(292a) CHAMP-D.D.I.R.: High Power Density Fuel Reforming Reactor for Hydrogen Generation

ABSTRACT
submitted to 2012 AIChE Annual Meeting  (Oct. 28 ? Nov. 2, 2012)

CHAMP-DDIR: High Power Density
Fuel Reforming Reactor for Hydrogen Generation

Thomas M. Yun, Peter A. Kottke,
and Andrei G. Fedorov*

Woodruff
School of Mechanical Engineering, Georgia Tech, Atlanta, GA30332-0405

*Corresponding
author: AGF@gatech.edu (email) / 404-385-1356 (phone)

Portable fuel cells are very attractive power generation
sources due to their potential for high energy density and high efficiency.
Particularly, portable hydrogen fuel cells have received significant attention
in the transportation industry, since they offer opportunities for
zero-emission power generation. However, the difficulty associated with compact
storage of hydrogen gas has been a major obstacle to achieving high energy
density in this technology. To this end, we are developing a new class of
catalytic microreactors which utilize the high volumetric energy density of
liquid hydrocarbon fuels (conventional or renewable) and serve as fuel
reformers to produce hydrogen on-demand.

 The direct droplet impingement reactor (DDIR) is a catalytic reformer of
liquid hydrocarbon fuels. DDIR has been developed as a novel method for portable
catalytic reforming of liquid fuels by atomizing the liquid with precise
control over droplet characteristics, having the droplets impinge in a
well-controlled manner on a heated catalyst to react and produce hydrogen.[1,2]
Compared to other portable liquid fuel reforming technologies, which are
commonly designed as scaled down versions of large reactor systems but
utilizing miniaturized components for process intensification, DDIR design
focuses on maximizing the power density of the reactor by understanding the
fundamental physics and chemistry of transport-reaction interactions and
optimizing the components in the system-level framework.

In order to achieve the goal of the maximum power density,
the fundamental concept of DDIR is embodied into various configurations which
offer some unique advantages for different modes of operation. The two types of
DDIR-based reactors that have been analyzed in this study are (i) an ideal
continuous-flow membrane reactor for steady-state operation and (ii) a batch
membrane reactor with a variable volume for transient operation. The latter
concept is based on adaptation of a hybrid configuration in which the DDIR is
coupled to the CHAMP (CO2/H2 Active
Membrane Piston) reactor pioneered by our research group[3].
The most unique feature of the CHAMP reactor is an active piston that enables
precise control of reactor volume (Figure 1). This attribute allows dynamic
control of residence time and reactor conditions to achieve optimal reaction
and separation processes [4].

The CHAMP-DDIR is
developed to combine the advantages of two reactor design concepts:

·        
DDIR
aspects: minimize the number of balance-of-plant components by DDIR's liquid
fuel atomization and direct on-catalyst-contact volatilization,

·        
CHAMP
aspects: achieve the highest power density conversion with the capability to
control operation in optimal conditions.

Figure 1 summarizes
activities in the development of various DDIR family reactors in this study.
The first activity is a comparison study between an ideal CF-DDIR that represents the
steady-state continuous-flow operation of the reactor and the baseline CHAMP-DDIR for transient, variable
volume operation. This first study revealed that the power density of both
reactors increases monotonically with pressure, and,
that for a given pressure, the CHAMP-DDIR power density at its optimal
temperature is always higher than that of the CF-DDIR at its own (different)
optimal temperature. The fundamental reasons for the superiority of the
CHAMP-DDIR is its ability to maintain an elevated pressure in the reaction
chamber due to an active volume control, irrespective of the rate of fuel
conversion and hydrogen permeation, along with relatively minimal power penalty
associated with active compression of reactants.

Based on insights gained
from the initial study, two new types of CHAMP-DDIR, which exploit a
dual-chamber design, were considered, and the models describing their behavior
were formulated and implemented. The key attractive characteristic of the
dual-chamber CHAMP-DDIR is a zero-dead-volume reaction chamber, and the two
reactors differ in terms of piston control: one has a free-piston, and the
other posses an actively controlled piston. To understand and quantitatively
assess if the reduction in dead volume translates to an increased power
density, a comparison study was performed for the three types of CHAMP-DDIR
(the baseline single-chamber/active-piston, a free-piston/dual-chamber, and an
active-piston/dual-chamber configurations). Finally, a new hybrid CHAMP-DDIR is proposed to integrate the key
insights and findings form this study.

Figure 1. Diagram of DDIR Family of Reactors a) Ideal
continuous flow DDIR, b) Baseline CHAMP-DDIR, c) Free Piston Dual Chamber CHAMP-DDIR, and d) Active Piston Dual Chamber CHAMP-DDIR

For the baseline comparison of CF-DDIR and CHAMP-DDIR, ideal
conditions for each reactor operation were assumed, and two models were
developed.  For the ideal CF-DDIR, an
isothermal plug flow reactor model is developed for advection-reaction
phenomena, including effects of hydrogen permeation (Figure 1. a).  For the CHAMP-DDIR, an isothermal variable
volume batch model is developed and the effects of hydrogen permeation were
included (Figure 1. b). Other important assumptions for both models are (1)
Transport limitations are not included within the reactor volume; (2) Catalytic
reactions proceed at their intrinsic rates as predicted by the detailed
kinetics model from the literature; (3) Separation of hydrogen is limited by the
intrinsic membrane permeability, which is a function of temperature and partial
pressures of all species present (4) Uniform pressure exists in reactor due to
fast equilibration (on the time scale dictated by the reaction kinetics).

A parametric comparison study was
performed with the developed models to achieve the level of
understanding and to develop insights necessary to identify the design
trade-offs for reactor optimization.  Focusing
on portable/mobile power applications, the primary performance metric was
chosen to be a volumetric power density of the reactor. Pressure, temperature,
steam-to-fuel ratio at the inlet, residence time and geometry of the reactor
are selected as independent parameters to find optimal conditions resulting in
the highest power density. The important findings from the study are:

·        
Tradeoff exists between power density and hydrogen yield, which can exploited depending on the
specific target application;

·        
Power density of CHAMP-DDIR is always higher than that
of the CF-DDIR, for the same pressures and operating each at its optimal
temperature; 

·        
Power
density increases monotonically with the pressure increase in the reactor, so
the reactor operation at elevated pressures is favorable from the performance
point of view; and

·        
For a
given geometry of reactors (Figure 1), maximum power density is achieved when the
volume (distance) separating the point of reactant injection and the catalyst
is the smallest.

The concept of dual-chamber CHAMP-DDIR was developed to
improve performance of a baseline ideal CHAMP- DDIR. The main motivation for
the dual-chamber variable volume design is to eliminate volume by designing a
reactor in which the reforming reaction occurs on both sides of the piston. Two
types of dual-chamber CHAMP-DDIRs are conceptualized based on the piston
movements ? free piston and active piston. Figure 1.c and  Figure 1.d show the simplified
configurations for both arrangements, respectively. The key operational
characteristics of the both dual-chamber CHAMP-DDIR is that each chamber runs
in different (lagging in time relative to one another) sequences linked by the
motion of piston, which affects each chamber's volume and pressure. In both
chambers, reaction occurs in a cycle with the same profile (i.e., time
evolution of pressure and volume), but a half-period phase lag exists between
cycles in two chambers.

Figure 2 shows the power density comparison for all types of DDIR
family reactors, including steady-state continuous flow (CF) operation and
different embodiments of CHAMP-DDIR. The results are shown for simulations with
maximum pressure at 8 bar, 88% hydrogen yield and
optimal temperature for each reactor. The baseline CHAMP and the
active-piston/dual-chamber CHAMP performances were nearly identical, and the
possessed the highest power density of all reactors under investigation.
Considering practical aspects of reactor design and operation, we conclude that
a
baseline CHAMP-DDIR is not only the best performing reactor in power density,
but it also simple in design and operation, as compared to its dual-chamber
versions with active and passive pistons.

Figure
2. Power density comparison for all
types of DDIR family reactors

REFERENCES:

1.       Varady, M. J. and Fedorov, A.G.,
Fuel reformation and hydrogen generation with direct droplet impingement
reactors: model formulation and validation, Ind. & Eng. Chem. Res.,
50, 9502-9513 (2011).

2.      
Varady,
M. J. and Fedorov, A.G., Fuel reformation and hydrogen generation with direct
droplet impingement reactors: parametric studies and design consideration for
portable methanol steam reformers, Ind. & Eng. Chem. Res.,
50, 9514-9524 (2011).

3.      
Damm,
D. L. and Fedorov, A.G., Comparative assessment of batch reactors for scalable
hydrogen production, Ind. & Eng. Chem. Res., 47 (14),
4665-4674 (2008).

4.      
Damm,
D. L. and Fedorov, A.G., Batch reactors for hydrogen production: theoretical
analysis and experimental characterization, Ind. & Eng. Chem. Res.,
48 (12), 5610-5623 (2009).

ACKNOWLEDGEMENTS:

The
authors acknowledge financial support of this work through NSF CBET Grant
0928716, which was funded under the American Recovery and Reinvestment Act of
2009 (Public Law 111-5). Any opinions, findings, and conclusions or
recommendations expressed in this publication are those of the authors and do
not necessarily reflect the views of the National Science Foundation.