the project weekdays. The average occupation rate of the parking lot is 23%,
correspondent to the last year with a negative cumulative dis- and its maximum occupation is 694 vehicles, i.e., about two-thirds
counted cash flow [years], and disc. stands for ‘discounted’. of the park capacity. Fig. 4 shows the average profile of the weekly
The IRR is the discount rate that nullifies the NPV. The IRR is occupation rate for the two weeks with maximum and minimum
determined using the discount rate adopted, and therefore the occupation along the year (in March and in December, respectively)
decision rule to invest is if the IRR is above this rate [45]. and for the entire year. As it was expected, during the weekends the
Finally, the discounted ROI (abbreviated below as ‘ROI’) is the occupation is significantly lower. The least occupied week is not
number of times that the investor receives the amount initially representative of the park use, as it was observed during Christmas.
invested. The ROI is calculated through Eq. (17), where it should be The comparison between the two extreme weeks and the differ-
considered the discounted cumulative at the end of the project net ence in occupation between workdays and the weekend show that
benefits and expenses, and the investment. the occupation varies a lot in different timescales, i.e., not only
P during the day, but also throughout the year.
ðrevenues expensesÞdisc:
discounted ROI ¼ P (17)
expensesdisc: þ investment
2.2.1. General assumptions
To test the methods, several assumptions had to be made. This
section describes those that are common to all the scenarios (Sec-
tion 2.2.3).
2.2. Case study The climate data was obtained through the TMY file for Lisbon
[32]. The surface of the parking lot is asphalt, thus the ground-
To test the methods above, the commuter parking lot from the reflectivity is assumed as 0.2 [36] (see Section 2.1.1). The tech-
train station Pragal in Almada (38.66ºN, 9.18ºW), Portugal, a city nical components that were assumed for the system and their
located south of Lisbon across the river Tagus (Fig. 3), was chosen. technical data are summarised in Table 2.
This location has one of the highest solar irradiation potential in One common assumption, unless otherwise stated, is that the
Europe, with an average daily irradiation of 5.15 kWh/m2 [46]. EVSPL occupation is only of EVs. The daily electricity demand of
This station is located in a municipality of 174,030 inhabitants each vehicle is considered to be 6 kWh (see Section 2.1.3). The
[47], and serves passengers traveling on the working days between charging rate (Pcharger ) is assumed as 3.7 kW, corresponding to Level
Almada and the capital, on a train line accounting for 25.6 million 1 AC Charging. The charging infrastructures are sized using the
trips every year [48]. observed maximum occupation of the parking lot.
For the study, the parking lot occupation data from the entire The project lifetime is 25 years (similar to the PV modules
year of 2014 was gathered. It consists of hourly records of the typical lifetime [43]). Since batteries typically have a lifetime of 10
number of arrivals and departures for both monthly subscriptions years [53], shorter than the project, the investment should account
and daily tickets, the two payment types that exist. The park has a for their replacement, and so a second battery had to be accounted
capacity for 1045 cars and is open every day from 7 a.m. to in the investment. Due to technology improvements, it has been
midnight. Each parking space uses around 12.2 m2 of area, which considered that the second battery would last 15 years.
R. Figueiredo et al. / Energy 140 (2017) 1182e1197 1187
Fig. 3. The parking lot and the metropolitan area where it is located.
0.7 they are assumed to be constant (in real terms) throughout the
project. The capital investment is considered to cover all the
Maximum occupation fraction
0.6
hardware, such as the photovoltaics modules, metallic structures,
0.5
charging stations, inverters, storage battery, cables, and labour.
0.4 It is also assumed that the EVs owners pay 0.25 V/kWh to charge
0.3
their cars, which is less than the average charging fee in Europe
(about 0.35 V/kWh, according to [59]). Even though home charging
0.2 could be cheaper (about 0.17 V/kWh [60]), charging at the park is
0.1 still attractive, e.g. due to convenience or the avoidance of the need
to have access to a charging station at home e a public one might
0
Monday Tuesday Wednesday Thursday Friday Saturday Sunday not exist and a private one is expensive and sometimes is not
Highest occupation Lowest occupation Average occupation possible to install.
It is also assumed that the EVSPL sells energy to the adjoining
Fig. 4. Occupation profile during the most and least occupied weeks and the average of
train station's facilities for the same price that the facility is
the entire year.
currently paying to the utility, which is better (0.072 V/kWh) than
the price at which the park sells energy to the grid (0.050 V/kWh)
In the end of the components life, their residual value is [61]. This selling is actually virtual for the case studied, since the
assumed to be 20% of the initial value [54]. Table 3 shows the in- operator of the train and parking facilities is the same. Additionally,
vestment costs' for the PV modules, stationary battery (see Section the parking lot purchases energy from the grid at the same price it
2.2.2.2 for the details on energy storage) and charging stations, both sells to the facilities (again, 0.072 V/kWh).
for 2015 and 2050 (2050 prices are given due to a delayed invest- Regarding annual expenses, the EVSPL operational and main-
ment and gradual implementation scenarios e see Sections tenance cost (cleaning, repairs, etc.) is considered to be 1% of the
2.2.2.4.2 and 2.2.2.4.3). investment, and the managing cost (administration, customer
As for the energy purchase prices between the stakeholders, support, etc.) is assumed to be 5% of the annual revenues. The
1188 R. Figueiredo et al. / Energy 140 (2017) 1182e1197
Table 2
System technical details.
Maximum power Standard Test Conditions e STD (STC - 1000 W/m2; 25 C; AM 1.5) 285 Wp
NOCT (800 W/m2; NOCT ¼ 46 C; AM 1.5) 213.1 Wp
Module area 1.67 m2
Nominal efficiency 17.17%
Temperature coefficient at the maximum power point, TC_PMPP 0.41%/K
Degradation rate [50] 0.8%
Table 4
Energy prices and other assumptions.
50 m
50 m
Fig. 5. Layout of the PV installation orientated southeast (above) and south (below).
Table 5
Characteristics and production of the photovoltaic installation.
Orientation Azimuth (º) Modules' tilt (º) # of PV modules Rated power [MWp] Energy generation [MWh/year] Energy generation per area [MWh/m2]
Table 6
Summary and description of the analysed scenarios in this work.
Scenarios Parking lot orientation Charging plan Other assumptions Energy Economic feasibility
storage [kWh]
PV layout I PV layout II Controlled Uncontrolled Delayed Gradual Base Flat fee Grid support
implentation deployment
2.2.2.1.2. PV layout II. Since in a parking lot yet in its design i.e., as if the park could be rotated to it, named ‘PV layout II’. The
phase the geographical orientation might possibly be chosen, it was optimisation procedure was as explained above, with the difference
decided to test an orientation corresponding to the optimum one, that the azimuth was also a variable. A boundary of þ15 and 15
1190 R. Figueiredo et al. / Energy 140 (2017) 1182e1197
Assumptions
Grid support
Frequency’s secondary regulation
provided by the park to the grid
1200 1600
a. b.
1400
1000
1200
800
Power [kW]
1000
Power [kW]
600 800
600
400
400
200
200
0 0
1 3 5 7 9 11 13 15 17 19 21 23 1 3 5 7 9 11 13 15 17 19 21 23
Daytime [h] Daytime [h]
PV gen. - Winter PV gen. - Spring PV gen. EV Load - Average day
PV gen. - Summer PV gen. - Autumn EV Load - Working day EV Load - Weekend
EV Load
Fig. 7. a: Average profiles of seasonal PV production and annual EV load; b: average profiles of EV load (working day and weekend) and annual PV power.
around south was imposed to the possible azimuth range of values. EVs when there is not enough sunshine. To simulate this, a lithium-
The optimum tilt and north-based azimuth were found to be 32 ion battery e the most promising battery technology for renewable
and 182 , respectively, enabling an energy generation of energy applications [65] e was considered, sized according to the
1491 MWh/year. total PV excess of energy, counted hourly, that occurs on average
The PV generation results for both approaches are presented in during the working days plus the energy losses in a charge/
the following resulted in Table 5. Both options enable the installa- discharge cycle (shadowed area in Fig. 7b). This was considered to
tion of 3480 photovoltaic modules (see Section 2.2.2.1), however be the appropriate storage size to maximise the solar EV driven
with optimal orientation the annual energy generation increases fraction, i.e., the distance driven by the cars with solar electricity
around 14.8%. (see Section 3.1.1); if the criteria were different, e.g. techno-
economic, the storage size could also be optimised using the
2.2.2.2. Energy storage. Since the EV load does not perfectly follow same procedure of Section 2.2.2.1.1.
the PV production, an energy storage device acting as an energy The round-trip efficiency of the battery was considered 92.5%
buffer between the PV system and the EVs can increase the solar EV [53]. Its size varies with the scenario (see Table 6 at the end of this
charging. It stores the surplus PV production, using it to charge the section, where the battery sizing presented is already
R. Figueiredo et al. / Energy 140 (2017) 1182e1197 1191
fuel, they charge their cars 50% more than those with daily tickets 3.1. Energy analysis
(9 kWh and 6 kWh, respectively).
2.2.2.5.3. Grid support. The V2G approach offers an opportunity Below the results for the scenarios considered to be the most
for the EV facilities to provide ancillary services to the grid, since representative, i.e., scenario A0, representing the uncontrolled
they represent controllable power capacity [8]. This business model charging approach with energy storage, and scenario A4, corre-
adds to the Base case the opportunity to offer secondary frequency sponding to the controlled charging approach without energy
regulation to the grid, being this a remunerated service. This type of storage are presented.
regulation contributes to maintain the grid voltage stable, and its The correlation between the PV production and the park occu-
provision does not imply significant energy trading. It considers as pation along the days is an important parameter for the profit-
bidirectional the EVSL-grid power flow. The contracts are based on ability of an EVSPL, because it translates to what extent it is possible
the power capacity offered during a certain period of time by the to directly solar-charge the EVs with on-site generated solar energy.
facility [68]. The power capacity offered by the EVSPL was assumed For each day, the correlation between the parking lot occupation
as being equivalent to the number of EVs parked multiplied by the and the PV production was determined, enabling the calculation of
connection power of each one (3.7 kW). To that, it is applied the the average correlation factor for several representative periods of
current secondary frequency regulation price paid by the grid, the year, Table 7.
which is of 25.6 V/MW/h [69]. As expected, the working days show a far better correlation than
This is considered as an additional revenue of the parking lot weekends, since the weekend occupation is flatter during the day,
(compared to the Base model), besides the energy sales. and thus farther from the PV production profile. As for the seasons,
the worst correlation is in the summer: since the park is southeast
oriented, it produces significantly more during the mornings in the
2.2.2.6. Summary. A summary of all the previous assumptions is summer (Fig. 7a), which results in a more asymmetric PV profile. In
presented below, Fig. 6. contrast, the production profile during the winter is more sym-
metrical (Fig. 7b) and therefore it relates better to the occupation
(see Fig. 4). As it was expected, the average correlation for the
2.2.3. Scenarios whole year is between these values.
Several scenarios were created concerning the EVSPL operation
and financial model; they are presented in Table 6.
The main concern when configuring the scenarios was to show 3.1.1. Uncontrolled EV charging plan
practical alternatives for an investor. To keep the study constrained, Despite of the importance of the occupation rate in the parking
not all the possible combinations of assumptions were tested. lot and its correlation with the photovoltaic production, the actual
Instead, the scenarios A0 to A5, based on the Base financial model, demand that the project must support is not correspondent to
were constructed as a sequential improvement on the project every vehicle in the park, but only from those charging. The
viability. The other ones (B0 and C0) are representative of alter- charging plan dictates the EV load profile, which is shown in Fig. 7a
native financing options, and were tested having as basis the for the ’Uncontrolled’ approach, along with several representative
Reference scenario. A scenario considering both controlled EV PV production profiles.
charging and storage was discarded because of redundancy, as the For the scenario A0, with storage, the energy demand, produc-
aim of both systems is the same. tion and surplus PV production is stated in Table 8 for the afore-
mentioned representative days.
Table 9 presents the fraction of the EV demand supplied by each
3. Results and discussion energy source, i.e., directly by the PV production, the battery and
the electric grid.
The results and their discussion are divided between the energy All periods of the year provide enough PV energy to satisfy the
and the financial analyses. For the sake of containment, the energy EV demand, except the winter. In the summer, 87.9% of the EVs'
analysis is shown just for the most illustrative scenarios. The demand is directly satisfied by PV generation, which means that the
financial analysis is shown for all the scenarios since it is the one battery in this period is mostly used to store energy produced
determinant for the EVSPL feasibility assessment. during the weekends; it is also in this period that most of the
1200
1000
800
Power [kW]
600
400
200
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Daytime [h]
PV gen. Uncontrolled charging
Controlled charging Controlled charging - Working days
energy selling to the facilities occurs, after the battery is full. During on the reference case (A0-A5), because they all have the same
winter, the direct solar EV charge is of 50.6%. For the whole year, it is revenues base, and then for the scenarios B0 and C0, based on a
of 72.1%, wherein battery supplies 27.8% of the EV load (this cor- different revenue model.
responds to the charging load that surpasses the PV generation Fig. 9 shows the discounted cumulative revenues, expenses and
between 8 a.m. and midday). The working and winter days are the profit for scenarios A0-A5.
only ones showing PV generation deficit, thus requiring grid pur- Comparing scenarios A1 and A2 with A0, the three scenarios
chases (9.1% and 10.2%, respectively). that consider energy storage, it is noteworthy the 35% reduction on
The energy storage, as described in Section 2.2.2.2, was sized in the investment costs in A2, which refers to a delayed project
order to maximise solar EV charging in a typical working day, and implementation for 2030, as explained above. As for the results
so its size corresponds to the shadowed area in Fig. 7b, which differences between A1 and A0, the optimal orientation from sce-
represents the amount of energy that the PV system cannot fulfil nario A1, as opposed of the southeast from A0 (they only differ in
directly (1700 kWh). The use of the energy storage is crucial to this regard), allows for fewer grid purchases, since the energy yield
allow a high solar EV driven fraction; on average, in a working day is higher, and thus higher profits (þ11%).
the solar EV driven fraction increases from 56.7% without storage to The scenarios without energy storage (A3 onwards) show clearly
90.1% with it. lower investments and O&M costs, due to the absence of the bat-
tery related expenses, and better overall results, except A5 (the
3.1.2. Controlled EV charging plan reason is addressed below). As expected, when an uncontrolled
charging without battery is assumed (A3), the grid dependence of
The controlled charging approach without storage (Scenario A4)
the EVSPL and the energy export to the facilities increase signifi-
allows a much better coincidence between the EV load and the PV
cantly. However, the profits show an improvement of 140% over A0,
generation profiles, as it can be seen in Fig. 8. The figure also shows
meaning that the use of energy storage allows for a greater fraction
the uncontrolled charging profile, to better perceive the differ-
ences. Even though both charging plans represent the same daily of solar EV charge, but financially is not a good option.
Compared to the uncontrolled plan without storage (A3), the
EV demand and PV generation (Table 8), the controlled charging
allows less grid or battery dependence by peak shaving the EV load controlled charging (A4) allows about the same profits (þ1%), 65%
less grid dependence and 40% less energy export to the facilities. If
according to the PV generation.
Table 10, analogous to Table 9, quantifies the EV charging by compared to A0, A4 has 142% better overall operational results,
even though grid imports increase (þ13%). This is caused by the EVs
source. As in scenario A0, the grid is mostly required during the
working winter days. The small amount of grid energy required in that arrive late, when PV is not enough to their supply.
Scenario A5 shows the lowest investment costs, because it does
the summer (on average 15 kWh/day) is due to a few vehicles that
arrive late in the day at the parking lot, when the PV generation is not consider storage and its gradual deployment allows the
investor to install the components at cheaper prices as the time
no longer able to supply them.
Even though in scenario A0 the working days EV demand is goes on. The lower EV load avoids the need to import energy from
the grid. Compared to A0, A5 exports more energy to the facilities
much larger than the PV generation until the middle of the day, the
grid dependence in scenario A4, without that difference, increases by a factor of three and has 44% lower profits. The significant lower
sales to EVs (less 88% than in scenario A0) are the main reason for
about 67.4%. This is because in scenario A0 the early morning PV
generation, a period during which the park is still unoccupied, is this improvement.
The results of the scenarios with alternative revenues models
stored for later use, whereas in scenario A4 it is sold to the facilities.
As a result, in A4 the grid has to supply the late afternoon EV (B0 and C0) are presented in Fig. 10, along with the results for the
base case, for comparison.
charging. Even though the controlled charging in the working days
increases the energy purchases to the grid (shadowed area in In A0 and C0, the main source of revenues is the energy sold to
the EVs, even though in C0 there are extra revenues coming from
Fig. 8), the direct solar EV charge is much higher, as it can be seen
when comparing Figs. 7b and 8. Replacing the energy storage by a the grid support (they represent 5% of the total income). In B0 the
daily and monthly fees are the main revenues source. Even though
controlled charging decreases slightly the solar EV driven fraction,
from 90.1% (see Section 3.1.1) to 84.8%, showing that this scenario there is an extra income source from A0 to B0, the grid imports also
increase (by a factor of four), since it is assumed that the monthly
does not compromise significantly the objectives, even though it
avoids a heavy investment on storage. fee subscribers charge more. Hence, the project's profit suffers a
strong decrease of 85%. Scenario C0 increases its income, which
leads to 28% higher profits. As for the energy export to the facilities,
3.2. Economic feasibility analysis they do not differ much between A0, B0 and C0.
The financial indicators concerning all the scenarios are pre-
The financial analysis is performed first for the scenarios based sented in Table 11.
It should start to be noted that all the scenarios are viable for all
Table 10 the indicators, being the payback time between 7 and 24 years e
EV charging sources for the controlled charging approach w/o storage (Scenario A4).
which is in agreement with the literature, e.g. Refs. [70] and [31], in
Averaged daily EV charging by its sources [kWh] which the payback time is 4e26 years and 17e26.5 years,
EV charging respectively.
The base case scenario, A0, presents a NPV of 0.87 MV and a
Directly from PV From the grid
payback of c. a. 16 years, which may be seen as average results.
Annual mean day 2805 0 These results only improve slightly in scenario A1 (e.g., the NPV
Winter 2441 289 raises c. a. 10%), due to the optimal PV layout configuration
Spring 2967 0 considered in the latter. In contrast, scenarios A2, A3 and A4 pre-
Summer 2731 15
sent all much better indicators than the predecessors, which can be
Autumn 2768 0
attributed to the absence of energy storage and the investment
Working day 3292 591 implied.
Weekend 222 0
Scenario A4, which considers a controlled charging plan, shows
1194 R. Figueiredo et al. / Energy 140 (2017) 1182e1197
4
Energy sales to EVs
3
2.08 2.10 Energy sales to train station
1.85
2 facilities
0.87 0.96
0.48 Grid purchases
1
[M€]
0 O&M&Management
-1
Investment
-2
Profit
-3
-4
A0 A1 A2 A3 A4 A5
Fig. 9. Revenues, expenses and profit for scenarios A0-A5 for the project lifetime. Positive values are revenues and negative are expenses.
the best results. They are however marginally better than the re- indicative that the grid support improves significantly the project
sults of scenario A3, which is similar but considering uncontrolled profitability.
charging. This was expected, since the prices of buying energy from Since an investor in this kind of project is greatly driven by a
the grid and selling energy to the train facilities e both the trans- short payback time, desirably less than 10 years, and typically 5e7
actions happen more in A3 e are considered to be the same, and so years [31], this puts scenarios A2, A3 and A4 as the appealing ones,
the small differences in the results may be attributed to different with paybacks of 7.3e7.4 years. Moreover, those scenarios also
efficiencies of directly using the electricity to charge the vehicles come with better NPV, IRR and ROI.
and send it and use it at the facilities. As for A5, considering gradual
deployment, its NPV is the lowest of the scenarios based on the 3.2.1. Sensitivity analysis
reference revenues model (A0-A5), but its IRR is better than some The EV public charging infrastructures have been widely sup-
of them (scenarios A0 and A1); this is due to different patterns of ported by the central state or local governments [71]. This support
cash flows e in A5, the initial energy sales to EVs are lower and to can be subsidies and lower taxes, benefiting all the parties
the train facilities higher, due to a higher PV surplus than in A0 and involved: from the users perspective, it allows them to have access
A1. to a wider EV charging network; from the promotor perspective, it
As for the scenarios considering alternative revenues models, allows lower investment and costs [72]; from the state perspective,
scenario B0, the one that considers a monthly flat fee, presents the the support of cleaner initiatives helps to achieve international
worst results, being profitable just by a small margin, since it has a commitments regarding GHG emissions, promoting societal well-
low NPV and an IRR slightly above 7%. This is due to the assumption being and economic development.
that the cars charge 50% more with this plan, and so it seems that Examples of such policies are in: Portugal, where the govern-
under these conditions the flat fee model is not appealing. How- ment supported in 2010 the installation of a network of 1300
ever, a flat fee business model can lead to different occupancy charging stations spread across the country [71] and, more recently,
patterns and rates, since it may be seen as a marketing strategy to its update and extension [73]; USA, where it is available a federal
get new customers. As for scenario C0, similar to A0 but with tax credit of 30% of the charging infrastructures costs [74]; Japan,
additional revenues from supporting the grid, it shows a 27.5% where governmental incentives cover two-thirds of the total cost of
bigger NPV and a 1.5-year lower payback than the latter, which is the charging stations (it allowed them to be the first country having
0
Grid purchases
-1
O&M&Management
-2
Investment
-3
Profit
-4
-5
A0 B0 C0
Fig. 10. Revenues and expenses for the main revenues models. Positive values are revenues and negative are expenses.
R. Figueiredo et al. / Energy 140 (2017) 1182e1197 1195
Table 11
Financial indicators.
Scenarios
A0 A1 A2 A3 A4 A5 B0 C0
NPV (M€) 0.87 0.96 1.85 2.08 2.10 0.48 0.13 1.11
discounted Payback (years) 16.1 15.5 7.4 7.4 7.3 14.3 24.3 14.5
IRR 10.08% 10.38% 17.36% 18.44% 18.53% 11.30% 7.04% 11.02%
discounted ROI (times) 1.07 1.13 1.59 1.53 1.75 1.25 0.76 1.17
12 18%
to the grid, such as frequency regulation, may improve significantly
10 15%
the project profitability. In addition, a financial incentive of 40% of
IRR [%]
8 12% the initial investment is required to reduce the payback time to 7
6 9% years, in line with what a private investor expects. The allocation of
4 6% such subsidies would thus have a large impact on the profitability
of this business, promoting the deployment of electric vehicles
2 3%
solar parking lots.
0 0%
0% 10% 20% 30% 40% 50%
This study intends to be demonstrative of the energy and
Subsidies [% of investment]
financial potential of a solar parking lot to charge EVs. It presents a
general methodology extendible to other cases. For further studies,
Payback time IRR it would be interesting to consider the uncertainty of the PV gen-
eration (as here it is assumed that the PV generation profile for the
Fig. 11. Sensitivity analysis on the subsidies of a solar parking lot, showing the payback
time and IRR. day is known) and exploit other business models, such as including
revenues coming from commercial advertising on the EVs charging
stations. Also, testing this methodology with parking facilities other
more EV charging points than gas stations) [75]. than park-and-ride, and therefore with potentially very different
These examples reflect the importance of government support occupation profiles, would be of interest. Additionally, the case
in EV adoption. This support can include subsidies to park operators study taken here is a typical park-and-ride facility. Different results
to install charging stations [31]. are expected for projects of other dimensions, which should
Since the more appealing scenarios may not be feasible in the improve with the project size due to economies of scale, and in
present, since they assume as electric all cars parking at the EVSPL, other regions, where the solar resource and financial conditions
an analysis to what extent a public co-financing of the initial in- may be different.
vestment could impact on the EVSPL project attractiveness under
current conditions was performed. And so, taking the scenario A5,
Aknowledgements
the most realistic one, a sensitivity analysis was done to a public
support from 0% to 50% of the total investment. Fig. 11 shows the
The authors gratefully acknowledge the financial support by the
results for the IRR and payback time.
MIT Portugal Program on Sustainable Energy Systems (PD/BD/
It can be seen that the IRR rises from 11.3%, which corresponds
114174/2016), the Portuguese Science and Technology Foundation,
to the scenario A5 with no subsidies, to around 22.5%, for the case of
grant PD/BD/114174/2016, project UID/GEO/50019/2013, and proj-
50% subsidised. As for the payback time, it decreases from 14 years
ect SusCity MITP-TB/CS/0026/2013.
to 6 years. If one considers that the top encouraging payback time
for an investor is of 7 years [31], a subsidy of 40% of the initial in-
vestment is required to make the EVSPL financially appealing. In References
this case, the IRR would be of 19%.
[1] International Energy Agency. World energy outlook 2016-executive summary.
2016. Paris.
4. Conclusions [2] International Energy Agency. Technology roadmap solar photovoltaic energy.
2014. Paris.
[3] International Energy Agency. IEA sankey diagram. 2014. https://www.iea.org/
This work presented the potential and viability of car parking Sankey/#?c¼World&s¼Final consumption. [Accessed 28 June 2017].
lots to solar-charge electric vehicles, using a park-and-ride facility [4] Coffman M, Bernstein P, Wee S. Integrating electric vehicles and residential solar
PV. Transp Policy 2017;53:30e8. https://doi.org/10.1016/j.tranpol.2016.08.008.
as a case study. Different operating and business models were [5] IEA. Snapshot of global PV markets. 2014. p. 2015.
designed and tested. [6] Poullikkas A. Sustainable options for electric vehicle technologies. Renew Sustain
A good correlation between the photovoltaic generation and the Energy Rev 2015;41:1277e87. https://doi.org/10.1016/j.rser.2014.09.016.
[7] Fattori F, Anglani N, Muliere G. Combining photovoltaic energy with electric
parking lot's occupation is crucial to directly use photovoltaics for vehicles, smart charging and vehicle-to-grid. Sol Energy 2014;110:438e51.
EV charging; for this park-and-ride, that correlation is of 52% on https://doi.org/10.1016/j.solener.2014.09.034.
working days. A controlled charging plan is found to create less [8] Honarmand M, Zakariazadeh A, Jadid S. Integrated scheduling of renewable
generation and electric vehicles parking lot in a smart microgrid. Energy
dependence on the grid as opposed to store solar electricity and
Convers Manag 2014;86:745e55.
using it later through an uncontrolled plan. [9] Delmas MA, Kahn ME, Locke SL. The private and social consequences of
The economic feasibility aimed to ascertain the most suitable purchasing an electric vehicle and solar panels: evidence from California. Res
business model for an investor. Considering a discount rate of 7%, all Econ 2017;71:225e35. https://doi.org/10.1016/j.rie.2016.12.002.
[10] Nunes P, Figueiredo R, Brito MC. The use of parking lots to solar-charge
the scenarios show viability, although some by a small margin. The electric vehicles. Renew Sustain Energy Rev 2016;66:679e93. https://
payback time is found to be as low as 7 years, although for current doi.org/10.1016/j.rser.2016.08.015.
1196 R. Figueiredo et al. / Energy 140 (2017) 1182e1197
[11] Bhatti AR, Salam Z, Aziz MJBA, Yee KP, Ashique RH. Electric vehicles charging 2012;15:219e28.
using photovoltaic: status and technological review. Renew Sustain Energy [38] Azimuth Angle | PVEducation. PVEducation n.d. http://pveducation.org/
Rev 2016;54:34e47. https://doi.org/10.1016/j.rser.2015.09.091. pvcdrom/properties-of-sunlight/azimuth-angle (Accessed 16 July 2015).
[12] Ashique RH, Salam Z. Bin Abdul Aziz MJ, Bhatti AR. Integrated photovoltaic- [39] Howey DA, Martinez-Botas RF, Cussons B, Lytton L. Comparative measure-
grid dc fast charging system for electric vehicle: a review of the architec- ments of the energy consumption of 51 electric, hybrid and internal com-
ture and control. Renew Sustain Energy Rev 2017;69:1243e57. https:// bustion engine vehicles. Transp Res Part D 2011;16:459e64. https://doi.org/
doi.org/10.1016/j.rser.2016.11.245. 10.1016/j.trd.2011.04.001.
[13] Xydas E, Marmaras C, Cipcigan LM. A multi-agent based scheduling algorithm [40] Fesharaki VJ, Dehghani M, Fesharaki JJ, Tavasoli H. The effect of temperature
for adaptive electric vehicles charging. Appl Energy 2016;177:354e65. on photovoltaic Cell efficiency. Emerg. Trends energy conserv. Tehran: ETEC;
https://doi.org/10.1016/j.apenergy.2016.05.034. 2011. p. 20e1.
[14] Forrest KE, Tarroja B, Zhang L, Shaffer B, Samuelsen S. Charging a renewable [41] Pasaoglu G, Fiorello D, Martino A, Zani L, Zubaryeva A, Thiel C. Travel patterns
future: the impact of electric vehicle charging intelligence on energy storage and the potential use of electric cars e results from a direct survey in six
requirements to meet renewable portfolio standards. J Power Sources European countries *. Technol Forecast Soc Chang 2014;87:51e9. https://
2016;336:63e74. https://doi.org/10.1016/j.jpowsour.2016.10.048. doi.org/10.1016/j.techfore.2013.10.018.
[15] Liu N, Zou F, Wang L, Wang C, Chen Z, Chen Q. Online energy management of [42] INE. Estatísticas dos transportes. 2009. p. 2010.
PV-assisted charging station under time-of-use pricing. Electr Power Syst Res [43] Solar World. Sunmodule plus SW 260-290 mono - data sheet. 2015.
2016;137:76e85. https://doi.org/10.1016/j.epsr.2016.04.002. [44] Mourin ~ o H. Metodos de avaliaç~ ao de projectos de investimento Me todos de
[16] Nezamoddini N, Wang Y. Risk management and participation planning of avaliaç~ao de projectos de investimento. 2012.
electric vehicles in smart grids for demand response. Energy 2016;116: [45] Hirth L. The market value of variable renewables. The effect of solar wind
836e50. https://doi.org/10.1016/j.energy.2016.10.002. power variability on their relative price. Energy Econ 2013;38:218e36.
[17] García-Trivi P, Torreglosa JP, Fernandez-Ramírez LM, Jurado F. Control and https://doi.org/10.1016/j.eneco.2013.02.004.
operation of power sources in a medium-voltage direct- current microgrid for [46] Photovoltaic geographical information system (PVGIS) - European Commis-
an electric vehicle fast charging station with a photovoltaic and a battery sion. JRC’s Inst Energy Transp n.d. http://re.jrc.ec.europa.eu/pvgis/info/faq.htm
energy storage system. Energy 2016;115:38e48. https://doi.org/10.1016/ (Accessed 19 September 2016).
j.energy.2016.08.099. [47] Ca^mara Municipal de Almada. Almada num Minuto n.d. http://www.m-
[18] Chandra Mouli GR, Bauer P, Zeman M. System design for a solar powered almada.pt/xportal/xmain?
electric vehicle charging station for workplaces. Appl Energy 2016;168: xpid¼cmav2&xpgid¼genericPage&genericContentPage_
434e43. https://doi.org/10.1016/j.apenergy.2016.01.110. qry¼BOUI¼5771022&actualmenu¼5770956 (Accessed 13 June 2016).
[19] Fathabadi H. Novel solar powered electric vehicle charging station with the [48] Fertagus. QUEM SOMOS n.d. https://www.fertagus.pt/pt/testee-quem-somos/
capability of vehicle-to-grid. Sol Energy 2017;142:136e43. https://doi.org/ quem-somos (Accessed 13 June 2016).
10.1016/j.solener.2016.11.037. [49] Regulamento de exploraça ~o de Parques. Fertagus n.d. https://www.fertagus.
[20] Fathabadi H. Novel grid-connected solar/wind powered electric vehicle pt/pt/regulamento-de-exploracao-de-parques (Accessed 27 August 2015).
charging station with vehicle-to-grid technology. Energy 2017;132:1e11. [50] Skoczek A, Sample T, Dunlop ED. The results of performance measurements of
https://doi.org/10.1016/j.energy.2017.04.161. field-aged crystalline silicon photovoltaic modules. Prog Photovoltaics Res
[21] Karan E, Asadi S, Ntaimo L. A stochastic optimization approach to reduce Appl 2009;17:227e40. https://doi.org/10.1002/pip.
greenhouse gas emissions from buildings and transportation. Energy [51] Sunny tripower inverters. Wind sun powering futur n.d. http://www.
2016;106:367e77. https://doi.org/10.1016/j.energy.2016.03.076. windandsun.co.uk/products/Inverters/SMA-Inverters/Sunny-Tripower-
[22] Hafez O, Bhattacharya K. Optimal design of electric vehicle charging stations Inverters#10066 (Accessed 22 July 2015).
considering various energy resources. Renew Energy 2017;107:576e89. [52] Mayrbaurl RM, Camo S. Guidelines for inspection and strength evaluation of
https://doi.org/10.1016/j.renene.2017.01.066. suspension bridge parallel-wire cables. 2004. Washington.
[23] Falahati S, Taher SA, Shahidehpour M. A new smart charging method for EVs [53] Tesla PowerWall Specs. Tesla power mot. 2016. https://www.teslamotors.
for frequency control of smart grid. Int J Electr Power Energy Syst 2016;83: com/en_EU/powerwall?redirect¼no. [Accessed 1 January 2016].
458e69. https://doi.org/10.1016/j.ijepes.2016.04.039. [54] Mayr Florian. How to determine meaningful, comparable costs of energy
[24] Usman M, Knapen L, Yasar AUH, Vanrompay Y, Bellemans T, Janssens D, et al. storage |. Energy Storage News; 2016. https://www.energy-storage.news/
A coordinated framework for optimized charging of EV fleet in smart grid. blogs/how-to-determine-meaningful-comparable-cost-of-energy-storage.
Procedia Comput Sci 2016;94:332e9. https://doi.org/10.1016/ [Accessed 10 November 2016].
j.procs.2016.08.049. [55] Mayer JN. Current and future cost of photovoltaics. 2015.
[25] Kavousi-Fard A, Khodaei A. Efficient integration of plug-in electric vehicles via [56] Bronski P, Creyts J, Guccione L, Madrazo M, Mandel J, Rader B, et al. The
reconfigurable microgrids. Energy 2016;111:653e63. https://doi.org/10.1016/ Economics of Grid Defection: When and where distributed solar generation
j.energy.2016.06.018. plus storage competes with traditional utility service. Boulder. 2014. https://
[26] Bhatti AR, Salam Z, Ashique RH. Electric vehicle charging using photovoltaic doi.org/10.1016/j.enpol.2015.03.005.
based microgrid for remote islands. Energy Procedia 2016;103:213e8. https:// [57] German J. Electric Vehicles : performance, cost, penetration. CCAP MAIN-Latin
doi.org/10.1016/j.egypro.2016.11.275. Am. Reg. Dialogue. The International Council on Clean Transportation - ICCT;
[27] Kamankesh H, Agelidis VG, Kavousi-Fard A. Optimal scheduling of renewable 2014.
micro-grids considering plug-in hybrid electric vehicle charging demand. [58] MCC16-Kiosk with satellites. Magnum Cap n.d. http://magnumcap.com/index.
Energy 2016;100:285e97. https://doi.org/10.1016/j.energy.2016.01.063. php/contacts?id¼24 (Accessed 31 July 2015).
[28] Tabatabaee S, Mortazavi SS, Niknam T. Stochastic scheduling of local distri- [59] Electric charging in Europe. New motion n.d. https://www.thenewmotion.
bution systems considering high penetration of plug-in electric vehicles and com/en/products/electric-charging-in-europe/(Accessed 20 August 2016).
renewable energy sources. Energy 2017;121:480e90. https://doi.org/10.1016/ [60] Tarifas Transito rias de Venda a Clientes Finais em Portugal Continental em.
j.energy.2016.12.115. 2017. p. 2017.
[29] Rahbari O, Vafaeipour M, Omar N, Rosen MA, Hegazy O, Timmermans J-M, [61] Entidade Reguladora dos Serviços Energe ticos (ERSE). Tarifas e preços para a
et al. An optimal versatile control approach for plug-in electric vehicles to energia eletrica e outros serviços em 2015 e par^ ametros para o período de
integrate renewable energy sources and smart grids. Energy 2017;134: regulaça~o 2015-2017. 2014.
1053e67. https://doi.org/10.1016/j.energy.2017.06.007. [62] Inflation forecast. OECD data. 2016. https://data.oecd.org/price/inflation-
[30] Mozafar MR, Moradi MH, Amini MH. A simultaneous approach for optimal forecast.htm#indicator-chart. [Accessed 2 September 2016].
allocation of renewable energy sources and electric vehicle charging stations [63] Sketchup Skelion solar design plugin renewable energy. n.d. http://skelion.
in smart grids based on improved GA-PSO algorithm. Sustain Cities Soc com/(Accessed 17 September 2016).
2017;32:627e37. https://doi.org/10.1016/j.scs.2017.05.007. [64] Ban~ os R, Manzano-Agugliaro F, Montoya FG, Gil C, Alcayde A, Go mez J.
[31] Robinson J, Brase G, Griswold W, Jackson C, Erickson L. Business models for Optimization methods applied to renewable and sustainable energy: a review
solar powered charging stations to develop infrastructure for electric vehicles. 2011;15. https://doi.org/10.1016/j.rser.2010.12.008.
Sustainability 2014;6:7358e87. https://doi.org/10.3390/su6107358. [65] Diouf B, Pode R. Potential of lithium-ion batteries in renewable energy. Renew
[32] LNEG - National laboratory of energy and geology n.d. http://www.lneg.pt/ Energy 2015;76:375e80. https://doi.org/10.1016/j.renene.2014.11.058.
iedt/unidades/25/(Accessed 8 December 2014). [66] Qamar Raza Muhammad, Nadarajah Mithulananthan, Ekanayake Chandima.
[33] Marion W, Urban K. User's manual for TMY2sdtypical meteorological years. On recent advances in PV output power forecast. Sol Energy 2016;136:
Gold Co Natl Renew Energy Lab 1995:55. 125e44. https://doi.org/10.1016/j.solener.2016.06.073.
[34] Gilman PSAM. Photovoltaic model technical reference SAM photovoltaic [67] Nunes P, Farias T, Brito MC. Day charging electric vehicles with excess solar
model technical reference. 2015. electricity for a sustainable energy system. Energy 2015;80:263e74. https://
[35] Noorian AM, Moradi I, Kamali GA. Evaluation of 12 models to estimate hourly doi.org/10.1016/j.energy.2014.11.069.
diffuse irradiation on inclined surfaces. Renew Energy 2008;33:1406e12. [68] Chukwu UC, Mahajan SM. V2G parking lot with pv rooftop for capacity
https://doi.org/10.1016/j.renene.2007.06.027. enhancement of a distribution system. IEEE Trans Sustain Energy 2014;5:
[36] Santamouris M. Environmental design of urban buildings - an integrated 119e27. https://doi.org/10.1109/TSTE.2013.2274601.
approach. London: Earthscan; 2006. [69] Preço da Banda de Regulaça ~o Secunda ria. Sist Inf Mercados Energ 2014. http://
[37] Benkaciali S, Gairaa K. Comparative study of two models to estimate solar www.mercado.ren.pt/PT/Electr/InfoMercado/InfOp/BandaSecundaria/
radiation on an inclined surface the radiometric station was installed on the Paginas/Preco_BandaSec.aspx. [Accessed 29 August 2016].
roof of Solar Radiation Laboratory of. Rev Des Energies Renouvelables [70] Tulpule PJ, Marano V, Yurkovich S, Rizzoni G. Economic and environmental
R. Figueiredo et al. / Energy 140 (2017) 1182e1197 1197
impacts of a PV powered workplace parking garage charging station. Appl 8 de junho de 2016 n.d. http://www.portugal.gov.pt/pt/o-governo/cm/
Energy 2013;108:323e32. comunicados/20160608-cm-com.aspx (Accessed 16 August 2016).
[71] REDE MOBI.E e Definiç~ ao e Pontos de Carregamento n.d. https://www.mobie. [74] Tax Credit for Electric Vehicle Stations - ChargePoint. Charg point. 2016.
pt/a-rede-mobi.e (Accessed 28 January 2015). http://www.chargepoint.com/station-incentives/. [Accessed 29 August 2016].
[72] Yang T, Long R, Li W, Rehman S. Innovative application of the publiceprivate [75] McCurry Justin. Japan now has more electric car charge points than petrol
partnership model to the electric vehicle charging infrastructure in China. stations. World news; 2016. The Guardian, https://www.theguardian.com/
Sustainability 2016;8:738. https://doi.org/10.3390/su8080738. world/2016/may/10/japan-electric-car-charge-points-petrol-stations.
[73] República Portuguesa. Comunicado do Conselho de Ministros do Ambiente de [Accessed 29 August 2016].