Diary

Renewable and Sustainable Energy Reviews 52 (2015) 802–814

Contents lists available at ScienceDirect

Renewable and Sustainable Energy Reviews

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

A review on energy management system for fuel cell hybrid electric
vehicle: Issues and challenges

N. Sulaiman a,n, M.A. Hannan b,nn, A. Mohamed b, E.H. Majlan a, W.R. Wan Daud a

a Fuel Cell Institute, Universiti Kebangsaan Malaysia, Malaysia
b Department of Electrical, Electronics & System Engineering, Universiti Kebangsaan Malaysia, Malaysia

article info abstract

Article history: Emerging issues on fuel price and greenhouse gas emissions have attracted attention on the alternative energy
Received 21 October 2014 sources, especially in transportation sector. The transportation sector accounts for 40% of total fuel
Received in revised form consumption. Thus, an increasing number of studies have been conducted on hybrid electric vehicles (HEVs)
19 May 2015 and their energy management system (EMS). This paper focuses on reviews of EMSs for fuel cell (FC) based
Accepted 28 July 2015 HEV in combination with battery and super-capacitor, respectively. Various aspects and classifications of fuel
cell–HEV EMS are explained in this paper. Different types of FC–HEV control models and algorithms derived
Keywords: from simulation and experiment are explained in details for an analytical justification for the most optimal
Energy management system control strategy. The performances of the various combinations of FC–HEV system are summarized in the table
Control strategy along with relevant references. This paper provides comprehensive survey of FC–HEV on their source
Fuel cell combination, models, energy management system (EMS) etc. developed by various researchers. From the
Battery rigorous review, it is observed that the existing technologies more or less are capable to perform well;
Supercapacitor however, the reliability and the intelligent systems are still not up to the mark. Accordingly, current issues and
Hybrid electric vehicle challenges on the FC–HEV technologies are highlighted with a brief suggestions and discussion for the progress
of future FC–HEV vehicle research. This review will hopefully lead to increasing efforts towards the
development of economic, longer lifetime, hydrogen viable, efficient electronic interface and well performed
EMS for future FC–HEV.

& 2015 Elsevier Ltd. All rights reserved.

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 802
2. FC–HEV classifications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 803

2.1. Fuel cell–battery HEVs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 804
2.2. Fuel cell–SC HEV . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 805
2.3. Fuel cell–battery–SC HEV . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 807
3. EMS control strategies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 809
4. Issues and challenges on FC–HEV . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 811
4.1. Hydrogen viability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 811
4.2. Fuel cell economic impact. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 812
4.3. Battery lifetime . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 812
4.4. Power electronic interface. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 812
4.5. Motor integration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 812
4.6. Energy management system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 812
5. Conclusion and suggestions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 813
Acknowledgment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 813
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 813

n Corresponding author.
nn Corresponding author.

E-mail addresses: [email protected] (N. Sulaiman), [email protected] (M.A. Hannan).

http://dx.doi.org/10.1016/j.rser.2015.07.132
1364-0321/& 2015 Elsevier Ltd. All rights reserved.

N. Sulaiman et al. / Renewable and Sustainable Energy Reviews 52 (2015) 802–814 803

1. Introduction between the main source and auxiliary sources [12]. This study
considers fuel cell as the main source, with the battery and super-
Hybrid electric vehicles (HEVs) have become a phenomenon since capacitor (SC) as the auxiliary sources of HEVs. A battery is an
the Toyota Prius and Honda Insight were introduced to the auto- electrochemical device that consists of electrodes separated by
motive industry in 1997 and 1999 [1], respectively. An electric motor electrolyte to convert chemical energy into electrical energy. Several
power rating of 33 kW was achieved by the first-generation Prius, types of batteries are available in the market. These types include lead
with the third generation having a 60 kW power rating as compared acid, lithium ion (Li-ion), nickel metal hydride (NiMH), alkaline nickel
with the 10 kW power rating of Insight [2]. Honda hybrids only use cadmium, and so on. Among them, NiMH and Li-ion are the most
one electric motor and one inverter, whereas Toyota hybrids embed commonly used in electric vehicles [7].
two electric motors and two inverters with split power. The config-
uration of Honda is simple but has limitations in power flow and SC is an electrochemical capacitor used to provide peak power
system optimization. Meanwhile, the Toyota design is more complex for short durations. The electrical characteristic of SC is similar to
because power can be fed to the motor by the battery alone, gasoline that of capacitors and consists of either electrical double-layer
alone, or a combination of both battery and gasoline [3]. Such design capacitors made of non-porous materials, such as activated carbon
provides continuously variable transmission and a larger room for [13], or pseudo-capacitors containing transition metal oxides,
electric motors because clutches are not used [1]. However, more than nitrides, and polymers possessing relatively high surface areas.
25% of complaints are made against the Prius [4]. The power electronic interface is the interconnection or integra-
tion between the energy source and the motor and usually
The idea of HEV started way back in early 1900s when the electric consists of power converters. Power converters can be a plain
vehicle failed to succeed in the market due to costly battery inverter or DC–DC converter (also known as chopper) with an
technology and limited driving range. Differences between electric inverter [10].
vehicles and hybrid electric vehicles may appear rather small. How-
ever, the difference in terms of impact of each vehicle and their However, the main concerns on EMSs are how efficient the used
controls are massively huge. An electric vehicle (EV) operates on strategy or control method. Apart from that, EMS in FC–HEVs faces
electric drives; consisting of batteries, electric motors and electric some issues or challenges in its development and application. This
generators, such as Honda EV Plus, GM EV1, and Toyota RAV4 EV. On paper analyzes and discusses the previous methods of EMS for
the other hand, a HEV operates using one electric motor and an FC–HEV in order to suggest the most efficient EMS for FC–HEV. The
internal combustion engine (ICE) due to its energy source consisting issues and challenges of developing FC–HEV are also discussed to
of battery and fuel, respectively [5]. The HEV reduces fossil fuel provide knowledge and information to the community as a whole.
consumption due to its operation of running on battery during low This is important for research involving future development of new
power demand and runs on fuel only during acceleration or high load EMS or improvement of previous EMS for FC–HEV.
power. Among HEV models that made its way into mass production
are Honda Insight, Toyota Prius, and recently Honda Civic Hybrid. 2. FC–HEV classifications
However, issues of pollution, global warming, and drastic rise of fuel
price, have forced automotive manufacturers to introduce fuel cell The FC–HEV are reviewed in terms of multiple sources of energy
vehicles. that have to be managed accordingly to ensure that the energy fed to
the electric motor is sufficient as per the demand or load power. This
Fuel cells have been attracting considerable attention for their zero is because under certain conditions, energy could be drawn from the
emission of greenhouse gases. Furthermore, the energy from a fuel battery or from the fuel cell or from both energy sources. In other
cell is drawn from the chemical reaction of hydrogen and oxygen from instances, the energy is drawn from both the fuel cell and battery, and
air, which is an abundant resource. Fuel cells directly transform at certain points, the battery and SC are charged [14]. The FC–HEV
chemical energy into electrical energy through an electrochemical operating conditions can be related to energy required or motor
process [6]. A fuel cell comprises two electrodes immersed in operation, such as starting, cruising, accelerating, and braking [15].
electrolytes and sandwiched together. Seven types of fuel cells are However, this study thoroughly reviews the EMS for current or
available in the market: proton exchange membrane fuel cell (PEMFC), previous fuel cell HEVs and then summarizes and proposes recom-
direct methanol fuel cell, alkaline fuel cell, phosphoric acid fuel cell, mendations for future research or improvement.
molten carbonate fuel cell, solid oxide fuel cell, and microbial fuel cell
[7]. These types of fuel cells are either used commercially or in A fuel cell vehicle fed with fuel cell alone, usually PEMFC, appears
research. Selecting the appropriate fuel cell is essential because of the to have constraints, such as slow dynamic properties [16]. Further-
different operating temperatures and power levels produced. Cost and more, the membrane electrode assembly of the PEMFC is open to
efficiency are important considerations in selecting the best fuel cell. failures, such as membrane breaks, internal gas leakage, and cell
flooding/drying [17]. Moreover, PEMFCs have slow dynamic response
Hydrogen can be supplied by methanol and propane from the with respect to load changes, which shortens the service life of the
reformation of hydrocarbon and biological processes. This gas can also fuel cell [18]. Performance, reliability, durability, cost, fuel availability,
be drawn from the electrolysis process by using an electrolyzer, which public satisfaction, and performance during transients [19] are essen-
breaks the chemical bond in water into hydrogen and oxygen and tial factors in developing fuel cell electric vehicles that can compete
then collects hydrogen in a gas tank [8]. Implementing renewable with the ICE vehicles that are monopolizing the streets. Thus,
energy resources, such as fuel cells, can give rise to such concerns as automotive manufacturers started to produce fuel cell HEVs. Toyota
efficiency, cost, and limitations. Efficiency depends on system config- has produced Toyota FCHV, whereas Honda produced Honda FCX-V4
uration, design, and component selection. Moreover, the cost of a and FCX Clarity [20]. Meanwhile, Hyundai has also produced its own
system strongly relies on its efficiency [9]. Other limitations, such as fuel cell vehicle, that is, TUCSON FCEV, which embeds a fuel cell and
the relation of vehicle speed with the power required to achieve a battery [21]. General Motors, on the other hand, released its Chevrolet
certain speed also need to be considered. Thus, the idea of a hybrid Equinox Fuel Cell Vehicle which feeds on hydrogen and battery in
fuel cell electric vehicle emerged. 2008 [22,23]. Daimler released its F-CELL B-CLASS in 2005 where else
Volkswagen released its Passat Lingyu FCEV in 2008 [24]. These FCEV
HEVs have more than one power source [10], which serves an of fuel cell and battery hybrid face the limitations in battery lifetime
important function in determining which power source should be though, of which is the reason supercapacitors are now being
activated or drawn [11]. An energy management system (EMS) considered by many automotive manufacturers [7].
controls the energy source to feed the electric motor. In other words,
the developed system serves as the power splitter of the energy

804 N. Sulaiman et al. / Renewable and Sustainable Energy Reviews 52 (2015) 802–814

Numerous studies have been conducted and published on HEVs, of Nevertheless, in all driving cycles it was presented that the battery
which EMS of the HEVs are reviewed in this paper. Sakka et al. [25] SOC does not change much which shows it is not fully used. This
reported that HEVs use two energy storage devices, one with high could extend the battery life span, but the cost is high considering cost
energy storage capability that extends the driving distance and ano- of battery is based on size and fixed structure. However, the DIRECT
ther with high power reversibility that aids in acceleration and algorithm was not elaborated and dynamic problems such as energy
regenerative braking. Regenerative braking is a condition in which starvation and anode–cathode pressure difference were ignored in
braking electrical energy is passed to a rechargeable battery for later this study.
usage, unlike the common dynamic braking in which motion energy
is converted into electrical energy [26]. Numerous studies have been Another study on EMS for fuel cell–battery HEV was reported by
conducted on fuel cell HEVs, either integrating fuel cell with battery or Xiao and Wang [14]. Xiao and Wang applied fuzzy control with four
SC or both, to come up with an efficient EMS. The reviews and operation modes on a 9 kW electric vehicle. The concept was to
comparative analysis of these studies are presented in the following calculate the total power required, considering the power negative
sections. when braking through the acceleration and braking pedal positions
and then allocating power from the fuel cell or battery to the fuel cell
2.1. Fuel cell–battery HEVs operating point and SOC. Triangular membership functions of the
variables for error or load power, battery power, fuel cell power,
Numerous studies have been conducted on the EMS approaches recovery power, and SOC were respectively set as [À 3 kW, 9 kW],
for fuel cell–battery HEVs. A battery can provide high current to start [À 3 kW, 6 kW], [0 kW, 5 kW], [0 kW, 3 kW], and [0.3, 0.9] for this
the motor and can serve as load limiting device that enables the fuel control strategy [14]. Fuzzy control usually embeds IF–THEN rules, but
cell to operate initially at low power and then progress into high- the rules are not visible in the report. This study focused on
power operation [10]. Shuang et al. [27] conducted a study on power interfacing the control strategy with the positioning of the accelerator
management strategy integrating serial regenerative braking. Power and brake pedal through the DSP320TM2812 hardware platform, but
management was performed by a controller, which consists of two the simulation only lasted for 15 min, and results on hardware
modules: (1) power distribution module to divide demand power into performance were unclear.
DC power and battery power and (2) braking strategy module that
can be either serial or parallel. The serial braking strategy was chosen Meanwhile, Ouddah et al. [28] conducted a study that compared
because the serial braking strategy has a higher percentage of two EMSs for another fuel cell–battery HEV. The first EMS was based
regenerative energy. Simulation was performed by using Matlab/ on power frequency splitting, in which the frequency decomposition
Simulink for the China Urban Bus Driving cycle. Results revealed that allocates the frequency components to each source. In this strategy,
relative to the objective of minimizing power loss, the revised optimal the fuel cell is allocated to supply for low frequency, and the battery is
strategy achieved a reduction of 79%, whereas the existing optimal allocated to supply for high frequency. Meanwhile, the other EMS is
strategy achieved a reduction of 60% [27]. However, the highlight of an optimal control strategy of which Ouddah et al. applied Pontrya-
this study was on minimizing power loss on the basis of serial gin's minimum principle, in which the system is represented by state
regenerative braking. Thus, the emphasis on power splitting algorithm space dynamic equations. In the optimal control strategy, the battery
was not apparent. Moreover, the results only showed simulation supplies transient energy that is not provided by the fuel cell. The
performance relative to the comparison between regenerative simulation that considered European Driving Cycle (ECE-15) was
and non-regenerative braking, as well as power loss and state of performed in Matlab/Simulink by using SimPowerSystems Toolbox.
charge (SOC) for the rule-base, existing optimal, and revised optimal Results showed that the load power was met and that hydrogen
strategies. consumption was reduced by the optimal control strategy. However,
the study was limited to simulation, and the speed and SOC were not
The general block diagram for the structure of a fuel cell–battery monitored. Losses in power converters were also neglected.
HEV, as illustrated by Alloui et al. [32] and Kelouwani et al. [33], is
shown in Fig. 1. The concept of the EMS involving fuel cell and battery Sundstrom and Stefanopoulou [29] reported on the optimal power
is that load sharing occurs between these two energy sources. The split of a fuel cell–battery HEV. The control strategy was based on
energy is then fed to the DC bus. The DC voltage will then be fed into deterministic dynamic programming. The method minimizes the cost
an inverter to convert the DC voltage into AC voltage, which can move function defined from the serial multiplication function of SOC
the AC motor. deviation, hydrogen consumption, and oxygen excess ratio. The
control strategy was simulated using ADVISOR simulation tool for
Li and Liu [12] reported a study on the fuel cell–battery with New York City Cycle, FTP-72, and SFTP drive cycles to represent mild,
combined design and power management optimization based on medium, and aggressive drive cycles, respectively. However, the
fuzzy logic controller. The combined optimization power management simulation only focused on the effects of battery size on optimal
which implies DIRECT algorithm was simulated for three driving control performance. However, The 80 kW power and 129 km/h
cycles: Urban Dynamometer Driving Schedule (UDDS), Highway Fuel system neglected energy losses in the power train, DC motor, inverter,
Economy Test (HWFET), and New European Drive Cycle (NEDC). The and DC/DC converter.
results showed highest efficiency in HWFET and highest DOH in
UDDS, and lowest efficiency in UDDS but lowest DOH in HWFET. The Based on a report by Xie et al. [30], the EMS for a fuel cell–battery
low DOH is explained with less braking, thus, less battery recharging. HEV was based on a neural network optimization algorithm. The
neural network structure was a three-layer forward structure with
three inputs: load power (PL), total load power interval time change

Fuel Cell Unidirectional DC Inverter Motor
Battery Converter bus

Bidirectional
Converter

Fig. 1. Structure of fuel cell–battery HEV.

N. Sulaiman et al. / Renewable and Sustainable Energy Reviews 52 (2015) 802–814 805

quantity (ePL), and battery SOC, as well as one output: DC output hybridization also has a longer life cycle and high frequency rate of
power (PDC). Suitable coding is essential because coding is the key to charging and discharging [16]. These features make SC suitable for
genetic algorithms. A fitness function provides information to guide hybridization with fuel cell as an energy source for HEVs. Thus, more
the search in the genetic algorithm and gradually approaches the studies have been conducted on fuel cell–SC HEVs. Some previous
optimum parameter combination. The EMS was simulated by using works on the EMS for fuel cell–SC HEVs are reviewed and summar-
ADVISOR, a tool based on Matlab for the Urban Dynamometer Driving ized in the following paragraphs.
Schedule (UDDS) drive cycle. Results of the comparison between fuzzy
logic and neural network showed a decrement of 7.23% in fuel Thounthong et al.[20] demonstrated the performance of a fuel
consumption, but dynamic performance was lower in the latter, cell–SC energy system for an electric vehicle using differential
which was evident in the increase of 0.4 s in accelerating time. flatness controls and proved that this system can circumvent the
fuel cell fast transition, thus reducing the fuel cell stack stress.
Meanwhile, Karunarathne et al. [31], reported on a study of a fuel Theoretically, this reduction indirectly extends the lifetime. Two
cell–battery hybrid unmanned aerial electric vehicle which adopts parallel boost converters with interleaving switching technique
Adaptive Neuro Fuzzy Inference System (ANFIS) based controller, to were adopted to minimize output current ripple. However, the
control the vehicle's propulsion system. This study focused on study only considered a small-scale 500 W, 50 A, 292 F, 30 V
simulation only which was performed in Matlab SimPowerSystems proton exchange membrane fuel cell [20].
toolbox and the energy management system (EMS) in Simulink. The
obtained results proved that the energy sharing is based on power A general block diagram for the structure of a fuel cell–SC HEV,
demand and SOC. However, it needs to be extended in real life as illustrated by Thounthong et al. and Rodatz et al., is shown in
application to ensure its true performance. Alloui et al. [32], on the Fig. 2. The operation is similar to that of the previous system, in
other hand, reported on energy management of a fuel cell–battery which energy is shared between the fuel cell and SC, with either
hybrid electric vehicle based on frequency separation. The simulation one or both of the energies being drawn and fed into the DC bus.
was applied in urban driving European cycle using a vehicle dynamic The DC voltage is then inverted via an inverter from DC voltage
model of 400 W load power energy management system. However, in into AC voltage, which will then be useful for the AC motor.
the report, there was no evidence on the simulation tool.
Tani et al. [37] conducted a study on an EMS for a fuel
Another study was reported by Kelouwani et al. [33] on an energy cell–ultracapacitor (UC) HEV. The strategy implemented in the
management system (EMS) of a fuel cell–battery plug-in hybrid study was based on the concept of bidirectional load power
electric vehicle (PHEV). The EMS applied optimal anticipatory power sharing using the new European drive cycle (NEDC) with poly-
splitting algorithm and expected that the study could analyze the EMS nomial control technique. The polynomial control technique was
designed in real-time. The results showed it ability of being able to compared with a PI controller, which is more robust and has better
generate required optimal control signals for less hydrogen consump- disturbance rejection. The expectations were that the EMS could
tion by 5% of fuel cell power source, preserves battery life by allocate the fuel cell average power and fluctuating power of the
maintaining it close to the minimum prescribed energy. The algorithm ultracapacitor (UC). Results showed that the fuel cell fed constant
requires no prior knowledge of entire driving cycle for the EMS to DC voltage and current, whereas the SC was the main contributor
trigger fuel cell operation. However, it is focused for a PHEV and the to the DC bus. However, the system had a limitation, that is,
hydrogen cost ratio influence towards the EMS was not apparent. fluctuations in the resultant graphs, which can be attributed to the
Benrabeh et al. [34], on the other hand, reported that the fuel cell– sensors for data acquisition used in the system.
battery HEV which was based on a digital current sharing method
integrated with a fuel cell interleaved boost converter (FC-IBC). Another study on fuel cell–UC HEVs was reported by Uzunoglu
Performance of the vehicle was simulated in Matlab/Simulink for and Alam [38], who applied a wavelet-based load-sharing algo-
Standard Driving Cycle in steady-state and dynamic conditions of the rithm. The UC used in this study consists of a string of 75 units of
HEV. The obtained results showed that the current for both inductors 2.5V, 2700F UCs connected in series to fulfill the requirement of
were the same. The report mentioned that the objective of the paper matching the PEMFC voltage of 188V. The idea of integrating a fuel
was to minimize size and harmonic content; however, there was no cell with UC was to enable the UC to aid the fuel cell during
evidence from the results presented in the report. In a report by Xiao acceleration or peak power demand and to ensure that the UC is
et al. [35], a fuel cell–battery hybrid electric vehicle EMS is applied recovered during deceleration or braking. The study considered
fuzzy logic with four operational modes. The vehicle was modeled charging the UC from the fuel cell system when the UC bank SOC
and simulated using ADVISOR; of which the vehicle was modeled decreases below 20%. However, in the simulation, the UC bank SOC
based on 25 kW fuel cell, 12 V battery and 75 kW motor, and with DC never decreased to below 20% [37]. The simulation results showed
voltage 308 V, performed 88.5 km/h acceleration speed. Unfortu- that the output voltage was low when the output power was high,
nately, the EMS was not described thoroughly. and the variation of DC load voltage was in the required range.
However, the study lacks real-time application.
Based on these reviews above, it is obvious that having fuel cell
with battery improves the fuel cell electric vehicle. However, batteries Meanwhile, Efstathiou et al. [39] reported on a study of a fuel
have longer charging time, and the life cycle depends on operating cell–UC HEV. The study was conducted to integrate the EMS with a
temperature and the number of discharge cycles [16]. Thus, some variable transmission gearbox. The EMS applied a control program
studies have examined fuel cell–SC HEVs, as explained in the next consisting of two fuzzy logic controllers (FLCs). The main FLC has four
chapter. Detail studies on EMS for fuel cell–battery HEVs are shown inputs such as road inclination, UC SOC (UC–SOC), hydrogen level
in Table 1. (H2L), average velocity, and three outputs, namely, rotational speed,
power required by fuel cell (FC-power), and power required by UC,
2.2. Fuel cell–SC HEV respectively. The second FLC has three input variables, namely,
UC–SOC, H2L, output current, and one output variable that is the
Super-capacitor (SC) is characterized by high power density and percentage of the fuel cell current. The study embedded an appro-
fast transient response which makes it preferable energy storage for priate generic rule to the FLC, and the vehicle performance was
vehicular applications for short instances and be recharged in short assessed specifically in terms of the hydrogen consumption and speed
period of time [36]. The combination of fast transient response of SC of the vehicle between the use of a constant 1/13 gear ratio and
with slow transient response of fuel cell is absolutely a practical continuous variable transmission (CVT) at a range of 1/17–1/10 [38].
alternative to improve the efficiency and performance of HEVs. This The vehicle was tested for zero-inclination track and real track for
both constant gearbox and CVT. The vehicle demonstrated lower
hydrogen consumption when using CVT without affecting the desired

806 N. Sulaiman et al. / Renewable and Sustainable Energy Reviews 52 (2015) 802–814

Table 1
Summary table of EMS for fuel cell–battery HEV.

Author EMS method/control strategy Simulation/ DC SOC Max Max Advantage Disadvantage
hardware bus speed power

Li and Liu [12] DIRECT algorithm based on Fuzzy logic Simulation: – 0.6– 30 m/s 30 kW Simulation shows highest Simulation tool was

control unknown tool 0.65 efficiency in HWFET not clear, dynamic

(UDDS, HWFET, problems were

NEDC drive cycles) ignored

Xiao and Wang Operational mode strategy (4 modes) Simulation: 300 V 0.67– 88.5 km/ 25 kW Effective braking energy Simulation only

[14] ADVISOR (UDDS 0.71 h recovery

cycle)

Xie et al. [15] Multi model control based on fuzzy Simulation: 300 V 0.64– 80 km/h 40 kW Fuel consumption per Simulation only

supervised rule Matlab/Simulink, 0.68 hundred km was lesser 7.2%

ADVISOR (UDDS, than fuzzy

EPA drive cycles)

He et al. [21] Comparison of 3 control strategies (FC Simulation: 450 V 0.55 90 km/h 40 kW Simulation comparison of Experimental setup

output power oriented, FCE loading and Matlab/Simulink three control strategies – based on one control

unloading, instantaneous power Hardware: Hybrid optimization strategy best strategy only

distribution optimization) bus for lowest hydrogen usage

Shuang et al. Comparison of 3 control strategies (rule- Simulation: – 0.5 60 km/h 130 kW Simulation focuses on Simulation only, focus

[27] based, existing-optimal, revised- Matlab/Simulink minimizing power loss on minimizing power

optimal) (China Urban Bus loss only

Drive Cycle)

Ouddah et al. Comparison of 2 strategies (power Simulation: 60 V – – 2 kW DC bus voltage is more Simulation only,

[28] frequency splitting, optimal control Matlab/Simulink smoothly regulated in losses in power

strategy – Pontryagin's minimum SimPowerSystem optimal control converter neglected

principle) (ECE-15 drive

cycle)

Sundstrom Optimal control strategy using Simulation: – 0.6 129 km/h 80 kW Comparison of results show Simulation only, focus

and Deterministic Dynamic Programming ADVISOR that hybridization is on performance with

Stefanopou- beneficial for mild drive battery size only,

lou [29] cycles energy losses

neglected

Karunarathne Adaptive Fuzzy Inference System Simulation: – 0.5 1200 rpm 2 kW Simulation results showed Simulation only

et al. [31] (ANFIS) Matlab/Simulink motor torque varying from

SimPowerSystem 2 N m to 5 N m in 0.1 s

Alloui et al. Frequency separation method using PI Simulation: 400 V – 50 km/h 8 kW Results showed DC bus Simulation tool was

[32] regulator unknown tool voltage smoothly regulated not clear

(European drive at 400 V

cycle)

Kelouwani Anticipatory Power Splitting Algorithm Simulation: – – o 40 m/s 4 kW Optimal energy Simulation tool was

et al. [33] unknown tool management system was not clear, for plug-in

(US06, NEDC, J1015 proven effective HEV

drive cycles)

Benrabeh et al. Digital current sharing method (4 Simulation: – 0.48– 120 km/h 60 kW Results showed regenerative Simulation only

[34] operating modes) Matlab/Simulink 0.5 energy was more in urban

(Standard Drive drive cycle

Cycle)

Xiao et al. [35] Fuzzy logic control (4 operating modes) Simulation: – 0.6– 88.5 km/ 75 kW Simulation gave efficiency of Fuzzy logic rules were

Matlab/Simulink 0.7 h fuel cell as 56% and battery not visible

as 80%

Fuel Cell Unidirectional DC Inverter Motor
Supercapacitor Converter bus

Bidirectional
Converter

Fig. 2. Structure of fuel cell–SC HEV.

speed. However, the study only focused on hydrogen consumption reversible energy. The control variable of this HEV is a power split
and not on the power feed. factor that determines the power path between fuel cell and SC or
that charges the SC. Simulations were performed in VP-SIM, a vehicle
Rodatz et al. [40] also reported on the EMS of a fuel cell–SC HEV. simulator programmed in Matlab/Simulink for driving cycles, namely,
The study aimed to construct an HEV with optimization in real-time NEDC, Federal Urban Driving Schedule, and Federal Highway Driving
operation. Thus, two constraints have to be considered, namely, the Schedule. After implementing a supervisory equivalent fuel consump-
lack or very limited knowledge of future driving conditions and the tion minimization strategy (ECMS) controller, the results were com-
sustainability of reversible energy source charge. Optimization is pared with the control variable. Less hydrogen was used and more
merely about cost function, fuel energy use, and variations of stored

N. Sulaiman et al. / Renewable and Sustainable Energy Reviews 52 (2015) 802–814 807

power was fed to fuel cell system when the applying supervisory discussed in the next chapter. Studies on the EMS for fuel cell–SC
ECMS controller. However, the limitation was that the parallel supply HEVs are shown in Table 2.
of reactant gases caused the system to be unstable when power was
above 30 kW. Thus, the power was limited to 27 kW to control the 2.3. Fuel cell–battery–SC HEV
acceleration rate of the vehicle. Other limitations were constraints in
the transient behavior and that the low SC efficiency at low voltage. HEVs with a secondary energy source, such as battery and SC,
Thus, the voltage was maintained at 180 V. are designed to provide power for short durations of peak power
demand, such as during driving uphill or acceleration. Good
Another report on fuel cell–SC EMS was made by Ates et al. [41]. acceleration requires both high power density and high energy
The EMS applied wavelet-Adaptive Linear Neuron (ADALINE), one of density [10]. Thus, having both battery and SC is essential because
the artificial neural network methods. The wavelet-ADALINE network the SC has high power density and low energy density, whereas
utilizes multi-level Haar wavelet transform wherein the simulation the battery has low power density and high energy density,
was performed using SimPowerSystems Toolbox in Matlab/Simulink respectively [16]. In this system, the energy is shared among the
for UDDS. In the study, the UDDS power demand was decomposed fuel cell, battery, and/or SC with the energy drawn from one, two,
into high-frequency and low-frequency components. The first half or even three energy sources. The drawn energy is then fed into
was used as perception, whereas the other half was used for testing in the DC bus and into the inverter, where the energy is converted
the ADALINE controller. Although the results of the study showed the from DC voltage to AC voltage. The AC voltage will then be able to
fulfillment of power demand, this work was only limited to a move the AC motor. Research on fuel cell–battery–SC HEVs is in
simulation. Thus, real-time application is required to prove vehicle progress, and the EMS is discussed in this chapter.
performance. Moreover, the results did not show the vehicle speed,
but showed the hydrogen consumption instead. Zheng et al. [42] also Hannan et al.[11], in his report, presented an EMS for light
reported a study on a fuel cell–SC HEV based on Adaptive Optimal electric vehicles, which also integrated the fuel cell, battery, and
Control Algorithm (AOCA) EMS. The control strategy embeds neural SC, respectively. The EMS utilized a PI regulator, which classifies
network computation for the formulation of lead and lag judgment the vehicular operation into 7 logic states based on 3 input
network, and was tested on fuel cell-lithium ion capacitor Hybrid conditions of pedal offset (PO), power duration load (PD), and
Power System (FHPS). Results showed that supercapacitor reduced the battery capacity (BC), respectively. The results of the vehicle speed
hydrogen consumption by 18.3% compared to using fuel cell only. were compared with that of the battery source only, multiple
However, the FHPS were not extensively described and simulation sources (fuel cell–battery–SC), and ECE-47 test drive cycle; the
tool of the EMS model was not enough evident. HEV displayed 94% efficiency compared with a battery-only
system, which demonstrated 84.9% efficiency, but only when the
However, based on review, it is said that SC is not a source of vehicle drove uphill. However, the study was limited to simulation
high energy density. Thus, some studies have combined fuel cell,
battery, and SC as energy sources for HEV. This combination is

Table 2
Summary of EMS for fuel cell-supercapacitor HEV.

Author EMS method/control strategy Simulation/hardware DC SOC Max Max Advantage Disadvantage
bus speed power

Thountong Comparison PI controller and Simulation: Matlab/ 42 V – – 720 W DC bus voltage of 42 V was of good Only showed simulation
et al. [20] differential flatness control Simulink 47 V – 150 rad/s 1 kW
Hardware: dSPACE 188 V – 56.7 mph 60 kW convergence from flatness control results, simulation results
DS1103
Simulation: Matlab/ did not consider speed
Simulink (NEDC drive
cycle) performance
Hardware:
Tani et al. Bidirectional load power PICI8F4431 Results showed and Semiconductor and
[37] sharing (polynomial control Simulation: Matlab/
technique) Simulink experimentally verified that wiring losses ignored,
SimPowerSystem,
ADVISOR (UDDS ultracapacitors provide the fluctuation from sensors,
drive cycle)
Hardware: VIA EPIA- dynamic components or transient reduced experimental
P700-10 Pico-ITX main
board current scale

Uzunoglu Wavelet-based load sharing Simulation: VP-SIM Results show how reliable an UC in Simulation only
and Alam algorithm (based on Matlab/
[38] Simulink) (NEDC, providing rapid switching of
FUDS, FHDS drive
cycles) positive and negative terminal
Hardware: Hy. Power
experimental vehicle voltage
Simulation: Matlab/
Efstathiou Two fuzzy logic controllers to Simulink – – 10 m/s – Continuous variable transmission Study was focused on
SimPowerSystem
et al. [39] gear box system (constant (UDDS drive cycle) showed less 5% usage of hydrogen gear box system
Simulation: unknown
gear ratio and continuous tool than constant gear performance

variable transmission)

Rodatz Supervisory ECMS controller 340 V 0.25– 120 km/ 60 kW ECMS showed 2% to 6.5% Fuel cell voltage unstable
above 30 kW, efficiency
et al. [40] with ECMS algorithm 0.8 h improvement in fuel efficiency of FC is low at low voltage

Ates et al. Wavelet ADALINE method 188 V 0.65– – 60 kW DC load voltage was maintained at Simulation only
[41] (ANN based method) 0.75
188 V, UC supplies charging and
Zheng et al. Adaptive optimal control – 0.35– –
[42] algorithm based on neural 0.95 discharging power demand
network
fluctuations successfully

100 kW Results showed that the sedan Simulation tool was not

using FC-UC hybrid source clear

consumes less 18.3% to 33.7%

hydrogen than FC alone

808 N. Sulaiman et al. / Renewable and Sustainable Energy Reviews 52 (2015) 802–814

only, and the charging and discharging performances of the that based on the simulated results, the vehicle could reach almost
battery and SC were not evident. In addition, the study was tested 129 km per hour.
only at a short distance, and a short duration of 120 s.
Meanwhile, Paladini et al. [47] reported a study on a HEV
Garcia et al. [43] reported on control strategies for high-power powered by a fuel cell–battery–SC. The vehicle in Paladini's study,
HEVs and extensively explained the control strategies for fuel cell– which embedded a PEMFC, nickel metal hydride battery, and
battery–SC HEVs. In the report, most common control strategies were Maxwell SC, was targeted to achieve an optimized HEV. The
compared and elaborated. These strategies include operation mode energy management to switch between the fuel cell, battery, or
control, cascade control, fuzzy logic control, equivalent consumption SC was simply using the ECoS code and traction control strategy,
minimization strategy control, and predictive control, respectively. which was simulated using the Matlab/Simulink software. The
The control strategies were analyzed by comparing their perfor- optimization of the system, which applied the Pareto front, was
mances on a 400 kW rated powertrain. The report showed that the tested on four driving cycles, namely, the NEDC, UDDS, High Way
equivalent consumption minimization strategy control demonstrates Fuel Economy Test, and Japanese driving schedule (10–15). Based
the lowest hydrogen consumption and the simplest control strategy. on the results reported, the system was evidently optimum on the
Meanwhile, fuzzy logic and predictive control were among the most NEDC, where there was only a 6.75 g/km hydrogen fuel consump-
complex control strategies. The study by Garcia provided a conclusive tion. However, the report focused more on the comparison of the
contribution as regards the comparison and analysis of common driving cycle testing and did not elaborate on the EMS.
energy system control strategies for a fuel cell–battery–SC HEV.
However, this study was only comparative in nature. Another study on the fuel cell–battery–SC HEV was reported by
Zandi et al. [48], where an EMS based on flatness control
Meanwhile, Martinez et al. [44] reported another study on the technique (FCT) and fuzzy logic control (FLC) was introduced.
fuel cell–battery–SC electric vehicle EMS. The strategy utilizes The concept was that the FCT would control the energy source
fuzzy logic to control the energy, where the slow dynamics of the between the fuel cell and the energy storage system (ESS), which
fuel cell and the SOC of the SC are considered. The maximal control comprises a battery and a SC that has two control variables, the
structure and practical control structure of the energetic macro- power of converter and the power of ESS. The FLC was used to
scopic representation testing of the performance of this system control the power share between the battery and the SC. More-
was done via an electrical chain component evaluation vehicle over, the FLC operates in three modes (normal, overload, and
testbed. However, in Martinez's [44], a battery is used as the main recovery mode) and has three input variables, the SC voltage,
source of energy, while the fuel cell and SC are additional sources. battery voltage, and output load power; however, only one output
The strategy is that when the SOC of the battery is greater than the variable, which is the percentage of power, was shared between
reference value, the fuel cell and SC will feed less energy, and vice the battery and the SC. The results of the control algorithm, as
versa. However, the fuzzy logic strategy was changed into the tested through the dSPACE hardware, showed that the control
type-2 fuzzy logic approach because of the limitations of the type- strategy was able to perform in different operating modes, but the
1 that was used in this study. load power observed was limited to 780W.

Another study by Schaltz and Rasmussen [45] comparing 10 cases Furthermore, Li et al. [49] reported a study that compared a
of the fuel cell, battery, and/or SC connections. In the 10 cases, the fuel cell–battery (FCþ B) and a fuel cell–battery–SC (FCþ Bþ SC)
energy source of the connections in Cases 1–6 was from the fuel cell EMS for HEV based on fuzzy logic. On the one hand, the FC þB
and either the battery or the SC. The utilization of all three energy system had 4 driving modes, namely: (1) starting, (2) fuel cell
sources was only evident in Cases 7–10. Considering that the fuel cell driving and battery charging, (3) fuel cell and battery combined
is the primary energy source in these 10 cases, the energy manage- driving, and (4) regenerative braking. On the other hand, the
ment strategy switches the input energy between the battery and the FC þB þSC system had 5 driving modes to cater to the fuel cell,
SC. When both battery and SC are used, the battery feeds energy, battery, and SC combined driving mode. Fuzzy logic, which used
which is determined by a low-pass filter, but with a higher bandwidth IF–THEN rules, was embedded in both hybrid configurations,
than that of the fuel cell. The short-term peak power during where the fuzzy reasoning rules and defuzzification were per-
acceleration and braking was fed into the SCs. Based on the results formed by using the Mamdani Inference Method and the centroid
of the comparison done on the volume, mass, efficiency, and days of method, respectively. Individual Simulink models of the fuel cell,
operation to the rated FC power for all the 10 cases, Cases 7–10 battery, and SC were modeled in the Simulink and tested in
evidently demonstrated lower volumes, lower masses, higher effi- ADVISOR, while considering 4 standard cycle conditions: (1) UDDS,
ciencies, and higher operating days to the rated FC power. These (2) Highway Fuel Economy Test, (3) Urban Schedule 06, and
results prove that the fuel cell, battery, and SC connections would (4) Economic Commission for Europe and Extra Urban Driving
provide not only a lower volume and mass, and a higher efficiency, Cycle. The performances of both configurations were competi-
but these connections will also provide a longer service life for the tively close, but the FC þ B þSC showed a better acceleration rate
battery. However, the current study was merely based on the com- and lower hydrogen consumption. The study, however, was only a
parison of the performances of the 10 connections of the HEV only. simulation, thus, more evidence of real-time application is needed.

In a report by Yu et al. [46], a fuel cell–battery–SC HEV power Meanwhile, Matopan et al. [50] reported on a research about an
allocation strategy was proposed through simulations done in the EMS for a fuel cell–battery–SC HEV, which compared five different
Matlab/Simulink software. Interestingly, the power control was control strategies: (1) state machine control strategy, (2) rule-based
separated into two, the fuel cell–battery and the battery-SC. The fuzzy logic strategy, (3) classical PI control strategy, (4) frequency
blocks were simulated in the Matlab/Simulink software and decoupling and fuzzy logic strategy, and (5) equivalent consump-
integrated with a certain algorithm, which allows the system to tion minimization strategy (ECMS). The simulations were per-
draw energy from the fuel cell, battery, or SC based on the power formed via Matlab/Simulink using the SimPowerSystems toolbox,
demand and the SOC condition of the battery and SC. The SOC of and the simulations were also tested in real-time through LabVIEW
both the battery and SC was set, thus, that of the battery was on NI-PXI 8108, which illustrated a DC bus voltage of 280 V, as rated
between 40% and 90%, whereas that of the SC was within 25–100% [50]. The research for a 15 kW fuel cell–battery–SC HEV generated
[46]. However, the current study was only concerned on sizing and results for all the strategies, focusing on the hydrogen consumption
monitoring urban and highway driving, both of which were flat and stress analysis of each strategy; however, there was no evident
roads. Thus, driving uphill was not considered. The report shows analysis on speed and load power fulfillment.

N. Sulaiman et al. / Renewable and Sustainable Energy Reviews 52 (2015) 802–814 809

In a report by Liu et al. [51], a fuel cell–battery–SC HEV Liu [12] was at simulation level, but the simulation tool was not
optimization was based on the configuration of the series and mentioned. Li et al. [49] and Xie et al. [15] who were studied on EMS
parallel stack of fuel cell and battery using rule-based energy system. adopting fuzzy logic with a maximum speed of 160 km/h, and a fuzzy
Despite it mentioning the system includes SC, there were not any supervised rule with 80 km/h, but only at simulation level. However,
descriptions on SC in the report. Moreover, it focused on optimization Efstathiou et al. [39], Martinez et al. [44], Zandi et al. [48] and Xiao
only hence the EMS was not elaborated. The research works et al. [35] tested their EMS on hardware, but there is still room for
conducted on fuel cell–battery–SC HEV are shown in Table 3. improvement because Xiao's [35] hardware results were not visible.
Zandi's [48] EMS was tested for 100 s only, Martinez's [44] results
3. EMS control strategies generated a very low result of only 20 km/h real-time drive, and
Efstathiou's [39] study was only done to monitor the hydrogen
Based on the previous research, quite a number of control consumption in different gearboxes, respectively.
strategies have been studied and even simulated and analyzed. Some
of these strategies were mentioned in previous research or have even Fuzzy logic EMS are based on IF–THEN rules and membership
been tested as being embedded in a hardware or real vehicle functions, which define certain situations and conditions that would
application. The control strategies are discussed in this chapter enable either one or two energy sources between the main and the
starting with fuzzy logic where most of the EMS adopted simulation auxiliary energy sources. Membership functions would resemble the
based on fuzzy logic controllers. As an example, the study by Li and difference between the target and the measured value in percentage,
which is usually labeled as Negative High, Negative Low, Zero, Positive
Low, and Positive High [52]. Designing fuzzy logic requires different

Table 3
Summary table of EMS for fuel cell–battery–SC HEV.

Author EMS method/control strategy Simulation/ DC SOC SOC Max Max Advantage Disadvantage
hardware bus (B) (SC) speed power

Hannan Operational mode control (7 Simulation: 120 V – – 47 km/h o 3 kW Better performance is seen in multi- Simulation only,
et al. [11] operational states) Matlab/Simulink 50 km/h
(ECE-47 test drive 750 V 0.65 0.65– source system compared with BEV short duration
Garcia et al. Comparison of 5 control cycle) 0.85 20 km/h
[43] strategies (operational mode, Simulation: 560 V 0.9 – especially when accelerating test
cascade control, fuzzy logic, Matlab/Simulink 0.6– 80 mph
Martinez equivalent consumption SimPowerSystem 42 V 0.6 0.97 (129 km/ 400 kW Comparison results showed ECMS Focused on fuzzy
et al. [44] minimization strategy, Hardware: Urbos 3 0.4– h)
predictive control) (urban street 300 V 0.4– 0.95 – 40 kW giving lowest hydrogen mass logic control only,
Schaltz and Type-1 fuzzy logic controller railway) 1.0 0.1– 6 kW
Rasmus- Hardware: ECCE 1.0 – 100 kW consumption of 3.82 kg generated
sen [45] Comparison of 10 configuration vehicle (electrical 48 V 0.6– 0.21– 60 kW
of fuel cell, battery, and/or chain component 0.81 0.76 160 km/ 780 W 400 kW but
Yu et al. supercapacitor evaluation) h
[46] Simulation not 42– – – 40 kW demand power
Comparison of optimal control clear 48 V –
Paladini strategies with thermostatic – showed 500 kW
et al. [47] control strategy Simulation: – 0.4– 120 km/
Matlab/Simulink 0.8 – h Energetic Macroscopic Simulation only
Zandi et al. Traction control strategy using (UDDS, US06 drive
[48] simulation code ECoS cycles) 280 V 0.5– – Representation of the vehicle is even though
Simulation: 0.7
Li et al. [49] Flatness control technique and Matlab/Simulink applied in this research mentioned on
fuzzy logic control (NEDC, UDDS, 300 V –
Matopan HWFET, 1015 drive hardware
et al. [50] Comparison of fuzzy logic in fuel cycles)
cell–battery (FC-B) and fuel cell– Simulation: Simulation of 10 configurations Simulation tool
Liu et al. battery-supercapacitor (FC-B-SC) Matlab/Simulink
[51] HEV Hardware: dSPACE showed that FC-B-SC system reduces was not clear
and PC test bench
Comparison of control strategies mass and volume and extends
(state machine control, rule- Simulation:
based fuzzy logic, classical PI, Matlab/Simulink, lifetime of system
frequency de-coupling þ fuzzy ADVISOR (UDDS,
logic, ECMS) HWFET, US06, Energy cost compared between Simulation only,
ECEþ EUDC drive
Simple-rule based strategy cycles) optimal and thermostatic control gradient/slope
Simulation:
Matlab/Simulink shows lower cost in optimal control not considered
SimPowerSystem
Hardware: NI- strategy
PXI8108
Fuel consumption for all four drive Simulation only,
Simulation:
Matlab/Simulink cycles stating lowest at 1015 drive focused on fuel
Pro@design
cycle with 37 g, battery showed less consumption and

percentage of usage for all cycles storage

enabling battery longer lifetime utilization only

SC and battery has power sharing Tested for 100 s

coefficients carried out by FLC only, cable

inductance and

capacitor serial

resistance

neglected

Results showed better performance Simulation only

of FCþB þ UC with 58.7 m distance in

5s and 16.9 s to reach 400 m

o 12 kW Classical PI control shows lowest Focused on
58 kW hydrogen consumption of 235 g and hydrogen
second lowest battery stress (22), consumption and
state machine control giving highest stress analysis
overall efficiency of 80.72% and only
lowest battery stress (21.91)
Determine the optimum size or Focus on sizing
parameter for the vehicle only

810 N. Sulaiman et al. / Renewable and Sustainable Energy Reviews 52 (2015) 802–814

Fig. 3. Fuzzy logic EMS (a) block diagram, (b) membership functions, and (c) fuzzy logic rules [50].

processes of fuzzification, rules (IF–THEN), and defuzzification [53]. An determined. The specification of the operational mode differs from
example of the block diagram, membership functions, and logic rules one designer to another, thus, the output power energy sources would
of the fuzzy logic EMS are shown in Fig. 3(a)–(c), respectively [49]. totally depend on the operational mode of specification [44]. The
Fuzzy logic EMS is evidently reliable and capable of increasing report by Azidin and Hannan proved the effectiveness of the EMS
efficiency based on the reduced usage of fuel when using fuzzy logic embedding the 7 operational states as simulated in Matlab/Simulink
controllers [39,52]. [54]. The example of a block diagram and operational states of the
operational mode controller is shown in Fig. 4. An EMS with PI
Rule-based fuzzy logic control strategies have been reported by controllers were embedded in a study conducted by Thounthong et al.
Shuang et al. [27] and Liu et al. [51], at maximum speeds of 60 km/h [20] to compare between the PI controller and the differential flatness
and 120 km/h operating condition. However, both of the EMS were control. They proved that the differential flatness control showed a
only simulated and not tested even in real-time driving cycle. Some of better DC voltage regulation compared to the PI controller. Meanwhile,
the EMS method or control strategies have conducted in operational a report by Jiang et al. [18] mentioned that the EMS being controlled
mode or state machine control as explained in Azidin et al. [13], Xiao via PI controller was reliable and feasible with an efficiency of about
et al. [14], and Benrabeh et al. [34], respectively. However, the study 57%, as shown in the simulation results.
was only simulated; thus, hardware or real-time testing is needed to
verify the validity of the simulation results. Xiao and Wang [14] and The optimal control strategy or equivalent consumption minimiza-
Benrabeh et al. [34] both reported on 4 operational modes with tion strategy (ECMS) is another type of EMS, which focuses on
maximum speeds of 88.5 km/h and 120 km/h, respectively, whereas calculating the optimal FC power to minimize the hydrogen con-
Azidin et al.[13] reported on an EMS with 7 operational states or sumption of the HEV [42]. ECMS usually embed formulations of
modes. The operational mode controller determines the fuel cell mathematical equations into the system. Sundstrom and Stefanopou-
reference power, of which the operation mode is then defined based lou [29], Yu et al. [46], Tritschler et al. [55], Zheng et al. [56] have
on the input parameters available. studied and reported the simulated results of the load sharing
between energy sources. Tritschler et al.[55] reported on a comparison
As an example, based on the demanded load power and battery or of the EMS rule-based strategy, gliding average strategy, and ECMS
SC SOC, the operational mode is defined and the reference power is

N. Sulaiman et al. / Renewable and Sustainable Energy Reviews 52 (2015) 802–814 811

SOCbat Operation Mode Control Pfc(opt)
Pm (OMC) Pbat(ideal)
Paux
Pbat,discharge
SOCsc,ref - PI Psc,discharge + Pbat(opt)
+ Psc,charge
Pbat,charge
SOCsc

State Super capacitor Fuel Cell Battery Condition
1 0 0 0 Off operation / safety features
2 0 0 1 BC is High, PD is Low, and PO is Low
3 0 1 0 BC is Low, PD is Low, and PO is Low
4 0 1 1 BC is High, PD is High, and PO is Low
- 1 0 0 - (Not possible)
5 1 0 1 BC is High, PD is Low, and PO is High
6 1 1 0 BC is Low, PD is Low, and PO is High
7 1 1 1 BC is High, PD is High, and PO is High

Fig. 4. Operational mode controller EMS (a) block diagram [43] and (b) operational states [11].

simulated in Matlab/Simulink, and proved that the ECMS showed the Pload Pfc* Ifc*
best performance in fuel consumption reduction and fuel cell peak SOC X
power demand reduction. Ouddah et al. [28] even compared the ECMS /
optimal control strategy, which used the Pontryagin's minimum pri-
nciple to power frequency splitting, and the results proved that the η Vfc
optimal control strategy used less hydrogen, which was better Fig. 5. Block diagram for ECMS EMS [50].
compared to power frequency splitting. Zheng et al. [42] proved that
the EMS of fuel cell–SC HEV embedding optimal control strategy, 4.1. Hydrogen viability
which was reported as the Adaptive Optimal Control Algorithm,
consumed hydrogen as high as 89.9% when compared with a fuel Hydrogen, being popular due to its high energy content by weight,
cell electric vehicle, and the system efficiently distributes the power has its' own limitations or issues that have to be considered and
load and captures the regenerative electricity during rapid speed discussed prior to development of fuel cell hybrid electric vehicles.
fluctuations. Fig. 5 shows an example of the block diagram for the Among the biggest considerations is the complex infrastructure which
ECMS EMS. includes hydrogen production, hydrogen storage, hydrogen transpor-
tation, environmental impacts of hydrogen, and hydrogen infrastruc-
Furthermore, Uzunoglu and Alam [38] and Ates et al. [41] reported ture cost [57]. Firstly, regarding hydrogen production, it is currently
on the wavelet-based frequency decoupling strategy load sharing being produced via steam reforming of natural gas because it is the
algorithm, which displays simulation results that shows the reliability cheapest method of about $1.50/kg for mass production of hydrogen.
of the EMS, but were not extended to a real-time application. Other Where else, hydrogen can be produced by more environmental
EMS include frequency separation method as reported by Alloui et al. friendly methods such as electrolysis of water using electricity from
[32], polynomial control as studied by Tani et al. [37], anticipatory solar, hydropower and wind technology or using photochemical [58].
algorithm by Kelouwani et al. [33], and supervisory ECMS as discussed It has been reported that capital cost of PEM electrolysis has been
by Rodatz et al. [40]. Among these EMS, only Rodatz and Tani reduced from $385/kW in 2007 to $171/kW in 2012 (for 1500 kg/day
extended their research to real-time application or hardware testing. hydrogen production) [59]. This shows that with growth in demand,
Rodatz showed the results of the reliability of the ECMS algorithm in the fuel cell vehicle infrastructure would reduce in cost. As per
the EMS with a maximum speed of 120 km/h [40]. hydrogen storage, it has been reported that the gravimetric and
volumetric capacities have increased 50% since 2007 [60]. Storing
4. Issues and challenges on FC–HEV hydrogen is another issue since it must be liquefied or compressed to
be safely and conveniently loaded or installed in a vehicle [61].
Developing a HEV, especially when more than two energy sources However, General Motor has proven that a fuel cell vehicle is safe
are involved, needs more considerations. There are several main as its 100 Chevrolet Equinox Fuel Cell has been in the market since
components of the HEV, which affect the reliability and efficiency of 2008 [23].
the EMS. However, several issues need to be addressed to overcome
the challenges on the reliability and efficiency of the HEV EMS. The Based on report on hydrogen production cost if using electrolysis,
issues and challenges with these components are discussed in the the cost would have to consider a system with water reactant delivery
following sub-sections and divided into the following categories. management system, oxygen gas management system, hydrogen gas
management system, and electrolyzer stacks. Apart from that, the cost

812 N. Sulaiman et al. / Renewable and Sustainable Energy Reviews 52 (2015) 802–814

would also consist of capital cost, operating cost and replacement cost. 4.4. Power electronic interface
The costing has been reported to provide guidance and information
for researchers or public for the development of hydrogen fuel The energy management, with three different energy sources,
infrastructure [59]. Considering hydrogen delivered by pipeline or the fuel cell produces low DC voltage by adjusting the fuel cell to a
tanker trucks to the hydrogen refueling stations, or in more general; level with the DC bus voltage by a suitable DC–DC converter [69].
the infrastructure, is a possible transformation with some modifica- Hence, power electronic interface is needed in a FC–HEV. However,
tions to the current infrastructure. There have been reports providing considerations have to be made prior to its development such as
guidance and information on the requirements of hydrogen vehicle open circuit faults draw increments of fuel cell stack ripple
infrastructure [62,63]. The infrastructure of hydrogen fuel stations for currents and add more stress to inductors. The possible malfunc-
fuel cell vehicles are expected to improve in future due to the fact that tions include driver failure, incorrect gate voltage, damage of
in 2009, seven of the top automobile manufacturers (General Motors, device, high voltage, or current and transients [17]. The fuel cell
Honda, Hyundai-Kia, Renault-Nissan, Toyota, Daimler, Ford) signed a system needs to boost the DC–DC converter and the storage device
joint letter of understanding. Among of the content was to urge the needs the bidirectional DC–DC converter [20]. By contrast, if the
setup of hydrogen infrastructure by 2015 especially for Europe, battery rating voltage does not tally with the DC bus voltage, a DC–
Germany, and then extending it to US, Japan and Korea. The hydrogen DC converter is still needed to interface the low-voltage batteries
infrastructure has already been in place since it has been reported that with the higher-voltage fuel cell-powered DC bus [70]. DC–DC
there were 227 hydrogen stations installed in 2011 [24]. converters for fuel cells should be able to buck or boost voltages in
one direction only because there is no charging of fuel cells.
4.2. Fuel cell economic impact Meanwhile, DC–DC converters for the battery and SC should be
able to buck and boost voltages for the charging and discharging
Despite the environmental advantage of fuel cell hybrid electric mode. Hence, buck–boost converters are usually used for the
vehicles, having almost zero-emission of greenhouse gasses [5], the battery and SC [44].
application of fuel cells in vehicles can still be improved, especially on
its efficiency, or technologies and policies related to costs. The well-to- 4.5. Motor integration
wheels (WTW) efficiency of a fuel cell vehicle has been reported as
only 14.2% [64]. In another report mentioned the WTW efficiency, FC– The motor would be the actuator for the conversion of electrical
PHEV for an example, considering grid electricity (53%), electrolysis energy into kinetic energy. The motor can either be an AC or a DC
(72%), fuel cell (55%), battery charging and discharging (96%), power motor depending on the input power fed into the machine. Among
control circuit and motor (80%), producing overall efficiency of only the common motors used for electric vehicles are induction motors,
14% [65]. Numerous research works have being conducted on permanent magnet brushless motors, and switched reluctance mot-
improvement of fuel cell efficiency for vehicles. Amongst them are ors. Switched reluctance motors have a simple and rugged cons-
studies on improving electrode efficiency by controlling Nafion truction, simple control, and high speed. However, this motor type is
ionomer volume in PEMFC and using multi-metallic anode catalysts noisy and has lower efficiency compared to permanent magnet
in PEM electrolyzer without affecting the performance or even motors. Induction motors are reliable because of the ability to operate
improving its efficiency [66,67]. in hostile environments with a simple construction and low main-
tenance and cost, but this motor type also has a lower efficiency
Cost problems with the fuel cell stack can be solved by compare to permanent magnet motors. Permanent magnet brushless
increasing the power density, reducing dependence on the plati- DC motors efficiently dissipate heat to its surrounding areas and
num loading in electrodes, and increasing the production volume. unlikely to suffer manufacturing defects or overheating, thus, this
The fuel cells commonly used for fuel cell vehicles are PEMFCs motor type has a higher efficiency and is mostly used in electric
because these fuel cells are efficient, compact, and have a quick vehicles [71]. However, permanent magnet brushless DC motors are
start. It is expected that the cost of fuel cell vehicles itself will expensive, and the mechanical strength does not encourage large
decrease, because in 2011, ten Japan oil and energy companies torques in motors [72]. Another type of motor used is the direct
signed a memorandum of understanding in the effort of reducing current brushless motor, but this motor type requires the driver to
FCEV manufacturing cost [24]. The price of transportation fuel cell change the phase of the electric converter output energy via an
has shown decline since 2002 with the price of $275/kW to $47/ inverter [73].
kW as in 2012 [68]. This comparison of 10 year duration shows a
reduction of almost 83%. It is mostly possible that the aim of 4.6. Energy management system
Department of Energy to lower the fuel cell cost to $30/kW as
achievable. In a vehicle energy system with a hybrid or a combination of
multiple energy sources, managing the power flow and meeting the
4.3. Battery lifetime high expectations of the market are crucial. The challenge is on the
configuration and controller design because of the complexity and
As mentioned previously, there are many types of batteries in the difficulty resulting from the integration requirement among the other
market, however, the most commonly used in electric vehicles would existing systems in the vehicle [74]. Power distribution among various
be the NiMH and Li-ion batteries [7]. On one hand, NiMH batteries are components becomes more challenging when the EMsS are expected
considered environmentally safe, low maintenance, has a high power to minimize the hydrogen consumption and increase the life expec-
and energy density, and considerably cheap. On the other hand, Li-ion tancy of the fuel cell, which pertains to the optimization of design.
batteries, are light, compact, have more energy storage than the Researchers have been expected to consider the limiting SOC, support
NiMH, but have a wide range of operational temperatures and there the fuel cell during high loads, and save the optimum energy during
are concerns regarding the availability of materials. The use of regenerative braking [75]. Power splits have become the main concern
batteries in fuel cell HEVs is necessary because the batteries provide also because of problems such as the frequent charge from the fuel
the energy storage for regenerative braking [10]. However, the service cell to the battery and fluctuations of the fuel cell output power, which
life of the battery, which depends on the SOC and depth of discharge, will affect the performance and service life of the battery and fuel cell
shall be thoroughly analyzed prior to application. There should be a [76]. There are continuous studies on energy management, some of
good compromise between the cost and the technical performance them are simple, but lacking in adaptability, while the others are able
when selecting batteries [18].

N. Sulaiman et al. / Renewable and Sustainable Energy Reviews 52 (2015) 802–814 813

to solve the optimal control, but not yet proven to be optimizing for [2] Kano Y, Inoue Y, Sanada M. Current specifications of vehicle motors. In:
any arbitrary drive cycles [77]. Furthermore, most of the previous Proceeedings of the 2013 IEEE ECCE Asia downunder (ECCE Asia); 2013. p.
research works were limited to computer simulation only [78]. Thus, 136–40.
the EMS still needs to be further explored and developed in terms of
hardware or real-time applications to ensure the reliability of a fuel [3] Kelly KJ, Mihalic M, Zolot M. Battery usage and thermal performance of the
cell HEV. toyota prius and honda insight during chassis dynamometer testing. In:
Proceedings of the 17th annual battery conference on applications and
5. Conclusion and suggestions advances; 2002. p. 247–52.

There are so many things that to be considered in developing a fuel [4] Ahsan K. Understanding trends of car recalls. In:Proceedings of the 2010 IEEE
cell hybrid electric vehicle. The HEV consists of traction or motor, international conference on industrial engineering and engineering manage-
motor control, inverter and dc converters, energy management system ment; 2010. p. 1123–7.
and not forgetting fuel or energy sources. In this paper, fuel cell,
battery and supercapacitor were among the energy sources consid- [5] Erjavec J, Arias J. Hybrid, electric and fuel-cell vehicles. United States of
ered. Each part or subsystem of the HEV has to be properly analyzed America: Thomson Delmar Learning; 2007.
and integrated in order to produce a reliably efficient vehicle. Apart
from the fuel cell vehicle itself, not to forget the cost and infrastructure [6] Ehsani M, Gao Y, Emadi A. Modern electric, hybrid electric and fuel cell
that should come along with the production of the vehicles, consider- vehicles: fundamentals, theory and design. 2nd edition. United States of
ing use of hydrogen fuel: hydrogen production, storage, delivery, and America: CRC Press; 2010.
refueling stations. It has been discussed though that the transforma-
tion to hydrogen fuel is possible with the modification of current [7] Pollet BG, Staffell I, Shang JL. Current status of hybrid, battery and fuel cell
infrastructure. Previously reported was also the statistics showing the electric vehicles: from electrochemistry to market prospects. Electrochem
reduction in fuel cell vehicle infrastructure. Despite that, collaboration Acta 2012;84:235–49.
or joint letter of understanding was also evident in the effort of
making hydrogen infrastructure most possibly soon. [8] Ali DM, Salman SK. A comprehensive review of the fuel cells technology and
hydrogen economy. In: Proceedings of the 41st international universities
Focusing on the objective of this paper, numerous studies have power engineering conference; 2006.
been discussed on the reports of the EMS for fuel cell HEV. However,
most of the studies were conducted only through simulation and [9] Barbir F. PEM fuel cells: theory and practice. United States of America:
there are insufficient research involving real time applications or real Academic Press; 2012.
life testing of the EMS. In particular, EMS such as fuzzy logic or
operational state strategy, which is actually simple rule-based EMS, is [10] Livint G, Horga V, Ratoi M, Albu M. Control of hybrid electrical vehicles.
simple, but not impressive when tested on hardware. This observation electric vehicles – modelling and simulation. Croatia: InTech; 2011.
is not surprising when looking into all the considerations during the
development of a HEV, especially when involving a fuel cell, battery, [11] Hannan MA, Azidin FA, Mohamed A. Multi-sources model and control
and SC, aside from the fuel cell size and efficiency, hydrogen con- algorithm of an energy management system for light electric vehicles. Energy
sumption and storage, battery and capacitor SOC, DC–DC converters Convers Manag 2012;62:123–30.
efficiency, inverter harmonics, and motor rotational losses. Further-
more, the EMS should be reliable and feasible, and based on the [12] Li C-Y, Liu G-P. Optimal fuzzy power control and management of fuel cell–
reviews made; the following are suggestions for the control strategy battery hybrid vehicles. J Power Sources 2009;192:525–33.
for the EMS in HEV:
[13] Azidin FA, Hannan MA, Mohamed A. Renewable energy technologies and
i) Embeds the optimal control strategy or the equivalent con- hybrid electric vehicle challenges. Prz Elektrotech 2013:100–6.
sumption minimization strategy because it resembles the most
outstanding strategy [14] Xiao D, Wang Q. The research of energy management strategy for fuel cell
hybrid vehicle. In: Proceedings of the international conference on industrial
ii) Has the ability to minimize hydrogen consumption without control and electronics engineering; 2012. p. 931–34.
affecting the power load distribution and fulfillment
[15] Xie C-J, Quan S-H, Chen Q-H. Control strategy of hybrid power system for fuel
iii) Tested on hardware or in a real-time application, with a cell electric vehicle based on neural network optimization. In: Proceedings of
deliberately acceptable speed range. the IEEE international conference on automation and logistics; 2008. p. 753–7.

The suggestions made would be such interesting and remark- [16] Hannan MA, Azidin FA, Mohamed A. Hybrid electric vehicles and their
able contributions to embed into the equivalent consumption challenges: a review. Renew Sustain Energy Rev 2014;29:135–50.
minimization strategy EMS in all fuel cell HEVs.
[17] Guilbert D Mohammadi A, Gaillard A, N'Diaye A, Djerdir A. Interactions
Acknowledgment between fuel cell and DC/DC converter for fuel cell electric vehicle applica-
tions: influence of faults. In: Proceedings of the 39th Annual Conference of the
This work was supported by the Grant LRGS/2013-UKM/TP/01 IEEE Industrial Electronics Society (IECON 2013); 2013. p. 912–7.
under Universiti Kebangsaan Malaysia.
[18] Jiang Z-L, Chen WR, Qu ZJ, Dai CH, Chen ZL. Energy management for a fuel cell
References hybrid vehicle. In: Proceedings of the power and energy engineering con-
ference; 2010. p. 1–6.
[1] Burress T, Campbell S. Benchmarking EV and HEV power electronics and
electric machines. In: Proceedings of the 2013 IEEE Transportation Electrifica- [19] El-Monem AAA, Azmy AM, Mahmoud SA. Performance analysis of polymer
tion Conference and Expo (ITEC); 2013. p. 1–6. electrolyte membrane fuel cells for electric vehicle applications. In: Proceed-
ings of the 2013 IEEE grenoble powertech (POWERTECH); 2013. p. 1–6.

[20] Thounthong P, Pierfedereci S, Martin J-P, Hinaje M, Davat B. Modelling and
control of fuel cell-supercapacitor hybrid source based on differential flatness
control. IEEE Trans Veh Technol 2010;59(6):2700–10.

[21] He H, Zhang Y, Wan F. Control strategies design for a fuel cell hybrid electric
vehicle. In: Proceedings of the IEEE vehicle power and propulsion conference
(VPPC); 2008.

[22] Emergency response guide: chevrolet equinox fuel cell. GM Service Technical
College. General Motors Corporation; 2007.

[23] Berman B. 〈http://www.hybridcars.com/chevrolet-equinox-fuel-cell/〉; 2009.
[24] Fuel cell electric vehicles: the road ahead. Fuel Cell Today; 2013.
[25] Sakka MA, Mierlo JV, Gualos H. DC/DC converters for electric vehicles. Electric

vehicles – modelling and simulations; 2011.
[26] Larminie J, Dicks A. Delivering fuel cell power. In: Fuel cell systems explained.

2nd edition. United Kingdom: John Wiley & Sons; 2003. p. 331–67.
[27] Shuang Y, Junzhi Z, Lifang W. Power management strategy with regenerative

braking for fuel cell hybrid electric vehicle. In: Proceedings of the Asia-pacific
power and energy engineering conference; 2009. p. 1–4.
[28] Ouddah N, Boukhnifer M, Raisemche A. Two control energy management
schemes for electrical hybrid vehicle. In: Proceedings of the 10th international
multi-conference on systems, signals & devices; 2013.
[29] Sundströ m O, Stefanopoulou A. Optimal power split in fuel cell hybrid electric
vehicle with different battery sizes, drive cycles, and objectives. In: Proceed-
ings of the international conference on control applications; 2006. p. 1681–8.
[30] Xie CJ, Quan SH, Chen QH. Multiple model control for hybrid power system of
fuel cell electric vehicle. In: Proceedings of the IEEE vehicle power and
propulsion conference (VPPC); 2008.
[31] Karunarathne L, Economou JT, Knowles K. Model based power and energy
management system for pem fuel cell li-ion battery driven propulsion system.
In: Proceedings of the 5th IET international conference on power electronics,
machines and drives; 2010.
[32] Alloui H, Becherif M, Marouani K. Modelling and frequency separation energy
management of fuel cell–battery hybrid sources system for hybrid electric
vehicle. In: Proceedings of the 21st mediterranean conference on control &
automation (MED); 2013. p. 646–51.

814 N. Sulaiman et al. / Renewable and Sustainable Energy Reviews 52 (2015) 802–814

[33] Kelouwani S, Agbossou K, Dubé Y, Boulon L. Fuel cell plug-in hybrid electric [55] Tritschler PJ, Bacha S, Rulliere E, Husson G. Energy management strategies for
vehicle anticipatory and real-time blended-mode energy management for an embedded fuel cell system on agricultural vehicles. In: Proceedings of the
battery life preservation. J Power Sources 2013;221:406–18. XIX international conference on electrical machines – ICEM 2010. Rome; 2010.
p. 1–6.
[34] Benrabeh A, Khoucha F, Herizi O, Benbouzid MEH, Kheloui A. FC/battery
power management for electric vehicle based interleaved dc-dc boost con- [56] Zheng C, Xu G, Zhou Y. Economic influence of prolonging fuel cell stack
verter topology. In: Proceedings of the 15th european conference on power lifetime of fuel cell hybrid vehicles based on optimal control theory. Energy
electronics and applications; 2013. Convers Congr Expo 2013:673–6.

[35] Xiao D, Qi W, Jin H, Fang J. Modeling and simulation for golf fuel cell electric [57] The Hydrogen Economy. The hydrogen economy and sustainable develop-
vehicle control system. In: Proceedings of the 2012 spring congress on ment. United Nations Environment Programme; 2006.
engineering and technology (SCET); 2012. p. 1–4.
[58] Lipman T. An overview of hydrogen production and storage systems with
[36] Yu A, Chabot V, Zhang J. Electrochemical supercapacitors for energy storage renewable hydrogen case studies. A clean energy state alliance report; 2011.
and delivery: fundamentals and applications. United States of America: CRC
Press; 2013. [59] Ainscough C, Peterson D, Miller E. Hydrogen production cost from PEM
electrolysis. DOE hydrogen and fuel cells program record; 2014.
[37] Tani A, Camara MB, Dakyo B, Azzouz Y. DC/DC and DC/AC converters control
for hybrid electric vehicles energy management-ultracapacitors and fuel cell. [60] Chu S. Report of the hydrogen production expert panel: a subcommittee of the
IEEE Trans Ind Inf 2013;9(2):686–96. hydrogen & fuel cell technical advisory committee. United States of America:
United States Department of Energy; 2013.
[38] Uzunoglu M, Alam MS. Modeling and analysis of an FC/UC Hybrid vehicular
power system using a novel-wavelet-based load sharing algorithm. IEEE Trans [61] Delucchi MA, Lipman TE. Lifetime cost of battery, fuel cell and plug-in hybrid
Energy Convers 2008;23(1):263–70. electric vehicles. In: Gianfranco Pistoia, editor. Electric and hybrid vehicles:
power sources, models, sustainability, infrastructure and the market. United
[39] Efstathiou DS, Petrou AK, Spanoudakis P, Tsourveloudis NC, Valavanis KP. Kingdom: Elsevier; 2010.
Recent advances on the energy management of a hybrid electric vehicle. In:
Proceedings of the 20th mediterranean conference on control & automation [62] Qin N, Brooker P, Srinivasan S. Hydrogen fueling stations infrastructure.
(MED); 2012. p. 896–901. United States of America: Electric Vehicle Transportation Center; 2014.

[40] Rodatz P, Paganelli G, Sciarretta A, Guzzella L. Optimal power management of [63] Ogden J, Yang C, Nicholas M, Fulton L. Next STEPS white paper: the hydrogen
an experimental fuel cell-supercapacitor-powered hybrid vehicle. Control Eng transition. United States of America: UC Davis Institute of Transportation
Pract 2005;13:41–53. Studies; 2014.

[41] Ates Y, Uzunoglu M, Erdinc O, Vural B. A Wavelet-ADALINE network based [64] Williamson SS, Lukic SM, Emadi A. Comprehensive drive train efficiency
load sharing and control algorithm for a FC/UC hybrid vehicular power system. analysis of hybrid electric and fuel cell vehicles based on motor-controller
In: Proceedings of the 2009 international conference on clean electrical efficiency modeling. IEEE Trans Power Electron 2006;21(3):730–40.
power; 2009. p. 591–94.
[65] Wu D, Williamson SS. A novel design and feasibility analysis of a fuel cell plug-
[42] Zheng CH, Lee CM, Huang YC, Lin W-S. Adaptive optimal control algorithm for in hybrid electric vehicle. In: Proceedings of the IEEE vehicle power and
maturing energy management strategy in fuel–cell/Li-ion- capacitor hybrid propulsion conference (VPPC); 2008. p. 1–5.
electric vehicles. In: Proceedings of the 2013 9th Asian control conference
(ACC); 2013. p. 1–7. [66] Bonifacio RN, Neto AO, Linardi M. Influence of the relative volumes between
catalyst and Nafion ionomer in the catalyst layer efficiency. Int J Hydrogen
[43] Garcia P, Torreglosa JP, Fernandez LM, Jurado F. Control strategies for high- Energy 2014;39:14680–9.
power electric vehicles powered by hydrogen fuel cell, battery and capacitor.
Expert Syst Appl 2013;40:4791–804. [67] Kokoh KB, Mayousse E, Napporn TW, Servat K, Guillet N, Soyez E, et al.
Efficient multi-metallic anode catalysts in a PEM water electrolyzer. Int J
[44] Martinez JS, Hissel D, Pera M-C, Amiet M. Practical control structure and Hydrogen Energy 2014;39:1924–31.
energy management of a testbed hybrid electric vehicle. IEEE TransVeh
Technol 2011;60(9):4139–52. [68] 2012 Fuel cell technologies market report. US Department of Energy; 2013.
[69] Nikzad MR, Radan A. Accurate loss modelling of fuel cell boost converter and
[45] Schaltz E, Rasmussen PO. Design and comparison of power systems for a fuel
cell hybrid electric vehicle. In: Proceedings of the IEEE industry applications traction inverter for efficiency calculation in fuel cell-battery hybrid vehicles.
society annual meeting; 2008. p. 1–8. In: Proceedings of the 1st power electronic and drive systems & technologies
conference; 2010. p. 218–23.
[46] Yu Z, Zinger D, Bose A. An innovative optimal power allocation strategy for [70] Rathore AK, Prasanna UR. Novel snubberless bidirectional ZCS/ZVS Current-
fuel cell, battery and supercapacitor hybrid electric vehicle. J Power Sources Fed Half-Bridge Isolated DC/DC converter for fuel cell vehicles. In: Proceedings
2011;196:2351–9. of the 37th annual conference on ieee industrial electronics society; 2011. p.
3033–8.
[47] Paladini V, Donateo T, Ad Risi, Laforgia D. Super-capacitors fuel-cell hybrid [71] Lungoci CM, Georgescu M, Calin MD. Electrical motor types for vehicle
electric vehicle optimization and control strategy development. Energy Con- propulsion. In: Proceedings of the 13th international conference on optimiza-
vers Manag 2007;48:3001–8. tion of electrical and electronic equipment (OPTIM); 2012. p. 635–40.
[72] Grilo N, Sousa DM, Roque A. AC motors for application in a commercial electric
[48] Zandi M, Payman A, Martin J-P, Pierfederici S, Davat B, Meibody-Tabar F. vehicle: designing aspects. In: Proceedings of the 16th IEEE mediterranean
Energy management of a fuel cell-supercapacitor/battery power source for electrotechnical conference (MELECON); 2012. p. 277–80.
electric vehicular applications. IEEE Trans Veh Technol 2011;60(2):433–43. [73] Liu V-T, Hong J-W, Tseng K-C. Power converter design for a fuel cell electric
vehicle. In: Proceedings of the 5th IEEE conference on industrial electronics
[49] Li Q, Chen W, Li Y, Liu S, Huang J. Energy management strategy for fuel cell– and applications; 2010. p. 510–5.
battery-ultracapacitor hybrid vehicle based on fuzzy logic. Electr Power [74] Thanapalan K, Zhang F., Premier, G. Maddy J., and Guwy A. Energy manage-
Energy Syst 2012;43:514–25. ment effects of integrating regenerative braking into a renewable hydrogen
vehicle. In: Proceedings of the UKACC international conference on control
[50] Motapon SN, Dessaint L-A, Al-Haddad K. A comparative study of energy (CONTROL); 2012. p. 924–8.
management schemes for a fuel–cell hybrid emergency power system of [75] Hemi H, Ghouili J, Cheriti A. A real time energy management for electrical
more-electric aircraft. IEEE Trans Ind Electron 2014;61(3):1320–34. vehicle using combination of rule-based and ECMS. In: Proceedings of the IEEE
electrical and power energy conference (EPEC); 2013. p. 1–6.
[51] Liu X, Diallo D, Marchand C. Design methodology of fuel cell electric vehicle [76] Yang Y-P, Guan R-M. Hybrid fuel cell powertrain for a powered wheelchair
power system. In: Proceedings of the 2008 international conference on driven by rim motors. In: Proceedings of the international conference on
electrical machines; 2008. p. 1181–6. power engineering, energy and electrical drives (POWERENG). Spain; 2011. p.
1–6.
[52] Hao RC, Kwok TC, Kim FT, Chung HS-H. Energy management of hybrid vehicles [77] Lv Y, Yuan H, Liu Y, Wang Q. Fuzzy logic based energy management strategy of
using artificial intelligence. In: Proceedings of the 2013 IEEE 2nd global battery-ultracapacitor composite power supply for HEV. In: Proceedings of the
conference on consumer electronics (GCCE); 2013. p. 65–7. first international conference on pervasive computing, signal processing and
applications (PCSPA); 2010. p. 1209–14.
[53] Mallouh MA, Surgenor B, Dash P, McInnes L. Performance evaluation and [78] Zheng CH, Lin WS. Self-optimizing energy management strategy for fuel–cell/
tuning of a fuzzy control strategy for a fuel cell hybrid electric auto rickshaw. ultracapacitor hybrid vehicles. In: Proceedings of the international conference
In: Proceedings of the 2012 American control conference. Fairmont Queen on connected vehicles and expo (ICCVE); 2013. p. 87–93.
Elizabeth, Montreal, Canada; 2012. p. 1321–6.

[54] Azidin FA, Hannan MA. Harvesting solar and energy management system for
light electric vehicles (LEVs). In: Proceedings of the 2012 international
conference on renewable energy research and applications (ICRERA); 2012.
p. 1–6.