(647c) High Temperature Microreactors with Inherent Dynamic Robustness

Richard Pattison and Michael Baldea

Department of Chemical Engineering

The University of Texas at Austin, 1 University Station C0400, Austin, TX 78712

email: mbaldea@che.utexas.edu

Hydrogen is produced industrially from light hydrocarbons (notably methane) via steam reforming.  In recent years, shale gas production and landfill gas production has caused methane capture to become wide spread and highly decentralized. As a result, there has been a technological push for localized processing of methane for hydrogen production [1]. Catalytic plate microreactors are a promising alternative for low cost, localized production of hydrogen via methane steam reforming. Millimeter channel heights and micron level catalyst thicknesses minimize transport limitations, resulting in a process that is controlled by reaction rates, rather than transport rates. Microreactors also afford the benefit of a greatly simplified scale-up: production rates can be increased by augmenting the number of units. These benefits notwithstanding, the widespread use of microreactors has been hindered by a number of design and operation challenges.

In hydrogen production microreactors, (endothermic) methane steam reforming is supported by the (exothermic) catalytic combustion of methane; the reactions occur in a layered structure with alternating microchannels. Counter-currentflow designs are advantageous due to simplified reactant distribution and potentially lower feed temperatures. However, in this case, consumption of the reactants in the combustion and reforming channels occurs in opposite spatial directions, and it is difficult to synchronize the respective rates of heat generation and consumption. Uneven heat fluxes can result in reactor extinction or localized “hot spots” that can damage the catalysts. Design modifications have been proposed for making the counter-current design feasible including distributed feed [2], multiple flow passes [3], and an offset catalyst distribution [4].

From an operational perspective, microreactors must be robust and capable of handling frequent demand fluctuations. For example, a rapid drop in reforming flow rate can result in temperature excursions detrimental to the catalysts along the reactor length. The implementation of feedback control mechanisms based on distributed temperature measurement devices and distributed actuators is limited for practical reasons (sensors and actuators cannot be implemented or simply do not exist at such small scales). Consequently, control is limited to a boundary control approach (i.e. manipulating feed flow rates) [5].

In this work, we propose an innovative approach for reactor temperature control based on the use of Phase Change Materials (PCMs). PCMs absorb energy in the form of latent heat when the environment reaches the material melting temperature. Confining a properly selected PCM layer between wall plates separating the reforming and combustion channels thus constitutes a potential means for distributed temperature control: the PCM melting process absorbs thermal energy at constant temperature, preventing the formation of hotspots. Metallic PCMs are also typically highly conductive, having the added benefit of rapidly dissipating thermal energy away from the hotspot. If the thickness of the PCM layer is properly selected, the melting dynamics are fast and represent an effective means for rejecting fast, transient fluctuations in the operation of the reactor. However, the temperature control effect provided by the melting process is limited in time by the quantity of PCM (once the material is melted, the temperature of the system will continue to rise).

In light of the above, we introduce a two-tiered control framework, consisting of i) a PCM layer that acts as a distributedlocal controller to mitigate the thermal effects of fast disturbances, e.g., fluctuations in hydrogen demand, and ii) a supervisory model-based output feedback controller aimed at rejecting the effect of slow, persistent disturbances by manipulating the flow rate to the combustion channel.

We also introduce a novel method for selecting the optimal geometry of the PCM layer (which amounts to “tuning” the fast component in the above hierarchical control framework). The proposed method relies on concepts from nonlinear system identification (notably, the use of multi-level binary sequences) and a recently developed time relaxation-based algorithm to determine the optimal thickness of the PCM layer in the presence of stochastic production disturbances.

The effectiveness of the proposed control scheme is demonstrated via simulations on a 2D first-principles microreactor model. The simulation results show the proposed control structure has excellent performance, completely avoiding the formation of hotspots in the reactor in the presence of significant disturbances in the production rate.

References

[1] N. Muradov, F. Smith, A. T-Raissi. Hydrogen production by catalytic processing of renewable methane-rich gases Int J Hydrogen Energy, 33: 2023–2035, 2008.

[2] G. Kolios, A. Gritsch, A. Morillo, U. Tuttlies, J. Bernnat, F. Opferkuch, and G. Eigenberger.  Heat integrated reactor concepts for catalytic reforming and automotive exhaust purification.  Appl. Catal. B: Environmental, 70(1-4):16-30, 2007.

[3] K. Venkataraman, JMRedenius, and LD Schmidt.  Millisecond catalytic wall reactors: dehydrogenation of ethane 1. Chem. Eng. Sci.,57(13):2335-2343, 2002.

[4] M. Zanfir, M. Baldea, and P. Daoutidis.  Optimizing the catalyst distribution for counter-current methane steam reforming in plate reactors.  AIChE J., 57(9):2518-2528, 2011.

[5] M. Mendorf, H. Nachtrodt, A. Mescher, A. Ghaini, and D. W. Agar. Design and Control Techniques for the Numbering-up of Capillary Microreactors with Uniform Multiphase Flow Distribution. Ind. Eng. Chem. Res., 49(21):10908-10916, 2010.