This study is focused on modeling, design and optimization various configurations of a cascade system, involving multiple adiabatic catalytic packed-beds and microchannel heat exchangers interconnected to each other, to supply hydrogen needed to drive a 1 kW PEM fuel cell. In this system hydrogen is produced in a number of Ni/MgO-Al2O3-packed adiabatic reactors by steam reforming of iso-octane whose heat requirement is supplied by exothermic combustion of methane running in a separate array of adiabatic reactors packed with a Pt/Al2O3 catalyst. Heat transfer between the two reaction systems is achieved by using microchannel heat exchangers that connect the effluent streams of the combustion and reforming beds. The heat-exchanged streams are then fed to the successive adiabatic beds and the reaction-heat exchange cycle is repeated until the desired level of hydrogen is produced. The concept of using interstage heating/cooling is well-known and applied in industrial reactors, but the difference of this system comes from coupling two chemical reactions as heat source and sink in microchannel heat exchangers allowing heat transfer to run at high efficiency in compact volumes as a result of improved heat transfer coefficients (>1000 W/sqm.K). A similar system on methane steam reforming has recently been proposed by Seris and coworkers [1].

This configuration, resulting in a reactor-heat exchanger cascade, has some benefits in terms of temperature control and ease of operation compared to an alternative system in which the two reactions run in parallel, wall-coated, catalytic microchannels and heat is transmitted across the solid wall between the catalytic flow paths: In the cascade system, the catalyst bed sizes in the combustion array can be adjusted to ensure heat release in a well-controlled scheme. This will result in temperature elevations below the light-off level and will prevent overheating of the endothermic reforming bed, which may lead to catalyst deactivation by thermal cracking of the hydrocarbons to carbonaceous species. This mechanism becomes more significant in the presence of heavier hydrocarbons, such as iso-octane, that are known to have higher tendencies to transform into coke at high temperatures. Apart from robust temperature control, catalyst replacement in the cascade configuration is straightforward as the deactivated bed(s) can easily be removed from the matrix and replaced with new ones. However, once catalyst deactivation happens in wall-coated parallel microchannels, either the catalyst layer needs to be destructed and removed from the microstructure and new catalyst should be coated on it, or the whole microstructure needs to be replaced. Obviously, either of these processes is much more cost and labor intensive.

The first part of this work involves the formulation of the mathematical models that is used in the design and optimization of the cascade configurations. Modeling of the reactors and heat exchanger parts are decoupled; one-dimensional pseudohomogeneous fixed-bed models are used to simulate the adiabatically operating beds. The transport of momentum and energy in the microchannels of the heat exchanger are quantified by using computational fluid dynamics techniques. The power-law type rate equations describing the kinetics of methane oxidation and iso-octane steam reforming over Pt/Al2O3 and Ni/MgO-Al2O3-respectively are obtained from literature [2,3]. Reactor models, which are in the form of multiple ODEs, are solved in MATLAB environment using stiff-ODE solver subroutines. Meshing of the microchannel geometry and solution of the transport equations are handled using COMSOL Multiphysics package. Outputs of the reactor simulations – temperature, molar flow rates and pressure – are fed to the CFD codes, whose results are then considered as the input sets for simulating the next set of catalytic beds. Computational tasks are executed through a Hewlett-Packard xw8400 workstation system.

The models formulated above are used in the second phase of this work including simulation, design and optimization of various configurations of the reactor-heat exchanger cascade. The basis of the system is taken as the production of 40 mol/h of hydrogen, which is typically needed to sustain 1 kW PEM fuel cell operation [4]. The effects of feed temperatures, steam-to-carbon and methane-to-oxygen ratios and catalyst-bed sizes are investigated to minimize the number of beds required to deliver the desired level of hydrogen. The parametric studies are conducted under a set of constraints dictating that the maximum methane conversion in each combustion bed should be lower than 15%, the maximum temperature in the steam reforming beds should be lower than 850 K and the steam-to-carbon ratio at the inlet of the reforming beds should be greater than 2.9. The first constraint ensures methane combustion below the light-off point [2], whereas the second and third constraints set the boundaries where carbon formation can be important during iso-octane stream reforming [5]. Number and dimensions of the rectangular cross-sections of the microchannels are investigated to yield an optimal balance between the heat transfer rate and pressure drop.

The simulation results give a saw-tooth type of temperature profile and indicate that the temperature of steam reforming can be adjusted so that it can be fixed around a desired temperature below 800 K which eliminates the risk of carbon formation. The optimum range of operating conditions are found to exist between 0.15-0.20 for methane-to-oxygen ratio, 2.9-3.0 for steam-to-carbon ratio, 900-950 K and 750-790 K for the methane combustion and iso-octane reforming feed streams, respectively. The combustion and reforming bed sizes are found to vary between 2.5-7.5 g and 10-30 g, respectively. Presence of these conditions will require the use of four combustion beds and four steam reforming beds in a cascade configuration to achieve the desired hydrogen production level.

References:

1. E. L. C. Seris, G. Abramowitz, A. M. Johnston and B. S. Haynes, Chem. Eng. J., 135S (2008) S9.

2. L. Ma, D.L. Trimm, C. Jiang, Appl. Catal. A: General, 138 (1996) 275.

3. Praharso, A.A. Adesina, D.L. Trimm, N.W. Cant, Chem. Eng. J., 99 (2004) 131.

4. J.M. Zalc, D.G. Löffler, J Power Sources, 111 (2002) 58.

5. Rostrup-Nielsen JR. Catalytic steam reforming. In: Anderson JR, Boudart M, editors. Catalysis, Science & technology. Berlin: Springer-Verlag, 1984. vol. 5. p. 1-117.