Diesel-Fueled Solid Oxide Fuel Cell Auxiliary

Power Units for Heavy-Duty Vehicles

9
Herbert H. Dobbs *
U. S. Army TACOM, AMSTA-TR-D/21 O (H. Dobbs) -e o
Warren, Ml 48397-5000 U. S. A. @$$
“(j)ti~
Theodore Krause, Romesh Kumar, and Michael Krumpelt
~~fi
Argonne National Laboratory /&d
Argonne, Illinois 60439-4837 U. S. A.

Abstract

This paper explores the potential of solid oxide fuel cells (SOFCS) as 3-1 O-kW
auxiliary power units for trucks and military vehicles operating on diesel fuel. It
discusses the requirements and specifications for such units, and the
advantages, challenges, and development issues for SOFCS used in this
application. Based on system design and analysis, such systems should achieve
efficiencies approaching 40% (lower heating value), with a relatively simple
system configuration. The major components of such a system are the fuel cell
stack, a catalytic autothermal reformer, and a spent gas burnen’air preheater.
Building an SOFC-based auxiliary power unit is not straightforward, however,
and the tasks needed to develop a 3-1 O-kW brassboard demonstration unit are
outlined.

Introduction

Solid oxide fuel cells could find application on trucks and military vehicles as
auxiliary power units (APUS) for operator quality-of-life and housekeeping needs
(e.g., heating, air-conditioning, and on-board electronics) and for cargo needs
(e.g., refrigeration). Historically, auxiliary power has been generated as parasitic
power drawn from propulsion power or by continuous engine idling or lead-acid
storage batteries (primarily for military vehicles) when propulsion power is not
required. Diesel-fueled auxiliary power units (APUs) have been developed for
trucks and military vehicles to improve fuel efficiency, reduce engine noise and
emissions, reduce engine wear, and meet the increased demand for
housekeeping power and quality-of-life features. Three major applications for
APUs are (1) overnight units to reduce noise and emissions, thus addressing
 DISCLAIMER

This report was prepared as an account of work sponsored

by an agency of the United States Government. Neither
the United States Government nor any agency thereof, nor
any of their employees, make any warranty, express or
implied, or assumes any legal Iiabiiity or responsibility for
the “accuracy, completeness, or usefulness of any
information, apparatus, product, or process disciosed, or
represents that its use wouid not infringe privateiy owned
rights. Reference herein to any specific commercial
product, process, or service by trade name, trademark,
manufacturer, or otherwise does not necessarily constitute
or imply its endorsement, recommendation, or favoring by
the United States Government or any agency thereof. The
views and opinions of authors expressed herein do not
necessarily state or refiect those of the United States
Government or any agency thereof.
 DISCLAIMER

Portions of this document may be illegible

in electronic image products. Images are
produced from the best available original
document.
 .,’,

. 2

anti-idling bans, (2) full-time units to replace engine-mounted alternators and

electrically driven components, and (3) units dedicated to cargo needs, such as
refrigeration. Although APUs for trucks and military vehicles can reduce engine
noise and emissions, as well as increase fuel efficiency, there is considerable
room for improvement. Furthermore, APUs are bulky, and for military applications,
they tend to be used on high-cost combat vehicles where the added cost can be
more easily adsorbed. For certain military applications, lead-acid storage
batteries are still used to provide auxiliary power when propulsion power is not
required. Problems with lead-acid storage batteries include deep discharge, high
cost, high maintenance, and the need to recharge the batteries.

New technologies that offer the potential for improvements in fuel economy and
reliability, as well as reduction in emissions and maintenance requirements, are
required. Fuel cells, which are essentially solid-state devices with few moving
parts, offer these benefits.

Among the various types of fuel cells, the polymer electrolyte cells (PEFCS) have
been the unanimous choice of the automotive companies for prime power in
vehicles. The main reasons are the ability to startup at ambient temperature
start-up capability and the mechanical robustness. As auxiliary power units,
however, the PEFCS may not be the optimal choice, because the cost of the fuel
processing train does not scale down well to the considerably smaller electrical
capacity, and because the prevalent fuel in trucks and military vehicles is diesel,
which is much harder to process than gasoline. Solid oxide fuel cells can operate
on mixtures of hydrogen and carbon monoxide, permitting the fuel processor to
be simple and inexpensive. Unlike the prime power application with its frequent
start-ups and shut-downs, an auxiliary power system may operate more
continuously. The longer start-up time of the SOFC compared with the PEFC
may weigh less heavily. In this paper, we examine a conceptual design for an
APU built around an SOFC to explore the system tradeoffs and heat and water
management issues.

Solid Oxide Fuel Cells as APUs

An APU of -5 kWe net power could be developed an SOFC system operating on

diesel or similar on-board fuel. A catalytic autothermal reformer (c-ATR) can be
used to convert the fuel to the fuel gas for the fuel cell. Such an APU offers
distinct advantages over an APU based on, for example, a diesel
engine/generator set. Along with these advantages, however, there are
developmental challenges and issues as well.
 3

Advantages

The major components of this power generator (fuel processor, fuel cell stack,
and air preheater) are solid state, while the total system has only a few moving
parts (air blowers, fuel, and water pumps). Using the c-ATR to convert the on-
board fuel to a gas mixture of H2, CO, C02, and N2, with small amounts of CH4
and possibly other species, the SOFC system can operate with the conventional
logistics fuel(s) with minimal fuel processing. This is because the SOFC can use
CO as a fuel, either by direct electrochemical oxidation of the CO, or by the
conversion of the CO to H2 and C02 by reaction with water vapor (the water-gas
shift reaction). These processes within the anode of an SOFC have been
modeled (1), as well as demonstrated experimentally on aviation fuel, JP-8 (2).

Any methane or other light hydrocarbons or alcohols produced by the

autothermal reforming reactions would be converted to hydrogen by reaction with
water vapor within the anode flow field of the SOFC, making use of the stack
waste heat for the endothermic steam-reforming process. The nickel used as the
anode electrocataiyst is an active steam-reforming catalyst, and the fuel cell
temperature, approximately 800°C, is high enough to obtain high reaction rates.
Further, the required water vapor is generated within the anode flow as a product
of the cell reaction.

By appropriate design and operation of the c-ATR, one can vary the level of
methane in the reformate to permit matching of the endothermic energy
requirement of the steam-reforming reaction with the exothermic energy (waste
heat) produced by the electrochemical cell reaction. Typically, this waste heat is
removed from the stack by using excess air as the cathode flow, which leads to
significant power consumption at the air blower (3). Using part of this waste heat
for the steam reforming of CHq and other light hydrocarbons reduces the air flow
requirement for cooling the stack, with a concomitant reduction in parasitic power
consumption in the fuel cell system. Further, since complete conversion of CH4 is
not required in the c-ATR, the amount of water fed to the c-ATR can be reduced
This reduction, in turn, improves the overall system efficiency.

The SOFC’S operating temperature of -800”C is close to the 750-850”C

operating temperature of a c-ATR for converting diesel and other hydrocarbon
fuels to a mixture of H2, CO, C02, N2, H20, and other species (4). This feature
permits close coupling of the reformer with the fuel cell, offering excellent
potential for thermal integration of the two key chemical and electrochemical
components in the APU system for efficient power generation. In other reformer-
fuel cell combinations, the operating temperatures of the reformer and the fuel
..’.
.
4

cell are very different (e.g., 200°C for the phosphoric acid fuel cell used in
demonstration transit buses, and 80°C for the polymer electrolyte fuel cell used in
automotive and transit bus demonstrations). A large difference between the
reformer and fuel cell operating temperatures requires significant heat exchange
components, leading to thermodynamic and operating inefficiencies,

The SOFC-based APUs will have very low emissions, certainly as compared to
an engine/generator-based APU. The pollutant emissions from the fuel cell APUs
are governed by the catalytic burner in the system, which combines the anode
exhaust with the cathode exhaust. The hydrogen and other fuel gases in the
anode exhaust are completely oxidized to H20 and C02 by the excess oxygen
available in the cathode exhaust, reducing the emissions of CO and unburned
hydrocarbons to trace levels. Particulate emissions are essentially eliminated.
Further, by appropriate system design, the burner temperature can be limited to
-800°C so that NOX levels areS100 ppm by volume in the APU exhaust.

The SOFC-based APUs are expected to reach system level efficiencies of z30Y0,
based on the lower heating value of the on-board fuel. This holds even for
relatively simple system configurations, such as the one shown in Fig. 1. In this
system, the fuel, preheated air, and steam are fed to the c-ATR, from which the
reformate flows directly to the SOFC. A major fraction of the preheated air is also
fed to the’ SOFC, Water needed for the c-ATR is condensed out of the anode
exhaust from the SOFC. Hydrogen and CO in the “dried” anode exhaust are
burned in the catalytic burner using the cathode exhaust, and the resulting hot
gas is used to preheat the process air as well as to generate steam for the c-ATR
before being vented from the system. The schematic diagram shown in Fig. 1 is
not meant to be representative of the actual hardware configuration. For
example, the catalytic burner and the air preheater shown as two separate
process components may be configured as one integrated piece of hardware in
practice. With improved thermal and process integration, efficiencies
approaching 40% could be achievable. This would come, however, at the
expense of added system complexity.

Challenges

Although the combination of the c-ATR and the SOFC operating at similar
temperatures has the potential of achieving high thermodynamic efficiencies,
attaining that potential requires the design of system configurations that
maximize thermal integration by minimizing irreversibilities, both in the heat
exchange components and in the fuel cell itself.
. . .
5

r . ----- ------ ------ ------ ------ -

1
I
I
~ ANODE
I
*
1 4 1
I I
I SOFC I c
FM; ; dh CATHODE 4 I 0
I -+ N
I 1 ~ D
1 i i E
I I ! N
I AIR
I ! s
I _, PREHEATER
E
I ..- -“
CATALYTIC -1 R
I ..- ----- /--
--- *
I
BURNER
1 ..- /--- 4 I
1 I
Q.” I / BOILER I
I I
-------
BL%ER ~ * .. .. ---- I
.-... - -----
..
I HOT BOX ---- -. HOT BOX :
------ ------ ----- - ------ ----

EXHAUST
vJ2-Lk
- WATER
PUMP
TANK

Fig. 1. Schematic system diagram for an SOFC auxiliaty power unit.

In Fig. 1, for example, the anode exhaust gas is cooled from -800°C to -60”C tO
condense and recover the water needed at the reformer. This water is then
vaporized in the boiler to generate the steam feed to the reformer. The heat
removed at the condenser is not effectively used to preheat the air and/or water
fed to the c-ATR or the fuel cell and, therefore, the system efficiency is degraded.
Due to the large temperature difference driving force available with this
arrangement, however, the condenser can be considerably smaller than if it is
placed at the system exhaust. One other practical advantage of the condenser
placement in Fig, 1 is that the catalytic burner temperature is lower than it would
be if the anode exhaust were not cooled, thereby decreasing the formation of
NOX at the burner.

Thus, various tradeoffs in system configuration, component sizes and weights,

and system packaging must be considered in order to minimize inefficiencies and
heat losses. At the same time, complexity must be minimized to permit ease of
control in normal operation.

As shown in Fig. 1, all of the high-temperature components of the system are

placed inside an insulated thermal enclosure (indicated by a dashed line),
..*. .
6

referred to as the “hot box.” Areas of concern include packaging these

components, minimizing the sizes of the flow lines between the components,
sealing, and thermal and electrical insulation.

To make such SOFC-based APUs viable, system startup protocols must be

developed that consume the least amount of time and fuel before the APU can
begin to deliver the power to the load. Also, the catalytic autothermal reforming
process uses water to decrease the potential for coking and to permit operation
at a lower temperature than would be feasible without water. For use in heavy-
duty vehicles, the necessary water must be recovered from within the process.
Effective approaches to doing so must be developed.

Development Issues

Several specific components must be developed for the SOFC-based APU for
trucks and military vehicles. These include (a) a fuel, (b) planar fuel cell stacks of
the requisite power ratings using diesel reformate, and (c) spent-gas burner/air
pre-heater.

Several organizations in the US. (5) and elsewhere are developing autothermal
fuel processors for use in transportation fuel cell systems. However, most of
these projects are on reforming gasoline for PEFC systems. As such, a major
part of these efforts is devoted to the conversion of CO to H2 by the shift reaction,
and final reduction of CO to trace levels by catalytic preferential oxidation or
other means. For the SOFC-based APU, the CO conversion and cleanup are not
required. The diesel (or other logistics) fuel, however, will require c-ATR designs
and operating conditions different from those for a gasoline reformer. Specific
issues to be addressed are: the need for minimizing or eliminating water/steam in
the feed; operating temperatures compatible with those of the fuel cell; tolerance
to sulfur at levels expected in the fuel; and size and packaging that permit a well-
integrated reformer-fuel cell subsystem.

Planar fuel cell stacks offering zO.4 W/cm2 have been designed and tested, but
not yet at the -5-kW power levels. As discussed later, a 5-kW stack could be
smaller than a 20-cm cube in overall dimensions, with an active cell area of
1.25 m2. However, specific issues related to stack design, operation on
reformate, projected lifetimes, and physical and mechanical ruggedness must be
addressed.
 e
.
.
7

In principle, the catalytic burner can be effectively combined with the air
preheater in q device similar to the fuel cell stack. Ceramic or metal-ceramic
designs may be considered for this application, and prototype hardware must be
developed. Finally, effective packaging concepts for the APU components, as
well as system control protocols for startup, steady-state, and transient operation,
must be developed.

System Design and Analysis of A SOFC-Based APU

Using Gctool (6), we have analyzed the APU shown schematically in Fig. 1 for
operating conditions and efficiencies. As discussed above, the major
components of this system are the fuel cell stack, the catalytic autothermal
reformer, a catalytic burner for the spent fuel, and an air preheater for the
cathode and reformer air. All of these components are housed in an insulated
thermal enclosure. Additional significant components are a condenser for water
recovery and a boiler to generate steam for feed to the c-ATR. Fuel and water
pumps and an air blower make up the rest of the system components.

Table 1 summarizes the results from modeling a system operating on diesel fuel.
For this simulation, diesel fuel was taken as a blend of 50% n-hexadecane, 20°/0
n-propylcyclohexane, and 10OA each of 1-hexadecene, n-butylbenzene, and
naphthalene (molar composition). As shown in Table 1, the cathode air feed rate
was set relatively high. This was done to limit the temperature rise from cathode
inlet to cathode outlet to -50°C and thereby minimize thermal stresses.

Table 1 shows that the overall energy conversion efficiency is 31% (based on the
lower heating value of the fuel, LHV). However, this system design is rather
conservative and has little thermal integration, other than the close coupling of
the c-ATR with the SOFC. For example, with the c-ATR efficiency of 77% and the
overall system efficiency of 31.1 ‘%, the balance-of-plant efficiency is -72Y0.

This balance-of-plant efficiency, as well as the overall system efficiency, can be

increased significantly by operational and design changes. For example, by
reducing the airflow to one-half of the base case value and increasing the fuel
utilization from 80% to 85%, the system-level efficiency can be increased to 34Y0.
If the fuel utilization is increased to 90%, and the airflow is reduced to one-third of
efficiency can be obtained by using the heat rejected at the anode condenser to
provide some of the air preheat, by increasing the fuel and oxidant utilization, etc.
Thus, it is clear that even the relatively simple system configuration can yield
high operating efficiencies, although some of these efficiency gains would be at
 8

the expense of system simplicity and, perhaps, additional stresses in the ceramic
stack components.

Table 1. Operi tinu ~arameters for the base-case 5-kW SOFC svstem.
Fuel Cell Stack
Cell voltage, V 0.750
Cell temperature, ‘C (average) 800
Cathode air feed rate, gh 63
Anode fuel gas feed rate, g/s 3
Cell efficiency, 5X0 56.2
Autothermal Reformer
Fuel flow rate, g/s 0.35
Air feed rate, g/s 1.95
Water/steam feed rate, gk 0.70
Steam-to-carbon, mole ratio 1.50
Reforming temperature, ‘C 760
Reformer efficiency, YO (LHV) 77

Overall System
Efficiency, % (LHV) 31.1

These analyses show that system efficiencies based on the net power production
can approach or exceed 40%. However, the design and operating parameters
influence the power density, size and weight (and cost), mechanical stresses,
and operational control, in addition to system efficiency. Compromises must be
made in designing the system for the heavy vehicle application. For example, the
fuel cell stack is likely to be the most expensive component in this system, and
obtaining the highest power density would minimize that cost. Although
increasing the efficiency, high fuel utilization on reformate fuel would reduce the
operating cell voltage, leading to lower power densities. On the other hand, the
catalytic burner/air preheater is likely to be the largest component (in terms of
weight and volume) in the system, other than the insulated hot box. The system-
Ievel power density may be increased by operating with high-temperature-
difference driving forces, which would. allow reduced size of the burner/preheaten
however, the result would bean increase in the thermodynamic inefficiencies and
reduced overall efficiency.
 9

The SOFC stack itself can be rather compact. Kim et al. (7) have reported power
densities of -1.8 W/cm2 for anode-supported single cells at 800”C at 0.5 V and
-1.0 W/cm2 at 0.75 V (7). Assuming stack-level power densities of 0.4 W/cm* are
achievable, a 5-kW SOFC would require an active area of 1.25 m2. This could be
obtained with 55 cells of 225-cm2 active area each. With three cells per
centimeter, the SOFC stack would approximate an 18-cm cube (assuming
square cells), for a stack power density of 0.86 kW/L of total stack volume.

Sizes of the heat exchangers range from 0.09 m2 active area for the anode
condenser to 4-8 m2 for the catalytic burner/air preheater (for the higher
efficiency case). Decreasing the extent of air preheat would shrink the size of the
air preheater, but there would be a small penalty in system efficiency.

Development of Brassboard Demonstration Unit

Planar SOFC stacks and other system components have been developed to the
stage where it appears practical to develop a prototype 3-1 O-kW brassboard
demonstration unit. Several issues still need to be addressed before this can be
accomplished, as discussed above. It is possible, however, to delineate the tasks
and steps that might be undertaken in pursuit of this objective. These include (i) a
more detailed engineering design of the system that takes into account the
various tradeoffs discussed above; (ii) fuel processor design for the specific
logistics fuel and determination of its operating envelope, such as permissible air-
to-fuel and water-to-fuel feed rates, air and water preheat temperatures, and
reformate exit temperatures and compositions; (iii) design and development of
planar SOFC stacks to generate -5 kW; (iv) design and development of the
catalytic burner/air preheater, including selection of low-cost materials for this
component; and (v) analysis and resolution of system integration and control
issues and protocols.

Discussion and Conclusion

The analysis shows that an SOFC-based APU would have few components and
would occupy a volume of probably less than 100 liters, with the air preheater
being the largest unit. The efficiency with diesel fuel, which has rather low
hydrogen-to-carbon ratios and is difficult to reform, would still be close to 40°/0
LHV.
 10

Planar solid oxide fuel cells of 3-1 Ok W capacity are available but have not yet
demonstrated the high reliability and durability of tubular SOFCS. Autothermal
reformers for diesel are also being tested in the specified size range, Air
preheater and condensers would have to be designed and tested for this
application and would not be expected to represent major challenges. Building
such systems is not straightforward, however, because critical unanswered
questions are the thermal stress tolerance during start-up of a system with major
ceramic components, and their toughness in a hostile environment. Considering
how much effort has been spent on PEFC system development, and how
relatively few resources have been allocated to addressing the critical issues for
SOFCS, it would seem appropriate to seriously consider the SOFC technology for
this application.

Acknowledgments

The authors are grateful to H. Geyer and X, Wang for their assistance in system
analysis. This work was supported by the U.S. Department of Energy, under
contract W-31 -1 09-Eng-38 with the University of Chicago,

References
(1) A.F. Massardo and F. Lubelli, J. Eng. Gas Turbines and Power, 122,27-35,
January 2000.
(2) N. Minh, A. Anumakonda, B. Chung, R. Doshi, J. Ferrall, G. Lear,
K. Montgomery, E. Ong, L. Schipper, and J. Yamanis, 1998 Fue/ Ccl/
SeminarAbstracts, November 16-19, 1998, Palm Springs, CA, pp. 262–265.
(3) S. Ahmed, C. McPheeters, and R. Kumar, J. E/ectrochem. Sot., 136 (9),
2712-2718, September 1991.
(4) C. Pereira, R. Wilkenhoener, S. Ahmed, and M. Krumpelt, Proc. 1999 U.S.
DOE Hydrogen Program Review, May 4–6, 1999, Denver, CO, Vol. 1, Report
No. NREL/CP-570-26938,
(5) FY 1999 Progress Repofi for Fuel Cc//s for Transpoflation, S. Chalk, Energy
Conversion Team Leader, Office of Advanced Automotive Technologies,
Office of Transportation Technologies, Energy Efficiency and Renewable
Energy, U. S. Department of Energy, pp. 23-66, October 1999.
(6) H. K. Geyer and R. K. Ahluwalia, “GCtool for Fuel Cell Systems Design and
Analysis: User Documentation,” Argonne National Laboratory Report,
ANL-98/8, March 1998.
(7) J.-W. Kim, A. V. Virkar, K.-Z. Fung, K. Mehta, and S. C. Singhal,
J. E/ectrochem. Sot., 146(1), 69-78, 1999.