Energy Conversion and Management 76 (2013) 746–752

Contents lists available at ScienceDirect

Energy Conversion and Management journal homepage: www.elsevier.com/locate/enconman

Hydrogen production by onboard gasoline processing – Process simulation and optimization
Vega Bisaria, R.J. Byron Smith ⇑
Process Systems Laboratory, School of Mechanical and Building Sciences, Vellore Institute of Technology University, Vellore 632014, Tamil Nadu, India

a r t i c l e

i n f o

Article history:
Received 5 September 2012
Accepted 2 August 2013

Keywords:
Onboard fuel processor
Hydrogen
Fuel cell vehicles
Process Simulation

a b s t r a c t
Fuel cell vehicles have reached the commercialization stage and hybrid vehicles are already on the road.
While hydrogen storage and infrastructure remain critical issues in stand alone commercialization of the technology, researchers are developing onboard fuel processors, which can convert a variety of fuels into hydrogen to power these fuel cell vehicles. The feasibility study of a 100 kW on board fuel processor based on gasoline fuel is carried out using process simulation. The steady state model has been developed with the help of Aspen HYSYS to analyze the fuel processor and total system performance. The components of the fuel processor are the fuel reforming unit, CO clean-up unit and auxiliary units. Optimization studies were carried out by analyzing the influence of various operating parameters such as oxygen to carbon ratio, steam to carbon ratio, temperature and pressure on the process equipments. From the steady state model optimization using Aspen HYSYS, an optimized reaction composition in terms of hydrogen production and carbon monoxide concentration corresponds to: oxygen to carbon ratio of
0.5 and steam to carbon ratio of 0.5. The fuel processor efficiency of 95.98% is obtained under these optimized conditions. The heat integration of the system using the composite curve, grand composite curve and utility composite curve were studied for the system. The most appropriate heat exchanger network from the generated ones was chosen and that was incorporated into the optimized flow sheet of the100 kW fuel processor. A completely heat integrated 100 kW fuel processor flow sheet using gasoline as fuel was thus successfully simulated and optimized.
Ó 2013 Elsevier Ltd. All rights reserved.

1. Introduction
The provision of reliable, affordable and clean energy services is critically important for achieving the objectives of sustainable development. Of the many sustainable energy pathways that have emerged, a hydrogen-based energy system has received particular attention. The hydrogen economy is regarded as a viable and preferable option for delivering high-quality energy services in an efficient, clean and safe manner while generating little or no polluting emissions at the point of use. The last two decades have seen significant progress in hydrogen as an energy carrier as a result of concerns about climate change and diversifying of energy resources. The development of fuel cell technology has opened a

Abbreviations: ATR, auto thermal reforming; FCV, fuel cell vehicle; HEN, Heat
Exchanger Network; H2O/C, steam to carbon ratio; H2O/CO, steam to carbon monoxide ratio; HTS, high temperature shift; LTS, low temperature shift; O/C, oxygen to carbon ratio; PEMFC, proton exchange membrane fuel cell; POX, partial oxidation; SR, steam reforming; WGS, water gas shift.
⇑ Corresponding author. Address: Tridiagonal Solutions Pvt Ltd., Pune 411007,
Maharashtra, India. Tel.: +91 7387284114.
E-mail address: byron.smith@tridiagonal.com (R.J.B. Smith).
0196-8904/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.enconman.2013.08.006 new opportunity for the use of hydrogen fuel. Further development of this technology will provide a wide range of applications from powering vehicles to supplying electricity and heat with environmental advantages over other energy technologies.
The success of hydrogen in the transportation sector is dependent on the development and commercialization of Fuel Cell Vehicles (FCV). In these vehicles a fuel cell generates onboard electricity needed to power an electric drive which in turn is fed with hydrogen. It converts stored energy within a fuel into usable energy using an electrochemical reaction to extract energy directly in the form of heat and electricity which can be utilized at the point of generation. Since it is a direct energy conversion process, they are able to achieve much higher conversion efficiencies. According to the National Academies of Science, fuel cell vehicles can reduce light duty demand for gasoline and reduce CO emissions by 80%.
Fuel Cell vehicles are 2–3 times more efficient than current internal combustion vehicles [1]. Major automobile manufacturers including General Motors, Volkswagen, Volvo, Chrysler, Nissan and Ford have announced plans to build prototype polymer electrolyte fuel cell vehicles operating on hydrogen, methanol or gasoline. IFC, Plug Power, Nuvera and Ballard Power Systems are involved in separate programs to build 50–100 kW fuel cell

V. Bisaria, R.J.B. Smith / Energy Conversion and Management 76 (2013) 746–752

systems for vehicle motive power. Other fuel cell manufacturers are also involved in similar vehicle programs.
In order to overcome the most critical issue of handling the hydrogen storage and distribution, on board fuel processors that would generate hydrogen from liquid hydrocarbon fuels have been proposed as possible near term solutions. Feed stocks that can be used for onboard hydrogen generation through fuel cells include n-heptane, gasoline, ethanol, methanol, diesel, methane, butane, propane, ethane, propanol and jet fuel. The liquid fuels show advantages in terms of high energy density, easy fuel handling and in the case of gasoline – an existing fuel infrastructure [2].
All these fuels are hydrocarbons or oxygenate that need to be reformed. A comparison of hydrogen, methanol and gasoline as fuels for fuel cell vehicles and their implications for vehicle design and infrastructure development was studied by Joan et al. [3]. Hydrogen productions from methanol reforming have been discussed by Shetian et al. [4] and Vinay and Srinivas [5]. Reuse et al. [6] have reported on hydrogen production with Steam Reforming (SR) using methanol in an auto thermal micro-channel reactor. Diesel as hydrogen source has been investigated by Lindstrom et al. [7]; Inyong et al. [8] and Philip et al. [9]. Ethanol reforming has been studies by Chiu et al. [10]. Ganesh and Baskar [11] have investigated the use of dry auto thermal reforming of gasoline fuel. Brian et al. [12] gives the insight of the Renault Nuvera program on onboard fuel processors which discusses the advanced catalysts employed in current technology, heat exchangers, and controls in a system that is small enough and powerful enough for use on a vehicle. A comparison of compact reformer configurations for onboard fuel processing using catalytic steam reforming of n-heptane was done by Mustafa Karakaya and Ahmet [13]. A review of energy sources and management systems for electric vehicles has been published by Siang and Chee [14].
The conventional fuels must be reformed in order to generate hydrogen on board. The fuel processor are based on any available reforming technology like steam reforming, autothermal reactor or partial oxidation followed by water gas shift reactors in a modular package. Due to the complex nature of fuel processors, the thermal management system is not very straightforward and analysis of the fuel processor system for vehicles has to be done in the context of the overall system. In this paper we analyze the process feasibility of a conventional fuel (gasoline) based onboard fuel processor to generate hydrogen for fuel cell vehicles using the steady state process simulator Aspen HYSYS.

2. Process flow sheet
The process for production of hydrogen consists of a fuel reformer, CO cleanup unit and a hydrogen separation unit. The balance of plant involves pumps, vapourizers, mixers and a series of heat exchangers that manage heat utilization in the fuel processor assembly. The conventional flow sheet for this process is depicted in Fig 1. In the reformer section, the higher hydrocarbon is broken into lower hydrocarbons and to end products. The reactions taking place in this section are endothermic in nature. Carbon monoxide is also formed here which acts as a poison to the anode of PEM fuel cell and hence to remove the CO content and enrich hydrogen, the
CO cleanup section follows the reformer. The ensuing gases are segregated in a membrane unit where hydrogen is separated and fed to the PEM fuel cell unit whereas the exhaust gases are used as heating medium inside the unit before being send out.
There are a number of fuel processing technologies for hydrogen generation from hydrocarbon fuels. The three major thermochemical reforming techniques used to produce hydrogen from hydrocarbon fuel are steam reforming, partial oxidation (POX) and autothermal reforming (ATR). Steam reforming can yield high

747

concentration of hydrogen (up to 70% on a dry basis), but is strongly endothermicand hence the overall configuration of steam reformer with heat exchangers makes the reforming system very bulky and heavy which is not suitable for an modular automobile fuel cell system. The POX process is exothermic and easily starts upon ignition even without the aid of a catalyst. It can raise the temperature to over 1000 °C, which permits adiabatic operation and promotes SR of the remaining fuel. However, POX produces high carbon monoxide concentration that is undesirable for PEMFCs. ATR combines the thermal effects of the POX and SR reactions by feeding the fuel, steam, and air together into the reactor. The two processes occur simultaneously in the presence of catalyst in the reactor. The thermal energy generated from POX is absorbed by SR and hence the overall temperature is lowered which is favorable to water–gas shift reaction to consume carbon monoxide and produce more hydrogen [15]. Hence the auto thermal reformer is more compact and practical for use with mobile fuel cells.
The carbon monoxide when present in concentration higher than 100 ppm poisons the proton exchange membrane fuel cells.
CO concentration can be reduced by using Pt/RHh catalysts or
Water Gas Shift Reaction (WGS). The use of platinum alloys is not as effective as the raising the operating temperature of a PEMFC is currently not feasible. Thus, setting up of water gas shift reactor appears to be the most practical method to mitigate the effect of CO poisoning. The WGS section in a fuel processor generally consists of low temperature shift (LTS) and high temperature shift
(HTS) reactors. The exit stream from the reformer goes to the
HTS which operates at about 250–400 °C and the outlet of HTS is sent to the LTS operating in the range of 150–250 °C [16]. The exit gases from this unit is send to a membrane separation unit were the hydrogen permeable membrane separates hydrogen and feeds it to the PEM fuel cell assembly. Thermal management is achieved by a battery of heat exchangers. A fuel processor of 100 kW power was to be simulated in this study as the fuel cell cars on road generally require this much power for their efficient running. Hence the selection of the operating parameters such as pressure, temperature, fuel utilization and the efficient placement of heat exchangers for thermal management were based on the fuel processor of 100 kW power output from the fuel cell.
3. Simulation and optimization
The hydrogen required for a fuel for 100 kW fuel processor was calculated based on a simple material balance and mole concept.
Based on the fuel cell handbook [17], for a typical PEM fuel cell the following parameters were chosen to calculate the hydrogen requirement for generation of 100 kW output power by the fuel cell. The parameters chosen were; fuel utilization efficiency = 60%, average cell voltage = 700 mV and current density = 0.5 A/cm2.
Using these parameters, the hydrogen requirement for the fuel cell was computed to be 8.96 kg/h. This quantity of hydrogen is to be produced by the fuel processor using gasoline as the feed stock.
Gasoline is a complex mixture of over 500 hydrocarbons of range
C5–C12. Iso-octane is the most prominent constituent of gasoline and is used in the simulation to represent gasoline.
3.1. Reformer
The ATR of octane involves a complex set of chemical reactions such as total oxidation, partial oxidation, steam reforming, methanation and CO2 reforming, cracking, and carbon gasification. When
SR and POX react simultaneously, the reformer efficiency also increases. The appropriate choice of O/C (Oxygen to carbon ratio) and H2O/C (Steam to carbon ratio) also allows the reaction to proceed without external heating. Important parameters for the

748

V. Bisaria, R.J.B. Smith / Energy Conversion and Management 76 (2013) 746–752

Fig. 1. Process flow sheet for onboard hydrogen production from gasoline.

reaction are temperature, pressure, O/C and H2O/C molar ratios.
These parameters should be chosen with the aim of avoiding the formation of carbonaceous deposits, maximum hydrogen production, hydrocarbon conversion and H2 + CO selectivity. In this study the ATR is approximated by Gibbs reactor as the stoichiometry of gasoline reforming is not fully understood. The condition that the
Gibbs free energy of the reacting system is at a minimum at equilibrium is used to calculate the product mixture composition in a
Gibbs reactor and it is not necessary to know the stoichiometry of the reaction taking place inside the reactor.
In autothermal reforming, the role of oxygen is to provide heat required to drive reforming reactions and to break fuel into smaller compounds. At high O/C ratio, the chemical energy of fuel may be converted to sensible heat rather than the desired H2 product and at low O/C ratio coke may form due to thermal cracking of hydrocarbons. As our desired product is hydrogen so a low value of O/C was to be considered [18]. Simulations were carried out by varying the O/C ratio between 0.1 and 1 by keeping the H2O/C ratio fixed at
0.5 and the exit gas composition from the reformer were plotted in
Fig. 2. The hydrogen and CO concentration were 0.2338 and 0.1099 respectively at O/C of 0.5 which is the maxima of hydrogen concentration and minima of CO concentration. Moreover, the steam content, methane, gasoline, carbon and CO mass fractions at the exit composition are all at their lower levels. The nitrogen from air and carbon dioxide from partial oxidation are higher and inevita-

ble. Thus an optimum O/C ratio of 0.5 is found from the simulation to be suitable for the operation of the ATR.
To analyze the influence of steam on the performance of the reactor, the O/C was kept as 0.5 and then the H2O/C was varied from 0.1 to 1 and again the exit gas composition from the reformer was plotted. These results are depicted in Fig. 3. The combined effect of H2O/C and O/C shows that the ATR should be operated at values of H2O/C = 0.5 and O/C = 0.5. The optimized results were in coherence with the study done by Aasberg Petersen et al. [19].
The reforming process favors high temperatures as it shifts the reforming reaction equilibrium towards the production of hydrogen and reduces methane. However, it is not advisable to operate the reformer above 1000 °C because the metallurgy of the catalyst tubes causes them to creep and bulge under the weight of the catalyst. In contrast, operating around 700 °C decreases hydrogen production and increases methane slip out of the reformer resulting in the waste of fuel. The only advantage of lowering temperature is a decrease in heat duty, which will reduce costs. The process is highly sensitive to temperature so it is desirable to operate reformer at a temperature as close to the metallurgical limit of 800 °C as possible in order to maximize H2 production [20]. The Fig. 4 shows the variation of the temperature of ATR with change in O/
C ratio and H2O/C ratio from 0.1 to 1 keeping the other values constant. The graph clearly indicates that when O/C ratio is increased the oxidation reactions predominates the steam reforming thereby

Fig. 2. Effect of O/C ratio on the exit gas composition of reformer for H2O/C = 0.5.

V. Bisaria, R.J.B. Smith / Energy Conversion and Management 76 (2013) 746–752

749

Fig. 3. Effect of H2O/C ratio on the exit gas composition of reformer for O/C = 0.5.

Fig. 4. Effect of varying O/C and H2O/C on the reformer exit temperature.

causing the temperature to increase continuously. The effect of variation of pressure on the composition of the outlet product gas was studied by varying the pressure from 1 to 4 atm. As can be seen from the graph in Fig. 5, the pressure does not play any major role as the compositions are almost constant. Hence the feed was input at 1 atm so as to reduce the extra energy needed for compression. 3.2. Water gas shift reactor
The ideal fuel for PEMFC is pure hydrogen, with less than
100 ppm carbon monoxide, as dictated by the poisoning limit of the Pt fuel cell catalyst. Therefore CO clean up section is crucial for the fuel processing. The operating variables are the H2O/CO ratio, reformate temperature and number of stages. The reaction is equi molar and therefore the effect of pressure on the reaction is

minimal. The equilibrium for H2 production is favoured by high moisture content and low temperature for the exothermic reaction. Steam to carbon monoxide ratio was analyzed in the range of 1.5–3.5. The exit gas compositions of the reactor were plotted against H2O/CO in Fig. 6. On analyzing the composition of the product gas after varying the H2O/CO from 1.5 to 3.5, the optimum value of 2.5 was taken.
The need for single stage WGS reactor and double stage WGS reactor were studied and a comparison was made between them with the hydrogen and CO compositions as the parameters. The double-stage WGS section consists of high-temperature shift
(HTS) and low-temperature shift (LTS) reactor while a single stage reactor consists of only LTS reactor. The resulting exit gas compositions by using a single reactor and two stage reactors showed that the composition of the double-stage WGS reactor consists of more amount of hydrogen and lesser amount of CO as desired.

750

V. Bisaria, R.J.B. Smith / Energy Conversion and Management 76 (2013) 746–752

Fig. 5. Effect of operating pressure on reformer exit gas composition.

Fig. 6. Effect of H2O/CO ratio on the exit gas composition of water gas shift reactor.

Moreover the single stage reactor unit was 82.06% efficient in removing CO while it was 94.38% for a double-stage unit. Hence the fuel processor with a double-stage unit was taken to further study. The exit gas composition of the shift reactor at various temperatures was plotted in Fig. 7. Based on the analysis, the feed temperature of the HTS is fixed at 350 °C and that of the LTS was fixed at 180 °C.
3.3. Heat integration
The heat integration on the process flow sheet is done by using the pinch analysis software Aspen Energy Analyzer. The unique features of Aspen Energy Analyzer are its ability to extract information from existing HYSYS cases. The heat integration in this software is based on pinch technique. Pinch Analysis is a Heat
Exchanger Network (HEN) optimization algorithm used for reducing energy consumption in processes by setting feasible energy tar-

gets and achieving them through optimizing the heat recovery systems, energy supply methods and process operating conditions for energy reduction.
The heat integration analysis includes the generation of the hot and cold composite curves and the grand composite curves. These curves help us in determining the quantity of hot/cold utilities required and the possible process to process heat transfer. The composite curve analysis showed that process to process heat transfer is sufficient for the heating load and cold utility is required to cool some streams. Based on the flowsheet for the onboard fuel processor, five alternative heat integrated flow sheet options were developed. The various designs were evaluated on the basis of minimum number of utility required and network cost index and performance. The selected network shows the requirement of 7 heat exchangers and two cold utility streams. The gases from the ATR are to be split into three streams and the LTS outlet split to two streams among the hot streams. The cold stream vapourizing

V. Bisaria, R.J.B. Smith / Energy Conversion and Management 76 (2013) 746–752

751

Fig. 7. Effect of temperature on exit gas composition from water gas shift reactor.

Fig. 8. Heat exchanger network analysis of the optimized process flow sheet.

water is split into two streams. This analysis provided the energy targets for the various streams available in the process flow sheet.
These energy requirements per stream from the HEN were used to rework the stream data in the process flow sheet to generate the heat integrated flow sheet for the fuel processor unit. The heat integrated process flow sheet is given in Fig. 8. This flow sheet is used to calculate the fuel and other inputs required by the fuel processor. 3.4. Optimized fuel requirement
Based on the optimized simulation it was estimated that for the production of 8.96 kg/h of hydrogen in the fuel processor which corresponds to 100 kW of power output from the fuel cell,
30.55 kg/h of gasoline, 94.02 kg/h of water and 146.94 kg/h of air are required. The performance of a fuel processor is measured by

its overall efficiency which is commonly defined as the ratio between the LHV of the hydrogen and carbon monoxide that are produced to the LHV of the fuel consumed. The efficiency of the fuel processor simulated by the process flow sheet was 95.98%. Thus a heat integrated process flow sheet for the production of hydrogen from gasoline fuel in a fuel processor was simulated under steady state and the operational parameters for the process were identified and reported. This requirement was equivalent to production of 100 kW of power output from the PEM fuel cell for the transportation sector.
4. Conclusion
This work discusses the steady state simulation of a 100 kW onboard fuel processing unit to generate hydrogen using conventional fuels such as gasoline. The fuel processor flow sheet

752

V. Bisaria, R.J.B. Smith / Energy Conversion and Management 76 (2013) 746–752

comprising the reforming and CO clean up section was generated using Aspen-HYSYS. The hydrogen generated from this fuel processor unit is assumed to be fed into a PEM fuel cell to generate
100 kW of power which is the requirement for driving a fuel cell powered vehicle. In order to obtain 100 kW power from this fuel processor, calculations were done and it was found out that
8.96 kg/h of hydrogen was needed to obtain 100 kW power. The
ATR section was optimized based on the operating variables temperature, pressure, H2O/C and O/C ratios. Both the H2O/C and O/C ratios were taken to be 0.5 because of high hydrogen and low CO concentration in the outlet composition. Also it was seen that pressure has no visible effect on the outlet composition so the system was run at atmospheric pressure in order to reduce the energy requirement for compression. Then the water gas shift section was optimized taking H2O/CO ratio, number of stages and temperature as the operating variables. The H2O/CO ratio was fixed as 2.5 based on the higher hydrogen and lower CO, water and nitrogen content. The extra steam needed for the reaction was injected directly to the system. The inlet temperature of HTS and LTS were taken as 350 °C and 180 °C based on higher hydrogen and lower CO content. The plant wide optimization was thus completed and the fuel processor was found to be 95.98% efficient in producing hydrogen. The actual fuel requirement to produce the desired amount of hydrogen was found to be 30.55 kg/h. Thus a heat integrated fuel processor of 100 kW capacity was successfully simulated and the optimized parameters for the operation of gasoline based onboard hydrogen generator were identified.
References
[1] Campanari Stefano, Manzolini Giampaolo, Iglesia Fernando Garcia de la.
Energy analysis of electric vehicles using batteries or fuel cells through well to wheel driving cycle simulations. J Power Sources 2009;186:464–77.
[2] Wang YH, Zhang JC. Hydrogen Production on Ni–Pd–Ce/Al2O3 catalyst by partial oxidation and steam reforming of hydrocarbons for potential applications in fuel cells. Fuel 2005;84:1926–32.
[3] Ogden Joan M, Steinbugler Margaret M, Kreutz Thomas G. A comparison of hydrogen, methanol and gasoline as fuels for fuel cell vehicles: implications for vehicle design and infrastructure development. J Power Sources
1999;79:143–68.

[4] Shetian L. Takahashi Katsumi, Uematsu Kazuo, Ayabe Muneo, ‘‘Hydrogen
Production by oxidative methanol reforming on Pd/ZnO’’. Appl Cat A: Gen
2005;283(1–2):125–35.
[5] Vadlamudi Vinay K, Palanki Srinivas. Modeling and analysis of miniaturized methanol reformer for fuel cell powered mobile applications. Int J Hydrogen
Energy 2011;36:3364–70.
[6] Reuse P, Ranken A, Santo KH, Oliver G, Schubert K. Hydrogen production for fuel cell application in an autothermal microchannel reactor. Chem Eng J
2004;101:133–41.
[7] Lindstrom, Karlsson JAJ, Ekdunge P, De Verdier L, Haggendal B, Dawody J, et al.
Diesel fuel reformer for automotive fuel cell applications. Int J Hydrogen
Energy 2009;34:3367–81.
[8] Kang1 Inyong, Carstensen Hans-Heinrich, Dean Anthony M. Impact of gasphase reactions in the mixing region upstream of a diesel fuel autothermal reformer. J. Power Sources 2011;196:2020–6.
[9] Engelhardt Philip, Maximini Marius, Beckmann Frank, Brenner Martin.
Integrated fuel cell APU based on a compact steam reformer for diesel and a
PEMFC. Int J Hydrogen Energy 2012;37(18):13470–7.
[10] Chiu Wei-Cheng, Horng Rong-Fang, Chou Huann-Ming. Hydrogen production from an ethanol reformer with energy saving approaches over various catalysts. Int J Hydrogen Energy 2013;38(6):2760–9.
[11] Kale Ganesh R, Kulkarni Bhaskar D. An alternative process for gasoline fuel processors. Int J Hydrogen Energy 2011;36(3):2118–27.
[12] Bowers Brian J, Zhao Jian L, Ruffo Michael, Khan Rafey, Dattatraya Druva,
Dushman Nathan, et al. On board fuel processor for PEM fuel cells’’. Int J
Hydrogen Energy 2007;32:1437–42.
[13] Karakaya Mustafa, Avci Ahmet K. Comparison of compact reformer configurations for on board fuel processing. Int J Hydrogen Energy
2010;35:2305–16.
[14] Tie Siang Fui, Tan Chee Wei. A review of energy sources and energy management system in electric vehicles. Renew Sustain Energy Rev
2013;20:82–102.
[15] Hagh Bijan F. Comparison of auto thermal reforming of hydrocarbons. Am
Chem Soc Div Fuel Chem 2004;49(1):144.
[16] Ersoz A, Olgun H, Ozdogan S, Gungor C, Akgun F, Tiris M. Auto thermal reforming as a hydrocarbon fuel processing option for PEM fuel cells. J Power
Sources 2003;118(1–2):384–92.
[17] Fuel cell handbook, 7th ed., EG&G Services Parsons Inc.; 2000.
[18] Wuyin Wang, Turn Scott Q, Keffer Vheissu, Douette Aurelien. Parametric study of auto thermal reforming of LPG. Am Chem Soc Div Fuel Chem
2004;49(1):142.
[19] Aasberg-Petersen Kim, Christensen Thomas S, Stub Nielsen Charlotte, Dybkjær
Ib. Recent developments in auto thermal reforming and performing for synthesis gas production in GTL applications. Fuel Chem Div Preprints
2002;47(1):96.
[20] Atilla Ersoza, Olguna Hayati, Ozdogan Sibel. Simulation study of a proton exchange membrane fuel cell system with auto thermal reforming. J Power
Sources 2006;154:67–73.