This article has been accepted for publication in a future issue of this journal, but has not been

fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TTE.2020.3015743, IEEE
Transactions on Transportation Electrification

A Comprehensive Review of DC Fast Charging

Stations with Energy Storage: Architectures, Power
Converters, and Analysis
Md Ahsanul Hoque Rafi, Student Member, IEEE and Jennifer Bauman, Member, IEEE

 drives without requiring very costly EVs with large batteries.

Abstract— Electric vehicle (EV) adoption continues to rise, yet Though typical home chargers of AC Level 1 or Level 2 are
EV sales still represent a small portion of vehicle sales in most well-suited for overnight charging for typical inner-city driving,
countries. An expansion of the DC fast charging (DCFC) network the need for recharging away from home on days with longer
is likely to accelerate this revolution towards sustainable
driving can cause inconvenience and uncertainty for drivers. A
transportation, giving drivers more flexible options for charging
on longer trips. However, DCFC presents a large load on the grid survey among 500 Nissan customers verifies this claim: 56% of
which can lead to costly grid reinforcements and high monthly customers placed high importance on fast charging, and 62%
operating costs – adding energy storage to the DCFC station can stated that the availability of charging infrastructure was their
help mitigate these challenges. This paper performs a primary concern regarding BEVs [4]. To help alleviate some of
comprehensive review of DCFC stations with energy storage, these concerns, most manufacturers offer BEVs capable of fast
including motivation, architectures, power electronic converters,
charging, as shown in Table 1 [5], with the top charging power
and a detailed simulation analysis for various charging scenarios.
The review is closely tied to current state-of-the-art technologies being 150 kW for the Audi e-Tron and Tesla Model 3. Though
and covers both academic research contributions and real energy these vehicles use 400 V battery systems, Aston Martin and
storage projects in operation around the world. Porsche are developing BEVs with 800 V batteries to make
even higher charging rates possible, due to reduced currents and
Index Terms— Battery chargers, energy storage, batteries, losses in the charging cables [6], [7]. Therefore, various DC fast
flywheels, hydrogen, vehicles. charging standards are emerging to serve this critical need for
faster EV charging. Table II shows a comparison of AC home
chargers and different DC fast charging options [8], assuming
I. INTRODUCTION
typical vehicle energy consumption is 177 Wh/km or 265

E LECTRIC vehicles (EVs) have the potential to

significantly reduce greenhouse gas emissions, regulated
emissions which cause local air pollution and negative health
Wh/mile and charger efficiency is 100%. The results show that
the higher charging rate options allow long-distance driving
with only minimal driving interruptions for charging.
effects [1], and society’s reliance on fossil fuels. Global EV However, fast and ultrafast charging directly from the grid
sales, which include battery electric vehicles (BEVs) and plug- poses a large and unpredictable load on the electric system [9]–
in hybrid electric vehicles (PHEVs), increased by 46% from the [11] and in rural areas with weak grid connections, (ultra)fast
first half of 2018 to the first half of 2019, with 1.134 million charging directly from the grid is not even possible [12]. Since
EVs delivered globally in the first half of 2019 [2]. However, much of the need for (ultra)fast chargers is on long drives,
global EV light vehicle market share was only 2.5% for the first possibly outside of cities, these charging stations should ideally
half of 2019 [2]. Thus, while the transition to sustainable be located along highways between cities. The electrical
personal transportation is accelerating, there is still much infrastructure along some of these highways is usually far from
improvement needed. the main distribution grid, and thus it has high impedance with
The main barriers to EV adoption are high vehicle costs, low short circuit power. Moreover, only a single-phase 120 V
range concerns, and charging infrastructure. Though range can or 240 V AC supply may be available at some rural locations
be improved with larger and more energy-dense batteries, there along these highways. Therefore, the installation of fast
is a limit to what can reasonably be achieved due to the charging stations is not possible at such locations without
associated increases in vehicle mass and cost – specifically, the upgrading the electrical service, which usually results in high
battery can account for up to 50% of the total BEV cost [3]. The installation cost [13]. Energy storage system (ESS) can
installation of comprehensive EV fast charging networks would mitigate the necessity of such high cost of service upgrading by
help alleviate range and charging concerns on longer inter-city acting as buffer in between the grid and the vehicle.

M.A.H. Rafi is with the Electrical and Computer Engineering Department, J. Bauman is with the Electrical and Computer Engineering Department,
McMaster University, Hamilton, ON, L8S4L8, Canada (e-mail: McMaster University, Hamilton, ON, L8S4L8, Canada (e-mail:
[email protected]). [email protected]).

2332-7782 (c) 2020 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
Authorized licensed use limited to: SUNY AT STONY BROOK. Downloaded on August 12,2020 at 11:54:55 UTC from IEEE Xplore. Restrictions apply.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TTE.2020.3015743, IEEE
Transactions on Transportation Electrification

TABLE I: TOP CHARGING POWER OF CURRENTLY AVAILABLE EVS [5] flywheel, hydrogen, and hybrid battery-flywheel) which
Top Charging Speed includes a new discussion on isolation requirements (Section
EV Make and
Charging Drops Above SOC
Model
Power (kW) (%) IV), (iii) a review of power electronic converters for use in fast
Audi e-tron 150 80 charging stations with energy storage, uniquely discussing the
Tesla Model 3 150 - option of single-phase grid connections for low grid power
Mercedes-Benz 110 40
Jaguar I-PACE 100 40
situations (Section V), (iv) modeling and analysis to
Hyundai Kona 77 58 quantitatively compare the ESS options for three station sizes
Kia e-Niro 77 58 in terms of efficiency, volume, and cost (including lifetime
Hyundai Ioniq 70 75
BWM i3 50 85
concerns), and considering multiple flywheel control strategies
Nissan Leaf 50 60 (Section VI), (v) simulation and detailed comparison of battery
VW e-Golf 40 80 ESS architectures (Section VII), and (vi) a discussion of future
trends in energy storage for DC fast charging stations (Section
TABLE II: SUMMARY OF CHARGING STANDARDS [8] VIII). Conclusions are summarized in Section IX.
Typical range Time to charge
gained per minute for 200 km
of charge (km) (minutes) II. MOTIVATION FOR FAST CHARGING STATIONS WITH
AC On-board Charging: ENERGY STORAGE
Level 1 (120V, 1.4kW) 0.13 1332
Level 2 (240V, 7.2kW) 0.68 260 A. Reinforcement of Grid Infrastructure
DC Off-board Charging:
Fast Charging (480V, 50kW) 4.70 37
DC fast charging (DCFC) stations, especially those providing
Tesla Supercharger (480V, 40kW) 13.15 13 charging power over 100 kW, can require high initial
Ultrafast Charging (800V, 400kW) 37.50 4.7 investment for grid reinforcement. For instance, a 350 kW
charger will require a considerably more expensive utility
Furthermore, an ESS is also often advantageous in fast upgrade compared to a 50 kW charger in case of a weak grid
charging stations with strong grid connections, as it prevents connection [23]. Moreover, it is not always possible to be
overloading the grid during charging [14]. Expensive grid directly connected to a strong grid, for instance, along the
reinforcement can be required with multiple fast charging ports highway and countryside roads – though these locations may be
at strong grid locations, but the use of ESSs can avoid or where DCFC stations are needed the most, i.e., between cities.
partially mitigate these infrastructure costs. An ESS will also Furthermore, according to California Public Utilities
help reduce the utility demand charges incurred due to fast Commissions’ studies, as the fast charging network expands,
charging and it is estimated that a charging station with six 350 the distribution lines and transformers will require more
kW chargers can save around $157,000 annually by using a upgrades than power plants, transmission lines, and substations.
battery storage system [15]. Further, the vision of sustainable For instance, a typical Electrify America highway charging
mobility will not be fulfilled until EVs are charged with station with two 350 kW and four 150 kW chargers will
renewable energy sources such as solar and wind. ESSs can introduce an additional 1.2 MW power demand and will need
store the energy generated from these intermittent renewable an upgraded or additional costly distribution transformer [14].
sources in order to charge EVs when needed. In Norway, it is estimated that $1.6 billion will be required for
This paper performs a comprehensive review and analysis of grid reinforcement to accommodate the uncontrolled charging
EV fast charging stations with energy storage. Previous behavior of the large numbers of EVs expected on the road by
literature has provided excellent reviews on EV charging in 2040 [24].
general [16], and EV fast charging with a focus on connection Although smart-charging strategies have been proposed in
to the medium voltage (MV) grid [17], [18], but no review [25]–[29] to garner more control over EV charging for grid
papers have focused on EV fast charging with energy storage. benefit, they are best applied to overnight home charging where
Research in this area is usually focused on a specific design. For there is ample time to shift individual vehicle charging times
example, [19] presents detailed design and implementation while still providing a full charge by the following morning. In
results of a fast charging station with battery energy storage, contrast, drivers requiring fast charging generally need high
and [20] proposes flywheel energy storage for public fast power charging immediately upon vehicle plug-in. A better
charging stations, but does not provide a detailed analysis. solution for DCFC stations is to use an ESS so that the power
Other prior reviews focus on grid energy storage in general required from the grid may be low enough to minimize or even
without considering the specific needs of fast charging stations: negate the need for grid reinforcement. The ESS can be charged
[21] performs detailed loss modeling of flywheels and when the electricity demand and price is low, and EVs can be
ultracapacitors for grid energy storage and [22] performs an charged with higher power from the ESS without creating a
excellent review of power electronic converters for batteries burden on the grid and at a predefined flat rate. In a grid with
used for general grid energy storage. Thus, this review paper lower grid capacity, such as those in remote areas, it can be less
fills a gap in the literature, and the main contributions are: (i) a costly and less complex to deploy an ESS-based charging
detailed discussion on motivations for using energy storage in station rather than reinforcing the grid infrastructure [30].
fast charging stations (Section II), (ii) a comprehensive
B. Operational Cost of DCFC Stations
presentation of architectures utilizing four ESS options (battery,
The high power draw of DCFC stations generates high

2332-7782 (c) 2020 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
Authorized licensed use limited to: SUNY AT STONY BROOK. Downloaded on August 12,2020 at 11:54:55 UTC from IEEE Xplore. Restrictions apply.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TTE.2020.3015743, IEEE
Transactions on Transportation Electrification

demand charges, which can make investment in and operation

of DCFC stations unappealing to investors. The utility charges
demand charges based on the highest monthly peak power
drawn per non-residential customer, including charging
stations, in order to recoup the fixed costs for generation,
transmission, and distribution. In the U.S., demand charges can
range from $2/kW to $90/kW [31]. For example, a DCFC
station may have three 50 kW charging ports, a fixed energy
price of $0.10/kW, a fixed monthly connection charge of $140
[10], and a monthly demand charge of $35/kW. If it is assumed
that three EVs charge at the station per day, and each EV
charges at 50 kW for 20 minutes (using 17 kWh of energy), then
three scenarios are possible: one, two, or three EVs are charging
at one time. These scenarios correspond to the monthly demand
charges of $1750, $3500, and $5250, respectively. Fig.1. Monthly operating cost of a DCFC with different charging scenarios
An ESS can help reduce the monthly cost incurred due to and payback period of the capital investment on energy storage
demand charges. A battery energy storage system (BESS) can
be installed such that all the energy required for charging three an ESS can be used as an intermediate energy buffer from the
EVs simultaneously at the station is supplied from the BESS. In renewable generation to the somewhat unpredictable EV
this particular example, the BESS would be characterized as charging needs.
high power density (capable of supplying power at a 3-C rate) Reference [33] provides a comprehensive review of the
and the efficiency of the BESS (from battery to EV) can be function of ESSs at different stages within power systems with
estimated at 90%. Thus, the capacity of the BESS should be a high penetration of renewable energy sources. An ESS can
around 56 kWh. If only three EVs per day are charged at the help reduce voltage and frequency instability and help peak
station, the BESS can be charged using a single phase 3.3 kW shaving by storing energy from large-scale photovoltaic (PV)
charger connected to the grid. Thus, the peak power plants [34]–[36]. One of the world’s largest battery storage
requirement from the grid to the charging station would be 3.3 projects of 60 MWh has been implemented to integrate solar
kW, resulting in a monthly demand charge of only $115.50. energy into the grid in Hokkaido, Japan [37]. Similar projects
To achieve this monthly savings, the additional cost of the are also initiated for more efficient integration of wind energy.
BESS must be considered. The initial cost of the BESS includes For instance, a 24 MWh BESS was integrated into a 156.6 MW
power electronic converters and the battery itself. For a DCFC wind farm in west Texas for storing power during off-peaks
station without energy storage, high power AC/DC and DC/DC hours [38]. Another similar but larger 32 MWh lithium-ion
converters are required (as discussed in Section IV). For a BESS was installed in southern California to provided voltage
DCFC station with energy storage, a low power AC/DC support and frequency regulation near a 4500 MW wind farm
converter and a high power DC/DC converter will suffice since [39]. Different Power to Gas (P2G) projects have been initiated
the AC connection will be rated for lower power. Thus, the to produce hydrogen with excess renewable energy. For
costs of the power electronic converters between the two DCFC instance, a 2 MW P2G installation has been established by
station types can be considered approximately equivalent, and E.ON, a German utility company in partnership with Swissgas
likely lower for the DCFC station with energy storage. Thus, in and Hydrogenics in Falkenhagen, Germany [40]. A
this example, only the additional cost of the battery itself will comprehensive list of other renewable power to hydrogen gas
be considered. If the battery cost is estimated at $390/kWh [32], projects in Australia, Canada, China, Germany, Japan,
then the initial capital cost for the battery is $21,840. This Netherlands and the U.S. can be found in [41]. Moreover,
capital cost of the battery will be paid back within 4 to 14 various ESSs are being used to provide peak-shaving,
months depending on the monthly demand charges as shown in frequency regulation, and improve the power quality and
Fig. 1. After payback, the monthly operating cost for the reliability of the electrical grid. For instance, a 24 MWh and a
charging station will include only the demand charge of 14 MWh sodium-sulfur BESS from PG&E were installed in
$115.50, and fixed and used energy costs. East San Jose and Vacaville, California, [42] to improve power
quality and reliability of the grid. A 5 MWh/20 MW flywheel
C. Use of Renewable Energy Sources and Services to the
Grid storage system was installed by Beacon Power in Hazel
Township, Pennsylvania to provide frequency regulation [43].
EVs have the potential to create a sustainable transportation As discussed previously, adding energy storage to DCFC
system with high efficiency and greatly reduce emissions – stations can reduce the initial investment required (through
however, this goal can only be achieved if the electricity used smaller or no grid reinforcements) and can reduce the
to charge EVs is generated from renewable sources such as
operational costs (through a reduction in monthly demand
solar and wind. A major challenge in achieving this goal is to charges). The real world installation examples above show that
match the charging demand with electricity generation from stationary grid energy storage can be beneficial for providing
unpredictable and/or intermittent renewable sources. However, services to the grid without even considering its use in DCFC

2332-7782 (c) 2020 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
Authorized licensed use limited to: SUNY AT STONY BROOK. Downloaded on August 12,2020 at 11:54:55 UTC from IEEE Xplore. Restrictions apply.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TTE.2020.3015743, IEEE
Transactions on Transportation Electrification

stations – the combination of using ESSs to provide grid TABLE III: ENERGY AND POWER DENSITY OF DIFFERENT BATTERY TYPES
[49]
services with DCFC operation provides even better financial
High Energy, Medium Energy Low Energy,
incentive to invest in and operate DCFC stations. Therefore, the Low Power & Power High Power
ESS can act as a virtual power plant coupled with solid-state Module Rack Module Rack Module Rack
electronics interfaces [44], playing a significant role as DCFC Energy
stations and renewable energy sources become widespread. Density 230 150 210 130 160 100
(kWh/m3)

III. ENERGY STORAGE OPTIONS Power

Density 120 80 210 130 480 300
This section discusses the most prominent energy storage (kW/m3)
technologies for use in DCFC stations found in the literature: Maximum
batteries, flywheels, and hydrogen. Other technologies have Discharge 0.5 C 1.0 C 3.0 C
been investigated, such as superconducting magnetic energy Rate
storage (SMES) [45], [46], yet the high cost and low energy
density of current SMES technology makes it infeasible for
near-term implementation.
Electrochemical batteries have high efficiency, high energy
density, and rapidly reducing costs. Lithium-ion technology is
a top contender in these areas [19], and thus will be considered
in this paper. A typical BESS is shown in Fig. 2, including
numerous battery packs, a battery management system (BMS),
and a thermal management system. Lithium-ion based BESSs
offer high efficiency in the range of 94% to 98% [47],
depending on series and parallel cell connections and cell type.
The calendar life often ranges from 10 to 15 years [48], but high
cycling rates can lower actual lifetime, presenting replacement
cost concerns. However, the use of second-life EV batteries can
reduce the initial investment required but adds new challenges Fig. 2. A general overview of the components in a BESS
for integration, as discussed in Section VIII. The power and
energy density for different types of cells and arrangements are If the flywheel is powered from grid AC voltage, then an
shown in Table III, for Samsung SDI battery systems [49]. AC/DC and DC/AC converter are required within the flywheel
BESSs are currently used in numerous grid storage and system [57]. Mechanical bearings contribute to high spinning
DCFC station applications. EVgo, the largest fast charging losses, but the alternative lower-loss magnetic bearings [58] are
network in the U.S., has BESSs at 11 of its DCFC stations generally too costly to be commercially feasible.
across the U.S. [50]. Volkswagen’s Electrify America has 1 2
announced the use of Tesla Powerpack BESSs in their DCFC E I (1)
2
stations [51]. In addition, Tesla has introduced Supercharger V3
stations with PV and BESSs [52]. In terms of published The main advantages of FESSs are high cycle lifetimes and
research, [9] and [53] find that a BESS helps reduce the grid high power density. In general, flywheels can complete
impact of EV charging. Reference [54] presents a cost-benefit between 175,000 to 200,000 cycles over a 20 year life span [59].
analysis of DCFC stations connected to the LV grid with BESS Moreover, FESSs require less maintenance over their lifetime
compared to MV grid connection, and finds that the short cycle compared to BESSs [48]. However, the energy density of
life of batteries is a barrier to achieving higher cost-benefit due FESSs is less than that of BESSs, and the high self-discharge
to the replacement cost of the batteries. However, the results rate must be considered. Amber Kinetics has developed a
show that BESSs provide better financial return compared to flywheel system with energy storage capability of more than
direct MV grid connection when the number of charging events four hours with minimal self-discharge losses [56]. However,
is from 10 to 325 per day. this technology suffers from low power output capability and
Flywheels generally have higher power density and lower would require multiple units to supply the necessary high power
energy density than batteries, and consist of a spinning rotor, of a DCFC station. Table IV shows the energy and power
electric motor, bearings, a power electronics interface, and a density of three commercial flywheels.
vacuum housing [55], as shown in Fig. 3. The stored energy in Although FESS-based DCFC stations have not yet been
the rotating rotor is proportional to the rotor’s moment of implemented in the real world, FESSs are used for grid storage
inertia, I, and the square of the rotating speed, ω, as shown in applications due to their high power density, and [60] outlines
(1). Since the motor runs on AC power, power electronics are a comprehensive list of FESSs for frequency regulation, voltage
integrated into the flywheel system to generate the specific AC support, and renewable energy integration. In [20], the use of
voltages needed to control the flywheel motor to store or release FESS in DCFC stations is found to give a higher payback than
energy. If the flywheel is powered from DC voltage, then only BESSs, and [61] presents a control algorithm for FESSs in a
a DC/AC converter is required within the flywheel system [56]. DCFC station. Furthermore, hybrid ESSs can be advantageous

2332-7782 (c) 2020 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
Authorized licensed use limited to: SUNY AT STONY BROOK. Downloaded on August 12,2020 at 11:54:55 UTC from IEEE Xplore. Restrictions apply.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TTE.2020.3015743, IEEE
Transactions on Transportation Electrification

discussed in the literature for fast charging stations. In AC-bus

stations, the secondary side of a MV-LV distribution
transformer is used as the common AC bus [18], [72] that the
EV chargers connect to. For instance, the commercially
available DCFCs from Charge Point and ABB operate with AC
input voltage of 400-480 V [73], [74] and for a multiport station
using these chargers, each charger will connect to the common
AC bus. Power electronic converters, which provide
rectification, power factor correction (PFC), isolation, and
voltage control, connect the main AC bus to the DC EV ports,
and to energy storage systems and/or renewable energy sources,
if used. The AC-bus architecture requires multiple power
conversions to connect to DC loads and sources, and has more
Fig. 3. A general overview of a Beacon Power flywheel system [57] complex control [75]. In DC-bus stations, the EV chargers are
all connected to a common DC bus, meaning only isolated
TABLE IV: Commercial Flywheel Technologies [56], [57]
DC/DC converters are required between the DC bus and the EV
Amber Beacon Beacon Power
Manufacturer Kinetics Power (Model 450 port. Thus, the DC-bus architecture generally has lower cost,
(Model 400) XP) smaller size, and better dynamic performance compared to the
Maximum Power (kW) 8 100 360 AC-bus architecture [76]. Therefore, this paper focuses on the
Energy (kWh) 32 25 36 DC-bus architecture. However, it should be noted that
Size (m3) 1.65 1.36 1.36
Energy Density(kWh/m3) 19.40 18.36 26.47
designing protection systems for high voltage DC-bus systems
Power Density (kW/m3) 4.85 75.53 264.71 is a challenge because DC current has no natural zero crossing
[77], [78] and the reliability of DC-bus architectures relies on
because the high power density of FESSs can be paired with the central rectification stage [76].
high energy density storage such as batteries. For example, [62] The selection of the nominal DC bus voltage will depend
proposes a hybrid flywheel-battery system to compensate the mainly on two factors: the expected EV battery voltage and the
power fluctuation of large scale wind farms. In [63], a hybrid maximum charging power (since higher powers will require
system is controlled in an islanded microgrid so that the FESS higher voltages to keep current levels reasonable). Furthermore,
provides support during transients and the BESS sustains long- when energy storage is to be connected directly to the bus, the
term load changes. Thus, this paper analyzes both a FESS and voltage range of the storage system must be designed or chosen
a hybrid flywheel-battery ESS. to match the desired bus voltage. For 400 V EVs (in which the
Hydrogen storage has the advantage of high energy density battery voltage may range from about 250 V to 450 V), the DC
and the possibility of dual use in fueling hydrogen powered bus voltage will generally be less than 1000 V [17], and [79]
vehicles as well as EVs. The energy density of compressed discusses a commercially-available DC/DC converter for 800
hydrogen gas at 700 bar is ~1300 kWh/m3 [64] and the energy V electric buses with input voltage (i.e., DC bus voltage) up to
density of current 700 bar storage tank systems is ~800 kWh/m3 1200 V.
[65]. Using an electrolyzer, electricity from the grid is The following subsections review the most promising
converted to hydrogen, compressed, stored in a high-pressure architectures for each considered energy storage system type,
gas tank, and then converted back to electricity when needed by and a detailed discussion of the power electronic converters that
fuel cells, as shown in Fig. 4. Electrolyzers can have 85% can be used in these architectures is provided in Section V. One
efficiency at full load [66] and fuel cell peak efficiency ranges
from about 50-60% [67], [68], thus, the round-trip efficiency is
lower than that of batteries.
Though there are numerous hydrogen storage projects
operating worldwide for storing excess renewable energy [40],
[41], hydrogen storage has not yet been used in DCFC stations.
However, there is a project on the combined operation of DCFC
with a hydrogen refuelling station by the Institute of
Transportation Studies at University of California, Davis [69].
Furthermore, [70] and [71] propose the use of hydrogen storage
for the dual purpose of filling fuel cell vehicles and providing
grid services, such as demand response.

IV. ARCHITECTURES OF DCFC STATIONS WITH ENERGY

STORAGE Fig. 4. Overview of the components in a hydrogen storage system with on-site
Two main architectures, AC-bus and DC-bus stations are hydrogen and electricity generation

2332-7782 (c) 2020 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
Authorized licensed use limited to: SUNY AT STONY BROOK. Downloaded on August 12,2020 at 11:54:55 UTC from IEEE Xplore. Restrictions apply.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TTE.2020.3015743, IEEE
Transactions on Transportation Electrification

common requirement across all architectures is that each EV One option to reduce components is to use an isolated
should be galvanically isolated from the DC bus so that there DC/DC converter that can also perform power factor correction,
are no safety concerns with ground current flowing between as shown in Fig. 6(a). Different isolated DC/DC converters with
multiple EVs plugged into the same DCFC station [17], [80]. PFC capability can be found in the literature for various power
Thus, all architectures include a high power isolated DC/DC levels [87]–[91]. With an isolated PFC stage, the overall size
converter between the DC bus and the EV. and cost of the AC-DC conversion stage is reduced while
improving the efficiency. In order to provide services to the
A. Architectures with Battery ESSs
grid, a DCFC station with energy storage must have a
A battery ESS DCFC architecture with common DC-bus bidirectional connection to the grid, as shown in Fig. 6(b). This
must provide AC-DC rectification, power factor correction, active front end can also provide power factor correction due to
voltage control, and isolation between the battery and the grid active control of the reactive power flow using pulse width
[81]. This isolation is recommended to prevent any ground modulation (PWM) control.
faults from flowing in the case of a degraded or compromised
battery casing [82], [83]. This requirement for isolation B. Architectures with Flywheel ESSs
between the grid and the storage battery is an additional Unlike BESSs, commercial FESSs have one or more power
constraint that is not present for DCFC stations without energy converters integrated into the system for controlling the
storage, where isolation is only required between the grid and flywheel motor. Thus, all commercial flywheel systems include
the EV. Fig. 5 shows three methods for achieving isolation a DC-to-high frequency AC converter to control the flywheel
between the grid and the storage battery. In Fig. 5(a), a motor. For a DC-input flywheel system, this is the only power
dedicated low-frequency transformer is used, along with a converter included in the module, and this flywheel system
standard diode rectifier and PFC DC/DC converter. This
architecture has the advantage of simplicity, as fewer power
electronic converters are required compared to the other
methods of achieving isolation. However, a large low-
frequency transformer can add cost and volume to the system.
Fig. 5(b) instead uses a high-frequency transformer in an
isolated DC/DC converter, which will reduce system volume.
In both of these architectures, the battery is directly connected
to the DC bus, meaning this bus voltage will vary as the battery
state-of-charge (SOC) changes and as the battery terminal
(a)
voltage fluctuates with charging and discharging events. Thus,
the PFC DC/DC converter in Fig. 5(a) and the isolated DC/DC
converter in Fig. 5(b) must be designed to allow the output
voltage swing to vary fully with the battery terminal voltage
swing. Furthermore, this direct battery connection to the DC
bus means the isolated DC/DC converter which connects to the
EV must have a wide input voltage range and a wide output
voltage range (corresponding to the battery voltage swing in the
EV as its SOC rises). The BESS voltage can vary widely
depending on the battery cell and module design, for example
(b)
with ranges of 730-946 V [49] and 486-756 V [84]. This
constraint complicates the design of the EV DC/DC converter
since soft switching is hard to achieve over a wide range of
input and output voltage [85], [86].
Fig. 5(c) shows an alternative where the DC bus voltage is
fixed due to the use of an isolated DC/DC converter between
the DC bus and the BESS. This fixed bus voltage simplifies the
design and control of the other DC/DC converters in the system.
However, in smaller DCFC stations that have a low-power
connection to the grid (e.g., 10kW-50kW), the isolated DC/DC
converter connected to the battery ESS would need to be sized
large enough to provide the high EV charging power (e.g., 150
kW). This would result in a larger and more costly isolated
DC/DC converter compared to that required in the lower-power (c)
path in Fig. 5(b). Each isolation strategy has advantages and Fig. 5. DCFC architecture with BESS where isolation is provided with (a)
dedicated low frequency transformer (varying DC bus voltage), (b) isolated
disadvantages, and the strategy shown in Fig. 5(b) will be used DC-DC converter (varying DC bus voltage), and (c) isolated DC-DC converter
in the discussion of other architecture options. (fixed DC bus voltage).

2332-7782 (c) 2020 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
Authorized licensed use limited to: SUNY AT STONY BROOK. Downloaded on August 12,2020 at 11:54:55 UTC from IEEE Xplore. Restrictions apply.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TTE.2020.3015743, IEEE
Transactions on Transportation Electrification

C. Architectures with Hydrogen ESSs

In a DCFC station with hydrogen storage, the electrolyzer
does not need to be galvanically isolated from the grid, yet the
fuel cell is a generation source, and thus should be isolated from
the grid [92]. It is common practice to use a DC/DC converter
at the output of a fuel cell because fuel cells have widely
varying output voltages dependent on load and fuel cells behave
like a current source with low voltage and high current [93].
Thus, the architectures considered in this paper use a DC/DC
converter at the fuel cell output.
(a) Fig. 8(a) shows a hydrogen storage DCFC architecture with
a dedicated low-frequency transformer at the AC grid
connection. In this case, a non-isolated DC/DC converter can
be used at the fuel cell output, which will be lower-cost and
higher-efficiency than an isolated DC/DC converter of the same
power rating. Since the fuel cell to EV path is the highest power

(b)
Fig. 6. DCFC architecture with BESS where (a) PFC is provided by an isolated
DC-DC converter (unidirectional) and (b) PFC is performed with active
switches along with rectification (bidirectional)

connects to a DC bus which can range in voltage from about

550 V to 750 V [56]. For an AC-input flywheel system, an
additional AC/DC converter (active front end) is included to
allow bidirectional power flow between the flywheel and the
AC grid [57]. The type of input connection of a commercial
flywheel system will determine where in the architecture the (a)
FESS should be connected.
As per IEEE 1547 standard [92], galvanic isolation should
also be provided between the FESS and the grid. Similar to the
BESS, this isolation can be provided by a dedicated low-
frequency transformer or an isolated DC/DC converter with
high-frequency transformer. Fig. 7(a) shows a DCFC
architecture for an AC-input FESS, where, due to the grid-
frequency AC input requirement, it is preferable to connect the
flywheel before the rectification stage rather than at the
common DC bus. A low-frequency transformer is preferred for
isolation in AC-input FESS-based stations to reduce the power
(b)
electronic conversion stages. A diode rectifier and PFC DC/DC
converter (or active front end) are used to create the DC bus for
EV connection.
DC-input FESS-based DCFC stations are shown in Figs. 7(b)
and 7(c). Fig. 7(b) shows the use of a low-frequency
transformer for isolation, and Fig. 7(c) shows the use of an
isolated DC/DC converter. For a smaller DCFC station with a
lower-power grid connection, the DC-input FESS is
advantageous because there are less high-power converters
required in the path between the flywheel and the EV: two
converters are in this path in Figs. 7(b) and 7(c), while five
converters are in this path in Fig. 7(a). Similar to the BESS case,
(c)
an active front end must be used in the DC-connected FESS Fig. 7. DCFC architecture with (a) AC-input FESS, (b) DC-input FESS with
architectures if the energy storage is also planned to provide dedicated LF transformer, and (c) DC-input FESS with isolated DC-DC
services to the grid. converter.

2332-7782 (c) 2020 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
Authorized licensed use limited to: SUNY AT STONY BROOK. Downloaded on August 12,2020 at 11:54:55 UTC from IEEE Xplore. Restrictions apply.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TTE.2020.3015743, IEEE
Transactions on Transportation Electrification

path for a DCFC station, it is advantageous to minimize the A. AC/DC Conversion Stage
number of isolated DC/DC converters in this path to reduce the The AC/DC conversion stage is composed of all converters
number of high-power transformers required. Fig. 8(b) shows between the AC grid and the DC common bus. If the ESS is not
another architecture option which also allows a non-isolated isolated from the DC common bus, the AC/DC conversion
DC/DC converter to be used after the fuel cell. An isolated stage must provide galvanic isolation, either with a dedicated
DC/DC converter is instead used in the grid interface, if a low- low frequency transformer or with a high frequency transformer
frequency transformer is not used. In Fig. 8(c), the grid in an isolated DC/DC converter. It must also provide AC/DC
connection is simplified, requiring only a rectifier and a PFC rectification, power factor correction to improve power quality
DC/DC converter, yet an isolated DC/DC converter is used at at the AC mains and reduce losses, and a DC bus voltage level
the fuel cell output. This architecture may be advantageous which is suitable for the ESS and EV connections. Optionally,
when the grid connection is higher power and the hydrogen the AC/DC conversion stage can provide bidirectional power
storage system is used to provide smaller boosts of power flow capability to allow the energy storage to provide grid
during EV charging. In all these architectures, a non-isolated services.
DC-DC converter is used at the electrolyzer terminal to control For unidirectional power flow and relatively low grid
the hydrogen production [94]. The active front end options of connection power ratings, a simple option is the single-phase
Fig. 8 are required for bidirectional power flow if it is desired full-bridge diode rectifier followed by a boost, buck, or buck-
to use the stored hydrogen energy to provide services to the boost DC/DC converter for improved power factor [96]. For a
grid. The DC-bus voltage can be controlled using the fuel cell 240 V grid voltage, a boost converter would most commonly be
DC/DC converter; for example, [95] allows any output voltage used after the diode rectifier, as shown in Fig. 10(a), in order to
between 375 V and 750 V. create a common DC bus with high enough voltage to connect
D. Architectures for Flywheel-Battery Hybrid ESSs
Similar to the DCFC architectures with BESS and FESS,
there must be isolation between the grid and energy storage
components in a flywheel-battery hybrid ESS. Fig. 9(a) shows
a hybrid architecture with a DC-input flywheel, where the DC
bus is allowed to vary with the battery voltage. Isolation is
provided by an isolated DC/DC converter at the grid side. Fig.
9(b) shows an alternative where a DC/DC converter connected
to the battery provides a fixed DC bus voltage, increasing losses
into and out of the battery, but reducing the complexity of the
DC/DC converter designs due to the fixed bus voltage. A low (a)
frequency transformer can also be used for grid isolation,
though this increases system volume.

V. POWER ELECTRONIC CONVERTERS IN DCFC STATIONS

WITH ENERGY STORAGE
There are three main differences to be considered in the
design of power electronic converters for DCFC stations with
energy storage compared to converters designed for DCFC
stations without energy storage. Firstly, for a small DCFC
station with only one or two vehicle charging ports, the AC/DC
conversion stage (between the grid and the common DC bus) (b)
can be of a lower power rating since the high power charging
comes from the ESS. Thus, single-phase topologies can be
considered for the lowest AC grid power connections, and the
size and cost of the AC/DC conversion stage can be reduced.
Secondly, due to the potential economic payback of utilizing a
DCFC ESS for providing grid services, bidirectional AC/DC
conversion stages may be merited. Thirdly, the isolated DC/DC
converter between the common DC bus and the EV port may
have a widely varying input and output voltage range, if a BESS
is connected directly to the DC bus, as in the architectures
shown in Figs. 5(a) and 5(c). With a focus on these issues,
Section V-A discusses promising power electronic converter (c)
Fig. 8. DCFC with hydrogen storage with on-site hydrogen and electricity
options for the AC/DC conversion stage, and Section V-B generation with (a) dedicated LF transformer, (b) isolated DC-DC converter at
discusses isolated DC/DC converter options. the grid side, and (c) isolated DC-DC at the fuel cell terminal.

2332-7782 (c) 2020 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
Authorized licensed use limited to: SUNY AT STONY BROOK. Downloaded on August 12,2020 at 11:54:55 UTC from IEEE Xplore. Restrictions apply.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TTE.2020.3015743, IEEE
Transactions on Transportation Electrification

the full-bridge topology as shown in Fig. 12(a) is a common

choice due to its simplicity and high power capability. In
addition, different three-leg topologies with additional
capacitor or inductor are also available to reduce the output DC
voltage ripple [96]. The topology with a third leg and an
additional inductor is shown in Fig. 12(b).

(a)

(a)

(b)
Fig. 10. Single-phase diode bridge rectifier with (a) boost and (b) buck-boost
(b) PFC
Fig. 9. DCFC architectures for hybrid flywheel-battery ESS with (a) varying
and (b) fixed DC bus voltage.

to a 300 V to 400 V or higher ESS. If the ESS operates at a

lower voltage, a single-switch buck-boost converter, as shown
in Fig. 10(b), could be used. The buck-boost topology could
also be based on the SEPIC converter, or another two-switch
topology [96].
Even though the diode rectifier with boost or buck-boost PFC
is capable of high power quality and power factor, these
converters are susceptible to low efficiency due to high
conduction losses in the diode bridge. To improve the (a)
efficiency of the AC/DC conversion stage, bridgeless
topologies can be considered. The basic bridgeless boost
topology suffers from large common-mode noise compared to
the traditional bridge topology [97], [98]. The most practical
bridgeless boost PFC converters are dual-boost bridgeless PFC
and back-to-back bridgeless PFC as shown in Figs. 11(a) and
11(b) respectively. Soft-switching of these converters while
operating in discontinuous conduction mode can further reduce
losses. The totem-pole bridgeless PFC is also emerging as an
excellent choice for low power if wide band-gap devices are
used to solve the issue of severe reverse recovery losses (b)
common with the use of silicon MOSFETS [99]. Fig 11(c)
shows the totem-pole bridgeless PFC with gallium nitride
(GaN) devices. Different multilevel topologies are also
available for further reduction of conduction losses or to deal
with higher voltages. One such three-level boost PFC is
presented in [98]. The bridgeless topologies of buck-boost
converter are still in development since the techniques
applicable to bridgeless boost topologies are more difficult to
apply in bridgeless buck-boost converters [100].
Among the active front-end boost topologies for reverse (c)
Fig. 11. (a) Dual-boost bridgeless PFC, (b) back-to-back bridgeless PFC, and
power flow capability, as in the DCFC architecture in Fig. 6(b), (c) totem-pole PFC with GaN switches

2332-7782 (c) 2020 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
Authorized licensed use limited to: SUNY AT STONY BROOK. Downloaded on August 12,2020 at 11:54:55 UTC from IEEE Xplore. Restrictions apply.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TTE.2020.3015743, IEEE
Transactions on Transportation Electrification

(a) (a)

(b) (b)
Fig. 12. Single-phase bidirectional converter with (a) four and (b) six active
switches

Similarly, for a higher power three-phase grid connection,

boost and buck-boost DC/DC converters can be used after a
three-phase diode rectifier. Fig. 13(a) shows a three-phase
diode rectifier with a boost PFC, where the line inductance is
used as the boost inductor, and Fig. 13(b) shows a three-phase
diode rectifier with a SEPIC-derived buck-boost PFC. The main
advantage of the Vienna rectifier, shown in Fig. 13(c), is the
three-level characteristics due to split DC-bus, and the switches,
therefore, only need to block half of the peak value of mains (c)
line-to-line voltage [101]. However, the Vienna rectifier suffers Fig. 13. Three-phase rectifiers with (a) boost and (b) buck-boost PFC and (c)
from the issue of voltage balancing of three-level converters, Vienna rectifier as a modification of boost topology
and the reactive power controllability depends on the output
voltage due to limited modulation vector [17], [101]. Bridgeless
topologies are not common for three-phase systems – in most
cases, three single-phase bridgeless topologies are paralleled if
required [100].
Three-phase active front-end converters are used for high-
power grid connections that require bidirectional power flow
between the grid and the ESS so that the ESS can also be used
to provide services to the grid such as frequency control,
voltage support, and peak power [33]. One of the most common
bidirectional AC/DC converter topologies is the two-level (a)
voltage source converter (2L-VSC), as shown in Fig. 14(a).
This 2L-VSC is a six-switch boost topology and insulated-gate
bipolar transistors (IGBTs) are commonly used for these higher
power levels. Since this power converter is based on active
switches, almost pure sinusoidal current can be generated at the
grid terminal along with the required power factor.
For medium voltage (MV) grid connections, the three-level
Neutral Point Clamped (3L-NPC) converter, shown in Fig.
14(b), is bidirectional and can solve the limited power
capability of the 2L-VSC as it can handle input voltages of 2.2-
6.6 kV [102]. The connection of two diodes in each phase leg
results in three-level voltage waveforms, unlike the 2L-VSC,
(b)
which have lower total harmonic distortion (THD). The Fig. 14. Three-phase bidirectional (a) 2-Level and (b) 3-Level NPC topology
presence of a neutral point needs attention and balancing is
required for proper operation. Moreover, a bipolar DC bus

2332-7782 (c) 2020 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
Authorized licensed use limited to: SUNY AT STONY BROOK. Downloaded on August 12,2020 at 11:54:55 UTC from IEEE Xplore. Restrictions apply.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TTE.2020.3015743, IEEE
Transactions on Transportation Electrification

operation is proposed in [103]. This will help reduce the

conversion efforts of DC/DC converters as the input DC voltage
will be reduced to half compared to unipolar DC bus operation.
However, this bipolar DC bus operation is challenging in a fast-
charging station due to the variable nature of the load; very
complex DC bus voltage balancing is required in such an
operation. Furthermore, higher levels such as five-level, seven-
level topologies are also possible but have not been used a lot
commercially.
The above-mentioned converters can be used in DCFC
architectures to create a common DC bus where isolation is (a)
provided by a dedicated low frequency transformer or separate
isolated DC/DC converters. Alternatively, previous research
has investigated various PFC topologies which can also provide
isolation, which can be used in the architecture of Fig. 6(a).
Even though these topologies are mainly used in industrial
motor drive systems, these converters can also be used as front-
end converters in DCFC stations. For instance, [87] presents a
single phase isolated bridgeless Cuk converter-based PFC
rectifier for DC motor drive, as shown in Fig. 15(a). Also, [90]
presents a three-phase isolated Cuk-based PFC rectifier which
uses three single phase isolated Cuk converters, as shown in
Fig. 15(b). These converters are capable of unidirectional
power flow. On the other hand, [90] describes a three-phase
matrix-type Dual Active Bridge (DAB) buck-boost isolated
PFC rectifier which can be used for bidirectional power flow
applications, as shown in Fig. 15(c).
B. Isolated DC/DC Converter Topologies
(b)
Isolation in DC/DC converters is provided with high
frequency transformers. A detailed review of various isolated
DC/DC converters is presented in [17]. Isolated DC/DC
converters can be used in the AC/DC conversion stage of a
DCFC station if isolation is not provided by a dedicated low
frequency transformer, as shown in Fig. 5(b). This converter
can be unidirectional or bidirectional if the ESS will be required
to provide grid services. A high power bidirectional converter
is required for connection with the battery ESS in the
architecture shown in Fig. 5(c). For the high power isolated (c)
DC/DC converter connecting from the DC bus to the EV Fig. 15. (a) Single-phase bridgeless Cuk topology based isolated PFC, (b) three-
battery, a unidirectional converter will generally suffice. While phase isolated PFC with three single-phase Cuk converter, (c) bidirectional
much research has investigated the possibility of EV batteries DAB based three-phase PFC.
providing services to the grid (V2G) [104], [105], this scenario
converter can be operated at high power and high switching
works best for idle vehicles that can be plugged in to
frequency resulting in high power density. Even though zero-
charge/discharge for longer periods of the day (e.g., while
voltage switching can be guaranteed around rated power, this
parked at home or at work). Since DCFC stations are inherently
capability is lost with light load in this converter topology.
used by drivers who need a fast charge to continue their driving
However, since the aim of a DCFC station is to top up the EV
plans, V2G grid services are not ideal for these on-the-go
battery as quickly as possible, it is expected that the converter
vehicles. This assumption simplifies the design of the high
will operate at rated power most of the time, and thus this
power isolated DC/DC converter that connects to the EV by
converter topology has proven to be a good candidate for DCFC
allowing it to be unidirectional. However, the design and
applications. For the potential bidirectional isolated DC/DC
control can be complicated by the fact that the input voltage
converter in the AC/DC conversion stage, the related DAB
(DC bus in Fig. 6(a)) and output voltage (EV battery voltage)
phase-shifted converter is suitable, as shown in Fig. 16(b).
can vary widely, depending on the particular SOC of the storage
Different phase-shift control methods such as single phase shift,
and EV batteries. Thus, these converters should have wide input
dual phase shift, and triple phase shift are possible with DAB
and output voltage ranges. The two prominent converter
converters, whereas only the single phase shift method is
topologies for this application are discussed below.
applicable for the former unidirectional topology. Moreover,
One suitable converter is the unidirectional phase-shifted
half-bridge topologies of these converters are also available
zero voltage transition full-bridge converter [106], as shown in
which are sometimes preferred over full-bridge ones for lower
Fig. 16(a). Due to the capability of zero voltage turn-on, this

2332-7782 (c) 2020 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
Authorized licensed use limited to: SUNY AT STONY BROOK. Downloaded on August 12,2020 at 11:54:55 UTC from IEEE Xplore. Restrictions apply.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TTE.2020.3015743, IEEE
Transactions on Transportation Electrification

(a)
(a)

(b)
Fig. 16. Full-bridge phase-shifted ZVS converter (a) unidirectional and (b) bi- (b)
directional. Fig. 17. Full-bridge LLC converter (a) unidirectional and (b) bi-directional

power grid connections due to lower size, weight, and cost [111]. Moreover, a half-bridge topology of the LLC converter
[106]. is also possible [106].
Another popular isolated topology for a DCFC station is the
full-bridge LLC (FB-LLC) resonant converter [17], [106] VI. COMPARISON OF STORAGE TECHNOLOGIES IN DCFC
because zero voltage switching (ZVS) is achievable over a wide STATIONS
load range. The unidirectional version of the FB-LLC converter
is shown in Fig. 17(a). A FB-LLC converter achieves voltage In this section, battery, flywheel, hydrogen, and hybrid
regulation by voltage division with frequency dependent flywheel-battery storage are compared in terms of overall
impedance, which provides a variable voltage gain. Therefore, system efficiency, volume, and cost for three different EV
the switching frequency varies for the purpose of line and load charging rates: 50 kW, 100 kW, and 150 kW in three different
regulation, which makes the design of passive components DCFC station scenarios. The first charging station scenario is a
complicated. Therefore, the gain curves should be as narrow as small one with only one vehicle charging event assumed per
possible, i.e., high gain variation within a small frequency range day. This type of charging station could be found along
to have a wide output voltage range. The variation in the gain highways in remote areas and in this study, it relies on a single-
curve at different operating points makes the design, operation phase 15 kW grid connection. The second charging station
and optimization of the FB-LLC complex. However, if the scenario has two charging ports and is assumed to charge five
converter operates close to resonant frequency, very high vehicles per day. Moreover, two of the vehicles are assumed to
efficiency is achievable due to ZVS operation. The LLC charge simultaneously during the peak charging hour of 5pm.
converter has some advantages over other ZVS topologies in This charging station is connected to the grid via a three-phase
terms of short-circuit protection, good voltage regulation at 50 kW connection. The third charging station scenario is a large
light loads, ZVS over a wide load range, ZCS for power diodes, one with five charging outlets similar to the average charging
and only a capacitor filter rather than the conventional LC filter outlets in an Electrify America DCFC station [112]. Various
[107]–[109]. studies have predicted there could be between 11 [15] and 25
A bidirectional LLC converter is shown in Fig. 17(b) and [113] daily charging events for this larger station type. Thus,
can be used within the AC/DC conversion stage, or for this analysis assumes the large station has 20 daily charging
connecting to the ESS in the architecture shown in Fig. 5(c). events, with up to five vehicles charging simultaneously. This
However, the gain curve of such an LLC converter reduces to a large charging station has a grid connection of 150 kW. Fig. 18
simple series resonant converter (SRC) during bidirectional shows the assumed EV charging power demands for each
operation, i.e., regeneration mode [110]. Since the efficiency of station in this analysis, where the demands in the medium and
an SRC reduces significantly as the operating switching large stations are estimated from the probability distributions in
frequency moves away from the series resonance, the efficient [114], [115]. Fig. 18 shows that higher EV charging rates cause
operation of the LLC converter in regenerative mode is limited. a smaller overlap in charging events as each EV is charged more
Even though some literature has proposed symmetrical quickly. Based on these assumptions, the three DCFC stations
operation of LLC converters, these converters lose their are modelled in MATLAB/Simulink.
symmetry as the input and output voltage vary. For In the MATLAB/Simulink model, the EVs are modelled as
bidirectional operation, a modified LLC converter with an Tesla Model S EVs with 85 kWh batteries. The battery
additional capacitor at the secondary side of the high frequency specifications are shown in Table V. It is assumed that the EVs
transformer is proposed, known as the CLLC converter [110], arrive at the DCFC station when the battery SOC is 20% and

2332-7782 (c) 2020 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
Authorized licensed use limited to: SUNY AT STONY BROOK. Downloaded on August 12,2020 at 11:54:55 UTC from IEEE Xplore. Restrictions apply.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TTE.2020.3015743, IEEE
Transactions on Transportation Electrification

(a)

Fig. 19. EV battery SOC change with different charging power

end t
1
Cbattery  3600 t 
SOCbattery  SOCinit  I ch arg ing dt (3)
start

Table VI shows the estimated efficiency and power density

of the required electrical components used in the DCFC station
models. Most power converters use a single average efficiency
since they will often run at a single operating point. However,
since the electrolyzer and battery operate at varying operating
points, the associated power converters efficiencies are
(b) modeled as shown in Fig. 20 [116]. Due to the lack of available
data, the power density of the diode bridge rectifier, a Vienna
rectifier in this case, is estimated at 2450 kW/m3.
A. DCFC with BESS
The Simulink model for the DCFC station with BESS is
based on the architecture in Fig. 6 (b). The block diagram of the
Simulink model is shown in Fig. 21. The storage battery is
modelled based on the AKASOL 15 AKM 46 NANO NMC
battery [84], as shown in Table VII. The charging and
discharging resistance of the storage system are based on the
data gathered from the company. The battery is allowed to
charge up to 95% SOC and discharge down to 20% SOC to
protect the health of the battery by limiting the depth of
discharge (DOD). The cycle lifetime is estimated at 7000 cycles
(c)
Fig. 18. Charging power demand along the day in the (a) small charging station, at 80% DOD [84], and the calendar life is estimated at 12 years
(b) medium charging station, and (c) large charging station [48]. Moreover, the battery never exceeds its maximum C-rate
during charging or discharging.
TABLE V: Specification of TESLA Model S Battery
Using trial and error, the minimum BESS energy
Parameter Value
Battery Nominal Energy 85 kWh requirement is found for each daily EV charging scenario in
Battery Nominal Voltage 375 V Fig. 18, such that the SOC limits and C-rate limits are not
Battery Capacity 226.67 Ah exceeded. BESS energy is increased by increasing the number
Battery Charging Resistance 1.2 mΩ of battery modules from Table VII in parallel, where non-
that they get charged up to 70%. For simplicity, the EVs are integer numbers of parallel modules are accepted so that the
charged with constant power, and the charging current is minimum energy requirement can be found regardless of the
calculated using (2) where 𝑃𝑐ℎ𝑎𝑟𝑔𝑖𝑛𝑔 and 𝑉𝑏𝑎𝑡𝑡𝑒𝑟𝑦 are the selected battery module size in Table VII. The BESS simulation
results are shown in Table VIII. The system volume is found by
charging power and battery terminal voltage, respectively.
summing the volumes of the battery and the power electronic
Using the charging current, the battery SOC is calculated using
converters calculated from battery energy density (Table VII)
(3) where 𝑆𝑂𝐶𝑖𝑛𝑖𝑡 is the battery initial SOC and 𝐶𝑏𝑎𝑡𝑡𝑒𝑟𝑦 is
and converter power densities (Table VI). In addition to the
battery capacity in Ah. The SOC changes of the EV battery at energy requirement and system volume, average daily
different charging powers are shown in Fig. 19. efficiency and the number of equivalent full cycles (EFC) per
Pch arg ing day is also calculated for use in the cost analysis. The EFC is
I ch arg ing  (2) calculated as (4), where battery capacity is adjusted to 80%
Vbattery because the available battery lifetime data is for 80% DOD
cycles. The efficiency results show that the BESS has high

2332-7782 (c) 2020 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
Authorized licensed use limited to: SUNY AT STONY BROOK. Downloaded on August 12,2020 at 11:54:55 UTC from IEEE Xplore. Restrictions apply.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TTE.2020.3015743, IEEE
Transactions on Transportation Electrification

TABLE VI: Specifications of Electrical Components The flywheel operates between 16,000 rpm (ωmax) and 8,000
Average Power Density rpm (ωmin); therefore, flywheel speed at a certain SOC can be
Component Efficiency (kW/m3)
(%)
found using (6). The flywheel system operates from 0% SOC
Transformer [117] 98% 242 to 100% SOC, where 0% SOC corresponds to 8,000 rpm.
Diode Bridge (Vienna) Rectifier [118] 98% 2450 Trial and error in the simulation is used to find the minimum
Active AC/DC Converter [119] 98% 2450 FESS energy capacity that satisfies each EV charging profile in
Non-isolated DC/DC Converter [116] 98% 9000
Isolated DC/DC Converter [120] 95% 1330
Fig. 18. FESS modules are connected in parallel to model
higher energy capacities. Table IX shows the flywheel
simulation results. The FESS system volume is calculated by
summing the flywheel volume (using energy density from
Table IV) and the associated converter power densities from
Table VI. The high spinning losses make the FESS unsuitable
for the small DCFC station due to low efficiency: the flywheel
is waiting charged most of the day for the EV charging event to
begin. For the medium station, two control strategies are
considered: (I) same as for the small station, the FESS is kept
at high SOC to be ready for any EV charge event (Fig. 24), and
(II) the flywheel is controlled to a lower SOC during less busy
times of the day, and only charged to high SOC during the
predicted busy charging times of the day (Fig. 25). Since
Fig. 20. Efficiency of electrolyzer-connected non-isolated DC/DC converter spinning losses increase at higher speeds, and speeds are higher
[116] and estimated efficiency of battery-connected isolated DC/DC converter at higher SOC, Control Strategy II increases system efficiency
efficiency, as expected, and that the system efficiency decreases by a few percentage points compared to Control Strategy I.
as the charging rate increases due to higher I2R battery internal Also, for either control strategy, system efficiency is highest at
losses. lower EV charge rates because most EV charging power can
come directly from the grid with less cycling of the FESS.
TotalEnergyThroughput Furthermore, the simulation results also show that the FESS is
EFC  (4) t
BatteryCapacity  0.8 1 end
SOC flywheel  SOCinit   Pflywheel dt
Ecapacity  3600 tstart
(5)
B. DCFC with FESS
For the FESS, the architecture from Fig. 7(c) is used as it has
an active front end for bidirectional power flow. The block
 flywheel  SOC  ( 2max   2min )   2min (6)
diagram of the Simulink model is shown in Fig. 22. The
flywheel model is based on the Beacon Power flywheel system
rated for 100 kW and 25 kWh from Table IV. The efficiency of
flywheel system with integrated power electronics is 90% [57].
The self-discharge (spinning loss) with respect to speed for one
flywheel module is presented in Fig. 23, as per communication
with Beacon Power. The SOC of the flywheel is calculated
using (5) where 𝐸𝑐𝑎𝑝𝑎𝑐𝑖𝑡𝑦 is the flywheel maximum usable
energy capacity and 𝑃𝑓𝑙𝑦𝑤ℎ𝑒𝑒𝑙 is the power into the flywheel.
TABLE VII: Specification of Stationary BESS [84]
Parameter Value
Module Nominal Energy 30.6 kWh
Module Nominal Voltage 666 V
Module Capacity 46 Ah Fig. 22. Block diagram of the Simulink model for the DCFC with FESS
C-rate 2.5
Energy Density 122.18 kWh/m3

Fig. 21. Block diagram of the Simulink model for the DCFC with BESS Fig. 23. Self-discharge (spinning losses) of the FESS with respect to rotational
speed

2332-7782 (c) 2020 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
Authorized licensed use limited to: SUNY AT STONY BROOK. Downloaded on August 12,2020 at 11:54:55 UTC from IEEE Xplore. Restrictions apply.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TTE.2020.3015743, IEEE
Transactions on Transportation Electrification

TABLE VIII: SUMMARY OF BESS SIMULATION RESULTS

DCFC Station Grid Connection EV Charging BESS Requirement System Efficiency System Volume
EFC/day
Size (kW) Rate (kW) (kWh) (%) (m3)
50 42.22 87.36 0.40 0.94
Small 15 100 52.63 86.37 0.52 0.93
150 57.83 85.45 0.60 0.92
50 141.98 87.97 1.25 1.51
15 100 163.10 87.49 1.50 1.55
150 177.48 87.08 1.70 1.53
Medium
50 63.95 87.99 0.66 1.24
50 100 94.25 87.26 0.98 1.94
150 108.02 86.59 1.17 2.07
50 111.38 92.51 1.27 1.79
Large 150 100 186.97 87.44 2.08 2.25
150 260.10 87.29 2.87 2.59

TABLE IX: SUMMARY OF FESS SIMULATION RESULTS

Grid FESS Energy FESS Power System
DCFC EV Charging Rate System
Connection Control Strategy Requirement Requirement Volume EFC/day
Station Size (kW) Efficiency (%)
(kW) (kWh) (kW) (m3)
50 42.00 53.07 14.78% 2.41 1.25
Small 15 I 100 47.50 112.87 13.98% 2.75 1.25
150 50.25 172.01 13.87% 2.95 1.25
50 63.75 80.57 33.40% 3.71 2.00
I 100 85.00 202.65 28.51% 4.97 2.54
Medium 150 94.50 321.89 27.33% 5.58 2.64
50
50 63.75 80.57 38.52% 3.71 2.00
II 100 85.00 202.65 30.78% 4.97 2.54
150 94.50 321.89 28.51% 5.58 2.64

not suitable for the medium station with 15 kW grid connection

(due to high spinning losses) and is not suitable alone for the
large station as the energy storage requirement indicates an
infeasible number of flywheel modules. Thus, the hybrid
flywheel-battery investigation will focus on the large DCFC
station.
C. DCFC with Hydrogen Storage
The architecture in Fig. 8(c) is considered in this analysis, as
it has an active front end and a low number of conversion
stages. The block diagram of the Simulink model is shown in
Fig. 26. The fuel cell system is modeled with the efficiency
curve in Fig. 27 [68] and power density of 162 kW/m3 [95]. An
alkaline electrolyzer is assumed, with 94%, 90.5%, and 85%
efficiency at 25%, 50%, and 100% load, respectively [66]. The
electrolyzer hydrogen production rate is estimated at 0.022 kg Fig. 24. FESS SOC and spinning losses for Control Strategy I with EV charging
rate of 50 kW
H2/hour per kW electrical power [66], where H2 is produced at
35 bar. The electrolyzer volume is estimated at 0.018 m3/kW
electrical power [66]. The compressor volume and power use is
modeled from Burkhardt Compression MD2.5-V [121] with
25% added to both metrics to estimate a 700 bar system. Thus,
the modeled compressor can process 3.4 kg of H 2/hour at 700
bar with an estimated volume of 5.5 m3 while consuming 15
kW from a secondary grid connection, which is included in the
efficiency calculation. Therefore, the volume of the required
compressor to process hydrogen produced by the electrolyzer
per kW grid connection at 100% load is 0.036 m3. The 700 bar
hydrogen tank density is 800 kWh/m3 [65].
The fuel cell is sized based on the maximum ESS power
requirement, which is the difference between the maximum EV
charging power and the grid connection power level. Then, the
electrolyzer and compressor sizes are chosen to use the full grid
connection power, except for the 150 kW grid connection case, Fig. 25. FESS SOC and spinning losses for Control Strategy II with EV
charging rate of 50 kW

2332-7782 (c) 2020 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
Authorized licensed use limited to: SUNY AT STONY BROOK. Downloaded on August 12,2020 at 11:54:55 UTC from IEEE Xplore. Restrictions apply.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TTE.2020.3015743, IEEE
Transactions on Transportation Electrification

kW electrolyzer was sufficient to keep the hydrogen tank

adequately filled for the daily charging profile. Furthermore,
based on the fuel cell efficiency curve, it is clear that fuel cell
efficiency could be increased by up to 10% if a larger fuel cell
was used, so that it could run at much less than maximum
power. However, this would have negative cost implications as
the fuel cell would be oversized for the expected load.
D. DCFC with Hybrid Storage
The architecture in Fig. 9(a) is considered in this analysis,
with active front end and isolated DC/DC converter at the grid
side. The block diagram of the Simulink model is shown in Fig.
28. The battery-flywheel hybrid system is only simulated for
the large DCFC with an EV charging rate of 150 kW, because
Fig. 26. Block diagram of the Simulink model for the DCFC with hydrogen the most demanding charging scenario would likely benefit the
storage most from a hybrid ESS. Although many sizing and control
strategies are possible, this analysis develops a strategy to
minimize battery cycling, and thus aging, as this is a key benefit
that can be exploited when combining a battery and flywheel,
due to the flywheel high power capability and high cycle
capability. Since battery calendar life is estimated at 12 years,
the control goal is to extend battery cycle life to 12 years so it
does not need to be replaced prematurely due to high cycling.
It is calculated that the battery can last for 12 years if its EFC
is limited to 1.6 cycles/day. To achieve this goal, a control
strategy is developed such that the BESS supplies 55% of the
storage power demand when this demand is less than the
BESS’s maximum power rating of 2.5 C. This allows the FESS
to deliver almost half of the required power when the power
Fig. 27. Fuel cell system efficiency at different loading condition [68] demand is low. When the FESS depletes to its minimum SOC,
where a 75 kW power rating is found to suffice for the the BESS supplies all the power. When the power demand is
considered EV charging scenarios. Then, the minimum size of greater than 2.5 C, the BESS supplies power at 2.25 C rate and
the hydrogen storage tank is found by the simulation using trial the FESS provides the rest until it is depleted.
and error. The fuel cell, electrolyzer, compressor, and hydrogen Choosing the sizes of components in a hybrid system has a
tank volumes are then found using the densities in the preceding complicated interrelation with the selected control strategy. In
paragraph and summed with the power electronic converter this case, the BESS size is selected to be the same as that from
volumes to obtain the system volume. Table X summarizes the the BESS-only case (260.1 kWh), so that the flywheel can be
simulation results. Similar to the FESS case, the medium DCFC used to reduce battery cycling (with a calculated FESS size
with hydrogen storage cannot supply the required energy with requirement of 57.5 kWh). It should be noted that other control
a 15 kW grid connection due to low efficiency. At lower EV and sizing strategies were investigated, but when the battery is
charging rates, the system efficiency is notably higher because made smaller, the flywheel system must be larger, and total
less energy must be processed in the electrolyzer-fuel cell path, volume and efficiency were generally worse than the selected
and more of the EV energy can come straight from the grid. case described here. For the selected case, system efficiency
Also, it was found that for the large station, the volume and cost over one day is 64.7% and the system volume is found to be
were optimized by using a smaller electrolyzer and compressor 6.12 m3. Fig. 29 shows the power and SOC profiles of the
than the full 150 kW available from the grid – in this case a 75 battery and flywheel for the proposed control strategy.

TABLE X: SUMMARY OF HYDROGEN ESS SIMULATION RESULTS

Maximum
Hydrogen Storage
Grid EV Charging Hydrogen System
DCFC Requirement (m3) Fuel Cell Power System
Connection Rate Production Rate by Efficiency
Station Size (Equivalent Energy Requirement (kW) Volume (m3)
(kW) (kW) Electrolyzer (%)
(kWh))
(kg/hour)
50 0.086 (69 kWh) 37.93 0.33 39.49 1.19
Small 15 100 0.106 (85 kWh) 90.56 0.33 35.52 1.58
150 0.116 (93 kWh) 143.19 0.33 34.38 1.95
50 0.126 (101 kWh) 56.26 1.10 63.50 3.34
Medium 50 100 0.190 (152 kWh) 161.53 1.10 45.24 4.13
150 0.216 (173 kWh) 266.79 1.10 41.05 4.88
50 0.256 (205 kWh) 116.16 1.65 73.46 5.38
Large 150 100 0.625 (500 kWh) 379.32 1.65 54.00 7.56
150 0.900 (720 kWh) 642.47 1.65 47.84 9.65

2332-7782 (c) 2020 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
Authorized licensed use limited to: SUNY AT STONY BROOK. Downloaded on August 12,2020 at 11:54:55 UTC from IEEE Xplore. Restrictions apply.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TTE.2020.3015743, IEEE
Transactions on Transportation Electrification

economic payback compared to the DCFC-only scenarios

considered here. The hybrid FESS could be feasible for a large
station that is also providing grid services. Hydrogen storage
has higher volume and lower efficiency than BESS in all three
metrics considered, however, the cost is only slightly higher
than that of the BESS. Thus, hydrogen storage would be best
considered only for DCFC stations that must also provide
hydrogen for refuelling fuel cell vehicles [122], such as
Fig. 28. Block diagram of the Simulink model for the DCFC with hybrid storage passenger vehicles or heavy duty transportation that requires
system high on-board energy storage for long distance trucking –
however, the cost premium compared to batteries is small. The
high efficiency of the battery options will translate into lower
operating costs due to less wasted electricity. Though the
battery packs may have to be replaced one or two times during
the 20 year life-span of the station depending on the EFC, the
battery costs (which include pack replacement costs) are still
lowest among the ESS options. Thus, as the BESS is the most
promising ESS type for stations that provide only DCFC, the
next section will perform a detailed technical analysis of BESS
architectures.

Fig. 29. Power delivered by BESS and daily charging and discharging cycle of VII. COMPARISON OF BESS ARCHITECTURES IN DCFC
BESS and FESS
STATIONS
E. Discussion This section compares the technical details of different BESS
The simulations provide important results on ESS sizing, architectures for a large DCFC station with the load profiles in
efficiency, volume, and EFC. However, to complete the Fig. 18(c), with a focus on efficiency and system volume. The
analysis, cost must also be considered. Costs have been five topologies considered in Figs. 5 and 6 are simulated using
estimated from a variety of sources, as shown in Table XI, and the power electronic average efficiencies in Table VI for four
include initial equipment costs and installation. The cost is DCFC large station scenarios: 50 kW EV charging rate with
estimated for a 20-year time frame. The ESS component 100 kW and 150 kW grid connections, and 150 kW EV
lifetimes are considered as noted in Table XI based on the charging rate with 100 kW and 150 kW grid connections. The
simulated EFC in each DCFC scenario. BESS energy requirement and volume are calculated using the
The results are shown in Fig. 30 for each station size. FESSs same procedure as in Section VI-A. The architectures in Fig.
are not suitable for small and medium DCFC stations due to 6(a) and 6(b) yield the same efficiency and volume results, so
high cost. The main advantage of FESS is high power they are listed as one architecture going forward. For the
capability, and for the scenarios considered here, more energy isolated DC/DC converter at the battery terminals, the
is needed such that the FESS power is not utilized to its fullest. efficiency curve in Fig. 20 is used, since the power rate through
Thus, FESSs would be best considered only for DCFC stations this converter fluctuates over the day.
that will be needing high power capability and high life cycles Fig. 31 shows the 50 kW EV charging rate results for both
to provide services to the grid, which will improve the 100 kW (circle markers) and 150 kW (triangle markers) grid

TABLE XI: ESS COST ESTIMATES (CAPITAL AND INSTALLATION)

Cost
Component Component Details Source Notes
(USD)
Isolated DC/DC Based on whole charger cost, assume 40% is removed for front
EV Charger $345/kW [146]
converter (EV port) end. Small station has 1 port, medium has 2, large has 5.
Calendar life 12 years. Attainable lifetime in each scenario
Battery packs $390/kWh [32]
calculated based on EFC. Per-year cost multiplied by 20.
Battery Includes power converters, etc.
Balance of plant $615/kW [32]
Installation $150/kWh [32]
Range is 650 – 1,600 $/kW. This analysis assumes the middle
Flywheel module (by
$1125/kW value. Lifetime is assumed 100,000 cycles with 80% DOD. Power
power) Discussions
electronics included.
with Beacon
Flywheel Flywheel module (by
$4500/kWh Power, [May Based on modeling a 100 kW/25 kWh flywheel module.
energy)
19, 2020]
Range is 20% - 30% of equipment cost. This analysis assumes the
Installation $281/kW
middle value.
Electrolyzer $700/kW [147] Lifetime 10 years. Includes AC/DC converter and installation cost.
Compressor $275/kW [69] Lifetime 15 years. $/kW grid connection.
Hydrogen
Tank $15/kWh [65] Lifetime 20 years
Fuel cell $260/kW [69] Lifetime 10 years. Includes DC/DC converter.

2332-7782 (c) 2020 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
Authorized licensed use limited to: SUNY AT STONY BROOK. Downloaded on August 12,2020 at 11:54:55 UTC from IEEE Xplore. Restrictions apply.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TTE.2020.3015743, IEEE
Transactions on Transportation Electrification

highest efficiency when the grid connection power level is high

(i.e., 150 kW). This is because during vehicle charging, power
for about three vehicles can come directly from the grid in this
scenario, meaning most charging power over the day comes
straight from the grid and does not cycle through the battery.
Thus, the high efficiency path from the grid to the EV in the
Fig. 5(c) architecture is well-suited to this DCFC scenario.
However, when more use of the energy storage is required, such
as if the grid connection is reduced to 100 kW, the architecture
in Fig. 5(c) has worse efficiency because the path in and out of
the battery has lower efficiency compared to all other
architectures. For both grid connection levels, the two-stage
architectures in Figs. 6(a) and 6(b) offer high efficiency and low
volume, due to a decreased number of converters.
Fig. 32 shows similar results for the 150 kW EV charging
rate at a large DCFC station. Again, the transformer option
(a) (green) gives the highest efficiency but also high volume.
Interestingly, the architecture in Fig. 5(c) has the lowest
efficiency for both grid connection levels because for the higher
EV charging rate, much more energy must be cycled in and out
of the battery, generating losses in the isolated DC/DC
converter connected to the battery during each charge and
discharge event. Thus, for the 150 kW EV charging rate, it is
most important to have a high-efficiency path from the storage
battery to the EV, since this is the path most charging energy
follows. Again, the architectures in Fig. 6(a) and 6(b) offer a
good balance of efficiency and volume for all considered cases.
For all four DCFC scenarios, the volume of the BESS increases
as the grid connection level decreases, since more energy
storage is required to meet the EV charging needs.
Fig. 33 shows the battery energy requirements for each
scenario and architecture. While the average kWh requirement
changes clearly between the four DCFC scenarios, there are
(b) also differences between the architectures for a given scenario,
which relate to the efficiencies simulated for each architecture.
For example, for 150 kW EV charging and 100 kW grid
connection, Fig. 5(c) requires an 11.6% larger battery than the
architecture in Fig. 6(a) because of its lower efficiency in this
scenario. Thus, the volume of the system is also partially related
to the efficiency of the system.

(c)
Fig. 30. Overall ESS type comparison with efficiency overlaid for (a) small
station with 15 kW grid connection, (b) medium station with 50 kW grid
connection, (c) large station with 150 kW grid connection

connections. The green markers represent the architecture in

Fig. 5(a), which uses a transformer for isolation between the
grid and the battery. This yields high efficiency but also high
volume due to the bulky 60 Hz transformer. However, the Fig. 31. Efficiency vs. volume for the BESS architectures for 50 kW EV
charging at large station
architecture in Fig. 5(c), which places an isolated DC/DC
converter between the battery and the high voltage bus, has the

2332-7782 (c) 2020 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
Authorized licensed use limited to: SUNY AT STONY BROOK. Downloaded on August 12,2020 at 11:54:55 UTC from IEEE Xplore. Restrictions apply.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TTE.2020.3015743, IEEE
Transactions on Transportation Electrification

boosting up voltage from Level 2 240 V chargers, and stepping

down voltage to the 12 V accessory bus).
To be compatible with 800 V EVs, the design of a DCFC
station with energy storage must be altered significantly
compared to the usual 400 V EV case. In the BESS architecture
of Fig. 5(a), the storage battery must either be increased in
voltage and/or the isolated DC/DC converter connecting to the
EV battery must use a transformer with a higher turns ratio. This
raises the question of how a DCFC station with energy storage
might be designed to accommodate either 400 V or 800 V
vehicles plugging in to the same port, while maintaining high
efficiency operation. The isolated DC/DC converter would
have to be designed to have a very wide output voltage range.
B. MV Connection with Solid State Transformers
As the demand for DCFC increases, so does the power
requirement from the grid. Although the use of energy storage
Fig. 32. Efficiency vs. volume for the BESS architectures for 150 kW EV
charging at large station
in DCFC stations can greatly reduce this power requirement,
for large DCFC stations with many ports, an MV connection is
often warranted. In this case, the usual MV-LV low frequency
transformer used to provide a LV AC bus for the fast chargers
increases the size of the system. Researchers have proposed
replacing the low frequency transformer with an solid state
transformer (SST) to reduce system size and increase efficiency
at low load [125]–[128]. The SST uses power electronic
switches and a high frequency transformer to provide the
voltage control and galvanic isolation required [17] for creating
the central DC bus of the DCFC station.
Energy storage can still be valuable in MV-connected DCFC
stations. Though high power may be available from the grid
connection, energy storage can still reduce demand charges,
provide the capability to provide services to the grid, and allow
even higher charging powers across numerous plugged-in EVs
than provided by the grid. Modular energy storage with SST-
based architectures have been investigated in [129]. A review
of different power converters for BESS connected to MV grid
Fig. 33. Battery energy requirement for BESS architectures at large DCFC is presented in [130].
station
C. Application of DCFC in Other Industries
VIII. FUTURE TRENDS The demand for fast charging is on the rise in other industries
A. High Voltage EV Battery as well. For instance, electrification of mining trucks is
accelerating due to the efficiency and underground air quality
Though DCFC stations are capable of charging EVs in
advantages compared to traditional diesel mining trucks [131],
shorter timeframes compared to Level 2 AC connections, it is
[132]. These heavy-duty mining trucks, particularly the haul
still slower than the refuelling of conventional gasoline
trucks, use high voltage batteries to provide the high power
vehicles. To charge EVs even faster, some vehicle
needs of hauling ore uphill. Fast charging of these trucks are
manufacturers, such as Aston Martin and Porsche, are
necessary because productivity in a mine is tightly linked to the
developing vehicles with 800 V batteries [6], [7] , as opposed
operation of these vehicles, thus down-time for charging is
to the usual range of 250 – 450 V. The main advantage of such
extremely costly. For example, ABB is developing a 2 MW fast
a high voltage battery system is providing high charging power
charger specifically for use in the challenging environment of
without violating the current limit of the connectors. Moreover,
mines (high temperature, conductive dust) [133]. For these very
using a high voltage DC bus in an EV can result in weight
high charge rates, incorporating energy storage into the fast
savings due to the reduced size of wiring required in the
charging system could significantly reduce peak grid power
powertrain [123]. The main challenges include: (i) extra cost
needs.
associated with using higher voltage power electronic switches
Additionally, the aviation industry is moving towards
(e.g. 1200 V switches instead of 600 V switches), (ii) higher
electrification [134]. For instance, NASA is developing an
voltage insulation requirements throughout the powertrain
electrified aircraft propulsion system [135]. These aircrafts
[124], and specifically in the high-voltage motor, (iii) more
would require high charging power due to large on-board
battery cell balancing required to equalize the charge of a larger
batteries. ChargePoint has revealed the concept of a 2 MW
number of cells connected in series, and (iv) more voltage
charger for electric aircrafts in partnership with Uber Elevate
conversions required onboard which contribute to losses (e.g.,

2332-7782 (c) 2020 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
Authorized licensed use limited to: SUNY AT STONY BROOK. Downloaded on August 12,2020 at 11:54:55 UTC from IEEE Xplore. Restrictions apply.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TTE.2020.3015743, IEEE
Transactions on Transportation Electrification

[136]. Due to the high power demand of these chargers, energy and the EV charging power. Future trends are also discussed to
storage systems can play a vital role in these DCFC systems. highlight ongoing areas of research and development related to
DCFC stations with energy storage.
D. Second-life Battery for Storage
It is evident from the analysis of different energy storage REFERENCES
systems that batteries will play an important role as storage in
[1] W. J. Guan, X. Y. Zheng, K. F. Chung, and N. S. Zhong, “Impact of air
DCFC stations. The use of second-life EV batteries can reduce pollution on the burden of chronic respiratory diseases in China: time
the initial investment required. Since the performance for urgent action,” Lancet, vol. 388, no. 10054, pp. 1939–1951, 2016.
requirements for stationary storage batteries are less demanding [2] R. Irle, “Global EV sales for the 1st half of 2019,” EV Volumes, 2019.
[Online]. Available: http://www.ev-volumes.com/country/total-world-
than batteries used in EVs, these second-life batteries with 60%
plug-in-vehicle-volumes/. [Accessed: 20-Nov-2019].
- 80% battery capacity are suitable for stationary BESSs [137]– [3] M. Fries et al., “An Overview of costs for vehicle components, fuels,
[139]. Moreover, the available second-life batteries could greenhouse gas emissions and total cost of ownership update 2017,”
exceed 200 GWh by 2030 and these batteries have a 30%-70% Researchgate, 2017.
cost advantage over new ones [140]. Second-life batteries can [4] S. Chakraborty, H. N. Vu, M. M. Hasan, D. D. Tran, M. El Baghdadi,
and O. Hegazy, “DC-DC converter topologies for electric vehicles, plug-
often be used for another 5-10 years depending on the required in hybrid electric vehicles and fast charging stations: State of the art and
yearly cycles [141]. However, the challenges with using future trends,” Energies, vol. 12, no. 8, p. 1569, 2019.
second-life BESS in DCFC stations are that they exhibit [5] Luc, “Vehicles & charging tips,” Fastned, 2019. [Online]. Available:
reduced power capability and cycling life compared to new https://support.fastned.nl/hc/en-gb/sections/115000180588-Cars-
charging-tips-.
cells, and complex BMSs will be required to integrate packs [6] F. Lambert, “Aston Martin unveils latest all-electric Rapide E prototype
from different manufacturers with different characteristics with 800v powertrain,” Electrek. [Online]. Available:
[139], [142]. Reference [142] discusses different modes of https://electrek.co/2019/01/25/aston-martin-electric-rapide-prototype-
800v-powertrain/.
operation of modular multiport converters to connect second
[7] M. Wienkötter, “The battery: Sophisticated thermal management, 800-
life batteries from different manufacturers in grid applications. volt system voltage,” Porsche. [Online]. Available:
The more frequent replacements required for second-life https://newsroom.porsche.com/en/products/taycan/battery-18557.html.
batteries must be included in the overall costing of the system. [8] DOE, “Enabling fast charging: technology gap assessment,” 2017.
[9] L. Richard and M. Petit, “Fast charging station with battery storage
Reference [138] performs a detailed analysis to calculate the
system for EV: optimal integration into the grid,” in 2018 IEEE Power
remaining useful life for second-life batteries in DCFC and & Energy Society General Meeting (PESGM), 2018, pp. 1–5.
other applications, providing environmental and financial [10] S. Knupfer, J. Noffsinger, and S. Sahdev, “How battery storage can help
benefits. Even though the widespread application of second life charge the electric-vehicle market,” McKinsey & Company. [Online].
batteries comes with integration challenges, universities and Available: https://www.mckinsey.com/business-
functions/sustainability/our-insights/how-battery-storage-can-help-
OEMs have initiated some projects. For instance, University of charge-the-electric-vehicle-market. [Accessed: 20-Sep-2011].
California, Davis [143] and Rochester Institute of Technology [11] S. Gallinaro, “Energy storage systems boost electric vehicles’ fast
[144] have deployed small scale second-life BESS research charger infrastructure,” Analog Devices, pp. 1–4, May-2020.
projects. BMW has partnered with Bosch and Vattenfall to test [12] B. Baatar, K. Heckmann, T. Hoang, R. Jarvis, and P. Sakhiya,
“Preparing rural America for the electric vehicle revolution,” 2019.
a 2 MW/2.8 MWh storage capacity built with 22 kWh and 33 [13] INL, “What were the cost drivers for the direct current fast charging
kWh second-life battery packs collected from 100 different installations?,” 2015.
BMW i3 cars [145]. Thus, the integration challenges of [14] M. Nicholas and D. Hall, “Lessons Learned on Early Fast Electric
Vehicle Charging Systems,” 2018.
combining batteries from different automotive manufacturers
[15] J. Francfort, S. Salisbury, J. Smart, T. Garetson, and D. Karner,
can be overcome by using batteries only from one manufacturer “Considerations for corridor and community DC fast charging complex
within one system. Furthermore, since second-life batteries system design,” 2017.
have lower energy density than new batteries, the system [16] M. Yilmaz and P. T. Krein, “Review of battery charger topologies,
volume benefits found in this study will be reduced compared charging power levels, and infrastructure for plug-in electric and hybrid
vehicles,” IEEE Trans. Power Electron., vol. 28, no. 5, pp. 2151–2169,
to that of new batteries. 2013.
[17] H. Tu, H. Feng, S. Srdic, and S. Lukic, “Extreme fast charging of electric
IX. CONCLUSION vehicles: A technology overview,” IEEE Trans. Transp. Electrif., vol. 5,
no. 4, pp. 861–878, 2019.
This paper provides a comprehensive review of DCFC [18] D. Ronanki, A. Kelkar, and S. S. Williamson, “Extreme fast charging
stations with energy storage, including a discussion of technology—prospects to enhance sustainable electric transportation,”
motivation, three energy storage options (battery, flywheel, and Energies, vol. 12, no. 19, p. 3721, 2019.
[19] D. Sbordone, I. Bertini, B. Di Pietra, M. C. Falvo, A. Genovese, and L.
hydrogen), system architectures, and power electronic Martirano, “EV fast charging stations and energy storage technologies:
converters suitable for each option. Furthermore, each option is A real implementation in the smart micro grid paradigm,” Electr. Power
modeled in Simulink to quantify efficiency, volume, and cost Syst. Res., vol. 120, pp. 96–108, 2015.
for a small, medium, and large DCFC station. The results show [20] A. Buchroithner, H. Wegleiter, and B. Schweighofer, “Flywheel energy
storage systems compared to competing technologies for grid load
that flywheels would be best considered if high power grid mitigation in EV fast-charging applications,” in IEEE 27th International
services will be also provided by the energy storage, and Symposium on Industrial Electronics (ISIE), 2018, pp. 508–514.
hydrogen storage would be best considered only if refueling of [21] J. Torres, P. Moreno-Torres, G. Navarro, M. Blanco, and M. Lafoz,
fuel cell vehicles is also required. Otherwise, batteries generally “Fast energy storage systems comparison in terms of energy efficiency
for a specific application,” IEEE Access, vol. 6, pp. 40656–40672, 2018.
exhibit the lowest cost, lowest volume, and highest efficiency. [22] G. Wang et al., “A review of power electronics for grid connection of
A comparison of battery architectures shows that the optimal utility-scale battery energy storage systems,” IEEE Trans. Sustain.
architecture is highly dependent on the grid connection power Energy, vol. 7, no. 4, pp. 1778–1790, 2016.
[23] M. Nicholas, “Estimating electric vehicle charging infrastructure costs

2332-7782 (c) 2020 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
Authorized licensed use limited to: SUNY AT STONY BROOK. Downloaded on August 12,2020 at 11:54:55 UTC from IEEE Xplore. Restrictions apply.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TTE.2020.3015743, IEEE
Transactions on Transportation Electrification

across major U.S. metropolitan areas,” 2019. [46] B. K. Kang, S. T. Kim, S. H. Bae, and J. W. Park, “Effect of a SMES in
[24] C. Molnar, “Norway needs $1.6-billion power grid upgrade to support power distribution network with PV system and PBEVs,” IEEE Trans.
EVs by 2040: study,” Driving. [Online]. Available: Appl. Supercond., vol. 23, no. 3, 2013.
https://driving.ca/auto-news/news/norway-needs-1-6-billion-power- [47] K. Li and K. J. Tseng, “Energy efficiency of lithium-ion battery used as
grid-upgrade-to-support-evs-by-2040-study. energy storage devices in micro-grid,” in IECON 2015 - 41st Annual
[25] E. Veldman and R. A. Verzijlbergh, “Distribution grid impacts of smart Conference of the IEEE Industrial Electronics Society, 2015.
electric vehicle charging from different perspectives,” IEEE Trans. [48] K. Mongird et al., “Energy storage technology and cost characterization
Smart Grid, vol. 6, no. 1, pp. 333–342, 2015. report,” Pacific Northwest Natl. Lab., 2019.
[26] R. Abousleiman and R. Scholer, “Smart charging: System design and [49] “ESS batteries by Samsung SDI,” Samsung SDI, 2019. [Online].
implementation for interaction between plug-in electric vehicles and the Available:
power grid,” IEEE Trans. Transp. Electrif., vol. 1, no. 1, pp. 18–25, https://www.samsungsdi.com/upload/ess_brochure/201902_Samsung
2015. SDI ESS_EN.pdf. [Accessed: 20-Sep-2011].
[27] O. Beaude, S. Lasaulce, M. Hennebel, and I. Mohand-Kaci, “Reducing [50] M. Blenner and B. Rosen, “EVgo balances EV fast charging with 14
the impact of EV charging operations on the distribution network,” IEEE battery storage systems across 11 EVgo fast charging stations,” EVGO,
Trans. Smart Grid, vol. 7, no. 6, pp. 2666–2679, 2016. 2019. [Online]. Available: https://www.evgo.com/about/news/evgo-
[28] B. Sun, Z. Huang, X. Tan, and D. H. K. Tsang, “Optimal scheduling for balances-ev-fast-charging-with-14-battery-storage-systems-across-11-
electric vehicle charging with discrete charging levels in distribution evgo-fast-charging-stations/.
grid,” IEEE Trans. Smart Grid, vol. 9, no. 2, pp. 624–634, 2018. [51] A. J. Hawkins, “Volkswagen will use Tesla battery packs in some of its
[29] M. H. Mobarak and J. Bauman, “Vehicle-directed smart charging US charging stations,” The Verge, 2019. [Online]. Available:
strategies to mitigate the effect of long-range EV charging on https://www.theverge.com/2019/2/4/18210439/volkswagen-tesla-
distribution transformer aging,” IEEE Trans. Transp. Electrif., vol. 5, battery-packs-ev-charging-station.
no. 4, pp. 1097–1111, 2019. [52] F. Lambert, “Tesla opens new V3 Supercharger with solar and battery –
[30] EASE, “Energy storage: A key enabler for the decarburization of the looks like EV charging station of the future,” Electrek, 2019. [Online].
transport sector,” Brussels, 2019. Available: https://electrek.co/2019/07/18/tesla-v3-supercharger-station-
[31] J. McLaren, P. Gagnon, D. Zimny-Schmitt, M. DeMinco, and E. Wilson, las-vegas-solar-power-battery/.
“Maximum demand charge rates for commercial and industrial [53] I. Pavic, T. Capuder, and I. Kuzle, “A comprehensive approach for
electricity tariffs in the United States. National Renewable Energy maximizing flexibility benefits of electric vehicles,” IEEE Syst. J., vol.
Laboratory.” [Online]. Available: https://dx.doi.org/10.7799/1392982. 12, no. 3, pp. 2882–2893, 2018.
[32] DNVGL/KEMA, “Battery energy storage study for the 2017 IRP,” [54] M. Gjelaj, S. Hashemi, C. Traeholt, and P. B. Andersen, “Grid
2016. integration of DC fast-charging stations for EVs by using modular li-ion
[33] M. Zidar, P. S. Georgilakis, N. D. Hatziargyriou, T. Capuder, and D. batteries,” IET Gener. Transm. Distrib., vol. 12, no. 20, pp. 4368–4376,
Škrlec, “Review of energy storage allocation in power distribution 2018.
networks: Applications, methods and future research,” IET Gener. [55] M. E. Amiryar and K. R. Pullen, “A review of flywheel energy storage
Transm. Distrib., vol. 10, no. 3, pp. 645–652, 2016. system technologies and their applications,” Appl. Sci., vol. 7, no. 3, p.
[34] A. Nagarajan and R. Ayyanar, “Design and strategy for the deployment 286, 2017.
of energy storage systems in a distribution feeder with penetration of [56] “Breakthrough 4+ hour kinetic energy storage system for high utilization
renewable resources,” IEEE Trans. Sustain. Energy, vol. 6, no. 3, pp. grid, microgrid, industrial & commercial applications,” Amber Kinetics.
1085–1092, 2015. [Online]. Available: https://www.amberkinetics.com/wp-
[35] X. Li, L. Yao, and D. Hui, “Optimal control and management of a large- content/uploads/2019/10/AK-Datasheet-M32.pdf.
scale battery energy storage system to mitigate fluctuation and [57] “Power electronics,” Beacon Power. [Online]. Available:
intermittence of renewable generations,” J. Mod. Power Syst. Clean https://beaconpower.com/power-electronics/.
Energy, vol. 4, pp. 593–603, 2016. [58] C. Zhang and K. J. Tseng, “Design and control of a novel flywheel
[36] H. Beltran, E. Bilbao, E. Belenguer, I. Etxeberria-Otadui, and P. energy storage system assisted by hybrid mechanical-magnetic
Rodriguez, “Evaluation of storage energy requirements for constant bearings,” Mechatronics, vol. 23, no. 3, pp. 297–309, 2013.
production in PV power plants,” IEEE Trans. Ind. Electron., vol. 60, no. [59] T. Aquino, C. Zuelch, and C. Koss, “Energy storage technology
3, pp. 1225–1234, 2013. assessment,” Omaha, Nebraska, 2017.
[37] A. Colthrope, “Round-Up: 60MWh Japan project, Northern Ireland’s [60] F. Goris and E. L. Severson, “A review of flywheel energy storage
10MW array and Imergy goes for Africa teleco.” [Online]. Available: systems for grid application,” in IECON 2018 - 44th Annual Conference
https://www.energy-storage.news/news/round-up-60mwh-japan- of the IEEE Industrial Electronics Society, 2018, pp. 1633–1639.
project-northern-irelands-10mw-array-and-imergy-goes-f. [61] B. Sun, T. Dragičević, F. D. Freijedo, J. C. Vasquez, and J. M. Guerrero,
[38] A. Ratnayake, “Notrees wind storage project description,” Duke Energy. “A control algorithm for electric vehicle fast charging stations equipped
[Online]. Available: https://www.sandia.gov/ess- with flywheel energy storage systems,” IEEE Trans. Power Electron.,
ssl/docs/pr_conferences/2011/3_Ratnayake_Notrees.pdf. vol. 31, no. 9, pp. 6674–6685, 2016.
[39] M. Montoya and K. Nuhfer, “Tehachapi wind energy storage project,” [62] H. Lee, B. Y. Shin, S. Han, S. Jung, B. Park, and G. Jang,
U.S. Department of Energy, 2012. [Online]. Available: “Compensation for the power fluctuation of the large scale wind farm
https://www.energy.gov/sites/prod/files/Tehachapi.pdf. using hybrid energy storage applications,” IEEE Trans. Appl.
[40] “E.ON inaugurates first 2 MW Power-to-Gas unit in Falkenhagen,” Fuel Supercond., vol. 22, no. 3, pp. 5701904–5701904, 2012.
Cells Bull., 2013. [63] A. A. Khodadoost Arani, G. B. Gharehpetian, and M. Abedi, “A novel
[41] D. Gielen, E. Taibi, and R. Miranda, “Hydrogen: A renewable energy control method based on droop for cooperation of flywheel and battery
perspective,” Tokyo, Japan. energy storage systems in islanded microgrids,” IEEE Syst. J., vol. 14,
[42] E. Wesoff, “Energy storage at grid scale: PG&E demo projects,” GTM. no. 1, pp. 1080–1087, 2020.
[Online]. Available: [64] DOE, “Hydrogen storage.” [Online]. Available:
https://www.greentechmedia.com/articles/read/energy-storage-at-grid- https://www.energy.gov/eere/fuelcells/hydrogen-storage.
scale-pge-projects. [65] “Physical hydrogen storage,” U.S. Department of Energy. [Online].
[43] J. Arseneaux, “20 MW flywheel energy storage plant,” Beacon Power, Available: https://www.energy.gov/eere/fuelcells/physical-hydrogen-
2014. [Online]. Available: https://www.sandia.gov/ess- storage.
ssl/docs/pr_conferences/2014/Thursday/Session7/02_Areseneaux_Jim_ [66] “HyProvideTM A-Series,” GreenHydrogen. [Online]. Available:
20MW_Flywheel_Energy_Storage_Plant_140918.pdf. https://greenhydrogen.dk/wp-content/uploads/2019/11/HyProvideTM-
[44] “StoraXe powerbooster,” adstec. [Online]. Available: https://www.ads- A-Series.pdf.
tec.de/en/energy-storage/industrial- [67] “DOE technical targets for fuel cell backup power systems,” U.S.
infrastructure/powerbooster/description.html. [Accessed: 20-Nov- Department of EnergyDepartment of Energy. [Online]. Available:
2019]. https://www.energy.gov/eere/fuelcells/doe-technical-targets-fuel-cell-
[45] Y. Liu, Y. Tang, J. Shi, X. Shi, J. Deng, and K. Gong, “Application of backup-power-systems.
small-sized SMES in an EV charging station with DC bus and PV [68] U. Eberle, B. Müller, and R. Von Helmolt, “Fuel cell electric vehicles
system,” IEEE Trans. Appl. Supercond., vol. 25, no. 4, pp. 1–6, 2015. and hydrogen infrastructure: Status 2012,” Energy Environ. Sci., vol.

2332-7782 (c) 2020 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
Authorized licensed use limited to: SUNY AT STONY BROOK. Downloaded on August 12,2020 at 11:54:55 UTC from IEEE Xplore. Restrictions apply.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TTE.2020.3015743, IEEE
Transactions on Transportation Electrification

2012, no. 5, pp. 8780–8798, 2012. [92] IEEE, “IEEE application guide for IEEE std 1547TM, IEEE standard for
[69] H. Zhao and A. Burke, “Sustainable hydrogen fueling/DC fast charging interconnecting distributed resources with electric power systems,”
systems at California highway rest stop areas,” California, 2016. IEEE Std 1547.2-2008, 2009.
[70] H. Khani, N. A. El-Taweel, and H. E. Z. Farag, “Supervisory scheduling [93] Y. Zhang et al., “Wide input-voltage range boost three-level DC-DC
of storage-based hydrogen fueling stations for transportation sector and converter with quasi-z source for fuel cell vehicles,” IEEE Trans. Power
distributed operating reserve in electricity markets,” IEEE Trans. Ind. Electron., vol. 32, no. 9, pp. 6728–6738, 2017.
Informatics, vol. 16, no. 3, pp. 1529–1538, 2020. [94] D. Guilbert, S. M. Collura, and A. Scipioni, “DC/DC converter
[71] N. A. El-Taweel, H. Khani, and H. E. Z. Farag, “Hydrogen storage topologies for electrolyzers: State-of-the-art and remaining key issues,”
optimal scheduling for fuel supply and capacity-based demand response Int. J. Hydrogen Energy, vol. 42, no. 38, pp. 23966–23985, 2017.
program under dynamic hydrogen pricing,” IEEE Trans. Smart Grid, [95] “FCeTM 80: Fuel cell engine,” US Hybrid. [Online]. Available:
vol. 10, no. 4, pp. 4531–4542, 2019. https://ushybrid.com/wp-content/uploads/2019/06/FCe80-
[72] S. Srdic and S. Lukic, “Toward extreme fast charging: challenges and Datasheet.pdf.
opportunities in directly connecting to medium-voltage line,” IEEE [96] B. Singh, B. N. Singh, A. Chandra, K. Al-Haddad, A. Pandey, and D. P.
Electrif. Mag., 2019. Kothari, “A review of single-phase improved power quality AC-DC
[73] “ChargePoint express plus,” ChargePoint. [Online]. Available: converters,” IEEE Trans. Ind. Electron., vol. 50, no. 5, pp. 962–981,
https://www.chargepoint.com/en-ca/products/commercial/express- 2003.
plus/. [Accessed: 20-Sep-2012]. [97] L. Huber, Y. Jang, and M. M. Jovanović, “Performance evaluation of
[74] “Multi-standard DC fast charging stations,” ABB. [Online]. Available: bridgeless PFC boost rectifiers,” IEEE Trans. Power Electron., vol. 23,
https://new.abb.com/ev-charging/products/car-charging/multi-standard. no. 3, pp. 1381–1390, 2008.
[75] I. S. Bayram, G. Michailidis, M. Devetsikiotis, and F. Granelli, “Electric [98] L. Huang, F. Chen, W. Yao, and Z. Lu, “Flexible mode bridgeless boost
power allocation in a network of fast charging stations,” IEEE J. Sel. PFC rectifier with high efficiency over a wide range of input voltage,”
Areas Commun., vol. 31, no. 7, pp. 1235–1246, 2013. IEEE Trans. Power Electron., vol. 32, no. 5, pp. 3513–3524, 2017.
[76] T. He, J. Zhu, D. D. C. Lu, L. Zheng, M. M. Aghdam, and J. Zhang, [99] Q. Huang, “Review of GaN totem-pole bridgeless PFC,” CPSS Trans.
“Comparison study of electric vehicles charging stations with AC and Power Electron. Appl., vol. 2, no. 3, pp. 187–196, 2017.
DC buses for bidirectional power flow in smart car parks,” in [100] Y. Chen, W. ping Dai, J. Zhou, and E. Hu, “Study and design of a novel
Proceedings IECON 2017 - 43rd Annual Conference of the IEEE three-phase bridgeless boost power factor correction,” IET Power
Industrial Electronics Society, 2017, pp. 4609–4614. Electron., vol. 7, no. 8, pp. 2013–2021, 2014.
[77] H. Kakigano, Y. Miura, and T. Ise, “Low-voltage bipolar-type dc [101] J. W. Kolar and T. Friedli, “The essence of three-phase PFC rectifier
microgrid for super high quality distribution,” IEEE Trans. Power systems- part I,” IEEE Trans. Power Electron., vol. 28, no. 1, pp. 176–
Electron., vol. 25, no. 12, pp. 3066–3075, 2010. 198, 2013.
[78] S. Beheshtaein, R. M. Cuzner, M. Forouzesh, M. Savaghebi, and J. M. [102] S. Kouro et al., “Recent advances and industrial applications of
Guerrero, “DC microgrid protection: A comprehensive review,” IEEE multilevel converters,” IEEE Trans. Ind. Electron., vol. 57, no. 8, pp.
J. Emerg. Sel. Top. Power Electron. (Early Access), 2019. 2553–2580, 2010.
[79] Zekalabs, “Electric buses DC charging stations.” [Online]. Available: [103] S. Rivera, B. Wu, S. Kouro, V. Yaramasu, and J. Wang, “Electric vehicle
https://www.zekalabs.com/applications/electric-buses-dc-charging- charging station using a neutral point clamped converter with bipolar
stations. DC bus,” IEEE Trans. Ind. Electron., vol. 62, no. 4, pp. 1999–2009,
[80] “Electric vehicle conductive charging system - part 23: Dc electric 2015.
vehicle charging station,” 2014. [104] T. S. Ustun, A. Zayegh, and C. Ozansoy, “Electric vehicle potential in
[81] A. Khaligh and S. Dusmez, “Comprehensive topological analysis of Australia: Its impact on smartgrids,” IEEE Ind. Electron. Mag., 2013.
conductive and inductive charging solutions for plug-in electric [105] S. S. Williamson, A. K. Rathore, and F. Musavi, “Industrial electronics
vehicles,” IEEE Trans. Veh. Technol., vol. 61, no. 8, pp. 3475–3489, for electric transportation: current state-of-the-art and future
2012. challenges,” IEEE Trans. Ind. Electron., vol. 62, no. 5, pp. 3021–3032,
[82] “Standard for the design of battery chargers used in stationary 2015.
applications (IEEE P2405),” 2014. [106] S. Dusmez, A. Cook, and A. Khaligh, “Comprehensive analysis of high
[83] “How galvanic isolation improves system performance and safety for quality power converters for level 3 off-board chargers,” in 2011 IEEE
solar and storage systems: Addressing the challenages and hazards Vehicle Power and Propulsion Conference, VPPC 2011, 2011, pp. 1–
found in new and existing solar and storage systems.” [Online]. 10.
Available: https://alenconsystems.com/whitepapers/galvanic-isolation/. [107] X. Fang et al., “Efficiency-oriented optimal design of the LLC resonant
[Accessed: 20-Nov-2019]. converter based on peak gain placement,” IEEE Trans. Power Electron.,
[84] “AKA system: 15 AKM 53 POC,” AKASOL. [Online]. Available: vol. 28, no. 5, pp. 2285–2296, 2013.
https://www.akasol.com/en/akasystem-akm-poc. [Accessed: 20-Sep- [108] W. Li, Q. Luo, Y. Mei, S. Zong, X. He, and C. Xia, “Flying-capacitor-
2011]. based hybrid LLC converters with input voltage autobalance ability for
[85] Z. Guo, D. Sha, X. Liao, and J. Luo, “Input-series-output-parallel phase- high voltage applications,” IEEE Trans. Power Electron., vol. 31, no. 3,
shift full-bridge derived DC-DC converters with auxiliary LC networks pp. 1908–1920, 2016.
to achieve wide zero-voltage switching range,” IEEE Trans. Power [109] H. Wang, S. Dusmez, and A. Khaligh, “Maximum efficiency point
Electron., vol. 29, no. 10, pp. 5081–5086, 2014. tracking technique for LLC-based PEV chargers through variable DC
[86] Y. Shi and X. Yang, “Wide range soft switching PWM three-level DC- link control,” IEEE Trans. Ind. Electron., vol. 61, no. 11, pp. 6041–
DC converters suitable for industrial applications,” IEEE Trans. Power 6049, 2014.
Electron., vol. 29, no. 2, pp. 603–616, 2014. [110] Z. U. Zahid, Z. M. Dalala, R. Chen, B. Chen, and J. S. Lai, “Design of
[87] V. Bist and B. Singh, “A unity power factor bridgeless isolated cuk bidirectional DC-DC resonant converter for Vehicle-to-Grid (V2G)
converter-fed brushless DC motor drive,” IEEE Trans. Ind. Electron., applications,” IEEE Trans. Transp. Electrif., vol. 1, no. 3, pp. 232–244,
vol. 62, no. 7, pp. 4118–4129, 2015. 2015.
[88] J. Duarte, L. R. Lima, L. Oliveira, M. Mezaroba, L. Michels, and C. [111] J. H. Jung, H. S. Kim, M. H. Ryu, and J. W. Baek, “Design methodology
Rech, “Modeling and digital control of a single-stage step-up/down of bidirectional CLLC resonant converter for high-frequency isolation
isolated PFC rectifier,” IEEE Trans. Ind. Informatics, vol. 9, no. 2, pp. of DC distribution systems,” IEEE Trans. Power Electron., vol. 28, no.
1017–1028, 2013. 4, pp. 1741–1755, 2013.
[89] M. Narimani and G. Moschopoulos, “A New interleaved three-phase [112] E. Wood, C. Rames, M. Muratori, S. Raghavan, and M. Melaina,
single-stage PFC AC-DC converter with flying capacitor,” IEEE Trans. “National plug-in electric vehicle infrastructure analysis,” 2017.
Power Electron., vol. 30, no. 7, pp. 3695–3702, 2015. [113] CCME, “Business case for investing in electric vehicle direct current
[90] S. Gangavarapu, A. K. Rathore, and D. M. Fulwani, “Three-phase fast charge station infrastructure,” 2016.
single-stage-isolated cuk-based PFC converter,” IEEE Trans. Power [114] T. S. Bryden, G. Hilton, B. Dimitrov, C. Ponce De León, and A. Cruden,
Electron., vol. 34, no. 2, pp. 1798–1808, 2019. “Rating a stationary energy storage system within a fast electric vehicle
[91] L. Schrittwieser, M. Leibl, and J. W. Kolar, “99% Efficient isolated charging station considering user waiting times,” IEEE Trans. Transp.
three-phase matrix-type DAB buck-boost PFC rectifier,” IEEE Trans. Electrif., vol. 5, no. 4, pp. 879–889, 2019.
Power Electron., vol. 35, no. 1, pp. 138–157, 2020. [115] T. S. Bryden, G. Hilton, A. Cruden, and T. Holton, “Electric vehicle fast

2332-7782 (c) 2020 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
Authorized licensed use limited to: SUNY AT STONY BROOK. Downloaded on August 12,2020 at 11:54:55 UTC from IEEE Xplore. Restrictions apply.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TTE.2020.3015743, IEEE
Transactions on Transportation Electrification

charging station usage and power requirements,” Energy, vol. 152, pp. and-semi/.
322–332, 2018. [137] D. Strickland, L. Chittock, D. A. Stone, M. P. Foster, and B. Price,
[116] Brusa, “Bidirectional 750 V DC/DC converter.” [Online]. Available: “Estimation of transportation battery second life for use in electricity
https://www.brusa.biz/fileadmin/template/Produkte/Wandler/BRUSA_ grid systems,” IEEE Trans. Sustain. Energy, vol. 5, no. 3, pp. 795–803,
Factsheet_BDC546.pdf. [Accessed: 30-Apr-2020]. 2014.
[117] Schneider, “Transformer dry type 150kVA.” [Online]. Available: [138] L. C. Casals, B. Amante García, and C. Canal, “Second life batteries
https://www.se.com/ca/en/product/EE150T1814H/transformer-dry- lifespan: Rest of useful life and environmental analysis,” J. Environ.
type-150kva-480d480y277/. [Accessed: 05-May-2020]. Manage., vol. 232, pp. 354–363, 2019.
[118] Microchip, “3 phase 30 kW Vienna PFC reference design,” 2020. [139] E. Hossain, D. Murtaugh, J. Mody, H. M. R. Faruque, M. S. H. Sunny,
[Online]. Available: and N. Mohammad, “A Comprehensive Review on Second-Life
http://ww1.microchip.com/downloads/en/DeviceDoc/MSCSICPFC- Batteries: Current State, Manufacturing Considerations, Applications,
REF5-3-Phase-30-kW-Vienna-PFC-Reference-Design- Impacts, Barriers Potential Solutions, Business Strategies, and Policies,”
DS50002952A.pdf. [Accessed: 30-Apr-2020]. IEEE Access, vol. 7, pp. 73215–73252, 2019.
[119] Zekalabs, “RedPrime AC-DC converter 100kW, 1200V.” [Online]. [140] H. Engel, P. Hertzke, and G. Siccardo, “Second-life EV batteries: The
Available: https://www.zekalabs.com/products/non-isolated-high- newest value pool in energy storage,” McKinsey Co., 2019.
power-converters/ac-dc-converter-100kw-1200v. [Accessed: 15-May- [141] Handbook on energy storage system. Aisan Development Bank, 2018.
2020]. [142] N. Mukherjee and D. Strickland, “Analysis and comparative study of
[120] “20kW full bridge resonant LLC converter,” Wolfspeed, 2016. [Online]. different converter modes in modular second-life hybrid Battery energy
Available: storage systems,” IEEE J. Emerg. Sel. Top. Power Electron., vol. 4, no.
https://www.wolfspeed.com/downloads/dl/file/id/931/product/0/20kw_ 2, pp. 547–563, 2016.
full_bridge_resonant_llc_converter.pdf. [143] “UC Davis RMI Winery microgrid project,” UC Davis. [Online].
[121] Burckhardt, “Oil-free high pressure diaphragm compressors.” [Online]. Available: http://mae.engr.ucdavis.edu/jwpark/DavisSite/pages/BWF-
Available: microgrid.html.
https://www.burckhardtcompression.com/solution/compressor- [144] K. Richa, “Sustainable management of lithium-ion batteries after use in
technologies/diaphragm-compressors/. [Accessed: 20-May-2020]. electric vehicles,” 2016.
[122] “Hydrogen advantages,” Nikolamotors. [Online]. Available: [145] F. Lambert, “BMW and Bosch open new 2.8 MWh energy storage
https://nikolamotor.com/hydrogen#hydrogen-advantages. facility built from batteries from over 100 electric cars,” Electrek, 2016.
[123] C. Jung, “Power up with 800-V systems: The benefits of upgrading [Online]. Available: https://electrek.co/2016/09/22/bmw-bosch-energy-
voltage power for battery-electric passenger vehicles,” IEEE storage-facility-built-from-batteries-from-over-100-electric-cars/.
Electrification Magazine, 2017. [146] C. Nedler and E. Rogers, “Reducing EV charging infrastructure cost,”
[124] A. Meintz et al., “Enabling fast charging – vehicle considerations,” J. 2019.
Power Sources, vol. 367, pp. 216–227, 2017. [147] IRENA, “Hydrogen from renewable power: Technology outlook for the
[125] X. Yu, X. She, X. Zhou, and A. Q. Huang, “Power management for DC energy transition,” 2018.
microgrid enabled by solid-state transformer,” IEEE Trans. Smart Grid,
vol. 5, no. 2, pp. 954–965, 2014.
[126] X. She, A. Q. Huang, S. Lukic, and M. E. Baran, “On integration of Md Ahsanul Hoque Rafi (M’20) received the
solid-state transformer with zonal DC microgrid,” IEEE Trans. Smart B.Sc. in electrical and electronics engineering
Grid, vol. 3, no. 2, pp. 975–985, 2012. from Middle East Technical University, Turkey in
[127] X. She, A. Q. Huang, and R. Burgos, “Review of solid-state transformer 2015. He received the M.Sc. degree from both
technologies and their application in power distribution systems,” IEEE KTH Royal Institute of Technology, Sweden and
J. Emerg. Sel. Top. Power Electron., vol. 1, no. 3, pp. 186–198, 2013. Eindhoven University of Technology, Netherlands
[128] A. Q. Huang, “Medium-voltage solid-state transformer: Technology for in electrical engineering under InnoEnergy, a
a smarter and resilient Grid,” IEEE Ind. Electron. Mag., vol. 10, no. 3, European consortium. He is currently with the
pp. 29–42, 2016. department of electrical and computer engineering
[129] M. Vasiladiotis and A. Rufer, “A modular multiport power electronic at McMaster University, Hamilton, ON, Canada as
transformer With integrated split battery energy storage for versatile a Ph.D. candidate. His research interests include design, control and
ultrafast EV charging stations,” IEEE Trans. Ind. Electron., vol. 62, no. optimization of high-power electronic converters, design and optimization of
5, pp. 3213–3222, 2015. on and off-road electric vehicles, and electrification of underground mining
[130] L. S. Xavier, W. C. S. Amorim, A. F. Cupertino, V. F. Mendes, W. C. system.
do Boaventura, and H. A. Pereira, “Power converters for battery
energystorage systems connected to mediumvoltage systems: a Jennifer Bauman (M’04) received the B.Sc. and
comprehensive review,” BMC Energy, vol. 1, no. 7, pp. 1–15, 2019. Ph.D. degrees in electrical engineering from the
[131] W. Jacobs, M. R. Hodkiewicz, and T. Bräunl, “A cost-benefit analysis University of Waterloo, Waterloo, Canada, in 2004
of electric loaders to reduce diesel emissions in underground hard rock and 2008, respectively. From 2009 to 2016, she was
mines,” IEEE Trans. Ind. Appl., vol. 51, no. 3, pp. 2565–2573, 2015. the Director of Research at CrossChasm
[132] B. Parkhideh, H. Mirzaee, and S. Bhattacharya, “Supplementary energy Technologies, where she led the modeling team on a
storage and hybrid front-end converters for high-power mobile mining wide variety of automotive projects. She is currently
equipment,” IEEE Trans. Ind. Appl., vol. 49, no. 4, pp. 1863–1872, an Assistant Professor of electrical engineering at
2013. McMaster University, Hamilton, Canada. Dr.
[133] B. Schmidt, “New 2MW electric vehicle chargers could revolutionise Bauman is a registered Professional Engineer and an
mining industry,” The Driven, 2019. [Online]. Available: Associate Editor for the IEEE Open Journal of Power Electronics. Her current
https://thedriven.io/2019/10/23/new-2mw-electric-vehicle-chargers- research interests include power electronic converters for electrified
could-revolutionise-mining-industry/. powertrains (including wide bandgap devices), vehicle design, modeling, and
[134] B. Sarlioglu and C. T. Morris, “More electric aircraft: review, control, and EV interactions with the smart grid.
challenges, and opportunities for commercial transport aircraft,” IEEE
Trans. Transp. Electrif., vol. 1, no. 1, pp. 54–64, 2015.
[135] “Electrified Aircraft Propulsion (EAP),” NASA. [Online]. Available:
https://www1.grc.nasa.gov/aeronautics/electrified-aircraft-propulsion-
eap/.
[136] “ChargePoint reveals new concept design for high-powered charging of
electric aircraft and semi-trucks,” ChargePoint, 2018. [Online].
Available: https://www.chargepoint.com/about/news/chargepoint-
reveals-new-concept-design-high-powered-charging-electric-aircraft-

2332-7782 (c) 2020 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
Authorized licensed use limited to: SUNY AT STONY BROOK. Downloaded on August 12,2020 at 11:54:55 UTC from IEEE Xplore. Restrictions apply.