lable at ScienceDirect

Energy 80 (2015) 331e339

Contents lists avai

Energy

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

Evaluation of an integrated methane autothermal reforming andhigh-temperature proton exchange membrane fuel cell system

Suthida Authayanun a, Dang Saebea b, Yaneeporn Patcharavorachot c,Amornchai Arpornwichanop d, *

a Department of Chemical Engineering, Faculty of Engineering, Srinakharinwirot University, Nakhon Nayok 26120, Thailandb Department of Chemical Engineering, Faculty of Engineering, Burapha University, Chonburi 20131, Thailandc School of Chemical Engineering, Faculty of Engineering, King Mongkut's Institute of Technology Ladkrabang, Bangkok 10520, Thailandd Computational Process Engineering Research Unit, Department of Chemical Engineering, Faculty of Engineering, Chulalongkorn University,Bangkok 10330, Thailand

a r t i c l e i n f o

Article history:Received 28 August 2014Received in revised form24 November 2014Accepted 27 November 2014Available online 23 December 2014

Keywords:HT-PEMFC (high-temperature protonexchange membrane fuel cell)Autothermal reformingMethanePerformance analysis

* Corresponding author. Tel.: þ66 2 218 6878; fax:E-mail address: Amornchai.A@chula.ac.th (A. Arpo

http://dx.doi.org/10.1016/j.energy.2014.11.0750360-5442/© 2014 Elsevier Ltd. All rights reserved.

a b s t r a c t

The aim of this study was to investigate the performance and efficiency of an integrated autothermalreforming and HT-PEMFC (high-temperature proton exchange membrane fuel cell) system fueled bymethane. Effect of the inclusion of a CO (carbon monoxide) removal process on the integrated HT-PEMFCsystemwas considered. An increase in the S/C (steam-to-carbon) ratio and the reformer temperature canenhance the hydrogen fraction while the CO formation reduces with increasing S/C ratio. The fuel pro-cessor efficiency of the methane autothermal reformer with a WGS (water gas shift reactor) reactor, asthe CO removal process, is higher than that without a WGS reactor. A higher fuel processor efficiency canbe obtained when the feed of the autothermal reformer is preheated to the reformer temperature.Regarding the cell performance, the reformate gas from the methane reformer operated at Tin ¼ TR andwith a high S/C ratio is suitable for the HT-PEMFC system without a WGS reactor. When considering theHT-PEMFC system with a WGS reactor, the CO poisoning has less significant impact on the cell perfor-mance and the system can be operated over a broader range to minimize the required total active area. AWGS reactor is necessary for the methane autothermal reforming and HT-PEMFC integrated systemwithregard to the system efficiency.

© 2014 Elsevier Ltd. All rights reserved.

1. Introduction

Due to environmental problems and the energy crisis, manystudies have aimed to develop both effective and clean technolo-gies. Fuel cells are one of the most attractive power generationtechnologies that do not release pollution. Compared to other typesof fuel cells, the PEMFC (proton exchange membrane fuel cell) issuitable for portable and automotive application as well as smallscale power generation, such as residential application, due to itslow weight, volume and operating temperature, which provide forthe shortest start-up time [1]. On-site and on-board hydrogenproductions are the best choices during the development ofhydrogen infrastructure and storage.

Currently, natural gas is the most common fuel employed forhydrogen production because of its cost effectiveness for industrial

þ66 2 218 6877.rnwichanop).

hydrogen. Natural gas primarily consists of methane, as well asnitrogen, CO2, ethane, propane, butane, pentane and traces of othercomponents. The high hydrogen-to-carbon ratio of methane resultsin a product with a high hydrogen concentration. The steamreforming of methane reportedly yields the highest hydrogencontent of 75e78 vol.% [2]. When synthesis gas with a higherhydrogen concentration is fed to fuel cells, the efficiency of the fuelcell system is improved because the hydrogen fraction directly af-fects the performance of the fuel cell.

Many studies have focused on fuel processors integrated withPEMFC systems [2e5]. Eszos et al. [6] concluded that steamreforming and autothermal reforming are the most competitive fuelprocessing options in terms of fuel processing efficiencies and heatintegration within the PEMFC systems to provide a high PEMFCsystem efficiency. Ouzounidou et al. [7] studied a combined meth-anol autothermal steam reforming and PEM fuel cell system in orderto develop a model for exploring the interactions of the integratedsystem. In addition, Salemme et al. [8] investigated a steamreforming and autothermal reforming integratedwith PEMFC system

Fig. 1. Autothermal reforming integrated with HT-PEMFC system without a water gasshift reactor (case 1): (a) Tin ¼ 298 K and (b) Tin ¼ Tr.

S. Authayanun et al. / Energy 80 (2015) 331e339332

with various types of fuel. The methane steam reforming-basedsystem achieved the highest efficiency. Furthermore, Specchia et al.[9] revealed that the efficiencyof a steam reforming integratedwith aPEMFC system was higher than that of an autothermal integratedsystem. However, steam reforming systems are more complex interm of heat integration, which impacts the system start-up time.The very good dynamic response of the autothermal reformingprocess makes it suitable for mobile applications [10,11]. Given therequired fast start up and size and weight limitation for automobiles,the ATR (autothermal reforming) has an initially higher reaction ratethan the steam reforming and can be operated without an externalheat-supplying unit.

Because the CO (carbon monoxide) tolerance of the PEMFC islow, sophisticated CO removal units are needed to reduce theamount of CO (carbon monoxide) to less than 10 ppm. Recently, aHT-PEMFC (high-temperature proton exchange membrane fuel cell)was developed as a promising technology to handle the COpoisoning problem found in a conventional PEMFC [12,13]. Whenoperating at a high temperature, the quantity of CO that adsorbs tothe Pt catalyst in the HT-PEMFC decreases, and thus, the CO removalunits of the HT-PEMFC system are less complex than those of aconventional PEMFC system [14]. Presently, the HT-PEMFC inte-grated with reforming processes has been widely reported. Garde-mann et al. [15] studied the compact ethanol autothermal processorintegrated with the HT-PEMFC system for small scale power gen-eration. The advantages of such system are its compactness andstartup reliability. It was found that heating of the shift reactor toremove CO in the fuel processor is the most time-consuming step.Wichert et al. [16] developed the LPG fuel processor integrated withan HT-PEMFC system. The results showed that main reactors of thesystems can be operated with stable performance and were easy tocontrol during long-term experiments. Arsalis et al. [17] investigateda 1 kWe HT-PEMFC integrated with a fuel processor consisting of amethane steam reformer and a water gas shift reactor for Danishsingle-family households. The waste heat was applied for spaceheating and producing hot water. The obtained electrical efficiencyof the partial (25%) and full loads was 45.4% and 38.8%, respectively.As the CO removal unit is related to the fuel processing process andthe HT-PEMFC performance, the effect of its inclusion on the HT-PEMFC system efficiency should be clearly analyzed.

The aim of this study was to analyze and design an integratedautothermal reforming and HT-PEMFC system fueled by methane.To investigate the effect of a CO removal unit, two HT-PEMFC sys-tems are considered in this study. The first system consists of theHT-PEMFC and methane autothermal reformer without the COremoval process. The second involves a HT-PEMFC whose fuelprocessor consists of a methane autothermal reformer and a WGS(water gas shift reactor) to remove CO. The mathematical models ofthe HT-PEMFC are based on the electrochemical model and thediffusion model of the gas diffusion layer and electrolyte film layer.The effects of the operating parameters, such as the reformertemperature, inlet temperature and S/C (steam-to-carbon) ratio onthe reformer efficiency, hydrogen yield and CO content areanalyzed. The operating parameters considerably affect not onlythe system efficiency but also the complexity of the system. Theoptimal operating conditions of the autothermal reformer thatprovide a suitable product gas for HT-PEMFC are given. The effi-ciency and performance of the HT-PEMFC systems were analyzedby considering the required total active area of the HT-PEMFC ofeach designed HT-PEMFC system.

2. System description

The autothermal reforming was selected as a hydrogen pro-duction process because of its advantageous fast start up and low

energy requirement. In addition, the autothermal reforming pro-cess can be operated without an external heat supply, and thiscondition is referred to as “the adiabatic condition”. Two HT-PEMFCsystems are considered here. The first involves a HT-PEMFC and amethane autothermal reformer without a CO removal process, asshown in Fig. 1. To enhance the hydrogen concentration andimprove the overall system efficiency, water gas shift reactors wereadded to an integrated HT-PEMFC system, as presented in Fig. 2.The methane and oxidants, which are air and water, are fed to themethane autothermal reformer when the air feed content iscontrolled to supply the required heat of reforming via an oxidationreaction. Therefore, a burner for supplying heat to the reformingprocess was not added to this system. The inlet temperature of theautothermal reformer feed stream was set at the reactiontemperature and standard temperature (298 K). The recovery heatfrom the HT-PEMFC is used to preheat the reactant of the auto-thermal reformer when the inlet temperature of the autothermalreformer feed stream is equal to the reaction temperature, asshown in Figs. 1(b) and 2(b). In contrast, the recovery heat from HT-PEMFC is not necessary when the inlet temperature is specified at298 K, as shown in Figs. 1(a) and 2(a). The target power output ofthe HT-PEMFC systems is 50 kW for a small vehicle.

3. Modeling of HT-PEMFC system

3.1. Fuel processor

The equilibrium composition of the reformate gas obtained fromthe methane autothermal reforming was calculated by directlyminimizing the Gibbs free energy, as shown in Eq. (1). For themethane autothermal reforming process, the gaseous componentsin the system are CH4, H2, CO, CO2, O2, N2 and H2O.

minni

�Gt�

T ;P ¼XCi¼1

niGi ¼XCi¼1

ni

G

�i þ RT ln

f if �i

!(1)

where C is the total number of components in the reaction system,and ni is the moles of each gaseous component. Regarding the

Fig. 2. Autothermal reforming integrated with HT-PEMFC system with a water gasshift reactor (case 2): (a) Tin ¼ 298 K and (b) Tin ¼ Tr.

Table 1Electrochemical model of HT-PEMFC.

Parameters Model equations Ref.

Reversible cell potentialEr ¼ �

�DHTnF � TDST

nF

�þ RT

nF ln

"ðRTÞ1:5CH2�PtC0:5

O2�Pt

aH2O

#[19]

Anode activation losshact;a ¼ RT

aFsinh�1

i

2i0;að1�qCOÞ2

![20]

i0;a ¼ iref0;aac;aLc;a

CPtCref ;a

!ga

exp

"� Ec;a

RT

1� T

Tref ;a

!#

qCO ¼ a*ln ½CO�½H2 � þ b*lnðiÞ*ln ½CO�

½H2 � þ c

Cathode activation losshact;c ¼ RT

aFsinh�1�

i2i0;c

�[19]

i0;c ¼ iref0;cac;cLc;c

CPtCref ;c

!gc

exp

"� Ec;c

RT

1� T

Tref ;c

!#

Ohmic losshohmic ¼

�smlm

�i [19]

sm ¼ AT exp

��BRðTÞ

�

Table 2Model parameters used for the HT-PEMFC model.

Parameters Value Unit

Anode reference exchange current density, iref0;a1440 A m�2

Cathode reference exchange current density, iref0;c0.0004 A m�2

Anode catalyst surface area, ac,a 64 m2 g�1

Cathode catalyst surface area, ac,c 32.25 m2 g�1

Anode catalyst loading, Lc,a 0.2 mg cm�2

Cathode catalyst loading, Lc,c 0.4 mg cm�2

Transfer coefficient at anode, aa 0.5e

Transfer coefficient at cathode, ac 0.75e

Reaction order at the anode, ga 1e

Reaction order at the cathode, gc 1.375e

Anode reference concentration, cref,a 0.0002 mol dm�3

Cathode reference concentration, cref,c 0.0004 mol dm�3

Anode activation energy, Ec,a 16900 J mole�1 K�1

Cathode activation energy, Ec,c 72400 J mole�1 K�1

Anode reference cell temperature, Tref,a 433.15 KCathode reference cell temperature, Tref,c 373.15 K

S. Authayanun et al. / Energy 80 (2015) 331e339 333

conservation of atomic species, ni must satisfy the element balancein Eq. (2).Xi

niaik � Ak ¼ 0 ðk ¼ 1;2; :::;wÞ (2)

where aik is the number of atoms of element k in component i, Ak isthe total number of atoms of element k in the reactionmixture, andw is the total number of elements.

Lagrange's undetermined method was used to solve the opti-mization problem mentioned above and find the amounts ofgaseous components (inmoles) in the system at equilibrium [18]. Inaddition, the WGS reactor was also modeled as an equilibriumreactor, and its operating temperature was specified at 473.15 K.Only a water gas shift reaction occurs in this reactor. The fuelprocessor efficiency that accounts for the methane autothermalreformer and WGS reactor (in the case of HT-PEMFC system withWGS reactor) can be calculated from Eq. (3) when the heat fromhigh temperature reformate gases is recovered to the preheatedreactant with a heat exchanger.

hFP ¼ nH2$LHVH2

nCH4$LHVCH4

þ QP þ QR(3)

where nH2and nCH4

are the molar flow rates of produced hydrogenand used methane, respectively. LHVH2

and LHVCH4are the lower

heating value of hydrogen and methane, respectively. QP is the heatrequirement of the reforming process that accounts for the heat topreheat the reactant (no external heat input to maintain auto-thermal reformer because it is operated at the adiabatic condition,and the required heat is supplied by the oxidation reaction thatoccurs within the reactor). QR is the heat recovery from the hotreformate gas.

3.2. HT-PEMFC

Based on the electrochemical model, the operating cell voltage(Ecell) can be calculated by subtracting the reversible cell potential(Er), the maximum voltage that can be achieved by a fuel cell atspecific operating conditions, from various voltage losses.

Ecell ¼ Er � hact;a � hact;c � hohmic (4)

where hact,a is the activation loss at the anode, hact,c is the activationloss at the cathode and hohmic is the Ohmic loss. The model

describing the reversible cell potential and voltage losses of the HT-PEMFC based on a phosphoric acid doped polybenzimidazolemembrane is given in Table 1 in detail.

To reach the catalyst active surface, the gaseous reactants need totransport from the gas diffusion and the electrolyte film covering thecatalyst agglomerates [19]. The concentration of hydrogen and oxy-gen can be obtained from StefanMaxwell equation (Eq. (5) and Fick'slaw (Eqs. (6)e(7)). It is noted that the validation of the HT-PEMFCmodel used in this study can be seen in our previous work [20].

dXi

dz¼ RT

P

XXiNj � XjNi

Deffij

(5)

where N is the molar flux of each component at the anode andcathode sides of HT-PEMFC and LT-PEMFC, X is the mole fraction ofeach component, and Deff

ij is the diffusion coefficient calculatedwith the SlatteryeBird correlation [21]. The parameters used tocalculate the diffusion coefficient are shown in Table 2.

NO2

SPt�cathode¼

�DO2

�CO2�pt � CO2ðdissolveÞ

�dcathode

(6)

Table 3Required oxygen molar flow rate to sustain the heat requirement of autothermalreforming at the adiabatic condition.

TR (K) Required oxygen molar flow rate (mole/s)

Tin ¼ TR Tin ¼ 298 K

S/C ¼ 2 S/C ¼ 3 S/C ¼ 4 S/C ¼ 2 S/C ¼ 3 S/C ¼ 4

673 0.0684 0.0891 0.1078 0.2102 0.2733 0.3335773 0.1681 0.2101 0.2440 0.3923 0.4896 0.5734873 0.3370 0.3640 0.3762 0.6331 0.7004 0.7578973 0.4145 0.4093 0.4049 0.7427 0.8032 0.86621073 0.4258 0.4183 0.4128 0.8266 0.9018 0.9799

S. Authayanun et al. / Energy 80 (2015) 331e339334

NH2

SPt�anode¼

�DH2

�CH2�pt � CH2ðdissolveÞ

�danode

(7)

where N is the molar flux, CPt is the reactant concentration on thecatalyst surface, Cdissolve is the equilibrium reactant concentrationin the electrolyte film at the studied temperature, SPt is the realplatinum surface area and d is the film thickness. CPt is used tocalculate the theoretical voltage and activation loss in the electro-chemical model.

To evaluate the performance of the HT-PEMFC, the power den-sity (PFC) and the thermal efficiency as given in Eqs. (8)e(9) aredetermined.

PFC ¼ iEcell (8)

hFC ¼ PFC$Uf_mH2

$LHVH2

(9)

Fig. 3. A flow chart of the simulation approach of the integrated autothermalreforming process and HT-PEMFC system.

where _mH2is the hydrogen molar flow rate that is used in elec-

trochemical reaction, and Uf is the fuel utilization, which is definedby the ratio of consumed hydrogen to supplied hydrogen.

The system efficiency can be defined as follows:

hsys ¼ hFP$hFC (10)

4. Numerical solution

In this work, the mathematical models of the fuel processor andHT-PEMFC were coded in Matlab. A flow chart of the simulation

Fig. 4. (a) H2 and (b) CO fraction of methane autothermal reformer at differentoperating condition: Tin ¼ 298 K (Black line) and Tin ¼ TR (Red line). (For interpretationof the references to color in this figure legend, the reader is referred to the web versionof this article.)

S. Authayanun et al. / Energy 80 (2015) 331e339 335

approach of the integrated reforming process and HT-PEMFC sys-tem is shown in Fig. 3. First, the methane molar flow rate andoxygen molar flow rate are initially guessed, and the equilibriumcomposition and enthalpy balance of the autothermal reformer arethen calculated. The oxygen molar flow rate changes until theenthalpy balance between the inlet and outlet stream of theautothermal reformer is equal to zero. For the HT-PEMFC systemwithout the WGS reactor (case 1), the obtained H2 and CO fractionis applied to calculate the voltage loss, cell voltage, and poweroutput directly, whereas the product composition of the auto-thermal reformer is used to calculate the H2 and CO fractions fromthe WGS reactor for the HT-PEMFC system with the WGS reactor(case 2). The voltage loss, cell voltage, and power output are sub-sequently calculated. The calculation is repeated until the calcu-lated voltage is equal to the specified voltage.

5. Results and discussion

Integrated methane autothermal reforming and HT-PEMFCsystems with and without a water gas shift reactor were consid-ered. First, the effects of the reformer temperature, inlet reformertemperature and S/C on the H2 fraction, CO fraction and fuel pro-cessor efficiency were studied. The required oxygen flow rate tosustain the heat requirement of autothermal reforming at theadiabatic condition (external heat input is not needed because theheat required for the autothermal reformer is supplied by the

Fig. 5. (a) H2 and (b) CO fraction of methane autothermal reformer with WGS reactorat different operating condition: Tin ¼ 298 K (Black line) and Tin ¼ TR (Red line). (Forinterpretation of the references to color in this figure legend, the reader is referred tothe web version of this article.)

oxidation reaction that occurs within the reactor) is shown inTable 3. The required oxygen flow rate increases when the auto-thermal process is operated at high temperatures, and the excesssteam is fed to the process. However, this flow rate decreases as theinlet reformer temperature increases. This trend is similar to thatreported in the other studies on the autothermal reformer [22,23].In addition, the performances of the HT-PEMFC fueled by reformategas from the methane autothermal reformer with and withoutWGS reactor at different operating conditions were investigated.The optimal conditions and required total active area of the HT-PEMFC that provide the target power output of all designed sys-tem are also reported.

5.1. Fuel processor

Methane autothermal reformers without and with a WGSreactor were studied to observe the effect of the reformer oper-ating conditions on the H2 fraction, CO fraction and fuel processorefficiency, which significantly affect the HT-PEMFC performanceand overall system efficiency. The H2 and CO fraction of methaneautothermal reformers without and with a WGS reactor atdifferent reformer temperatures and S/C ratios are shown inFigs. 4 and 5, respectively. Two inlet reformer temperatures(Tin ¼ TR and Tin ¼ 298 K) were investigated in this work. For theautothermal reformer without a WGS reactor in Fig. 4, the

Fig. 6. The fuel processor efficiency of methane autothermal reformer at S/C ¼ 3 anddifferent reformer temperatures and inlet reformer temperatures: (a) without WGSreactor and (b) with WGS reactor.

Fig. 7. Performance of HT-PEMFC systemwithout WGS reactor at different Tin, TR and S/C ratios of the methane autothermal reformer: Tin ¼ 298 K (black line) and Tin ¼ TR (Red line).(For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

S. Authayanun et al. / Energy 80 (2015) 331e339336

hydrogen production and CO formationwere lower when the inletreformer temperature was equal to 298 K because some of the fuelwas used to sustain the adiabatic autothermal reformer, and theenergy needed for the reformer at low inlet reformer tempera-tures exceeds that at high inlet reformer temperatures. At Tin ¼ TR(Red line), an increase in the operating temperature and S/C ratioenhanced the hydrogen fraction, whereas the CO fractiondecreased at low temperatures and high S/C ratios. At Tin ¼ 298 K(Black line), increasing the S/C ratio increases the hydrogen frac-tion at low reformer temperatures, but the hydrogen fractiondecreases in response to increasing S/C ratios at high tempera-tures. This relationship arises because the heat required tomaintain the the autothermal reactor at Tin ¼ 298 K is higher thanthat at Tin ¼ TR, and thus the high amount of methane is consumedby the oxidation reaction instead of producing hydrogen via thereforming reaction. The hydrogen production from the methaneautothermal reformer with theWGS reactor shows the same trendas the other system, as shown in Fig. 5. However, the CO fractionseems to decease at high reformer temperatures when the inletreformer temperature is 298 K. Compared to methane auto-thermal reforming without a WGS reactor, the addition of a WGSreactor to the fuel processor improves the quality of the productgas for HT-PEMFC by means of increasing the hydrogen fractionand reducing the CO concentration. The optimal reformer tem-perature of both the autothermal reformer without and with aWGS reactor that provides the maximum hydrogen yield is in therange of 773e873 K.

The fuel processor efficiencies of the methane fuel processorswithout and with WGS reactor at different operating conditiondesigns are shown in Fig. 6(a) and (b), respectively. Heat is notneeded to preheat reactant at Tin ¼ 298 K. However, heat isrequired to preheat the reactant at Tin ¼ TR, and thus the fuelprocessor efficiency is calculated by considering heat recoveryfrom the hot reformate gases and without considering the heatrecovery (external heat is required). The simulation results

indicate that the fuel processor efficiency is higher for the auto-thermal reformer with a WGS reactor because of the highhydrogen production of this process. At high reformer tempera-tures (TR � 973 K), the methane autothermal reformer operated atTin ¼ TR with heat recovery shows the highest fuel processor ef-ficiency. However, the methane autothermal reformer should beoperated at TR ¼ 873 to achieve the highest fuel processor effi-ciency when the inlet temperature is set to 298 K for bothmethane autothermal reformers with and without a WGS reactor.The heat recovery can enhance the fuel processor efficiency,especially for high reformer temperature operation.

5.2. HT-PEMFC system

The product gases obtained from the hydrogen production unitat different operating conditions are fed to the HT-PEMFC, and theperformance of the HT-PEMFC was investigated while consideringthe CO poisoning effect. The polarization curve of the HT-PEMFCfueled by hydrogen from the methane autothermal reformerwithout a WGS reactor is presented in Fig. 7. The simulation resultsshow that the performance of the HT-PEMFC is higher at 773 K anda high S/C ratio due to the low CO concentration obtained from themethane reformer at this condition. Compared to the reformate gasfrom the methane reformer operated at Tin ¼ 298 K, the reformategas from themethane reformer operated at Tin¼ TR is more suitablefor a HT-PEMFC run at a voltage exceeding 0.3 V. However, theperformance of the HT-PEMFC operated on reformate gas from amethane reformer with Tin ¼ TR was lower than that of the othersystem at low voltage operation (high current density) andreformer temperature of 873 K due to the higher CO concentration,which significantly affects the cell performance, especially at a highcurrent density. Conversely, the HT-PEMFC system with a WGSreactor showed similar performances when the methane reformerwas operated at all studied reformer temperatures and S/C ratios(see Fig. 8). This similarity arises because the CO concentration in

Fig. 8. Performance of HT-PEMFC system with WGS reactor at different TR and S/C ratios of the methane autothermal reformer: Tin ¼ 298 K (black line) and Tin ¼ TR (Red line). (Forinterpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

S. Authayanun et al. / Energy 80 (2015) 331e339 337

the reformate gas from themethane reformerwith aWGS reactor islow and insignificantly impacts the HT-PEMFC. However, the sys-tem performed better at Tin ¼ TR than at an inlet reformer tem-perature of 298 K.

Figs. 9 and 10 show the required total active area of the HT-PEMFC system without and with a WGS reactor at differentreformer operating conditions to achieve the target power output(50 kW) when the cell is operated at a voltage of 0.6. Fig. 9 showsthat the studied HT-PEMFC system without a WGS reactor requiresa higher total active area to reach the target power output when theinlet reformer temperature is 298 K. In addition, the minimum totalactive area requirement is approximately 70 m2 for a HT-PEMFCsystem without a WGS reactor operated at TR ¼ 773e873 K, S/C¼ 3e4 and Tin ¼ TR. Operating the methane autothermal reformerat temperatures of 673 and 973 K results in a large total active arearequirement due to the low hydrogen and high CO concentration ofthe reformate gas from the methane autothermal reformer run atthese temperatures.

Fig. 9. The required total active area of 50 kW HT-PEMFC system without a WGSreactor at different reformer operating conditions.

For the HT-PEMFC systemwith aWGS reactor, Fig. 10 shows thatthe required total active area is in a similar range (70e85 m2) at allstudied reformer operating conditions because of the similar cellperformance obtained from the HT-PEMFC system with a WGSreactor. Furthermore, theminimum total active area requirement ofthe HT-PEMFC system with a WGS reactor is approximately 70 m2

for a system operated at TR ¼ 773e973 K, S/C ¼ 2e4 and Tin ¼ TR.The simulation results indicate that the HT-PEMFC systems withand without a WGS reactor require the same minimum total activearea to reach the target power output, but the HT-PEMFC systemwith a WGS reactor can be operated over a broader range.

At the desired power output, the system efficiencies of the HT-PEMFC system without and with a WGS reactor at variousreformer operating conditions are shown in Figs. 11 and 12,respectively. For the HT-PEMFC system without a WGS reactor, thesystem efficiency increases with the reformer temperature until itreaches an optimal level, and it then deceases with the reformer

Fig. 10. The required total active area of 50 kW HT-PEMFC system with a WGS reactorat different reformer operating conditions.

Fig. 11. System efficiency of HT-PEMFC systemwithout WGS reactor at different TR andS/C ratios of the methane autothermal reformer: (a) Tin ¼ 298 K and (b) Tin ¼ TR.

Fig. 12. System efficiency of HT-PEMFC system with WGS reactor at different TR and S/C ratios of the methane autothermal reformer: (a) Tin ¼ 298 K and (b) Tin ¼ TR.

S. Authayanun et al. / Energy 80 (2015) 331e339338

temperature for Tin ¼ 298 K and remains constant for Tin ¼ TR. Inaddition, the S/C ratio positively affects the system efficiency whenthe reformer temperature is lower than the optimal point. Whenthe temperature exceeds this point, the effect of the S/C ratio on thesystem efficiency is insignificant. The optimal efficiency is near 24%and can be obtained at a temperature of 873 K and all studied S/Cratios for Tin¼ 298 K (see Fig. 11). For Tin¼ TR, the optimal efficiencyis near 25% and can be obtained at temperatures at or above 873 K.

A similar trend was observed for the HT-PEMFC system with aWGS reactor, and the optimal operational range that yields thepreferable system efficiency is a temperature of 873 K and above inthe absence of a preheating reaction before it is fed to adiabaticautothermal reformer (Tin ¼ 298 K). However, a temperature of973 K and S/C ratio of 2 are preferred when the reactant of theautothermal reformer is preheated to the reformer temperature. Inaddition, the system efficiency of the HT-PEMFC systemwith aWGSreactor at both Tin ¼ TR and Tin ¼ 298 K is maximized near 30%.

6. Conclusions

A methane autothermal reformer integrated with a HT-PEMFCwas designed and analyzed in this work. Effect of the inclusion ofa CO removal unit in the HT-PEMFC system was studied. Thesimulation results showed that the inlet reformer temperaturestrongly affects the efficiency because it is a key factor for

controlling the energy use of the autothermal reformer. Thehydrogen production and CO formation are lower at an inletreformer temperature of 298 K for the autothermal reformerwithout a WGS reactor. The S/C ratio has a positive effect by bothincreasing the hydrogen fraction and decreasing the CO forma-tion, while an increase in the operating temperature can enhancethe hydrogen concentration. An almost identical trend was alsoobserved for the methane autothermal reformer with a WGSreactor, but CO the fraction decreased at high reformer temper-atures when the inlet reformer temperature was 298 K. The fuelprocessor efficiency of the methane autothermal reformer with aWGS reactor was higher than that for the system without a WGSreactor, and reformer operation at Tin ¼ TR is preferable to ach-ieve a high fuel processor efficiency. Considering the cell per-formance, the reformate gas from the methane reformeroperated at Tin ¼ TR and a high S/C ratio is suitable for HT-PEMFC.The minimum total active area required to reach the target po-wer output is similar for both HT-PEMFC systems with andwithout a WGS reactor, but the HT-PEMFC system with a WGSreactor can be operated over a broader range. The highest effi-ciencies of the HT-PEMFC systems without and with a WGSreactor are 25% and 30%, respectively. Considering the systemefficiency, it is found that a WGS reactor is necessary for themethane autothermal reforming and HT-PEMFC integratedsystem.

S. Authayanun et al. / Energy 80 (2015) 331e339 339

Acknowledgments

The support from the Thailand Research Fund, the Srinakhar-inwirot University and the Ratchadaphiseksomphot EndowmentFund of Chulalongkorn University (RES560530067-EN) is gratefullyacknowledged.

Nomenclatures

Cdissolve equilibrium concentration, mole cm�3

CPt concentration on the catalyst surface, mole cm�3

CrefPt reference concentration on the catalyst surface, mole

cm�3

Deffij binary diffusion coefficient, m2 s�1

Ecell cell voltage, VEr reversible cell potential, VF Faraday constant, 96,485C mol�1

G Gibbs free energy, J/molH enthalpy, J mol�1

i current density, A m�2

i0 exchange current density, A m�2

Lc catalyst loading, mg cm�2

LHV lower heating value, kJ mol�1

lm membrane thickness, mm molar flow rate, mol s�1

N molar flux, mol s�1 m�2

P pressure, atmPFC power output of fuel cell, WQP heat required for the steam reforming process, J s�1

QR heat recovery for the hot reformate gas, J s�1

R gas constant (¼8.314), J mol�1 K�1

SPt real platinum surface areaT cell temperature, KTin inlet reformer temperature, KTR reformer temperature, KX mole fraction

Greek lettersa transfer coefficientg reaction orderqCO CO coverageqH H2 coveragesm proton conductivity, S cm�1

hohmic ohmic loss, Vhact activation loss, Vhsys system efficiency

Subscripts and superscriptsa anodec cathodem membranei,j components “i” and “j”in inlet streamout outlet stream

References

[1] Tang Y, Yuan W, Pan M, Li Z, Chen G, Li Y. Experimental investigation of dy-namic performance and transient responses of a kW-class PEM fuel cell stackunder various load changes. Appl Energy 2010;87:1410e7.

[2] Heinzel A, Vogel B, Hubner P. Reforming of natural gas-hydrogen generationfor small scale stationary fuel cell system. J Power Sources 2002;105:202e7.

[3] Chutichai B, Authayanun S, Assabumrungrat S, Arpornwichanop A. Perfor-mance analysis of an integrated biomass gasification and PEMFC (proton ex-change membrane fuel cell) system: hydrogen and power generation. Energy2013;55:98e106.

[4] Silva AL, Dick LFP, Muller IL. Performance of a PEMFC system integrated with abiogas chemical looping reforming processor: a theoretical analysis andcomparison with other fuel processors (steam reforming, partial oxidationand auto-thermal reforming). Int J Hydrogen Energy 2012;37:6580e600.

[5] Stamps AT, Gatzke EP. Dynamic modeling of a methanol reformerdPEMFCstack system for analysis and design. J Power Sources 2006;161:356e70.

[6] Ersoz A, Olgun H, Ozdogan S, Gungor C, Akgun F, Tırıs M. Autothermalreforming as a hydrocarbon fuel processing option for PEM fuel cell. J PowerSources 2003;118:384e92.

[7] Ouzounidou M, Ipsakis D, Voutetakis S, Papadopoulou S, Seferlis P.A combined methanol autothermal steam reforming and PEM fuel cell pilotplant unit: experimental and simulation studies. Energy 2009;34:1733e43.

[8] Salemme L, Menna L, Simeone M. Calculation of the energy efficiency of fuelprocessor-PEM (proton exchange membrane) fuel cell systems from fuelelementar composition and heating value. Energy 2013;57:368e74.

[9] Specchia S, Cutillo A, Saracco G, Specchia V. Concept study on ATR and SR fuelprocessors for liquid hydrocarbons. Ind Eng Chem Res 2006;4:5298e307.

[10] Samsun RC, Pasel J, Janben H, Lehnert W, Peters R, Stolten D. Design and testof a 5 kWe high-temperature polymer electrolyte fuel cell system operatedwith diesel and kerosene. Appl Energy 2014;114:238e49.

[11] Sopena D, Melgara A, Briceno Y, Navarrob RM, Alvarez-Galvan MC, Rosa F.Diesel fuel processor for hydrogen production for 5 kW fuel cell application.Int J Hydrogen Energy 2007;32:1429e36.

[12] Chandan A, Hattenberger M, El-kharouf A, Du S, Dhir A, Self V, et al. Hightemperature (HT) polymer electrolyte membrane fuel cells (PEMFC) e a re-view. J Power Sources 2013;231:264e78.

[13] Singdeo D, Dey T, Ghosh PC. Modelling of start-up time for high temperaturepolymer electrolyte fuel cells. Energy 2011;36:6081e9.

[14] Kim J, Kim M, Kang T, Sohn YJ, Choi KH, Song T. Degradation modeling andoperational optimization for improving the lifetime of high-temperature PEM(proton exchange membrane) fuel cells. Energy 2014;66:41e9.

[15] Gardemann U, Steffen M, Heinzel A. Design and demonstration of an ethanolfuel processor for HT-PEM fuel cell applications. Int J Hydrogen Energy2014;39:18135e45.

[16] Wichert M, Men Y, O'Connell M, Tiemann D, Zapf R, Kolb G, et al. Self-sus-tained operation and durability testing of a 300 W-class micro-structured LPGfuel processor. Int J Hydrogen Energy 2011;36:3496e504.

[17] Arsalis A, Nielsen MP, Kær SK. Modeling and off-design performance of a 1kWe HT-PEMFC (high temperature-proton exchange membrane fuel cell)-based residential micro-CHP (combined-heat-and-power) system for Danishsingle-family households. Energy 2011;36:993e1002.

[18] Smith JM, Van Ness HC, Abbott MM. Introduction to chemical engineeringthermodynamics. 6th ed. McGraw-HILL International Edition; 2005.

[19] Mamlouk M, Sousa T, Scott K. A high temperature polymer electrolytemembrane fuel cell model for reformate Gas. Int J Electrochem 2011;2011:1e18.

[20] Authayanun S, Mamlouk M, Arpornwichanop A. Maximizing the efficiency ofa HT-PEMFC system integrated with glycerol reformer. Int J Hydrogen Energy2012;37:6808e17.

[21] Slattery JC, Bird RB. Calculation of the diffusion coefficient of dilute gases andof the self-diffusion coefficient of dense gases. AIChE J 1958;4:137e42.

[22] Souza TL, Rossi CCRS, Alonso CG, Guirardello R, Cabral VF, Fernandes-Machado NRC, et al. Thermodynamic analysis of autothermal reforming ofmethane via entropy maximization: hydrogen production. Int J HydrogenEnergy 2014;39:8257e70.

[23] Authayanun S, Arpornwichanop A, Paengjuntuek W, Assabumrungrat S.Thermodynamic study of hydrogen production from crude glycerol auto-thermal reforming for fuel cell applications. Int J Hydrogen Energy 2010;35:6617e23.