i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 9 ( 2 0 1 4 ) 1 3 2 7 6 e1 3 2 8 4

Available online at www.sciencedirect.com

ScienceDirect

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

Dynamic performance assessment of the efficiency

of fuel cell-powered bicycle: An experimental
approach

Azadeh Kheirandish, Mohammad Saeed Kazemi, Mahidzal Dahari*

Department of Mechanical Engineering, Faculty of Engineering, University Malaya, Kuala Lumpur, 50603, Malaysia

article info abstract

Article history: Zero-emission fuel cell driven systems are regarded as promising technological advances
Received 26 March 2014 in the future of the transportation industry that have the potential to replace internal
Received in revised form combustion engines. The design, performance, and efficiency properties of a vehicle are
24 June 2014 often stated to be some of the key challenges in its commercialization. This paper high-
Accepted 26 June 2014 lights a polymer electrolyte membrane fuel cell (PEMFC)-powered system of an electric
Available online 22 July 2014 bicycle. The system consists of a 250-W fuel cell, ECU, battery pack, DC/DC converter,
electric motor, and other supporting equipment. After introducing the different parts of the
Keywords: bicycle, its overall efficiency will be discussed in great detail. The efficiency of fuel cells is
PEM fuel cell not specific; it is a subordinate to the power density where the system operates. Experi-
Efficiency analysis mental work has been conducted to measure the values of the efficiency and energy flow.
Electric vehicle The results indicated a maximum fuel cell efficiency of 63% and an overall system effi-
ciency of 35.4%. The latter value is expressed with regards to the Lower Heating Value
(LHV) of hydrogen. All measurements were taken for the cruising conditions of the vehicle
and its corresponding to power consumption. The results are superior to those of a stan-
dard internal ignition engine. The fuel cell performance is least efficient when functioning
under maximum output power conditions.
Copyright © 2014, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights
reserved.

shown remarkable promise. In addition, a high efficiency,

Introduction near zero emissions in producing electricity, and recyclability
are other unique futures of these devices [2]. While hydrogen
Due to the increasing adverse effects of global warming, the and oxygen atoms undergo an electrochemical reaction in a
depletion of energy resources, and the environmental impact fuel cell, they release a considerable amount of power in the
of conventional energy sources such as fossil fuel, the devel- form of electrical currents. The by-product of these reactions
opment of alternative energy sources has become a consid- are water molecules [3]. Currently, fuel cells are categorized
erable subject of interest, especially in the fields of portable into six different types, including the [4] Direct Methanol fuel
and stationary power generation [1]. In the context of these cell (DMFC), which consumes methanol for functionality. The
problems and purposes, commercial hydrogen fuel cells have utilization of these fuel cells is limited to applications in which

* Corresponding author. Tel.: þ60 133599300; fax: þ60 3 79675317.

E-mail address: mahidzal@um.edu.my (M. Dahari).
http://dx.doi.org/10.1016/j.ijhydene.2014.06.138
0360-3199/Copyright © 2014, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved.
 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 9 ( 2 0 1 4 ) 1 3 2 7 6 e1 3 2 8 4 13277

the efficiency is superseded by the power density in terms of work is more interested in the ability of a light vehicle that is
importance. To act as a fuel, methanol should be diluted in powered by a hydrogen fuel cell to meet daily transportation
water at low concentrations. Hence, the utilization of DMFCs requirements. Numerical and theoretical models regarding
is not as common as that of other fuel cells [5]. Solid Oxide fuel the efficiency of fuel cell applications in light vehicles vastly
cells (SOFC) are another type of fuel cell; they contain a solid outnumber experimental reports. To study the behavior and
oxide or ceramic electrolyte. A high temperature threshold the efficiency of such a system, we decided to study, test, and
between 600 and 1000  C constitutes the main concern for the analyze the features of an electric-assisted bicycle that is
utilization of SOFC in vehicles [6,7]. Molten carbonate fuel cells powered by a PEM fuel cell as a common and suitable type of
(MCFCs) are based on a mixture of molten salts that act as an power source for transportation systems.
electrolyte. These fuel cells can perform only at temperatures
that are higher than 650  C [8]. MCFCs also require gasified coal
and natural gas, which is another disadvantage of this system, PEM fuel cell principles
because global warming remains a major concern [9]. Alkaline
fuel cells (AFC) are currently popular; they contain an alkaline Similar to other types of fuel cells, PEMs also consist of three
solution derived from an alkaline electrolyte. They can oper- significant parts: a cathode and an anode that act as electro-
ate over a wide range of temperatures that depend on the fuel lytes formed by platinum-catalysis and the membrane [2]. In a
cell application, which constitutes the main advantage of PEM fuel cell reflex, the hydrogen oxidation and oxygen
these fuel cells over other types. However, the significant reduction reactions occur simultaneously at the anode and
technical disadvantage of this type of fuel cell is the carbon cathode [17]. Fig. 1 shows the single cell of a fuel cell.
dioxide poisoning of the electrolyte [10]. Phosphoric acid fuel At the anode, the stream of hydrogen molecules is dis-
cells (PAFCs) use liquid phosphoric acid as an electrolyte. The articulated into protons and electrons as follows:
operating temperature of these devices is approximately
150e200  C [11]. PAFCs are used in both stationary power and
mobile applications, such as large vehicles. The pre-heating H2 / 2Hþ þ 2e Anode reaction
requirement, and its open-ended structure, which requires
the careful control of hydrogen flow, are some of its draw- Electrons are released from hydrogen, and move along the
backs [12]. Finally, proton exchange membrane fuel cells external load circuit to the cathode; therefore, the flow of
(PEMFC), or solid polymer fuel cells (SPFC), are one of the most electrons creates the electrical output current.
promising systems, especially regarding automotive and The electron arriving at the cathode from the external
vehicular transport [4]. In recent years, the high efficiency of circuit concurrently reacts with oxygen molecules that are
PEMFCs has provided impressive capabilities for the trans- joined with a platinum catalyst of electrode and two protons
portation sectors. For example, the maximum theoretical ef- (which have moved through the membrane) to create water
ficiency of these fuel cells is 83% at standard conditions, molecules; this reduction is represented by the following:
although the experimental efficiency is considerably lower
due to internal resistance issues [13].
Detailed theoretical efficiency calculation methods have 1/2O2 þ 2Hþþ2e / H2O Cathode reaction
been presented in many available publications. Low efficiency
and high production cost are some of the most serious chal-
lenges for previous fuel cell technologies. Therefore, many H2 þ 1/2O2 / H2O Overall reaction
research groups have focused on this area as key challenges
The chemical reaction is now complete. Despite the reac-
for the commercialization stage. Z Qi (2009) [14] demonstrated
tion, a portion of the energy is expended in the form of heat
various vehicular applications, such as bicycles, wheelchairs,
released from the respective redox reaction as a byproduct.
forklifts, and scooters, that utilize PEM fuel cells instead of
PEM fuel cells are used in many applications without
batteries. Fuel cells are expected to generate power for longer
periods and at higher efficiencies compared to batteries.
Meiyappan Siva Pandian (2010) [15] stipulated that enhancing
the power output of PEM would reduce the efficiency, which
would be detrimental to the economical aspect of the system.
Furthermore, Barbir and Gomez (2007) [16] surmised that the
economics and operating efficiency of a fuel cell are interre-
lated. Yongping Hou et al. (2007) [13] proposed models that
detail the efficiency of fuel cells and investigated their theo-
retical and experimental efficiency while evaluating several
influencing parameters that are related to the efficiency of a
fuel cell.
Furthermore, the reduction of hydrogen production costs
renders the cost of hydrogen comparable or even cheaper
than the cost of fossil fuels in the near future. Although such
big-picture studies showed that the hydrogen economy and
fuel cells can the future of the transportation system, this Fig. 1 e Schematic of fuel cell's single cell.
13278 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 9 ( 2 0 1 4 ) 1 3 2 7 6 e1 3 2 8 4

geographical restrictions, and their superior efficiency is being because the bicycle tank had been directly charged from a
capitalized on in automobiles. The works in Refs. [18,19] buffer tank of an electrolyzer system. The energy content of
discuss this fact in more detail. the hydrogen LHV is 241.98 kJmol1, or 120.1 MJkg1. Hydrogen
Using a proton conductive polymer membrane as an elec- can release this energy via ignition heat. Hence, the energy of
trolyte leads to lower operating temperatures, which renders consumed hydrogen for lower heat values is calculated using
the fuel cell viable for both portable and stationary applica- the following equation:
tions. Factors such as being lightweight, a minuscule amount
of corrosive fluid, a long stack lifetime, the generation of zero
emissions, and higher efficiencies render this fuel cell perfect Energy of consumed H2, kJ ¼ 241.98  DP  V/(R  T) (3)
for automobile applications [16].
A low temperature and high efficiency are two dis- where DP is the pressure drop (atm) of the electrolyzer
tinguishing features that set PEMFCs apart from other fuel storage tank after charging the bicycle metal hydride tank, V is
cells [20]. The operating temperature range of PEMFCs is the volume (L) of an electrolyzer buffer tank, T is the tem-
50e100  C, which leads to a very quick commissioning ability. perature (K) of the buffer tank, and R is ideal gas constant,
The total cost is also rather low because cheaper materials are 0.08206 L atm mol1 K1. Because 1W.h ¼ 3.6 kJ, the numerator
viable at low temperature settings, as the operations carries of Eq. (2) can be defined as follows:
less risks. The efficiency of a PEMFC is also much higher
compared to that of an internal combustion engine in vehicles
while direct hydrogen acts as its input. Furthermore, the small Generated electric energy, kJ ¼ 3.6  S(V  I  Dh) (4)
load on a PEMFC translates into higher efficiencies. For a
During the transfer of two electrons from water molecules
normal driving time, a vehicle will require only a small
in fuel cells, the ideal cell voltage or standard potential to
amount of nominal engine power, and a PEMFC is especially
produce water in a liquid state (25  C, 1 atm) is 1.23 V, while the
poignant in this regard because its efficiency is maximized
voltage drops to nearly 1.18 V for producing water in gaseous
when the loads are small. This efficiency peak stands in
state or at operating temperatures exceeding approximately
contrast to that of an internal combustion engine [21].
80  C [3]. Therefore, the theoretical or maximum efficiency of
fuel cell were calculated to be 83% and 94.1% for HHV and LHV,
Efficiency of PEMFC respectively, by substituting the standard voltage in Eq. (1).
Nevertheless, the experimental efficiency is expected to be
The efficiency properties of the fuel cell stack and each single lower than the theoretical efficiency for emerging losses, even
cell are similar because all similar single cells are connected in without being connected to any external load. Furthermore,
series. The fuel cell efficiency was computed from the fuel cell's the efficiency is expected to steeply decline when the load is
stack operating voltage (Uoper) using the following equation. For connected because the voltage drops due to inevitable losses
a more comprehensive explanation, please refer to [22]: generated by the electrical current in a closed circuit.
Furthermore, a fraction of the power generated by a primary
fuel cell is used by the required auxiliary components, which
Fuel cell efficiency, % ¼ 100%  fuel cell Uoper/(1.482 V  No. of include the control unit, blowing fan, cooling fan, and battery.
cells in series) (1)

where 1.482 V corresponds to the hydrogen higher heating

value (HHV) [22]. For hydrogen, the HHV is used when the Experimental
amount of released hydrogen heat energy is measured by
cooling its combustion vessel to 25  C (initial temperature of Overall system design
vessel). On the contrary, 1.254 V is used in Eq. (1), which cor-
responds to the lower heating value (LHV) when the vessel In this section, the experimental components of an electric-
cooling is halted at 150  C. The hydrogen LHV is used when the assisted bicycle are presented and explained. The commer-
efficiency will be compared to internal combustion engine for cial electric-assisted bicycle that we used in this research is
transportation. However, many research groups reported fuel shown in Fig. 2.
cell efficiencies using hydrogen HHV. A fuel cell stack forms the base of the system and is inte-
Eq. (1) cannot be used to calculate the efficiency of the grated with other major components to generate electricity.
entire system because this formula provides the instant effi- The block diagram of the fuel cell-powered electric bicycle is
ciency of the stack that is proportional to the stack's voltage. shown in Fig. 3. The H2 from a metal hydride tank and O2 from
The system's efficiency during the experimental time is the air blower are fed to the fuel cell, where they are combined
computed using the following equation: to form water, a process that consequently produces
electricity.
Fig. 4 shows a complete fuel cell system, which consists of
Fuel cell-powered bicycle efficiency, % ¼ 100%  Generated the fuel cell, one cooling and one reaction air blower, the
electric energy/Energy of consumed H2 (2) Electronic Control Unit (ECU), and the necessary auxiliaries to
manage the hydrogen flow. The installed fuel cell is a 250-W
The mass of consumed hydrogen stored in a metal hydride PEMFC, which contains a 22-cell air-cooled stack that pro-
tank for our experiment is calculated using the ideal gas law vides a nominal stack voltage and current at 14 V and 20 A. It
 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 9 ( 2 0 1 4 ) 1 3 2 7 6 e1 3 2 8 4 13279

Fig. 2 e Fuel cell-powered electric bicycle.

generates a peak power of 250 W to a 25-kg bicycle with a rated coupled with the motor in tough conditions, such as cycling
consumption of 3.5 normal liters of H2 per minute. The spec- uphill.
ifications of the fuel cell are listed in Table 1. The 150-W The entire system is controlled by an ECU designed by MES
electric motor drives the bicycle via the regulated voltage S.A., which monitors all of the system's parameters and
from DC/DC converter, which is connected to a 13.2e20.2 V operates the system on command. The ECU features an
unregulated output voltage of the fuel cell. The DC/DC accessible remote control outside of the main box, and the
convertor is 300-W convertor from Zahn Elect with efficiency driver can switch the main power on/off and start the fuel cell
of 92% (at Vin ¼ 12 V, Iout ¼ 8.4 A) and recommend input voltage on demand. The ECU operates the fuel cell during the start,
of 10e20 V. The experimental results show that a motor with a normal operation, and shutdown procedures by monitoring
wide operating voltage range would decrease the cost of the various parameters, such as the stack temperature, hydrogen
system and improve its efficiency for an eliminating DC/DC pressure, stack voltage, stack current, battery voltage, and
converter. A brushless DC electric motor was installed in the ambient condition. When the ECU detects an abnormal con-
rear axle hub of the bicycle, and this unique design created dition, it will alert the driver via the LED indicators placed on
some extra space. The motor was activated by a handlebar- the remote control and handlebar throttle. If the condition for
mounted throttle, just like on most motorcycles or scooters. a security shut off (TRIP) is fulfilled and remains unchanged
In other words, this vehicle was operated on a power-on- for a certain security time, the system will shut off automat-
demand basis; the electric motor was only engaged and ically. The condition to enter in the Trip is at least one of the
operated manually using a throttle, while the pedaling can be following:

Fig. 3 e Block diagram of the fuel cell-powered electric bicycle system.

13280 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 9 ( 2 0 1 4 ) 1 3 2 7 6 e1 3 2 8 4

that can store large amounts of hydrogen relative to its size.

The installed-in tank (Ø54  L350 mm) in this experimental
bicycle can store more than 200 NL of hydrogen. Due to the
exothermal reaction, it releases heat when the vessel is
charging, which makes the charging-time dependent on the
surrounding temperature.
Moreover, the system can utilize the external battery or
capacitor not only for providing burst power for difficult sit-
uations, such as uphill cycling, but also to minimize the size
and weight of the fuel cells.

DAQ system

The DAQ system and its software were designed by MES S.A.
and utilized the serial connection between fuel cell ECU and
the monitoring software. As shown in Fig. 5, all of the pa-
Fig. 4 e Fuel Cell and auxiliary components. rameters of the fuel cell are monitored and could be logged
accordingly. The data were logged with a sample time of 1 s to
a designated computer. The data collection error that was
faced during experimental was outliers. The outliers which
T.1 Stack voltage below 12 V for more than 3 s. were eliminated from collected data were 0.12% of 9527 data
T.2 Stack temperature above 67  C for more than 3 s. sets.
T.3 Voltage of the starting power source below 8.6 V for
more than 3 s.
T.4 Stack current above 35 A for more than 3 s. Results and discussion
T.5 The hydrogen supply pressure is:
Fig. 6 shows an example of VeI and PeI curves for our PEM
_ less than 0.2 bar for 0.1sec; stack, which were obtained at an approximately constant fuel
_ out of the range 0.55e0.8 bar and the stack current is > 5 A cell temperature and relative humidity averaging 37.6  C and
for 3 s; 48.8%, respectively. This plot illustrates the typical VeI
_ out of the range 0.55e0.9 bar and the stack current is < 5 A behavior; the voltage monotonically decreased as the current
for 10 s. increased.
While the load circuit is open, the current is close to zero,
The system features an external rechargeable 12-V battery and the stack voltage is estimated to at least be close to the
as a 10-W power source to supply the ECU, the cooling fan, and theoretical value. However, the voltage is often considerably
the hydrogen control valves. Because the acceptable external less when a fuel cell is utilized. The following key points are
power supply lies in the range of 8.8e10.8 V, the ECU charges evident from the graph of the cell voltage against the current:
the battery at 9.6 V. The fuel cell receives highly pure
hydrogen from a storage tank after passing the pressure  The open circuit voltage is less than the theoretical value
regulator to reduce hydrogen pressure to approximately  A sharp initial voltage decrease
0.8 bar. The hydrogen storage tank is a metal-hydride tank  The voltage then decreases less rapidly and more linearly

This characteristic is due to various losses, such as [12]

activation losses, fuel crossover and internal current, resis-
Table 1 e Nominal fuel cell specifications. tive or ohmic losses and concentration losses. Activation
Parameter Value losses are most effective with regards to nonlinear voltage
drops after the start-up of the system, as evident from the
Number of single cells 22
Maximum total power output (W) Approximately 250
beginning of the curve. Activation losses are caused by slow-
Nominal stack voltage (V) 14 ness of the reaction that takes place on the electrodes' surface.
Nominal stack current (A) 20 (max 28) A proportion of generated voltage is lost in driving the
Weight of the stack (kg) 0.7 chemical reaction transferring electrons between electrodes.
Overall dimensions of the fuel cell approx. 230  220  150 In addition, activation losses are important at low current and
system (blowers Inc.) (mm3)
reflect the fact that cell requires energy to initiate an elec-
Ambient temperature range ( C) >0 up to þ35
trochemical reaction because of to the catalyst; a lower acti-
Ambient relative humidity range (%) 30e80
(recommended) vation energy requires superior catalysts. Further research on
Hydrogen supply nominal pressure (bar) 0.65 superior catalyst materials could improve the performance of
Hydrogen supply minimum flow rate 3.5 fuel cell systems after start-up. Fuel crossover and internal
guaranteed (Nl/min) current losses are insignificant in PEMFCs due to low-
Maximum purged hydrogen flow rate 0.07 temperature cells. However, this lower temperature also
(Nl/min)
causes some hydrogen to diffuse into the crossing via the
 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 9 ( 2 0 1 4 ) 1 3 2 7 6 e1 3 2 8 4 13281

Fig. 5 e The panel of monitoring software for fuel cell System.

electrolyte, from the anode to the cathode. Furthermore, concentration of gaseous reactants at the electrode's surface.
Ohmic losses are linear and outweigh the current density. As Mass transport losses also define this phenomenon.
the VeI curve shows, the pattern after the sharp voltage drop The efficiencyepower curve of a fuel cell stack is shown in
of the start-up experiences a steady linear decline when the Fig. 7, which was plotted using Eq. (1) for the hydrogen HHV.
current is increased. Ohmic losses are driven by the combined As expected, the maximum efficiency of the fuel cell's stack is
membrane resistance and electrical resistance of the fuel available in the low power or current range as the highest
cell for transferring electrons and protons. At the final stage, stack voltage at an open circuit. Accordingly, while the stack
the concentration losses are caused by a change in the power or current is increased, the efficiency will decrease. As
previously mentioned, this characteristic is the opposite of the
efficiency curve over power in an internal combustion engine,
rendering the fuel cell superior in terms of city transportation,
where at most only a small fraction of nominal power is
required.
Next, the efficiency of the entire system is analyzed. In this
experiment, the data collection is performed on a stationary
bicycle while the tyre could spin freely on traditional Kick-
stand. As explained, the ratio between the energy output and
input forms the basis of calculation of the efficiency in any
energy transformation system. This ratio in the fuel cell falls
between the output electrical power and the hydrogen energy
that is consumed. The consumed hydrogen energy for this
experiment is calculated using Eq. (3) as follows:
Energy of consumed H2 ¼ 241.98  4.96  20/
(308.65  0.08206) ¼ 947.75 kJwhere the pressure of a 20-liter
electrolyzer buffer tank decreased from 19.91 bar to
Fig. 6 e Plot of the voltage‒current and powerecurrent 14.88 bar for 3.92 mol of hydrogen. Thus, the fuel cell used
curves of Fuel Cell stack at temperature average of 37.6  C. 947.75 kJ as the system input, and the electric motor used the
13282 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 9 ( 2 0 1 4 ) 1 3 2 7 6 e1 3 2 8 4

period of approximately 159 min. Hence, the overall system

efficiency for this experiment was calculated using Eq. (2):
The overall system efficiency ¼ 100  335.82/
947.75 ¼ 35.43%
Fig. 8 shows the Sankey diagram, which visualizes the
different energy flow and power losses in an analyzed electric-
bicycle system.
The following important point should be considered. As
previously explained, the efficiency of the fuel cell stack is not
constant and depends on the stack voltage (bicycle usage
condition). Thus, this value can change for different experi-
mental conditions for the same bicycle. For this experiment,
we attempted to keep the bicycle at the cruise condition
(constant speed) with a fuel cell power average of 35.29 W, and
fuel cell stack efficiency average of 48.45%. The specifications
of our efficiency study are listed in Table 2.
Fig. 7 e Plot of FC stack efficiency versus FC stack output
Fig. 9 shows the plot of the fuel cell power output
power for relatively average temperature 37.6  C over
throughout the efficiency experiment. As the curve shows, we
experiment period.
attempted to keep the output power constant for small frac-
tions of the nominal fuel cell power as needed at the cruise
condition. The logged power data in this curve show a bold
line at the bottom of the power lane. The power differences
between these higher points on top of this line and the points
on the line are due to auxiliaries, such as 10-W cooling blower,
which the ECU altered to maintain its speed and compensate
for the temperature of the fuel cell at the designated working
temperature. When the fan is working at a minimum rate, the
load power and other constant auxiliaries form the bold line in
the curve, and the other variable components form the
approximately 15-W lane.
In addition, the stack at higher power production or cur-
rent density releases more heat if the system design is
imperfect; the overheating that occurs during tough working
situations can disturb the normal operation of system. Fig. 10
shows that the ECU can perfectly control the temperature of
the stack and keep it almost constant. The humidity curve
Fig. 8 e Flow diagram for PEMFC-powered electric bicycle. illustrates the humidity ratio of the input air.
This diagram illustrates the various energy flows in During the experiment, it was observed that the ECU al-
system. ways automatically shuts down the system prior to reaching
the full output power, which is important to the performance.
After investigating, we discovered that the stack voltage of the
fuel cell fell below the minimum value (12 V) before reaching
FC power output. The output electric power of the system was the maximum power due to not sufficient hydrogen flow at
calculated by measuring the power consumption of the bi- stack, and the ECU automatically shut down the system to
cycle's electric motor over time using Eq. (4), which is 335.82 kJ protect the fuel cell. As it was mentioned, the temperature
for a sample time of 1 s (Dh ¼ 1/3600) over an experimental of Metal Hydride tank decrease during de-charging, the

Table 2 e Fuel cell-powered electric bicycle parameter measurements from data demonstrated in Fig. 9.
Parameters Experiment Hydrogen Load energy System Maxe Min of FC Mean of FC Mean of FC
duration consumptiona consumptionb efficiencyc efficiencyd efficiencye powerf
Value 9515 s 3.917 mol 335.82 kJ 35.43% 63.88% 48.45% 35.29 W
(947.75 kJ) 42.69%
a
From Eq. (3).
b
Consumption electric power obtained from the sum of 9515 load power values that logged each second.
c
Efficiency of fuel cell-powered electric bicycle base on experimental data using Eq. (2).
d
The maximum and minimum of fuel cell stack efficiency during this experiment.
e
Average of FC stack efficiency over 9515 individual efficiency values calculated for each second of experimental time.
f
The mean was calculated from 9515 individual data points shown in Fig. 9.
 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 9 ( 2 0 1 4 ) 1 3 2 7 6 e1 3 2 8 4 13283

operation will be halted when the current or voltage exceeds

the limit. These phenomena need to be investigated further in
order to improve the stack behaviors at extreme conditions.

Conclusion

In this paper, we described and experimentally tested a

commercial electric bicycle powered by a PEM fuel cell to
investigate its performance and efficiency in real-life opera-
tions. The polarization curve was obtained by using logged
experimental data. The stack voltage rapidly declined during
the start-up of the system due to losses, which strongly
influenced later stages. The stack voltage also subtly reduces
while the current increases. This finding proves that
Fig. 9 e Plot of fuel cell stack power measured during improving the stack behavior during start up ameliorates the
efficiency experiment. efficiency in future fuel cells. Furthermore, the overall system
efficiency was measured throughout normal daily life opera-
tions. The results demonstrate that similar to internal com-
bustion engines, the efficiency of the fuel cell is not a single
canister's surface temperature dropped from 27  C to 0  C in value and directly depends on the output power of the system.
less than 15 min and this leads to affect the maximum However, the efficiencyepower curves of fuel cells and inter-
hydrogen flow rate, which could potentially provide for fuel nal combustion engines showed opposite behaviors. The fuel
cell. In other word, when the fuel cell of bicycle works for a cell-powered systems are more efficient at lower output
while, Metal Hydride tank cannot supply the minimum 3.5 Nl/ powers, which is the exact inverse of the behavior of internal
min of hydrogen in the time of high power demand of bicycle. combustion engines. The overall system efficiency was
Consequently, the voltage of fuel cell drops under the limita- calculated to be approximately 35% for close to cruising op-
tion and ECU shuts down the stack. To prevent this circum- erations, while the stack efficiency varied 43e64% which is
stance, there are two possible solutions. The first solution is significantly higher than that of an internal-combustion en-
installing heater around hydrogen hydride tank in order to gine and also compare to other similar research which effi-
keep the temperature at room-temperature and consequently ciency of 18.8e38% was recorded [23] is acceptable.
the flow rate of hydrogen at nominal point. However this so- Consequently, any improvement in the system's perfor-
lution could potentially decrease the efficiency of the system mance to amend the efficiency could lead to the growth of the
due to heater power consumption. The Second proposed so- fuel cell systems in transportation. For instance, adding a
lution is using bigger tank (or pressurized tank), which still regenerative breaking system to future designs can significantly
could supply enough hydrogen flow rate for fuel cell in low increase the system's efficiency, which could also save up to
temperature of de-charging condition. This behavior demon- 28% of energy in a similar project involving electric buses [24].
strates the limitation of metal hydride tank and its effects on
the system are important parameters that may ultimately
affect the system's performance, or simply, the system's
Acknowledgments

The authors gratefully acknowledge that this project is funded

by the Faculty of Engineering, University of Malaya under the
High Impact Research program, (project UM.C/HIR/MOHE/
ENG/23).

references

[1] Mori D, Hirose K. Recent challenges of hydrogen storage

technologies for fuel cell vehicles. Int J Hydrogen Energy
2009;34(10):4569e74.
[2] Cook B. An introduction to fuel cells and hydrogen
technology. Eng Sci & Educ J 2002;11(6):205e16.
[3] Barbir F, Gomez T. Efficiency and economics of proton
exchange membrane (PEM) fuel cells. Int J Hydrogen Energy
Fig. 10 e Plot of Stack temperature and air humidity ratio 1997;22(10e11):1027e37.
versus current density of Fuel Cell. [4] Larminie J. Fuel cell systems explained. 2nd ed. 2003.
13284 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 9 ( 2 0 1 4 ) 1 3 2 7 6 e1 3 2 8 4

[5] Sundarrajan S, Allakhverdiev SI, Ramakrishna S. Progress [14] Dyer M, Ogumi, Rand, Scrosati. Encyclopedia of
and perspectives in micro direct methanol fuel cell. Int J electrochemical power sources; 2009.
Hydrogen Energy 2012;37(10):8765e86. [15] Pandian MS. Efficiency and economics analysis of proton.
[6] Chick LA, Marina OA, Coyle CA, Thomsen EC. Effects of IPEC, 2010 Conference Proceedings 2010:875e80.
temperature and pressure on the performance of a solid [16] Yilanci A, Dincer I, Ozturk H. Performance analysis of a PEM
oxide fuel cell running on steam reformate of kerosene. J fuel cell unit in a solarehydrogen system. Int J Hydrogen
Power Sources 2013;236:341e9. Energy 2008;33(24):7538e52.
[7] Jung G-B, Chen J-Y, Lin C-Y, Sun S-Y. Fabrication of hydrogen [17] Sharifi Asl SM, Rowshanzamir S, Eikani MH. Modelling and
electrode supported cell for utilized regenerative solid oxide simulation of the steady-state and dynamic behaviour of a
fuel cell application. Int J Hydrogen Energy PEM fuel cell. Energy 2010;35(4):1633e46.
2012;37(20):15801e7. [18] Marr C, Li XG. Composition and performance modelling of
[8] Song SA, Kim HK, Ham HC, Han J, Nam SW, Yoon SP. Effects catalyst layer in a proton exchange membrane fuel cell.
of marine atmosphere on the cell performance in molten J Power Sources 1999;77(1):17e27.
carbonate fuel cells. J Power Sources 2013;232:17e22. [19] Baschuk JJ, Li XH. Modelling of polymer electrolyte
[9] Mizuhata M, Ohashi T, Be le
ke
 AB. Electrical conductivity and membrane fuel cells with variable degrees of water flooding.
related properties of molten carbonates coexisting with J Power Sources 2000;86(1e2):181e96.
ceria-based oxide powder for hybrid electrolyte. Int J [20] Wee J-H. Applications of proton exchange membrane fuel
Hydrogen Energy 2012;37(24):19407e16. cell systems. Renew and Sustain Energy Rev
[10] Verhaert I, Verhelst S, Huisseune H, Poels I, Janssen G, 2007;11(8):1720e38.
Mulder G, et al. Thermal and electrical performance of an [21] Salemme L, Menna L, Simeone M. Analysis of the energy
alkaline fuel cell. Appl Therm Eng 2012;40:227e35. efficiency of innovative ATR-based PEM fuel cell system with
[11] Wannek C, Konradi I, Mergel Jr, Lehnert W. Redistribution of hydrogen membrane separation. Int J Hydrogen Energy
phosphoric acid in membrane electrode assemblies for high- 2009;34(15):6384e92.
temperature polymer electrolyte fuel cells. Int J Hydrogen [22] Feroldi D, Basualdo M. Description of PEM fuel cells system;
Energy 2009;34(23):9479e85. 2012. pp. 49e72.
[12] Vanhanen J, Kauranen PS, Lund PD. Operation experiences of [23] Moore RM, Hauer KH, Ramaswamy S, Cunningham JM.
a phosphoric acid fuel cell in a solar hydrogen energy Energy utilization and efficiency analysis for hydrogen fuel
system. Int J Hydrogen Energy 1997;22(7):707e13. cell vehicles. J Power Sources 2006;159(2):1214e30.
[13] Hou Y, Zhuang M, Wan G. The analysis for the efficiency [24] Folkesson A, Andersson C, Alvfors P, Alaküla M, Overgaard L.
properties of the fuel cell engine. Renew Energy Real life testing of a hybrid PEM fuel cell bus. J Power Sources
2007;32(7):1175e86. 2003;118(1e2):349e57.