5298
Ind. Eng. Chem. Res. 2006, 45, 5298-5307
Concept Study on ATR and SR Fuel Processors for Liquid Hydrocarbons Stefania Specchia,* Angelo Cutillo, Guido Saracco, and Vito Specchia Dipartimento di Scienza dei Materiali ed Ingegneria Chimica, Politecnico di Torino, Corso Duca degli Abruzzi 24, 10129 Torino, Italy
Models for auxiliary power units (APU) based on-board hydrocarbon fuel processors (FP) for hydrogen production for polymer-electrolyte-membrane (PEM) fuel cells were coded in Matlab/Simulink software considering the following: (i) either an autothermal (ATR) or a steam reforming (SR) unit fed with several fuels (road distribution net fuels: gasoline, light diesel, heavy diesel, and biodiesel; reference fuels in internal combustion enginer (ICE) applications: isooctane for gasoline and cetane for diesel oils); (ii) the secondary units for the CO cleanup process (water gas shift and CO preferential oxidation reactors: WGS and COPROX); (iii) the auxiliary units for the balance of plant of the whole system (afterburner, heat exchangers, water recovery radiators, air compressor, and water and fuel pumps), necessary to properly operate the FP; and (iv) the PEM fuel cell. First, a study on the feasibility of the substitution during the simulations of the real liquid fuels available for automotive traction (mixtures of various organic compounds) with their reference fuel counterparts (pure substances normally used as references for ICE applications) enlightened that isooctane and cetane are not satisfactory substitutes of gasoline and diesel oils, respectively. Then, beyond the prevalent goal of comparing the ATR and the SR options, particular attention was paid to the models as design tools for the optimization of the FP scheme to obtain the highest possible energy conversion efficiency within the constraints imposed by, e.g., catalyst and materials durability and self-sustainability of the system as concerns energy and water balances. From the simulation results, gasoline emerges as the most effective fuel, whereas the highest performance, in terms of efficiency, seems to belong to the SR APU system (efficiency up to 39%), which is, though, characterized by a plant complexity higher than that of its ATR counterpart (efficiency slightly exceeding 36%). 1. Introduction A sustainable high quality of life cannot be kept apart from the worldwide supply of clean, safe, reliable, and secure energy, taking into account that the energy demand is growing more and more. The European “World Energy Technology and Climate Policy Outlook” (WETO) predicts, for primary energy, an average growth rate worldwide of 1.8% per year for the period 2000-2030.1 The increased demand will be largely satisfied by fossil fuel reserves that, however, (i) emit either greenhouse gases or other pollutants, (ii) are not unlimited, and (iii) will then become increasingly expensive. Besides, a prevalent issue in handling energy remains safety. Moreover, a reexamination of petroleum-fueled internal combution engine (ICE) vehicles as the basis for road transportation throughout the world is being pursued on the grounds of (i) the increasing concern about the environmental consequences of fossil fuels used for vehicles propulsion; (ii) the global earth warming; (iii) the growing dependence of the industrialized countries on fossil oils; and (iv) the increasing global awareness of how human activities influence the environment and how a sustainable development can be achieved with the increasing world population. Although modern cars emit far less toxic pollutants including hydrocarbons, nitrogen oxides, carbon monoxide, and particulates, their increasing number is resulting in increasing automobile pollution. This brought about always further and further restrictive emission legislations all over the world, in line with the Kyoto Protocol,2 which paves the way to the introduction of zero-emission vehicles (ZEVs). This results in ever-increasing attention on the hydrogen utilization as energy vector.3 Moreover, fuel cells (FCs) help * Corresponding author. Tel.: +39 011 5644608. Fax: +39 011 5644699. E-mail:
[email protected].
diminish poisonous emissions into the atmosphere, since FCs have generally higher electrical efficiency compared to ICE vehicles.4 FCs fed with pure hydrogen produce only water, thus eliminating local emissions. As far as the actual lack of infrastructure for hydrogen storage and distribution is concerned, vehicles equipped with FCs fed with hydrogen produced by onboard reforming of a fossil fuel to generate at least auxiliary power represents a valid and interesting alternative to overcome such an obstacle, waiting for further development of infrastructures. The conversion of fossil fuels to hydrogen for auxiliary power generation via a FC will help in reducing the production of CO2 gas by ∼33%, and there will be virtually no other polluting gases.5 In this context, research and development of several fuels reforming systems for fuel processors has gained a prevalent role in the perspective of solving these problems. In the present study, a fuel processor (FP) for the vehicle on-board hydrogen production to be fed to the polymerelectrolyte-membrane (PEM) FC stack of an auxiliary power unit (APU) was considered, with its several units: the main one (the catalytic reformer reactor of the primary fuel) for hydrogen (the derived fuel) gross production; the secondary units for both the CO cleanup process and the simultaneous increase of the hydrogen flow rate (water gas shift and CO preferential oxidation catalytic reactors); the auxiliary units for the balance of plant of the whole system (afterburner for the combustion of the hydrogen exhaust gas from the FC, heat exchangers for the internal heat recoveries, water recovery radiators, air compressor, and water and fuel pumps), necessary to properly operate the FP; and the PEM FC itself. In line with the main part of EU industrial projects in the field,6 the whole system was simulated by modeling with Matlab/Simulink software. This software has the prevalent advantage that both steady-state and dynamic simulations are feasible and that its codes, once optimized, can
10.1021/ie050709k CCC: $33.50 © 2006 American Chemical Society Published on Web 06/15/2006
Ind. Eng. Chem. Res., Vol. 45, No. 15, 2006 5299 Table 1. Physical-Chemical Properties of the Primary Fuels and of Their Reference Compounds Employed in the Simulation
fuel
formula
HCR
molecular weight (kg/kmol)
gasoline isooctane light diesel heavy diesel cetane biodiesel
C7.93H14.83 C8H18 C12.32H22.17 C14.6H24.8 C16H34 C18.74H34.52O2
1.870 2.250 1.800 1.699 2.125 1.842
110 114 170 200 226 291.6
heat of vaporization (kJ/kg)
liquid specific heat (kJ/kg)
vapor specific heat (kJ/kg)
higher heating value (MJ/kg)
lower heating value (MJ/kg)
305 308 270 230 358 232.6
2.4 2.5 2.2 1.9 2.2 2.47
1.7 1.6 1.7 1.7 1.6 1.7
47.3 47.8 44.8 43.8 47.3 39.5
44 44.3 42.5 41.4 44 37
be readily converted into the electronic board units (ECUs) used for APU control through specific codes (e.g., D-Space). The first part of the work focuses only on the reforming process, because the primary role of an FP is played by the reformer. A comparison between the autothermal (ATR) and steam-reforming (SR) reactors was accomplished based on several fuels: the possible feedstocks for on-board hydrogen production (the real road distributed fuels, mixture of different organic compounds: gasoline, diesel oils, and biodiesel) and their reference “single-molecule” fuels used in ICE applications (isooctane and cetane for gasoline and diesel oils, respectively).7-13 The model results obtained by using the real fuel mixture compounds or the pure substances, assumed for simplicity as their simulating molecules, enlightened differences so significant as to dissuade us from the use of reference molecules instead of real feedstocks: the reference fuel molecules, in fact, overestimate the performance of the reformer. On the basis of these results, the second part of the work deals with the comparison between two different APU systems, considering both the FP units (for the production of hydrogen-rich fuel gas starting from gasoline, diesel oil, and biodiesel, based, respectively, on ATR and SR processes and the related CO cleanup technologies) and the integrated PEM FC stack, to which the obtained hydrogen-rich gas is fed. On the basis of a series of simulations, the two systems models were compared in terms of reformer, FP, and APU efficiency, hydrogen concentration fed to PEM FC, water balance management possibility, and process scheme complexity. 2. Hydrocarbons Reforming Hydrogen production from hydrocarbons can be carried out mainly by two reactions, partial oxidation (POX)14 or steam reforming (SR):15
CnHm + 0.5nO2 f nCO + 0.5mH2
(1)
SR reaction: CnHm + nH2O f nCO + (n + 0.5m)H2
(2)
POX reaction:
The SR reaction is endothermic and requires an external heat supply; conversely, the POX reaction is exothermic but produces a gas containing less hydrogen with a lower concentration, also because of the dilution of the nitrogen present in the air fed to the reformer. The third possibility is to combine the two SR and POX reactions in one single unit, to perform an autothermal process ATR, with the exothermic POX driving the endothermic SR:16 in this particular case, the reaction products react together with the remaining water according to the exothermic water gas shift (WGS) reaction equilibrium:
WGS reaction: CO + H2O T CO2 + H2 (∆HR ) -41.2 kJ/mol at 298 K) (3) The methanation reaction (eq 4), which provokes an undesired hydrogen consumption, is not taken into account: at tempera-
tures above 500 °C, and for high steam-to-carbon ratio (SCR) values, the reforming reaction is predominant and the amount of undesired product like methane is minimized.17
Methanation reaction: CO + 3H2 T CH4 + H2O (∆HR ) -217.0 kJ/mol at 298 K) (4) During the catalytic autothermal reforming of n-isooctane into hydrogen (catalyst based on noble metals), low methane yield (0.1-0.45%) at maximum bed temperatures of 700 °C was observed,8 due to high initial selectivity of the catalysts to partial oxidation products. Moreover, the presence of Ni as catalyst helps in reducing to nearly zero the methanation reaction above 500 °C.18 Heat management for the reformer can be fine-tuned to obtain a thermally self-sustaining system through suitable values of the steam-to-carbon (SCR) and oxygen-to-carbon (OCR) ratios. 3. Simulations Results of the Reforming Stage A comparative analysis, based on simulations carried out with Matlab/Simulink code starting from several primary hydrocarbon fuels, is presented at first in terms of hydrogen production through both the ATR and SR routes. In the modeling tool, the primary mixture fuels are usually substituted by pure substances assumed as reference fuels (isooctane for gasoline and cetane for diesel oils). Conversely, no reference fuel has been defined for biodiesel, yet. Biodiesel, an alternative diesel fuel, was selected because it is a renewable and environmental friendly fuel, made from renewable biological sources (vegetable oils and animal fats), absolutely biodegradable, and nontoxic.19,20 The primary real and reference fuels used in the simulations and their properties are shown in Table 1.21,22 The biodiesel properties were calculated as average values of the main components of the soybean oil (composed by the following fatty acid methyl esters: palmitic ) 12 wt %, stearic ) 5 wt %, oleic ) 25 wt %, linoleic ) 52 wt %, and linolenic ) 6 wt %) from which biodiesel can be derived.23 The reference fuels show a hydrogen-to-carbon ratio (HCR) higher that that of their real counterparts. Hence, it is expected that these reference fuels might allow one to overestimate the hydrogen production, as well as the reformer efficiency, when compared with the corresponding real fuels. It goes without saying that isooctane and cetane are reference fuels for internal combustion engine applications, but their use in other contexts, as the reforming unit, might indeed be misleading. The first series of simulations were carried out by considering the reformer reactor alone without the APU boundary conditions and materials and catalyst stability constraints. The following conditions were also considered: primary fuel overall feed flow rate mfuel ) 1 kg/h; reformer inlet temperature Tin ) 700 °C; no limitations for the reformer outlet temperature Tout; and the same molar amount of oxygen and steam per mole of carbon (OCR ) 0.35, SCR ) 3) for both ATR and SR, assumed as general standard feed conditions. These values are the typical
5300
Ind. Eng. Chem. Res., Vol. 45, No. 15, 2006
Table 2. Simulated ATR Outlet Conditions for Different Primary Fuels: SCR ) 3, OCR ) 0.35, Tin ) 700 °C mfuel (1 kg/h) gasoline isooctane light diesel heavy diesel cetane biodiesel
(mH2)th Tout mH2 mCO mH2O mCO2 mN2 ηR (kg/h) (°C) (kg/h) (kg/h) (kg/h) (kg/h) (kg/h) (%) 0.135 0.158 0.130 0.124 0.150 0.119
892 764 788 771 794 876
0.245 0.276 0.253 0.251 0.265 0.208
1.074 0.895 0.918 0.885 0.960 0.896
2.896 2.718 2.808 2.803 2.779 2.664
1.484 1.676 1.747 1.822 1.601 1.421
2.656 2.581 2.670 2.690 2.604 2.369
67.8 75.6 72.4 73.6 73.3 68.2
average ones which allow one to obtain acceptable reactions kinetics within the reformer, without a major contribution of side reactions. The OCR for the SR represents the oxygen fed to the burning side (AFB), where a fraction of the overall primary fuel is burned off to sustain the endothermic SR reaction. Table 2 shows the simulation results (temperature and gas composition) at the ATR outlet for the different primary fuels employed, at the fixed Tin value. The (mH2)th represents the theoretical hydrogen mass flow rate achievable from the primary fuel, taking into account all the H atoms in the fuel molecule. The hydrogen produced in the reformer is higher, owing to the crucial contribution of the WGS reaction (eq 3), which produces further hydrogen from the H2O molecules fed as steam. A first comparison among the primary fuels, after the reforming stage, can be made by calculating the reformer efficiency as follows (LHVH2 ) 121.5 MJ/kg):
ηR )
mH2,refLHVH2 mfuelLHVfuel
(5)
The reformer efficiency values listed in the last column of Table 2 show some discrepancies, especially when the real fuels are compared to their reference molecules. In the case of gasoline vs isooctane, an overestimation in the hydrogen produced from isooctane is clear, as well as an underestimation in the outlet reformer temperature (Tout). The difference is quite evident from the standpoint of the ηR values: 75.6% for isooctane as opposed to 67.8% for gasoline. Therefore, as suspected, notwithstanding the same LHV and vaporization heat of the real and reference fuel, isooctane is not a satisfactory choice as a reference molecule for the gasoline reforming context. In the case of reforming reactions, the HCR ratio plays indeed a major role; thus, a reference hydrocarbon molecule with an HCR ratio closer to that of gasoline should be preferable. Conversely, in the twin case of light-heavy diesel vs cetane, the small overestimation in H2 production rate with cetane seems to be compensated by its slightly higher LHV value (see Table 1), and hence, only limited differences among the related ηR values are shown. However, the significantly higher vaporization heat of cetane could in some way create problems for the thermal sustainability of the system during simulation and entail misleading values for the thermal duty of the fuel evaporator and its related design. Therefore, cetane does not appear to be a satisfactory reference molecule for diesels oils in the reforming contest. Finally, biodiesel shows rather low values of both produced H2 (0.208 kg/s) and ηR (68.2%) compared to the other diesel oils. This is due to the oxygen present inside the biodiesel fuel: the global OCR is, in fact, equal to 0.457 with about a 30% increase compared to the external OCR. Notwithstanding, the actual practical interest derives mostly from the fact that it is an environmental friendly fuel.
The reforming conditions being the same (OCR ) 0.35 and SCR ) 3), it seems interesting that gasoline and biodiesel show the highest Tout values and the lowest ηR values (lowest hydrogen mass rates produced): the higher the Tout, the more significant is the partial combustion mechanism with respect to the reforming and WGS ones; the less contribution of the WGS equilibrium is also confirmed by the lower CO2 produced at higher Tout. The lower Tout (isooctane, cetane, and light/heavy diesel), the higher is ηR; thus, considering the ATR unit as separated from the FP-APU system, it has to be prevented from reaching too high of Tout values. In the whole system, however, high Tout values allow thermal system self-sustainability due to the higher internal heat recoverable from the outlet reformate gas. A tradeoff Tout value should be determined. The key parameter to control Tout is OCR, which discriminates the prevalence of the partial combustion vs the WGS reaction. Moreover, taking into consideration the existing technology for the ATR reactors, Tout values in the range of 700-800 °C must be assured: temperatures >800 °C can indeed severely reduce the ATR catalyst durability. On the other hand, temperatures
