2020 IEEE/ASME International Conference on

Advanced Intelligent Mechatronics (AIM), Boston, USA

(Virtual Conference), July 6-9, 2020

Modeling and Control of Fuel Cell Power System with Varying Load and
Temperature
Shafiqul Islam

Abstract— In this paper, we develop proportional-integral- uncertain dynamical behavior of the cell electrochemistry,
derivative (PID) control algorithm to examine the dynamic thermal energy balance and reactant flow. The performance
behavior of the proton exchange membrane fuel cell (PEMFC) of PEMFC system depends on the response time of the fuel
power generation system in the presence of thermal and load
variation. The model is based on empirical equations combining cell for the given operating load. The dynamic response of
three complex dynamics as electrochemical, reactant flow dy- the cell electrochemistry and thermodynamics is relatively
namics and hydrogen gas reformer. We first model the fuel cell slower than the load variation. Therefore, it is essential to
power system dynamics along with the hydrogen gas reformer. analyze the static and dynamic phenomenon of the FC system
The fuel cell dynamical model is based on empirical equation in the presence of the load demand. To investigate uncertain
of the cell electrochemistry and reactant flow dynamics. The
hydrogen gas reformer is modeled by using second order dynamical property of the PEMFC system, different state
transfer function. The PID controller is employed to control models have been reported in the literature [14, 15, 17-
the hydrogen, oxygen and water vapor flow rate by regulating 25, 28]. These models can be divided into two groups
the methane flow rate of the gas reformer. We then examine as physical and empirical models. The empirical model
dynamic phenomenon of PEMFC power system in the presence pioneered by Amphlett and his research group has been
of the sudden changes in the load and stack temperature.
Finally, we evaluate the proposed model and controller on a considered to be the most popular in the fuel cell research
5KW PEMFC power system to analyze the impact of the load over the past decade. Authors in [16] introduced small signal
and temperature variation on the output voltage. linearized model to analyze dynamic property of the PEMC
system. The model, however, only considers the effect of the
I. I NTRODUCTION cell electrochemistry and reactant flow dynamics. Recently,
F uel cell (FC) is considered as one of the most at-
electrical circuit based dynamic model for PEMFC proposed
in [22]. Authors in this paper combined thermodynamic
tractive renewable energy source as it produces electricity
characteristics using thermal energy balance condition with
without emitting greenhouse gases. Motivated by environ-
the cell electrochemistry and reactant flow model. In [41],
mental friendly power generation, FC technology has been
authors have analyzed steady-state and dynamic behavior of a
extensively studied for future distributed power generation
PEMFC system. Most recent advances and challenges of fuel
systems in portable [5, 6], stationary [9, 10] and trans-
cell based power system can be found in [42]. The dynamic
portation applications [2, 3, 4]. U.S. Department of Energy
modeling and simulation analysis of the above stated models
[1, 13] and automobile industry spent significant amount
was only used for the prediction of the output response
of research grant for advancing FC technology. Automo-
characteristics and for the experimental verification of the
bile industries, such as, Honda, GM, Ford, Mazda, Toyota,
analytical model. For the control design and analysis of
Suzuki, Daimler Chrysler, are specifically interested for FC
PEMFC system, a ninth-order state space control oriented
powered fuel cell vehicles for mobile applications to replace
dynamic model of the air-flow delivery system has been
conventional combustion engine based vehicles to reduce the
introduced by Pukrushpan in [21]. Authors in this paper
environmental problem and the cost of the oil [9, 11, 30,
derived state feedback controller by using well-known LQR
31, 32, 33, 34]. FC technology transfers hydrogen energy
algorithm. The design combines nonlinear feed forward and
directly into electrical and thermal energy with water as
feedback terms to control air supply systems. The PEMFC
a byproduct. The electrical power is generated from the
model used in reported control algorithm and performance
electrochemical reaction between hydrogen fuel and oxidant.
analysis is assumed that the slower dynamics associated
Among various types of fuel cells, phosphoric acid fuel cell
with the temperature variations and heat dissipation in FC
(PAFC), molten carbonate fuel cell (MCFC) and solid oxide
stack is negligible. Among various FC operating conditions,
fuel cell (SOFC), proton exchange membrane fuel Cell or
load demand and temperature variation plays significant
polymer electrolyte membrane fuel cell (PEMFC) has been
role to achieve reliable and efficient power response. As
attracted because of its relatively small size, light weight,
a matter of fact, one of the most important barriers for
quick start-up, low operating temperature, easy to build and
commercialization of the FC system is its dynamic limitation
high power density, low green house gas emission and clean
including uncertain dynamic behavior of the slow start up
energy production [12, 13]. PEMFC is a very complex elec-
process, load and temperature variation. In practice, the
trochemical nonlinear system which depends on the coupled
stack temperature in PEMFC system, such as Ballard MK5-
Dr. Islam is with Xavier University of Louisiana, 1 Drexel Drive, Box E stack based PGS-105B system [15, 35] is kept below
28 New Orleans, LA 70125. 80o C by using externally regulated coolant flow, which is

978-1-7281-6794-7/20/$31.00 ©2020 IEEE 1236

Authorized licensed use limited to: UNIVERSITY OF ROCHESTER. Downloaded on August 29,2020 at [Link] UTC from IEEE Xplore. Restrictions apply.
usually done by using constant flow of water circulating/a
water circuit between specific group of cells. When the
temperature reaches closed to 75o C at the oxygen outlet, then
the system’s water is transferred to a radiator. The radiator
fans are used to control the stack temperature based on the
number of fans that are running at each instant of time.
However, operating temperature may vary significantly with
varying load current. The warming or cooling the stack at the
start up or current variation associated with the time delay
Fig. 1. PEMFC System [39],[16],[40]
to reach optimal operating temperature (70o C-80o C) of the
FC system. Generally, stack power dissipation increase with
the increase of the load current increasing stack temperature.
An increase in stack temperature may cause membrane pressure inside the stack. Water management unit plays an
dehydration and oxygen depletion reducing the stack voltage important role in achieving optimal performance of the fuel
degrading the performance of the FC system. Therefore, it cell system. Humidifiers on cathode and anode are usually
is very important to analyze the dynamic property of the essential to regulate the membrane humidity of the cell.
PEMFC system in the presence of the load and temperature In practice, such as MK5-E Ballard fuel cell system [35]
variation. In this paper, we develop PID control algorithm reactant is humidified internally by circulating water stream
to explore the dynamical property of the PEMFC system in through radiator. If the temperature becomes very high, then
the presence of the load and temperature variation. We first membrane may dry out which increase the resistance to the
model FC system that couples complex cell electrochemistry conduction of the hydrogen ions degrading the output voltage
and fluid dynamics with the gas reformer. The PID controller and damaging electro catalyst property of the FC system.
is designed to regulate the air and hydrogen flow rate by Fuel Cell Structure: FC unit consists of solid polymer
controlling the input flow rate of the gas reformer. Then, membrane electrolyte, electrodes, gas diffusion layer, gas
we investigate the dynamic behavior of 5KW PEM fuel cell flow channel and bipolar plates as shown in Fig. 1 [16,
system under various loading and temperature conditions. It 39, 40]. Solid polymer membrane is made up from Nafion
is shown in our analysis that the membrane resistance may coated with highly dispersed catalyst particles. Membrane
increase with the increase of the dehydration level of the electrolyte is placed between cathode and anode electrodes.
membrane electrolyte. The membrane resistance increases Electrode consists of platinum catalyst particles. The solid
due to the slow response of the fuel or oxidant flow rate graphite materials are usually used to make the bipolar
from the variation of the load current and temperature as plates for better chemical reaction stability and electrical and
they cannot response instantly increasing the dehydration thermal conductivity. The solid graphite materials also have
level of the membrane electrolyte. As a result, large voltage strong chemical and physical reaction stability. They deliver
dropped may appear for short time degrading the PEMFC the reactant gas to the surface of the electrodes from gas flow
performance. This paper is organized as follows. A detailed channels. The gas diffusion layers are made from porous
dynamical modeling analysis is derived in section II. In and electrically conductive material such as carbons. They
section III, 5KW PEM fuel cell system is used to evaluate the diffuse the reactants to and from the polymer membrane.
model and control derived in section II with the presence of Gas reformer is an important part of the fuel processing
varying load and temperature. Finally, section IV concludes unit in FC based power generation system. It usually takes
the paper. hydrocarbons, such as, propane, natural gas and methanol as
an input and separates them into hydrogen molecule. Natural
II. DYNAMICS AND C ONTROL OF PEMFC S YSTEM or propane gas usually preferred in stationary applications as
PEMFC power system comprises various components as they can be easily accessible in various stationary applica-
fuel cell stack unit, fuel processing unit, air delivery system, tions, such as, residence, portable electronics, telecommuni-
water and thermal management and power conditioning unit. cation installation facilities and distributed power generation
Fuel processing unit consists of various auxiliary devices, stations. For transportation applications, such as, hybrid
such as, gas processor, storage tank, flow meter, filter, electric/fuel cell vehicle, methanol is preferred as methane
pressure regulator, and humidifier. In reformer, hydrogen gas has a high power density and is easy to transport.
molecules are separated from gas through vaporizing liquid To generate hydrogen fuel from methane gas, mixer of
hydrocarbons with water and then stored in a storage tank. the liquid methanol and water is first vaporized and then
Dry hydrogen gas from pressurized tank is then feed into feed through the heated chamber. This chamber uses highly
anode side of the fuel cell stack via flow meter and pressure dispersed catalyst particles, such as, platinum. When water
regulator. The flow meter can measure consumed hydrogen vapor strikes catalyst particles, it separates into hydrogen
flow rate. The regulator can regulate required gas pressure and oxygen gas. At the same time, methane molecules strike
in fuel cell tank [36]. In air supply system, purified air from the platinum catalyst and split into carbon monoxide and
a air compressor is delivered to the cathode side of the stack hydrogen gas. Then, the carbon monoxide combines with the
through a air flow controller which maintains desired air freed oxygen from water molecules to form carbon dioxide.

1237

Authorized licensed use limited to: UNIVERSITY OF ROCHESTER. Downloaded on August 29,2020 at [Link] UTC from IEEE Xplore. Restrictions apply.
The output dynamics of the reformer model can be expressed dynamics. Therefore, the power generated by PEMFC relies
by the following transfer function [18, 26] on various factors, such as, operating temperature, operating
Fcf qmeth load, membrane hydration, reactant and oxidant concentra-
qh2 = (1) tion and pressure. The thermal equilibrium potential for the
τr1 s + τr2 s + 1
H2/O2 fuel cell reaction depends on temperature, pressure
where Fcf is the conversion factor, qh2 is the molar flow of and current. Actual cell reaction potential generated by
hydrogen, qmeth is the methane flow rate and τr1 , τr2 are H2/O2 reaction is reduced by the voltage drop resulting
the reformer time constant. Notice from (1) that the reformer from ohmic, reactant concentration and activation losses. The
input qmeth controls the hydrogen flow rate qh2 . To meet the output potential generated by a single FC can be written by
load demand, one requires to control the hydrogen flow rate the sum of the thermodynamic potential, ohmic voltage drop
qh2 by controlling the input methane flow rate qmeth . The and activation voltage drop as
required hydrogen can be related to the demand of the load
as [18] Vsf = (EN ernst − ηac − ηohm ) (7)
demand N if where EN ernst is the open-circuit reversible voltage, ηac is
qh2 = (2)
2F U the activation voltage drop and ηohm is the ohmic voltage
Where U is the utilization rate, if is the stack current and loss. Applying Nernst’s equation and voltage drops due to
F is the Faraday’s constant. Using the required hydrogen, kinetic and ohmic loss, the overall voltage can be presented
the control input can be designed for the methane flow rate. for the N number of cells as
This flow rate can be regulated by the following PID control
law Vstack = N Vsf (8)
Z
qmeth = k1em + k2ėm + k3 em dt (3) Thermodynamic Voltage Equation:The reversible cell volt-
age or thermodynamic potential in the fuel cell system is
where ki > 0 with i = 1, 2, 3, em = measured at open circuit condition. This voltage is calcu-
demand


qh2 − qh2 + qmethr , qh2 is the actual hydrogen lated at the thermodynamic equilibrium balance condition
flow rate and qmethr is the reference of the methane. of H2/O2 fuel cell reaction. Modified Nernst’s equation is
The input for the oxygen flow rate can be calculated by commonly used to represent the thermodynamic potential
hydrogen-oxygen flow ratio rhof in the reformer [18]. generated by overall reaction of hydrogen and oxygen at the
EMFC based power can be viewed as one of the most thermodynamic balance state as [15, 13, and 28]
greenest power solution as its low operating temperature, RT h 0.5
i
high power density, very low noise, low emission of carbon, EN ernst = Eo + ln Ph2 (Po2 ) (9)
2F
nitrogen and sulfur oxides, light construction and clean
energy production. In PEMFC, polymer membrane is used where R is universal gas constant, Eo defines standard
to divide the anode and cathode part of the cell. Pure reversible cell potential (1.22V), Po2 is the oxygen partial
hydrogen from pressurized tank supplied from gas reformer pressure and Ph2 is the hydrogen partial pressure. The stan-
is continuously feed into the cathode electrode and diffuse dard reversible potential Eo may vary with cell temperature
through the porous electrode into the three phase interfaces and load. Therefore, Eo can be represented with the stack
formed by catalyst particles, reactants and membrane. The temperature T by the following equation
key feature of the membrane electrolyte is that it only Eo = 1.229 − 8.510−3(T − 298.15) (10)
allows positive charged particle to travel from anode to
cathode chamber. Freed electrons are then transferred to the The internal voltage given in (10) is generated based on
cathode side through the load. At the same time, oxygen is the electrochemical reaction between hydrogen and oxygen.
supplied to the cathode and combined with the hydrogen When the external load is connected to the fuel cell system,
ions diffused through the membrane to produce water. In then the open circuit voltage is reduced significantly due
PEMFC system, the following chemical reaction at the to the presence of over voltages. These voltage losses vary
anode electrode is taken place significantly with the load current and stack temperature.
Overvoltage losses in PEMFC system are the activation
H2 → 2H + + 2e− (4)
losses and resistive losses. The activation voltage drop ap-
The oxidation reaction takes place at the cathode side to form pears due to the presence of slow electrode kinetics. There-
water fore, activation losses can be minimized by maximizing the
catalyst contact area for electrochemical reactions. Voltage
4H + + 4e− + O2 → 2H2O (5) drop due to activation losses can be defined by the following
Then, overall reaction in PEMFC can be written as equation

2H2 + O2 → Heat + 2H2O + Energy (6) ηact = −Bln(Cif ) (11)

This power generation process depends on the dynamics of where B and C are the activation voltage constants. The
the cell electrochemistry, gas reformed and reactant flow resistive loss is caused by current flow through the resistance

1238

Authorized licensed use limited to: UNIVERSITY OF ROCHESTER. Downloaded on August 29,2020 at [Link] UTC from IEEE Xplore. Restrictions apply.
of the whole electrical circuit including the membrane elec- If the relative humidity is constant, then the water flow across
trolyte, electrodes and various interconnections. The mem- membrane can be derived as a function of the stack current
brane humidity directly influence the values of the internal as
resistance Rin. The conduction of the hydrogen ions through  
1.2684N
the polymer membrane depends on the humidity of the qh2ombr = if (18)
F
membrane electrolyte. The values of Rin increase with the
increase of the dryness of the membrane. The values of the Now, we use (17), (18) and the general relationship between
Rin also varies with the variation of the stack temperature. the mass flow rate of a gas through the valve and partial
The ohmic voltage drop due to membrane resistance can be gas pressure inside the channel [28]. We then derive the
expressed by the following equation derivative of the partial pressures of hydrogen, oxygen and
water inside the fuel cell can be derived as
ηohomic = −if Rin (12) RT
−5 −5 Ṗh2 = (qh2in − Kh2 Ph2 − ζ1 if ) (19)
Rin = 0.01605 − 3.5 × 10 T + 8 × 10 if(13) Va
RT
Reactant Flow Dynamics: In PEMFC, the gaseous fuels are Ṗo2 = (qo2in − Ko2 Po2 − ζ2 if ) (20)
Vc
directly feed into anode from hydrogen storage tank. The RT
hydrogen gas is supplied from gas reformer. It is important Ṗh2o = (ζ1 if − KH2o Ph2o + ζ3if ) (21)
Vc
to note that the electrolyzer can also be used to produce N

N



hydrogen from electricity. The input energy in electrolyzer where ζ1 = 2F , ζ2 = 4F , ζ3 = 1.2684N
2F
, KH2 , KH2o
can be taken from the renewable sources, such as, wind and and Ko2 are the valve molar constant for hydrogen, water
solar power. The fuel and oxidant consumption in fuel cell and oxygen, respectively.
system directly depends on the load demand. So, when the III. S IMULATION R ESULTS
power is supplied from the cell, the oxygen from cathode
In this section, we evaluate the dynamic model and
is consumed. The consumption of the fuel and oxygen also
control introduced in previous section on a 5KW PEMFC
depends on the volume of the electrodes, input and output
power system. The stack current for this PEMFC system is
of the mass flow. The flow meter can provide accurate
considered as 95 [A] based on the power rating. The rated
measurement of the input and output flow rate. If the input
output voltage is 50 [V]. It is assumed that the pressurized
and output flow rates are available, then the partial gas
dry hydrogen from a storage tank is applied to the anode. The
pressure can be derived by using gas and the principles of
purified air from compressor is also delivered to the cathode
the mass conservation law. The time derivative of the partial
side. The hydrogen and oxygen flow regulator designed by
pressure of the hydrogen molar flow rate can be balanced
PID controller maintains desired cathode and anode pressure
by the input flow rate, reacted hydrogen flow rate and the
in PEMFC system. In our evaluation, the gas reformed is
output flow rate as 2
modeled as qh2 = 4s+4s+1 with Fcf = 2, τr2 = 4 and τr1 =
RT 4. The gains for PID controller is considered as k1 = 0.25,
Ṗh2 = (qh2in − qh2out − qh2react) (14)
Va k2 = 0.05 and k3 = 0.125. Notice from equation (7), (9),
with qh2in is the inlet flow rate; qh2out is the outlet flow rate; (11), (12), (13) that the output response of the fuel cell stack
qh2react is the used or reacted hydrogen flow rate; Va is the mainly depends on the reactant pressure, load current and
volume of the anode [l]. Similarly, the time derivate of the stack temperature. In our evaluation, the hydrogen, oxygen
partial pressure of the oxygen is governed by and water vapor pressure is assumed to be [Link],
any variations in the load communication directly effect the
RT fuel cell current. Heat generation/dissipation in the stack
Ṗo2 = (qo2in − qo2out − qo2react) (15)
Vc increase with the increase of the load current which in
where Vc is the volume of the cathode; qo2in is the inlet turn increase the system temperature. To predict the fuel
flow; qo2out is the outlet flow; qo2out is the used or reacted cell stack temperature, a number of thermal model can be
oxygen flow rate. The time derivate for the partial pressure found in the literature [37,38]. These lumped models are
of the hydrogen inside the fuel cell can be written as based on the thermal energy balance principle, where the
RT input hydrogen energy into fuel cell stack is consumed by
Ṗh2o = (qh2ogen − qh2oout + qh2ombr ) (16) electrical load, heat exchanger and fuel cell stack. However,
Vc
actual values of the empirical parameters used in existing
with qh2ogen is the generated water flow; qh2oout is the outlet thermal models may not be possible to realize in real time
flow; qh2ombr is the mass flow of water across the membrane. operation. High stack temperature may increase the ohmic
Using the property of the basic electrochemical reaction, the voltage drop, degrading the stack performance, reducing the
rate of oxygen and hydrogen reacted and water generated realizability and lifetime of the fuel cell system. Therefore,
inside the fuel cell during electrochemical reaction can be maintaining an optimal operating stack temperature plays
calculated as an important role in achieving desired output performance
 
N of the PEMFC system. So, in our evaluation, we aim to
qh2react = 2qo2react = qh2ogen = if (17)
2F examine the dynamic limitation of the PEMFC system in

1239

Authorized licensed use limited to: UNIVERSITY OF ROCHESTER. Downloaded on August 29,2020 at [Link] UTC from IEEE Xplore. Restrictions apply.
 0.05 58 800
1000

0.04 56
600 800

Nerns’ts Voltage
Ohomic Voltage

Load Current

Temperature
54

Variation

Variation
0.03 600

ENernst
ohm
η 52 400
400
0.02 50
200
48 200
0.01
46 0 0
0
0 10 20 30 40 50 10 20 30 40 50 0 10 20 30 40 50 0 10 20 30 40 50

800 58 0.2 0.2

56
600 0.15 0.15

Output Voltage
Load Current

54
Variation

H2 Flow

O2 Flow
400

Rate

Rate
52

Vsf
0.1 0.1
50
200
48 0.05 0.05
0 46
0 0
0 10 20 30 40 50 10 20 30 40 50 0 10 20 30 40 50 0 10 20 30 40 50
Time in sec. Time in sec. Time in sec. Time in sec.

Fig. 2. The output voltage response in [V] and the load current [A] Fig. 5. The output voltage response in [V] and the load current [A]
variation. variation.

0.03 60

0.02

previous evaluation. The obtained results are shown in Fig.

55
Ohomic Voltage

Nerns’ts Voltage

0.01 50
Nernst
ηohm

3. In third case, both the temperature and load current is

0 45

−0.01 40

−0.02
0 10 20 30 40 50
35
10 20 30 40 50
considered to be time varying as shown in the first row of
1000

800
60

55
the Fig. 4. The model and control design parameters were
Output Voltage

kept similar to our last evaluation. The simulation results are

Temperature

50
Variation

600
sf
V

400 45

200

0
40

35
given in Fig. 4 and Fig. 5. These evaluation results show the
0 10 20 30 40 50 10 20 30 40 50

significant impact of the input temperature and load current

on the output performance of the PEMFC power system. It is
Fig. 3. The output voltage response in [V] and the load current [A]
variation.
noticed from our evaluations that the fuel cell output response
reaches to a steady state value at about two sec. From Figs.
2 to 4, we can observe that the Nernst’s voltage decreases
the presence of the temperature and load variations. To with the increase in temperature and load current which in
explore that, we consider three cases. In first case, the load turn reduces the fuel cell output voltage. On the other hand,
current is kept constant as rated value as 95 [A], while the the resistive voltage drop decreases with the increase of the
temperature variation is given in Fig. 2. The parameters used temperature and current variation. It can be seen from Figs.
in this evaluation is given as follows: Faraday’s constant 2 to 4 that the resistive losses can be reduced significantly
(F) 96484600 [C/kmol], Universal gas constant (R) 8314.47 by operating the fuel cell stack at an optimal operating point
[J/(kmol K), No load voltage, (Eo ) 1.229 [V], Number of of the load current and temperature. This implies that the
cells (N) 42, Hydrogen valve constant (KH2 ) 4.22 × 10−5 fuel cell should be operated at the rated current and higher
[kmol/(s atm], Oxygen valve constant (KO2), 2.11 × 10−5 temperature than the standard value of temperature 298 [K].
[kmol/(s atm], Water valve constant (KH2O ) 7.716 × 10−6
IV. C ONCLUSION
[kmol/(s atm)], Hydrogen time constant (KH2 ) 3.37 [sec.],
Water time constant 18.418 [sec.], Response time for water In this paper, we have developed PID algorithm to analyze
flow 6.74 [sec.], Utilization factor (U) 0.8, Activation voltage the impact of the temperature and load current variation
constant (B) 0.04777 [A−1], Activation voltage constant on the output dynamic characteristics of the PEMFC power
(C) 0.0136 [V], Ratio of hydrogen to oxygen (rhof ) 1.168, system. Our evaluation showed that the membrane resistance
Methane reference signal (qmethr ) 0.15. With these above increase with the increase of the dryness/dehydration level
design parameters, we then simulate the model and control of the membrane electrolyte. The membrane resistance may
algorithm introduced in section II. The evaluation results are also increase for a short time for the slow response of the
presented in Fig. 2. Let us now assume that the temperature fuel or oxidant flow rate to the change in the load and
is constant as 298 [K]. In this evaluation, the load current is temperature. This may be because of the fact that the fuel
assumed to be time varying as depicted in Fig. 3. The design or oxidant flow cannot response instantly. As a result, the
parameters for the modeling and simulation are similar to our existence of the time delay to the change of the current and
temperature increase the dehydration level of the membrane
electrolyte which in turn increase the voltage dropped for
0.05

0.04
0.1
short time degrading the PEMFC performance. The key
Activation Voltage

0.05
Ohomic Voltage

0.03
problem with the variation of the membrane dehydration
ohm

act

0
η

0.02

0.01
−0.05
level is that it may severely and permanently damage the
0 −0.1
0 10 20 30 40 50 0 10 20 30 40 50
electro catalyst reducing the lifetime and makes unreliable
60 60

55 55 operation of the FC system.

Nernst’s Voltage

Output Voltage

50 50
Enernst

Vsf

45 45

40 40 R EFERENCES
35 35
10 20 30 40 50 10 20 30 40 50
Time in sec. Time in sec.
[1] US Dept. of Energy, NETL Inc. Study, Fuel cell Handbook, National
Energy Technology Laboratory, Morgantown, WV,October 2000.
Fig. 4. The output voltage response in [V] and the load current [A] [2] C. C. Chan, The State of the Art of Electric, Hybrid, and Fuel Cell
variation. Vehicles, Proceedings of the IEEE, Vol. 95, No. 4, 2007.

1240

Authorized licensed use limited to: UNIVERSITY OF ROCHESTER. Downloaded on August 29,2020 at [Link] UTC from IEEE Xplore. Restrictions apply.
 0.05

0.04
0.1
stack simulation model for control study on direct reforming molten

Activation Voltage
0.05
Ohomic Voltage
ηohm 0.03
carbonate fuel cell power plant, IEEE Trans. on Energy Conversion,

ηact
0
0.02 Vol. 14 , no. 4 , 1999, pp. 1651-1657.
−0.05
0.01

0 −0.1
[18] M. Y. El-Sharkh, A. Rahman, M. S. Alam, P. C. Byrne, A. A. Sakla,
0 10 20 30 40 50 0 10 20 30 40 50
T. Thomas, A dynamic model for a stand-alone PEM fuel cell power
60 60 plant for a stand-alone PEM fuel cell power plant for residential
55 55
applications, Journal of Power Sources, Vol. 138, pp. 199-204, 2004.
Nernst’s Voltage

Output Voltage
50 50
nernst

[19] J. Padulles, G. W. Ault, and J. R. McDonald, An Approach to

Vsf
E

45 45

40 40 the Dynamic Modeling of Fuel Cell Characteristics for Distributed

35
10 20 30
Time in sec.
40 50
35
10 20 30
Time in sec.
40 50 Generation Operation, IEEE PES Meeting, 2000, pp. 134-138.
[20] S. Pasricha, and S. R. Shaw, A dynamic PEM fuel cell model, IEEE
Trans. Energy Conversion, Vol. 21, Issue 2, 2006, pp. 484-490.
Fig. 6. The output voltage response in [V] and the load current [A] [21] J. Purkrushpan, A. G. Stefanopoulou, and H. Peng, Modeling and
variation. control for PEM fuel cell stack system, Proc. Amer. Control Conf.,
Anchorage, 2002, pp. 3117-3122.
800
[22] C. Wang, M. H. Nehrir, and S. R. Shaw, Dynamic odel and model
1000

600 800
validation for PEM fuel cells sing electrical circuits, IEEE Trans.
Load Current

Temperature

Energy Convers., Vol. 20, no. 2, 2005, pp. 442-451.

Variation

Variation

600
400
400
200
200
[23] K. Agbossuo, R. J. Chahine, F. Hamelin, J. Laurencelle, Renewable
0

0 10 20 30 40 50
0

0 10 20 30 40 50
energy systems based on hydrogen for remote applications, Journal of
Power Sources, 2001, Vol. 96, pp. 168-72.
0.2 0.2
[24] D. J. Hall and R. G. Colclaser, Transient modeling and simulation of
0.15 0.15
tubular solid oxide fuel cells, IEEE Trans. Energy Conversion, vol.
H2 Flow

O2 Flow
Rate

Rate

0.1 0.1
14, pp. 749-753, 1999.
0.05 0.05
[25] E. Achenbach, Three-dimensional and time-dependent simulation of a
0 0
0 10 20 30
Time in sec.
40 50 0 10 20 30
Time in sec.
40 50
planar SOFC stack, J. Power Sources, vol. 49, 1994.
[26] K.-H. Hauer, Analysis tool for fuel cell vehicle hardware and soft-
ware (controls) with an application to fuel economy comparisons of
Fig. 7. The output voltage response in [V] and the load current [A] alternative system designs, Ph.D. dissertation, Dept. Transportation
variation. Technology and Policy, University of California Davis, 2001.
[27] M. D. Lukas, K. Y. Lee, and H. Ghezel-Ayagh, An explicit dynamic
model for direct reforming carbonate fuel cell stack, IEEE Trans.
[3] K. Rajashekara, J. Grieve, and D. Daggett, Hybrid fuel cell power in Energy Conversion, vol. 16, pp. 289-295, 2001.
aircraft, IEEE Ind. Appl. Mag., vol. 14, no. 4, pp. 54-60, JulyAug. [28] J. Padulls, G. W. Ault, and J. R. McDonald, An approach to the
2008. dynamic modeling of fuel cell characteristics for distributed generation
[4] A. Emadi, Y. J. Lee, and K. Rajashekara, Power electronics and motor operation, J. Power Sources, 2000.
drives in electric, hybrid electric, and plug-in hybrid electric vehicles, [29] L. J. Blomen and M. N. Mugerwa, Fuel Cell Systems, New
IEEE Trans. Ind. Electron., vol. 55, no. 6, pp. 2237-2245, 2008. York:Plenum, 1993, pp. 37-69.
[5] S. Eccarius, F. Krause, K. Beard, and C. Agert, Passively operated [30] L. Guzella and A. Sciarretta. Vehicle Propulsion Systems: Introduction
vapor-fed direct methanol fuel cells for portable applications, J. Power to Modeling and Optimization. Springer, 2005.
Sources, vol. 182, no. 2, pp. 565-579, 2008. [31] C. Lina, J. Kang, J. Grizzle, and H. Peng. Energy management strategy
[6] C. Kim, K. J. Kim, and M. Y. Ha, Investigation of the characteristics for a parallel hybrid electric truck. In SAE 2000 World Congress,
of a stacked direct borohydride fuel cell for portable applications, J. Detroit, Michigan, USA, 2000.
Power Sources, vol. 180, no. 1, pp. 114-121, 2008. [32] K. Wipke, M. Cuddy, and S. Burch, A user-friendly advanced power-
[7] R. Von Helmolt and U. Eberle, Fuel cell vehicles: Status 2007, J. train simulation using a combined backward/forward approach, IEEE
Power Sources, vol. 165, no. 2, pp. 833-843, 2007. Transactions on Vehicular Technology, Vol. 48, no. 6, pp. 1751-1761,
1999.
[8] D. Lecocq, Tomorrows energy today with Axane, Fuel Cells Bull.,
[33] K. Wipke, T. Markel, and D. Nelson. Optimizing energy management
vol. 2004, pp. 13-15, 2004.
strategy and degree of hybridization for a hydrogen fuel cell suv, In
[9] S. Varigonda and M. Kamat, Control of stationary and transportation
International Electric Vehicle Symposium EVS 18, Berlin, Germany,
fuel cell systems: Progress and opportunities, Comput. Chem. Eng.,
2001.
vol. 30, no. 1012, pp. 1735-1748, 2006.
[34] Z. Rahman, K. Butler, and M. Ehsani. A comparison study between
[10] S. Y. Choe, J. W. Ahn, J. G. Lee, and S. H. Baek, Dynamic simulator
two parallel hybrid control concepts. In Proceedings of the American
for a PEM fuel cell system with a PWM dc/dc converter, IEEE Trans.
Control Conference, Arlington, VA, USA, 2001.
Energy Conversion, vol. 23, no. 2, pp. 669-680, 2008.
[35] Ballard Power Systems Inc., Canada, [Link]
[11] S. Jain and V. Agarwal, An integrated hybrid power supply for
[36] G. Hoogers, Fuel Cell Technology Handbook. Boca Raton, FL: CRC
distributed generation applications fed by nonconventional energy
Press, 2003.
sources, IEEE Trans. Energy Conversion, vol. 23, no. 2, pp. 62-631,
[37] J. C. Amphlett, R. F. Mann, B.A. Peppley, P. R. Roberge, A. Rodrigues,
2008.
A model predicting transient response of proton exchange membrane
[12] M. C. Pera, D. Candusso, D. Hissel, and J. M. Kauffmann, Power
fuel cells, Journal of Power Sources, Vol. 61, pp. 183-188.
generation by fuel cells, IEEE Ind. Electron. Mag., vol. 1, no. 3, pp. [38] F. Laurencelle, R. Chahine, J. Hamelin, K. Agbossou, M. Fournier, T.
28-37, 2007.
K. Bose, A. Laperriere, Characterization of a Ballard MK5-E Proton
[13] J. Larminie, A. Dicks, Fuel cell systems explained, John Wiley and Exchange Membrane Fuel Cell Stack, Fuel cell, Vol. 1, no. 1, 2001.
Sons Ltd, West Sussex, England, 2001. [39] Fuel Cells 2000s Fuel Cell Basics,[Link]
[14] R. F. Mann, J. C. Amphlett, M. Hooper, H. M. Jensen, B. A. Peppley, [40] FuelCell Today, [Link]
P. R. Roberge, Development and application of a generalized steady cells/technologies/pemfc.
state electrochemical model of a PEM fuel cell, Journal of Power [41] S. M. Sharifi Asl, S. Rowshanzamir, M. H. Eikani, Modelling and
Sources 2000, Vol. 86, pp. 173-180. simulation of the steady-state and dynamic behaviour of a PEM fuel
[15] J. C. Amphlett, R. M. Baumert, R. F. Mann, B. A. Peppley, P. R. cell, Energy, Vol. 35, no. 4, pp. 1633-1646, 2010.
Roberge, T. J. Harries, Performance modelling of the Ballard mark- [42] V. Das, S. Padmanaban, K. Venkitusamy, R. Selvamuthukumaran, F.
IV solid polymer electrolyte fuel cell, Journal of the Electrochemical Blaabjerg, P. Siano, Recent advances and challenges of fuel cell based
Society, Vol. 142, pp. 9-15. power system architectures and control-A Review, Ren. and Sust. Ene.
[16] L.-Y. Chiu, B. Diong, and R. S. Gemmen, An Improved Small-Signal Re.h, Vol. 73, pp. 10-18, 2017.
Model of the Dynamic Behavior of PEM Fuel Cells, IEEE Trans. on
Industry Applications, Vol. 40, no. 4 , July-Aug. 2004 pp. 970-977.
[17] M. D. Lukas, K. Y. Lee, and H. Ghezel-Ayagh, Development of a

1241

Authorized licensed use limited to: UNIVERSITY OF ROCHESTER. Downloaded on August 29,2020 at [Link] UTC from IEEE Xplore. Restrictions apply.