Professional Documents
Culture Documents
micro-CHP System
Based on High Temperature Proton Exchange
Membrane Fuel Cell Technology
Alexandros Arsalis
Dissertation submitted to the Faculty of
Engineering and Science, Aalborg University
in partial fulfillment of the requirements for the degree of
DOCTOR OF PHILOSOPHY
in
Mechanical Engineering
Thesis
assessment
committee:
Research
program:
Mechanical Engineering
Contents
List of Publications
Papers included in the thesis
Relevant papers not included in the thesis
9
9
10
Acknowledgements
11
Abbreviations
13
1
1.1
1.2
1.3
1.4
1.5
Introduction
15
Literature Review
15
1.1.1 Fuel Cell Fundamentals
15
1.1.2 Proton Exchange Membrane Fuel Cells
16
1.1.3 High Temperature PEMFC
18
1.1.4 Comparison of Fuel Cell Technologies
19
1.1.5 Combined Heat and Power
20
1.1.6 Micro-CHP Technology
21
1.1.7 Micro-CHP Classification based on the Conversion Process
23
Comparison Analysis Between the Proposed micro-CHP System and a
Centralized CCGT Power Plant/Heater Combination
27
1.2.1 Efficiency
28
1.2.2 Environmental Considerations
29
1.2.3 Cost
30
Background and Motivation
30
Definition of the Research Question
31
Objectives of the Research Project
31
2
2.1
2.2
35
35
36
37
40
41
3
3.1
3.2
3.3
3.4
Description of Publications
Paper I
Paper II
Paper III
Paper IV
43
43
43
44
45
3.5
3.6
Paper V
Overall Evolution of the Configuration Topology
46
48
51
5
5.1
5.2
61
61
63
References
65
List of Publications
Papers included in the thesis
This thesis is based on the work contained in the following five journal
articles1, referred to by Roman numerals in the text:
I
Arsalis, A., Nielsen, M.P. & Kr, S.K. (2011). Modeling and off-design
performance of a 1 kWe HT-PEMFC (high temperature-proton exchange
membrane fuel cell)-based residential micro-CHP (combined-heat-andpower) system for Danish single-family households. Energy 36(2), 9931002.
II
Arsalis, A., Nielsen, M.P. & Kr, S.K. (2011). Modeling and parametric
study of a 1 kWe HT-PEMFC-based residential micro-CHP system.
International Journal of Hydrogen Energy 36(8), 5010-5020.
III Arsalis, A., Nielsen, M.P. & Kr, S.K. (2012). Application of an improved
operational strategy for a high temperature-proton exchange membrane
fuel cell-based micro-combined heat and power system for Danish singlefamily households. Energy. (submitted)
IV Arsalis, A., Nielsen, M.P. & Kr, S.K. (2012). Modeling and optimization
of a 1 kWe HT-PEMFC-based micro-CHP residential system.
International Journal of Hydrogen Energy. (in press)
1. Every research study topic matches exactly with every corresponding journal article
publication. Therefore the terms paper/article/study topic are used interchangeably throughout this
thesis work.
9
V Arsalis, A., Nielsen, M.P. & Kr, S.K. (2012). Optimization of a high
temperature PEMFC-based micro-CHP system by formulation and
application of a process integration methodology. Energy Conversion and
Management. (submitted)
Articles I and II are reproduced with the permission of the publishers.
10
Acknowledgements
This PhD research project was supported by Danish companies Dantherm
Power and Danfoss. It was also co-funded by the HyFC (Hydrogen Fuel Cell
Academy) and Aalborg University.
Thanks to my academic supervisors Mads Pagh Nielsen and Sren Knudsen
Kr for providing valuable advice, comments and suggestions throughout the
research study.
Special thanks go to Anders Korsgaard for providing simulation models in
the initial stages of the project, and also prompt feedback, when needed,
throughout the project.
I would also like to thank my colleagues at the Fuel Cell Research Program
for exchanging fruitful ideas on the research project.
Finally, I am indebted to my family who has helped me, while I devoted
time and energy to this research work.
Alexandros Arsalis
Aalborg
October 2011
11
(This is an example of an empty page. A white rectangle is drawn on top of the page number.)
12
Abbreviations
AFC
BOP
CHP
EES
GA
GAMS
HEN
HHV
HT
ICE
IRR
LHV
LT
MCFC
MEA
MINLP
PAFC
PBI
PEMFC
RC
RE
RH
SC
SE
SMR
SOFC
VI
WGS
13
(This is an example of an empty page. A white rectangle is drawn on top of the page number.)
(This
14
Introduction
This chapter presents an overall literature review on the fuel cell and microCHP technologies. Further on, the background and motivation for the research
study are explained, and the project objectives are outlined. Finally, the
research question is posed and analyzed.
15
PAFC
Electrolyte
Polymer
membrane
Mobile ion
MCFC
SOFC
Molten
carbonate
Ceramic
H+
H+
OH-
CO32-
O2-
Operating
temperature
60-200
205
65-220
650
600-1000
Catalyst
Platinum
Platinum
Platinum
Nickel
Nickel
Cell
components
Ceramicbased
Fuel intake
H2, CH3OH
H2
AFC
H2
H2, CH4
(1)
The most common material for the polymer membrane of the ionically
conductive material is NafionTM(Barbir, 2005). The membrane must be thin,
2. This section refers primarily to conventional Nafion-based PEMFCs.
16
flexible and coated with a platinum-based catalyst, while the electrode support
material must be a porous carbon. The total assembly is a sandwich structure,
namely the Membrane Electrode Assembly (MEA), connecting the anodecatalyst-membrane-catalyst-cathode components. The fuel cell operating
temperature for NafionTM-based fuel cell stacks cannot exceed 90C, because
the membrane requires continuous hydration to maintain its conductivity
capabilities. Therefore a rigorous water management is required throughout the
operation.
e-
load
H2
AFC
OH-
H 2O
H2O
H2
PAFC
PEMFC
O2
O2
H+
H2O
H2
MCFC
(CO) CO2
CO3=
CO2
H 2O
SOFC
(CO)
H2
(CH4)
H 2O
O2
O=
O2
fuel in
oxidant in
anode
electrolyte
cathode
Figure 1 Schematic of the operating principle for different fuel cell types [based on the
description found in (Barbir, 2005)].
17
cost of this fuel cell technology will remain high. On the other hand, power
densities are higher than other fuel cell types, ranging from 300 to
1000*mW/cm2 (OHayre et al., 2009). Further on, on-off operations and startups are also faster than other fuel cell technologies (e.g., SOFC). For these
reasons, PEMFC is the primary fuel cell technology for vehicular and portable
applications (<1 kWe) (Mench, 2008). In addition, due to their high sensitivity
and low tolerance to sulfur and carbon monoxide, hydrogen is usually
preferred as the system fuel, although effort has been made to use other more
convenient fuelling options, such as methanol. Usage of fossil fuels, such as
natural gas, can add a further cost burden to the fuel cell systems capital cost,
because more complicated fuel processing will be required.
1.1.3 High Temperature PEMFC
18
19
materials. Also, degradation and lifetime issues have been reported in the
literature (OHayre et al., 2009).
SOFC is a promising fuel cell technology (Zink et al., 2007) and therefore
still under development. Nevertheless, current SOFC-based systems have
several disadvantages, as compared to HT-PEMFC technology. These include
slower start-up times (due to their higher operational temperature), need for air
and fuel preheaters and more complex cooling systems (Larminie & Dicks,
2003; OHayre et al., 2009). Also the need for higher temperature operation
suggests greater heat losses and thereby more expensive insulation is required.
Finally, it requires pressurization of fuel and air, which suggests greater BOP
component (e.g. air compressor) power losses. On the other hand, SOFC-based
systems allow the use of carbon monoxide content in the reformate gas as fuel,
with the use of an internal reforming process (without the need of a separate
unit) (Arsalis, 2007; Calise, Dentice dAccadia, Palombo, et al., 2006), which
suggests a significant advantage over other fuel cell technologies. Also the
high reaction rates achievable by the SOFC technology allow the use of
cheaper catalysts and thereby reducing their capital cost (Larminie & Dicks,
2003).
1.1.5 Combined Heat and Power
The conventional method of covering electrical, heating (e.g. hot water) and
cooling (e.g. space cooling) load demands is by purchasing electricity from the
electricity network grid and with a fossil fuel-fired boiler. A different method
of covering these loads is combined-heat-and-power (CHP) (or cogeneration),
which can aid in the reduction of running (fuel) costs and in effect increase the
total efficiency. Therefore, CHP is defined as the combined generation of
electrical power and heat from a single chemical energy source.
The useful recovery of the biggest portion of waste heat can increase the
system efficiency from 30-40%, up to 75-90%, depending on the application
and the size of the system. In addition to the efficiency increase, CHP can also
lead in the reduction of emissions, since a smaller amount of fuel is required.
Therefore CHP is the preferred choice of generating electricity and heat,
provided the capital cost is within viable limits in terms of lifetime and
payback time.
Cogeneration systems can be distinguished into two main configurations.
The first configuration is the production of a high temperature fluid product,
which can be used to generate electricity (e.g. gas turbine), while the exhausted
heat is at a low temperature and can be used throughout several thermal
processes for district heating purposes, or in some cases for additional
production of electricity (Arsalis, 2008). Alternatively, the high temperature
20
fluid product (e.g. from a waste-to-energy power plant) can be used in a heat
recovery steam generator to produce superheated steam for a Rankine cycle
(steam turbine) process. In some cases it is also possible to use the hot gases
directly into a Brayton cycle (gas turbine), without the need of a heat recovery
boiler.
1.1.6 Micro-CHP Technology
For the purposes of this research study, micro-CHP (Dentice dAccadia, Sasso,
Sibilio, & Vanoli, 2003; Ferguson, 2004; Gunes & Ellis, 2003; Hawkes &
Leach, 2005a, 2005b, 2007, 2008, 2009; Hawkes et al., 2007, 2006; Hubert,
Achard, & Metkemeijer, 2006; Pehnt et al., 2006) is defined as the
simultaneous production of electricity and heat in a residential application for
systems up to 5 kWe. Micro-CHP systems typically have (or expected to have)
a lifetime of ten to twenty years, which is comparatively lower than large-scale
CHP systems. They are designed to exhibit minimum total efficiencies of 75%.
In order for these systems to make a breakthrough to the power and heat
market, many requirements should be satisfied. The main requirements are (a)
low cost, (b) compact volume and size, (c) easy installation and (d) automated
operation without demanding routine maintenance checks.
Depending on the load profile, an appropriate heat-to-power ratio must be
selected based on the demand. For example, a system with a high heat-topower ratio will be inappropriate for a household requiring a high electrical
load. Nowadays, most newly-built households have efficient insulation and are
more electricity-demanding than heat-demanding, as compared to the past.
Therefore, in general, even lower heat-to-power ratios will be required in the
future. Also the production of additional heat, if needed, can be provided by
condensing gas-fired boilers, which have become very efficient in recent years,
with efficiencies near or above 90% (Hawkes & Leach, 2005a). Therefore,
focus is primarily given on systems exhibiting high electrical efficiencies,
rather than thermal efficiencies.
For the better utilization and distribution of power and heat, micro-CHP
systems must be grid-interconnected, although micro-CHP systems can also
operate in a stand-alone (island) mode. The reason for grid-interconnection is
that a stand-alone micro-CHP system will be required to operate continuously
within an electricity-led operational strategy. As a consequence of this practice,
large amounts of heat will have to be vented to the atmosphere. Otherwise a
thermal storage tank of massive dimensions will be required, which is
inappropriate and unacceptable for a residential design. Another consequence
is also the rate of heat loss, which increases with the increasing size of the
21
Japan has been the leading market for micro-cogeneration in the last decade,
but recently a rapid growth of the technology has also been observed in
Europe. An increase in the sales of micro-CHP systems by around 25%,
between 2009 and 2010, suggests that the market will grow even at higher
proportions in the near future. 20,000 to 70,000 micro-CHP units are expected
to be sold by 2015 (Brown, 2011).
3. Although theoretically all the heat can be recovered to cover the space heating and hot water
loads, it should be noted that a realistic system will not be able to completely recover this heat.
22
23
Company
Type
el ,net (%)
tot (%)
Market status
RE
26
92
CA
Yanmar
RE
N/A
N/A
CA
RE
28
88-91
CA
Energetix Kingston
RC
10
N/A
CA
OTAG
RC
N/A
N/A
CA
BAXI Ecogen
SE
14
90
CA
WhisperGen
SE
10-12
90
CA
Bosch ENATEC
SE
10
N/A
CA
SE
21
85
Demo
SOFC
30
90
Demo
Ceres Power
SOFC
N/A
N/A
Demo
Osaka Gas-Kyocera
SOFC
N/A
N/A
Demo
SOFC
60
N/A
Planned
ClearEdge Power
LT-PEMFC
N/A
90
Demo
LT-PEMFC
40
90
Demo
LT-PEMFC
32
83
Demo
Plug Power
HT-PEMFC
32
90
Demo
Dantherm
HT-PEMFC
40
85-90
Demo
Stirling Engines
A Stirling engine (SE) is an external combustion engine operating on the
Stirling thermodynamic cycle, which includes four consecutive processes: (a)
isothermal expansion, (b) isochoric cooling, (c) isothermal compression, and
(d) isochoric heating. The SE typically includes three pistons, two outer and
one displacer. The latter circulates the cylinder products into the chamber,
which is cooled or heated by their respective outer streams. The two outer
pistons can be used to regulate the capacity of the combustion process.
The main characteristics of Stirling engines, as compared to other heat
engines, include adequate part-load performances, low emission rates and low
vibration and noise levels. In addition, the external combustion, closed cycle
24
25
Electricity
inverter
PEMFC micro-CHP system
heat exchanger
hot water
space heating
natural gas
electricity import/export
26
based PEMFC and SOFC. It is divided into three phases, and two fuel types are
being tested: hydrogen and natural gas. The choice of fuel depends on the fuel
cell technology and its availability at the installation site. A preliminary
assumption concerning the operational control of the system, indicated that the
electrical load must be fulfilled, but also the design should secure the right
amount of heat will be available on demand for the space heating and hot water
loads. Further on, an auxiliary burner for peak production of thermal energy is
under consideration to be integrated to the end-user system (Danish MicroCHP, 2009; Korsgaard, Nielsen, & Kr, 2008).
The HT-PEMFC-based micro-CHP units are developed by Dantherm Power
and operate on natural gas. Experimental tests, with pure hydrogen fuel,
showed start-up times of 30 to 60 minutes, with an electrical efficiency of 40%
(hydrogen fuel, LHV-based). From the operating point of the finalized design,
an efficiency of 50% is expected (hydrogen fuel, LHV-based). Further
experimental tests and calculations showed a potential system efficiency of 85
to 90% (LHV-based) (Danish Micro-CHP, 2009). A fast adaption to load
variations was also observed. In the first phase, Danish micro-CHP developed
unit prototypes with PEMFC and SOFC technologies. In the second phase, ten
micro-CHP units are scheduled to be installed and tested at selected consumer
households in the municipalities of Lolland and Snderborg. In the third and
last phase, micro-CHP systems will be installed and demonstrated at around
100 households in the two aforementioned municipalities. This will allow the
Danish micro-CHP system project to gather and analyze realistic experiences
related to installation, operational procedures, maintenance needs and also
consumer satisfaction (Danish Micro-CHP, 2009). Currently, HT-PEMFC
activities are on standby, pending durability issues to be resolved.
27
resources (e.g. natural gas) will improve. In other words, energy availability
and reliability will improve, because dependency to the network grid will
decrease, and therefore grid congestions and blackout events will diminish
(Chicco & Mancarella, 2009; Praetorius et al., 2009). In addition, the current
centralized power production system status-quo will shift towards a more
competitive regime, with obvious benefits (Praetorius et al., 2009). Therefore,
it should be strongly emphasized that the purpose of this research work is not
the investigation of a proposed replacement of centralized CCGT power plants,
but rather the replacement of aging natural gas boilers4, as already noted.
Therefore the discussion below is only provided for comparison purposes.
Centralized power generation is considered more conventional and power
plants are available in various configurations and based in different
technologies. The most significant ones are the following: Steam turbines (ST),
gas turbines (GT) and CCGT. The latter type is considered to be the most
advanced technology, exhibiting overall efficiencies close to 60% (Lund, 2008;
Praetorius et al., 2009). It combines the gas turbine (Brayton) and steam
turbine (Rankine) thermodynamic cycles, where the former is the topping cycle
and the latter the bottoming one. In addition, some other, more novel, central
power plants designs have been under investigation in recent years. These
include hybrid systems, which combine high temperature fuel cells with
turbine cycles (i.e. SOFC-GT, SOFC-ST, SOFC-GT-ST and MCFC-GT).
Therefore the following hypothetical question must be answered:
Can the proposed system compete with centralized CCGT power plants and
heat-only boilers which are in widespread use in single-family households
today?
To answer the above question, a number of comparison parameters are
considered in order to distinguish the characteristics, including advantages and
disadvantages, of the two technologies. The following analysis is performed in
terms of efficiency, environmental considerations and cost.
1.2.1 Efficiency
The efficiency of a CCGT power plant is lower than the efficiency of the
proposed system in many ways. At nominal load, modern CCGT systems can
exhibit electrical efficiencies up to 60%5 (Ang, Fraga, Brandon, Samsatli, &
Brett, 2011), which is higher than the expected electrical efficiency of the
4. This is also the target of the Danish micro-CHP project (Danish Micro-CHP, 2009).
5. One of the most modern CCGT power plants in Denmark is the Silkeborg CHP Plant with a
power output of 105 MWe and an electrical efficiency of 50% (DONG Energy, 2011).
28
proposed system. On the other hand, both systems must operate both at
nominal and part-load. At part-load, the electrical efficiency of the CCGT will
decrease6 (Kehlhofer, Bachmann, Nielsen, & Warner, 1999), while the
electrical efficiency of the proposed system will increase. For example, a
CCGT plant with a nominal load efficiency of 60%, will have a decreased
efficiency of 52% at 50% part-load (estimation based on empirical data found
in (Kehlhofer et al., 1999)).
Further on, due to distribution and transmissions losses, around 6 to 10% of
the efficiency is lost in the case of CCGT (Al-Sulaiman, Hamdullahpur, &
Dincer, 2011; Lund, 2008), while in the case of the proposed system, there are
practically no such losses, because the power is produced and consumed onsite (Raven & Verbong, 2007). It should be noted though that this assumption
is only valid provided electricity is not exported by the proposed system to the
grid (electricity-led operation).
In terms of overall efficiencies, the proposed system can reach efficiencies
up to 91%, which is significantly higher than the corresponding efficiencies for
CCGT, especially in the case where the exhausted heat of the power plant
cannot be usefully utilized (Ang et al., 2011; Westner & Madlener, 2011).
Finally innovative control techniques, such as virtual control of many
micro-CHP systems (virtual power plant), can aid in a more efficient operation
of the micro-CHP systems, since neighboring micro-CHP systems can operate
jointly, transmitting heat and power, as needed (Delta Energy & Environment
Ltd, 2010; Praetorius et al., 2009; Raven & Verbong, 2007).
1.2.2 Environmental Considerations
Fuel cell technology features high electrical and total efficiencies, which in
terms of environmental benefits translate to lower fuel consumption and
decreased CO2 emissions, as compared to a CCGT/boiler combination
(Praetorius et al., 2009). The Kyoto agreement has prompted the EU to adopt
energy policies that favor power and heat production with reduced greenhouse
emissions (De Paepe & Mertens, 2007; Praetorius et al., 2009; Ropenus,
Schrder, Costa, & Ob, 2010). Therefore, it is expected that the proposed
system will be offered with several benefits, such as lowered taxation,
subsides, etc. Therefore energy policies can further favor the adoption of
decentralized systems, such as the proposed system, even though their (actual)
total cost will remain higher than CCGT/heater per kWh.
6. In CCGT power plants, without supplementary firing, the total plant efficiency primarily
depends on the degree of heat recovery, the GT efficiency and the size of the plant (Kehlhofer et
al., 1999).
29
The main problem with the proposed system, in terms of emission rates, is
that the emissions will be exhausted in urban areas. Thereby CCGT is more
emissions attractive, at least locally, in the sense that centralized power plants
are typically located in rural areas (Chicco & Mancarella, 2009).
1.2.3 Cost
31
32
To eliminate possible confusion, while reading this thesis work, the following
assumptions are given:
The fuel cell type was predetermined to be of the HT-PEMFC type. It
is therefore assumed to be a fixed, non-decision variable, since no
evaluation process took place to select which fuel cell technology
type is more applicable (e.g. in terms of efficiency) for a fuel cellbased micro-CHP system;
The computational model development varied throughout the
projects evolution, based on decisions taken by the author and the
supervising committee, or due to knowledge and information gained
while developing the models.
33
(This is an example of an empty page. A white rectangle is drawn on top of the page number.)
34
35
The purpose of the fuel processing subsystem is the generation of a hydrogenrich reformate gas, with a low CO-content (composition depends on the
tolerance of the fuel cell type). Additionally other poisonous substances, such
as H2S must be removed. To accomplish these tasks various processes must
take place in appropriate components (e.g. reactors) in a series of several steps.
A simplified configuration of a typical fuel processing subsystem arrangement
is shown in Figure 7. The main components include heat exchangers, chemical
reactors, a burner, pipelines and extraction equipment (Jahn & Schroer, 2005;
Jannelli, Minutillo, & Galloni, 2007; Kolb, 2008; OHayre et al., 2009). A
steam generator converts liquid water into superheated steam. Steam
generation is necessary for the fulfillment of the chemical reactions (e.g. steam
methane reforming) in the reactors. The compressed natural gas is also
preheated to accelerate and facilitate the reforming reaction. Steam generation
and fuel preheating requirements are usually easily accomplished by waste heat
generated by either the fuel cell stack or fuel processing. Then the
methane/steam mixture enters the reformer, where it reacts at a high
temperature (600-700 ), in the presence of a catalyst (to accelerate chemical
kinetics), resulting in a hydrogen-rich reformate gas. The reformate gas then
enters the water gas shift reactor, which increases the quantity of hydrogen in
the stream and decreases the CO content.
The catalytic burner serves into the complete combustion of fuel remnants
in the fuel cell exhaust exiting from the fuel cell anode and cathode. Based on
the fuel processing arrangement the generated heat can be used in an
endothermic reforming reaction (e.g. steam methane reforming) and/or provide
heating for other system needs (e.g. steam generation). If heat is still available
it can be used externally (e.g. thermal storage tank). Depending on the
operating fuel utilization of the fuel cell stack, hydrogen-rich gas may be
depleted at the exiting streamline of the fuel cell stacks anode and used in the
burner. In some cases, the steam generated by the combustion of the hydrogenrich gas in the burner can be reused in other parts of the system (OHayre et al.,
2009). Finally after the burner, a condenser can be utilized to convert steam
back to liquid water by cooling this stream and capturing the latent heat of
condensation. In a fuel cell system, a condenser is important for both
37
recapturing heat and recovering liquid water to achieve neutral system water
balance12. Finally it should be noted that very careful insulation and integration
of the fuel processing components is necessary to minimize heat loss rates.
Depleted fuel
Natural gas
BURNER
air
WGS
Reformate fuel
SMR
Reformate fuel
Flue gas
FUEL CELL STACK
air
DESULFURIZER
Natural gas
exhaust
Steam Reforming
Steam reforming involves the reaction process of a hydrocarbon with steam. It
is highly endothermic and proceeds at high temperature over a nickel catalyst
(Kolb, 2008; Xu & Froment, 1989a, 1989b) as follows,
1
C x H y xH 2 O(g) xCO y x H 2
2
CO, CO2 , H 2 , H 2 O
(2)
12. A neutral system water balance describes the situation where all the water consumed by
system components is produced by other system components, without any external water addition
(OHayre et al., 2009).
38
(3)
All the main chemical reactions associated with steam methane reforming are
summarized in Table 3. The most common steam reforming design
configuration is the tubular type (e.g. shell-and-tube) and includes a furnace
containing tubes filled with catalysts, through which the steam reforming
reactants flow (OHayre et al., 2009). In a typical shell-and-tube arrangement,
the reforming reaction takes place inside the tubes (tube-side), which are
heated by the hot flue gas flowing in the shell-side. Shell-and-tube is not easily
applicable in low-scale fuel cell-based systems, because integration becomes
more difficult in compact design applications.
Table 3 Steam reforming reactions.
0
hrxn
kJ mol
Reaction #
Reaction Type
Stoichiometric Formula
Steam reforming
CH 4 H 2O(g) CO 3H 2
+206.4
CO H 2O(g) CO 2 H 2
-41.2
Evaporation
H 2O(l) H 2O(g)
+44.1
nCO
n
(4)
13. The other main reforming options are partial oxidation and autothermal reforming. Partial
oxidation is less efficient than steam reforming in terms of hydrogen production. In particular for
methane, it is only able to generate two moles of hydrogen per mole of methane, while steam
reforming can generate three moles (Barbir, 2005).
39
where nCO is the number of moles of CO in the reformate stream. The WGS
reaction can reduce the CO-content up to around 0.2%, with the presence of a
shift catalyst (Kolb, 2008).
The reversible WGS reaction is moderately exothermic. At high
temperatures the balance is shifted towards the reactants (CO and H2O), where
at low temperatures, the balance follows the opposite direction (CO2 and H2)
(OHayre et al., 2009). Therefore, at low temperatures, the reaction increases
its H2 yield (equilibrium), but on the other hand, at high temperatures the
reaction kinetics proceeds faster. To accomplish both tasks, the WGS process
may proceed in two (or more) stages, utilizing initially a high temperature
reactor, and afterwards another low temperature reactor, downstream of the
first. Typical low temperature shift (LTS) reactors and high temperature shift
(HTS) reactors operate at around 200-300C and 375-450C, respectively
(Kolb, 2008). Because of the temperature difference between the two shift
reactors, an intermediate cooling heat exchanger is required to maintain the
desired input temperature to the LTS reactor. The most common catalyst
material for LTS reactors is copper/zinc oxide (Kolb, 2008), while HTS
reactors incorporate iron oxide-based catalysts (Keiski et al., 1996; Kolb,
2008). A PBI-based PEMFC can tolerate up to 3% of CO-content, and
therefore a single stage WGS reactor is usually adequate for CO-cleaning
(Serenergy, 2011).
2.2.2 Fuel Cell Subsystem
The H2-rich reformate gas is fed to the anode side of the fuel cell stack, while
simultaneously compressed air is fed to the cathode side. The fuel cell reaction
produces DC electrical power and exhaust heat. The total system electrical
efficiency is lower than the fuel cell electrical efficiency, due to the parasitic
power required to operate pumps or compressors (OHayre et al., 2009).
Therefore a high efficiency can be maintained, only if the system can operate
near ambient pressure. The fuel cell efficiency increases with an increasing
voltage, but higher voltage values result into a decrease in power density.
Therefore in this case a larger fuel cell stack, with an associated higher
purchase cost, must be utilized for the same power output (Barbir, 2005).
Equivalently, part-load fuel cell efficiencies are higher than the
corresponding efficiencies at nominal load. It should be noted though, that a
high fuel cell efficiency does not necessary correspond to a high overall system
efficiency, because other system components may perform poorly under partload conditions. For example, the SMR reactor will be less efficient at partload due to the increase of heat losses, or balance-of-plant components may
require more power input (comparatively to power output) at these conditions.
40
The net electrical efficiency of the fuel cell subsystem R ,SUB is defined as the
ratio of the net electrical power of a fuel cell subsystem Pe ,SUB and the HHV of
H2 in the inlet gas (OHayre et al., 2009),
R,SUB
Pe,SUB
H
(5)
HHV,H 2
The thermal management subsystem recovers waste heat from the system for
both internal (system) use and also external use (OHayre et al., 2009), if
applicable. Therefore this subsystem must integrate the system streamlines in
such a way that will allow both smooth and efficient operation. This is
typically accomplished by the use of heat exchangers. The selection of a heat
exchanger network can be a very difficult process, because many important
parameters are involved. These parameters include the heat exchanger
purchase cost, the heat exchanger type, the exchanger heat transfer area, the
number of heat exchangers and also the heat exchanger integration with other
system components.
A fuel cell-based system can be very demanding in terms of thermal
integration, especially in the case of an integrated fuel processing subsystem
(Barbir, 2005). A series of heat exchangers will be required to either heat (e.g.
steam generation and fuel preheating) or cool (e.g. cooling of reformate fuel
between reactors and before the fuel cell anode inlet) the various system
streamlines. Typically heat can be recovered from the catalytic burner,
exothermic chemical reactor, reformate cooling processes and the fuel cell
stack exhaust. On the other hand, heat must be supplied to the steam generator,
fuel preheater, thermal storage tank, etc.
A very rigorous process for the configuration of heat exchangers is
required, in order for the system to accomplish an economically feasible
operational threshold, without the need for extensive external heating and/or
cooling. An optimum heat exchanger network can be designed with the aid of
advanced integration techniques, such as pinch analysis and process
integration, which allow the determination of an idealized configuration, based
on the system designers requirements. Therefore the objective function for
such a procedure may involve cost and other parameters, usually within a
constrained optimization regime (Kemp, 2007).
41
(This is an example of an empty page. A white rectangle is drawn on top of the page number.)
42
Description of Publications
In this chapter all published and submitted journal articles are introduced and
described in terms of hypothesis, methodology and results. In addition, a
description of the logical and scientific progression of the work is explained.
3.1 Paper I
In the first paper (I), the HT-PEMFC-based micro-CHP system is designed and
modeled in LabVIEWTM. A detailed literature review, including an
introduction to stationary fuel cell systems, micro-CHP and HT-PEMFC
technology, is outlined. Also, an introduction to the Danish micro-CHP project
is given, including state-of-the-art, project schedule and future objectives.
The system is designed and simulated in LabVIEWTM, with the intention of
providing easiness in user usage and control of the model, and to allow future
experimental testing capabilities with Data Acquisition hardware.
A representative averaged load profile of a single-family household in
Denmark is used to simulate the model within design and off-design
conditions. The system is divided into subVIs (subsystems) and connected and
controlled with the main VI. The model includes component models for the
fuel cell stack, SMR and WGS reactors, mixers, by-pass valves, heat
exchangers, combustor, steam generator and water pump. The model is then
validated and results are extracted to describe its characteristics for 25 to 100%
operational loads.
3.2 Paper II
In the second paper (II), the micro-CHP system is modeled in EES. The reason
is that modeling in EES is relatively easy and most importantly it includes
many built-in features, such as parametric tables, uncertainty analysis,
43
44
In the results and discussion section the averaged Danish load profile is
fulfilled (partly or fully) by use of the three operational strategies. Based on the
results from the two conventional operational strategies, the improved strategy
is formulated and applied to the simulation model. Emphasis is given on
constraining the system operation within high efficiency regimes, reduction of
frequent interaction with the grid, and avoiding the production of excess heat
that cannot be stored in the thermal storage tank. Also carbon dioxide
emissions are analyzed and compared with conventional systems. In addition,
heat losses from the thermal storage tank during different loads (and time
periods) are analyzed. Finally, an overall analysis of the three operational
strategies is presented and analyzed through comparisons.
3.4 Paper IV
After the development of the simulation model (paper II), and the investigation
of efficient operational strategies (paper III), the system is optimized using the
EES built-in GA optimization method (Godat & Marechal, 2003; Palazzi,
Autissier, Marechal, & Favrat, 2007; Weber, Marechal, Favrat, & Kraines,
2006). The purpose of this research study topic is the maximization of the net
electrical efficiency by variation of nine decision variables. In this system a
different subsystem simulation model for the SMR reactor (Georgopoulos,
2002) is modeled and coupled to the simulation model, since the previous
experimental model, used in study topics II-III does not allow simulation
flexibility (discretized model) and geometrical optimization. All other
components are modeled similarly as in II-III.
The paper includes the theoretical background of the optimization strategy,
including the modeling and optimizing assumptions used in the simulation
model, and the built-in EES min/max function. This function is based on
PIKAIA, which is a public domain, general purpose GA-based optimization
subroutine (Charbonneau, 2002). Internally, PIKAIA seeks to maximize a userdefined function f ( x) in a bounded n -dimensional space,
x x1 , x2 ,...xn , xk 0.0,1.0 k
(6)
Parameter values are restricted in the above range to allow greater flexibility
and adaptability across problem domains. Maximization is carried out on a
population made up of N p individuals, while the population size remains fixed
throughout the evolution. Instead of evolving the population until some
tolerance criterion is satisfied, the evolution is carried over a user-defined,
preset number of generations N g . Since breeding involves the production of
two offspring, the inner loop executes N p 2 times per generational iteration,
45
where N p is the population size. All parameter values defining the individual
members of the initial population are assigned a random number in the range
above, extracted from a uniform distribution of random deviates. This ensures
no initial bias is introduced by the initialization.
The values of the fixed parameters are given. Then the decision variables
are described, including their initial, minimum, maximum and the calculated
optimum values. The details on the optimization results including the number
of the needed generations and iterations to reach the optimum value are
described in detail. The behavioral pattern of the objective function and every
decision variable throughout the optimization procedure is illustrated and
analyzed in detail.
3.5 Paper V
The final study topic (paper V), includes a more advanced optimization
methodology, using process integration techniques. The purpose of this study
is to further increase the net electrical efficiency of the system using pinch
analysis (Kemp, 2007; Linnhoff & Flower, 1978a, 1978b; Linnhoff &
Hindmarch, 1983; Smith, 2005), and also redesign the HEN of the micro-CHP
system, using a MINLP problem formulation (Biegler et al., 1997; Grossmann,
2004; Ponce-Ortega, Jimenez-Gutierrez, & Grossmann, 2008; Viswanathan &
Grossmann, 1990; Yee & Grossmann, 1990). The objective function (PonceOrtega et al., 2008) is defined as the minimization of the total yearly cost,
which includes the cost of utilities and the fixed and variables costs of the
exchangers,
min
CCUqcu
iHPS
iHPS
iHPS
cu, j
jCPS
CHUqhu
jCPS
CF
jCPS kST
CF
jCPS
i , j i, j ,k
CF
iHPS
i , cu
zcui
(7)
zhu j
qi , j , k 1 hi , k 1 h j , k
C
i, j LMTD
kST
i, j ,k
where C is the area cost coefficient, CCU is the unit cost of cold utility, CHU
is the unit cost of hot utility, CF is the fixed charge for exchangers, CPS is the
{j|j} cold process stream, HPS is the {i|i} hot process stream, cu is the cold
utility, hu is the hot utility, h is the fouling heat transfer coefficient, qi , j , k is the
heat exchanged between hot process stream i and cold process stream j in stage
46
k, qcui is the heat exchanged between cold utility and hot stream i, qhu j is the
heat exchanged between hot utility and cold stream j, ST is the {k|k stage in the
superstructure, k=1, . . ., NOK}, zi,j,k is the set of binary variables for match (i,j)
in stage k, zcui is the set of binary variables for the match between the cold
utility and the hot stream i, zhuj is the set of binary variables for the match
between the hot utility and the cold stream j, is the exponent for area in the
cost equation and is a small number. Subscripts i and j denote the hot and
cold process streams, respectively, while k is the index for the stage (1,,
NOK) and temperature location (1,, NOK+1).
A general introduction on pinch analysis and process integration techniques
is given to explain the purpose of the study. Then, a literature review on fuel
cell-based systems using different methods of pinch analysis and process
integration (Autissier, Palazzi, Marechal, van Herle, & Favrat, 2007; Godat &
Marechal, 2003; Palazzi et al., 2007; Verda & Nicolin, 2010; Wallmark &
Alvfors, 2002; Weber et al., 2006) is given. Finally, the objectives of the
research study topic are given in steps. The simulation model is described in
detail. In general, the simulation model is the same as in IV, although there are
a few changes. The first modification is the removal of the condenser from the
system configuration. A condenser is not included, after personal
communication with the HT-PEMFC stack manufacturer (Serenergy) and the
Jlich Research Center (Korsgaard, 2011; Stolten, 2011)14. The second
modification is the inclusion of heat losses considerations for the fuel cell
stack, SMR reactor and combustor. This is done in order to make the system
simulation closer to an actual system, and also to investigate the extent of heat
loss rate from different system components.
The general pinch analysis and process integration methodology is
formulated and analyzed in detail, with all necessary steps followed in the
research study. The steps include definition of the cold and hot data set,
removal of all heat exchangers included in the initial configuration,
reevaluation of the HEN using process change techniques, and HEN
optimization to obtain a minimum total annual cost. The pinch analysis results
are first found for the initial stream data set. Based on these preliminary results
and by means of process change, the GA optimization method in EES is used
to optimize the system in terms of maximization of the net electrical efficiency.
The pinch analysis is then re-applied to the optimum stream data set, and used
in the MINLP optimization model in GAMS to minimize the total HEN annual
cost. Based on this result, a new HEN configuration is designed with all
14. Regarding the steam-content in the reformate fuel, the communication concluded that
steam could remain in the reformate fuel, without causing problems in the fuel cell stack
operation.
47
Methane
Reformate
Catalytic
Combustor
Hotfluegas
(in)
Preheated
Water
air
17
Air
Nodes 135
16
19
15
SteamGenerator
25
steam
H2O
CH4
steam
mixture
5
2
Desulfurizer
Naturalgas
I
1
CH4
WaterGas
Shift
8
II
28
10
III
13 14
Waterinjection
18
Preheater
20
Hotfluegas
(out)
24
4
Fuel
SteamMethane
Reformer
AirBlower
23
Reformatefuel
(H2Ocontentremoved)
Waterpump
29
26
Condenser/
Waterknockout
11
35
22
Exhaust
21
V
30
HotWater
Air(+H2 Ocontent)
34
IV
31
Closedcircuit
H2 Omixture
(Space Heating)
33
Exhaust
32
VI
exhaustmixture
(airwater,etc.)
AnodeCathode
27
HotWater
Depleted
fuel
12
Electricity
(togrid)
48
BypassValve
Mixer
HeatExchanger
Methane
Reformate
Water
Air
Fluegas
Catalytic
Combustor
127 Nodes
17
Fluegas
13
16
AirBlower
SteamGenerator
14
25
6
CH4steam
5
Desulfurizer
2
Naturalgas
SteamMethane
Reformer
7
Reformatefuel
II
WaterGas
Shift
I 4
III
10
15
18
3
24
19
20
23
Condenser/
Waterknockout
Waterpump
11
27
21
Totalfluegas
26
Heatloss
Air
FuelCellStack
Anode
Exhaust
Cathode
12
Depleted fuel
Thermal
Storage
Tank
Heating
demand
Electricity
(togrid)
22
Exhaustfluegas
13
Catalytic
Combustor
AirBlower
128 Nodes
14
17
Fluegas
16
SteamGenerator
15
25
19
6
CH4steam
5
Desulfurizer
SteamMethane
Reformer
I 4
II
18
2
Naturalgas
7
Reformatefuel
WaterGas
Shift
III
10
27
24
28
IV
26
Exhaust
23
Condenser/
Waterknockout
Waterpump
11
20
21
FuelCellStack
Anode
Heatloss
Air
Cathode
12
Depleted fuel
Thermal
Storage
Tank
Heating
demand
Electricity
(togrid)
22
Exhaustfluegas
49
BypassValve
Mixer
HeatExchanger
Methane
Reformate
Water
Air
Fluegas
13
Catalytic
Combustor
AirBlower
17
Fluegas
127 Nodes
14
16
SteamGenerator
15
25
19
2
1
6
CH4steam
5
Desulfurizer
3
I 4
SteamMethane
Reformer
7
Reformatefuel
18
II
WaterGas
Shift
III
27
10
24
Naturalgas
11
IV
26
Exhaust
23
Waterpump
21
20
Heatloss
FuelCellStack
Anode
Thermal
Storage
Tank
Air
Cathode
Heating
demand
12
Depleted fuel
Water
Tank
Electricity
(togrid)
22
Exhaustfluegas
WaterSeparator
50
51
Figure 13 Annual variation of electricity (production and demand) and heat (production, demand
and auxiliary).
53
due to the need for heat dumping. Further on, efficient operation has a positive
effect on CO2 emissions, which have been lowered to 2609 kgCO2, as
compared to 2931 and 4653 kgCO2 for the electricity- and heat-led operational
strategies, respectively.
Table 4 Performance of the three considered operational strategies: Overall comparison
Variable
Description [unit]
Electricityled
Heat-led
Improved
ELHV ,in
15630
26475
13805
Eel , prod
4984
7164
4533
Eel ,imp
753
1127
Eel,exp
2933
676
Eheat , prod
8136
12639
7334
Eheat ,aux
6502
5305
Eheat ,dump
sys
el ,net
71.2
74.8
85.9
27.1
32.8
th
39.3
47.7
53.1
eCO2
2931
4653
2609
54
The objective function was improved greatly, 40.9% compared to the initial
32.6%, suggesting that there was indeed a large space of improvement, with
respect to the selected objective function. The maximization evolution of the
Table 5 Initial, minimum, maximum and optimum values of the decision variables used in the
optimization procedure.
Variable
Description (unit)
Initial
Min
Max
Optimum
SC
Steam-to-carbon ratio
4.000
2.500
4.000
2.908
Hydrogen stoichiometry
1.500
1.500
1.550
1.509
275
275
300
297
1.800
1.800
2.000
1.994
0.350
0.350
0.500
0.465
1173
1050
1174
1095
400
399
530
530
499
470
500
474
550
670
620
SMR
tubes
SMR
reactor
WGS
reactor
Comb
fg ,out
SMR
ref ,in
WGS
ref ,in
FP
fg ,in
55
Figure 15 Hot and cold composite curves extracted from the initial process stream data set.
56
Description (unit)
T6
Initial
Min
Max
Optimum
400
600
587
T11
500
700
597
m 17
6.50E-3
6.87E-3
6.55E-3
SC
2.50
4.00
2.83
626
3.19
Pinch analysis is then repeated for the new stream data set obtained by the
optimization strategy. The procedure accomplished minimization of the cold
utility requirement from 1.78 to 1.54 kW. In a threshold problem it is necessary
to distinguish its characteristics. The system performance is then tested at partload operation to examine the response of the model at off-design conditions15.
The variation of efficiency at different net electrical power outputs is shown in
Figure 16. As the load decreases, the net electrical efficiency increases linearly,
while the thermal efficiency and total efficiency decrease because of heat
losses.
Since component16 temperatures and areas remain constant at all loads, heat
losses have an increasing effect at lower operational loads (see Figure 17).
Therefore, system operation at low, and especially critical, loads may not be
favored at periods of high demand on cogeneration heat. The greater loss is due
to the SMR reactor, which although it is heavily insulated, it contributes 50%
of the losses. This is due to the high temperature flows (reformate and flue gas)
occurring in the SMR reactor. Finally, the MINLP model is applied using
15. In a realistic system, part-load conditions may not allow smooth operation without the use
of BOP components, such as by-pass valves between heat exchangers.
16. Heat losses are considered for three components: (a) Fuel cell stack, (b) Combustor and (c)
SMR reactor. Heat losses for the WGS reactor are neglected, since the process is slightly
exothermic. Also, heat losses for heat exchanger units are not considered, because the HEN
configuration is not known a priori.
57
GAMS. The compact nature of the micro-CHP system does not favor the use
of stream splitting, and therefore this is restricted in the code. In addition to the
potentially more complex operation, stream splitting would require an
additional investment cost for a control valve.
58
The grid diagram representation, shown in Figure 18, shows the heat transfer
operation of the optimum HEN. Hot streams are at the top running from left to
right, while cold streams are located at the bottom running from right to left.
Heat exchange matches are represented by vertical lines joining two arrows on
the two streams being matched.
Figure 19 Configuration of the proposed micro-CHP system after the application of the
simultaneous HEN optimization.
59
The coupling of the optimized HEN configuration with the micro-CHP system
is shown in Figure 19. The new configuration is significantly different than the
original one, due to the cost reduction requirements imposed in the objective
function. Therefore, the optimum configuration encompasses the need for heat
exchangers with a cheaper structure, such as a minimum heat transfer area. The
heat exchangers must allow a more cost efficient heat transfer distribution
along the streams, although the total number of heat exchangers remains the
same.
Overall, the efficiency was improved from an initial value of 31.9% to an
optimum value of 35.2% (at full-load). The highest net electrical efficiency
was 43.6%, while the highest total system efficiency was 91.1%. The MINLP
optimization strategy, managed to reach a minimized HEN total cost of
$8,147/year. The new HEN configuration is significantly different than the
initial one, indicating the level of improvement made after the application of
the optimization strategy. Finally, it should be noted that a more complete
analysis should include part-load results for the new HEN configuration. This
can be accomplished by modification of the original micro-CHP system
simulation model, to verify that HEN operability is indeed feasible.
60
This thesis work has defined and completed a series of tasks to address the
main research question associated with the modeling, design and optimization
of a HT-PEMFC-based micro-CHP system. Components and subsystems are
modeled from first principles or modified previously developed models are
used to build up the total simulation system model. The obtained simulation
system models (see papers I-II) are validated successful using a comparable
model from the literature (Korsgaard et al., 2008).
5.1 Conclusions
A posteriori, it can be concluded that the structure of the five study topics
could have been completed in a slightly different order. After topic #2, topic #3
could have been left as the final topic of the research project. This is because
the system is optimized during topics #4-5, suggesting that better results could
have been obtained if an optimum system configuration was used to reach an
improved operational strategy. Further on, the first study topic proved to be
very time consuming and not so influential in the overall progression of the
research project. Therefore, if the project was repeated, this part would have
been omitted. If the two modeling methods, LabVIEWTM and EES-based
simulations, are compared, it can be concluded that EES modeling is easier and
possesses more capabilities, as illustrated throughout this research work.
Moreover, although LabVIEWTM simulations can provide a high degree of
easiness to the end user in terms of calculating manipulation, it is difficult to
model and modify highly complicated models, due to the graphical nature of
the program. Also LabVIEWTM has a limited number of parameters that can be
transferred from a subVI to the main VI (Virtual Instrument). Therefore,
LabVIEWTM modeling is more appropriate for single component dynamic
systems.
61
62
63
References
Ahlstrom-Silversand, A. F., & Odenbrand, C. U. I. (1999). Modelling catalytic
combustion of carbon monoxide and hydrocarbons over catalytically
active wire meshes. Chemical Engineering Journal, 73(3), 205-216.
doi:10.1016/S1385-8947(99)00029-7
Al-Sulaiman, F. A., Hamdullahpur, F., & Dincer, I. (2011). Trigeneration: A
comprehensive review based on prime movers. International Journal of
Energy Research, (March 2010), 233-258. doi:10.1002/er
Ang, S. M. C., Fraga, E. S., Brandon, N. P., Samsatli, N. J., & Brett, D. J. L.
(2011). Fuel cell systems optimisation Methods and strategies.
International Journal of Hydrogen Energy, 36(22), 14678-14703.
Elsevier Ltd. doi:10.1016/j.ijhydene.2011.08.053
Arsalis, A. (2007). Thermoeconomic modeling and parametric study of hybrid
SOFCgas turbinesteam turbine power plants ranging from 1.5 to 10
MWe. M.S. thesis, advisor: M.R. von Spakovsky, Virginia Polytechnic
Institute & State University, Blacksburg, Virginia.
Arsalis, A. (2008). Thermoeconomic modeling and parametric study of hybrid
SOFCgas turbinesteam turbine power plants ranging from 1.5 to 10
MWe. Journal of Power Sources, 181(2), 313-326.
doi:10.1016/j.jpowsour.2007.11.104
Autissier, N., Palazzi, F., Marechal, F., van Herle, J., & Favrat, D. (2007).
Thermo-Economic Optimization of a Solid Oxide Fuel Cell, Gas Turbine
Hybrid System. Journal of Fuel Cell Science and Technology, 4(2), 123.
doi:10.1115/1.2714564
BAXI. (2011a). Baxi SenerTec Dachs - Internal Combustion Engine. Retrieved
November 11, 2011, a from http://www.baxi.co.uk/products/DACHS.htm
65
66
67
68
70
71
Ropenus, S., Schrder, S. T., Costa, A., & Ob, E. (2010). Support Schemes
and Ownership Structures: The Policy Context for Fuel Cell Based
Micro-Combined Heat and Power. Ris, Denmark. Retrieved November
25, 2011, from
http://www.bigscience.dk/sitecore/content/Risoe_dk/Home/Knowledge_b
ase/publications/Reports/ris-r-1730.aspx?sc_lang=en
Salcines, D. S., Estbanez, C. R., & Herrero, V. C. (2004). Simulation of a
solar domestic water heating system with different collector efficiencies
and different volumen storage tanks. International Conference on
Renewable Energies and Power Quality. Barcelona.
Serenergy. (2011). serenergy.dk. Retrieved February 2, 2011, from
http://www.serenergy.dk/
Sieniutycz, S., & Jezowski, J. (2009). Energy Optimization in Process Systems
(1st ed.). Amsterdam: Elsevier.
Slowe, J. (2010). Can small make it big? The progress of micro-CHP in world
mass markets. Retrieved November 20, 2011, from
http://www.cospp.com/articles/print/volume-7/issue-4/features/cansmall-make-it-big-the-progress-of-micro-chp-in-world-massmarkets.html
Smith, R. (2005). Chemical Process Design and Integration. Chichester, UK:
Wiley.
Stolten, D. (2010). Hydrogen and Fuel Cells: Fundamentals, Technologies and
Applications (1st ed.). Weinheim, Germany: Wiley-VCH.
Stolten, D. (2011). Personal communication concerning water-content in the
fuel cell stack anode inlet. Aalborg, Denmark.
Thomas, B. (2008). Benchmark testing of Micro-CHP units. Applied Thermal
Engineering, 28(16), 2049-2054.
doi:10.1016/j.applthermaleng.2008.03.010
Tokyo Gas Ltd. (2011). Tokyo Gas and Panasonic to Launch New Improved
Ene-Farm Home Fuel Cell with World-Highest Generation Efficiency
at More Affordable Price. Retrieved November 20, 2011, from
http://www.tokyo-gas.co.jp/Press_e/20110209-01e.pdf
72
73
74
Appendix
The appendix includes all journal articles mentioned in this thesis work (I-V).
Contents
Journal
Journal
Journal
Journal
Journal
Article
Article
Article
Article
Article
I
3
II 15
III 29
IV 47
V 61
APPENDIX
APPENDIX
Journal Article I
APPENDIX
APPENDIX
Energy
journal homepage: www.elsevier.com/locate/energy
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 10 May 2010
Received in revised form
25 November 2010
Accepted 7 December 2010
Available online 8 January 2011
A novel proposal for the modeling and operation of a micro-CHP (combined-heat-and-power) residential
system based on HT-PEMFC (High Temperature-Proton Exchange Membrane Fuel Cell) technology is
described and analyzed to investigate its commercialization prospects. An HT-PEMFC operates at
elevated temperatures, as compared to Naon-based PEMFCs and therefore can be a signicant candidate for cogeneration residential systems. The proposed system can provide electric power, hot water,
and space heating for a typical Danish single-family household. A complete fuel processing subsystem,
with all necessary BOP (balance-of-plant) components, is modeled and coupled to the fuel cell stack
subsystem. The micro-CHP system is simulated in LabVIEW environment to provide the ability of Data
Acquisition of actual components and thereby more realistic design in the future. A part-load study has
been conducted to indicate performance characteristics at off-design conditions. The system is sized to
provide realistic dimensioning of the actual system.
2010 Elsevier Ltd. All rights reserved.
Keywords:
PBI membrane
HT-PEMFC systems
Micro-CHP
Residential systems
Fuel processing
1. Introduction
Fuel cell-based stationary power generation offers a great
market opportunity, because the fuel cell technology is capable of
achieving higher efciencies, with lower emissions as compared to
conventional power systems [1e3]. Residential fuel cell systems
can be grid-interconnected to allow power ow from/to the grid as
needed. This design offers greater exibility than a stand-alone
system, and is very attractive if the incoming power is produced by
renewable energy sources, such as wind power [3,4]. This means
that when cheap wind power can be produced, the fuel cell system
can operate at a minimum load and therefore reduce the fuel
consumption. A fuel cell-based micro-CHP (Combined-Heat-andPower) system converts on-site the chemical energy in a fuel, e.g.
natural gas, into electrical power and heat, as required by the
household demand. The range of energy demand for a Danish
household in terms of electricity and heat is 0.5e5.0 kWe, and
2e10 kWth, respectively. Therefore, for this system conguration,
a fuel-processing unit must be coupled with the fuel cell stack, to
allow conversion of natural gas to hydrogen. Several BOP (BalanceOf-Plant) components are needed for the controlling and smooth
operation of the system, while heat exchangers are necessary for
the thermal management of the system. The thermal management
of the system includes heating/cooling of components (e.g. steam
reforming), and also heat recovery to satisfy the residential load
prole (e.g. space heating).
The operating temperature in a fuel cell stack is considered an
important factor to the efciency and the degradation of the
membrane. High operating temperatures reduce the cooling
requirements, simplify water management and lessen contamination problems. An HT-PEMFC (high temperature-proton exchange
membrane fuel cell) utilizes a PBI (Polybenzimidazole) membrane,
operating at temperatures between 160 C and 200 C. An HTPEMFC is therefore an ideal match for a micro-CHP system, because
not only the rates of electrochemical kinetics are enhanced and
water management and cooling is simplied, but also useful waste
heat can be recovered, and lower quality reformed hydrogen may
be used as fuel [5]. However, the frequent changes of demand, in
terms of electrical power and/or heat, require operation in varying
partial-load conditions. A simple and compact design, with efcient
adaptability to load changes must be accomplished, which is vital
for this type of application.
0360-5442/$ e see front matter 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.energy.2010.12.009
APPENDIX
994
The Danish micro-CHP project is managed by a national consortium consisting of nine Danish energy companies. The consortium combines all the competencies necessary to develop, test, and
demonstrate micro-CHP systems. The project cooperates with
a wide range of political and specialist parties, such as the Danish
Energy Agency. It is also nancially assisted by the Danish Ministry
of Climate and Energy by 40% of the total cost of the project. It is
scheduled to run from 2006 to 2012, and includes fuel cell systems
based on three different technologies: LT-PEMFC (low temperatureproton exchange membrane fuel cell), HT-PEMFC and SOFC (Solid
Oxide Fuel Cell). It is divided into three phases, and two fuel types
are used: hydrogen and natural gas. The choice of fuel depends on
the fuel cell technology and its availability at the installation sites.
A preliminary assumption concerning the operational control of the
system, indicated that the electrical load must be fullled, but also
the design should secure the right amount of heat will be available
on demand for the space heating and hot water loads. Further on,
an auxiliary burner for peak production of thermal energy is under
consideration to be integrated to the end-user system [3,6].
The HT-PEMFC-based micro-CHP units are developed by Dantherm
Power and operate on natural gas. Experimental tests, with pure
hydrogen fuel, showed start-up times of 30e60 min, with an electrical
efciency of 40e58% (LHV (Lower Heating Value)-based). From the
operating point of the nalized design, an efciency of 50% is expected.
Further experimental tests and calculations showed a potential system
efciency of 85e90% (LHV-based) [6]. A fast adaption to load variations
was also observed. In the rst phase, Danish Micro Combined-Heatand-Power developed micro-CHP unit prototypes with PEMFC and
SOFC fuel cells. In the second phase, ten micro-CHP units are scheduled
APPENDIX
995
the purposes of this research study, the entire load prole, shown in
Table 1, has been simplied into three time segments, i.e. winter,
summer, and spring (autumn).
By observation, it is clear that the winter segment appears to be
the one with the most stringent for electrical and space heating
loads, while hot water load is most stringent during spring season.
Therefore the system must be designed to fulll the maximum
winter electrical load 0.95 kWe, since the proposed micro-CHP
system assumes an electricity-led operation. It should be noted that
at short time segments the electrical power demand will exceed
1 kWe. In this case, electrical power will be imported from the grid.
On the other hand, the remaining heat load demand can be satised
with external means, e.g. auxiliary burner. It should be noted that
these are beyond the scope of this study, and only the percentage of
heating load fullled from the proposed system is analyzed.
The various efciencies should be dened, before the modeling
details are given in the next section. The cogeneration thermal efciency is dened as the ratio of the actual waste heat recovered by heat
exchangers IV, V, and VI and the chemical energy input to the system,
hthermal cogen:
Q_ HEx
cogen:
(1)
Time segment
Winter
Summer
Spring
Mean
Max
Mean
Max
Mean
Max
540
1450
330
950
1930
1080
380
70
230
650
140
1420
460
600
310
920
1070
1620
helectrical net
_
W
electrical net
_ CH4 ;in LHVCH4
m
(2)
hsystem
cogen:
_
W
electrical
net
Q_ HEx
cogen:
(3)
APPENDIX
996
(4)
(5)
hohmic iRohmic
The total cell voltage is given by,
Vcell V0 ha hc hohmic
i i
i
hc
Rdiff
ln 0
lair 1
4acathode F
i0
RTcell
Wcell Vcell i
(10)
(11)
(12)
Istack
Pstack
Vstack
(13)
DH298 206:2kJ=mol
DH298 41:1kJ=mol
DH298 164:9kJ=mol
(14)
(15)
(16)
The heat transfer model is based on [10], where the shell-andtube heat integration model is based on a transient model. For the
current research study, the model was simplied in a steady-state
one, where the partial differential equations are reduced into
ordinary differential equations. The convective heat transfer coefcient of the inside tube wall is based on a semi-empirical relation
for spherical packing [11].
The heat transfer coefcient for a packed-bed tubular reactor is
obtained from the following empirical correlation,
hw 2:03Re0:8
p
kref
6dp
exp
Di
Di
(17)
(6)
(9)
RTcell
i
ha
arcsinh
aanode F
2keh qH2
(8)
(7)
Deq Gs 0:55 Cp;s ms 1=3 ms 0:14
ks
hs 0:36
ms
mwall
Deq
ks
APPENDIX
(18)
997
phs Do Tw Ts
dTs
_ s Cp;s
dz
m
(19)
Tw
Do hs Ts Di hw Tref
Do hs Di hw
(20)
CO H2 O#CO2 H2
(21)
KT exp
4400
4:063
T
(22)
Ea
RT
KWGS k0 exp
(23)
3.5. Mixers & by-pass valves
dxCO
3600rCO
urwiremesh
n_ CO;i
dz
(24)
The mixers and by-pass valves used in the micro-CHP system are
necessary for the operation and the regulation of the system. The
modeling of each mixer and by-pass valve is done with simple mass
and energy balances, assuming no pressure losses.
(25)
rwiremesh
dT
dz
rg cp;mass us rCO dHr;1
3.4. Plate heat exchangers
APPENDIX
998
(26)
(27)
_ pump
_ pump
m
m
out
in
(28)
_ pump m
_ in hout hin pump
W
(29)
4. System validation
The validation of the SMR reactor, the WGS reactor, the HTPEMFC stack, and the overall system were compared with reference
models from the literature [3]. Before validation is explained, it
should be noted that mass balances for every component input/
output were performed; the mass balances veried the validity of
the results. All three individual subsystems showed only minor
discrepancies between the reference models and the models under
study.
Although an analytical comparison between the research under
study and a reference system is not possible, due to signicant
differences between the current system and the systems in the
literature, some values can be compared to investigate validity, and
also the sources of discrepancy. The chosen reference system is the
HT-PEMFC-based micro-CHP system developed by [3]. The results
are given in Table 2. The most signicant differences between the
two systems are the following:
Table 2
Overall system validation.
Input values
Fuel cell active area, Acell
Results
_ net
Electrical power output, W
Korsgaard, 2008
Proposed system
10
135 cm2
Number of cells
65
1500 W
666 W
669 W
0.45
0.4455
0.88
0.95
APPENDIX
999
Table 3
Simulation results for full-load and part-load operation.
Node
T( C)
Full-load operation
1
10.0
7
602.5
8
250.0
9
296.6
10
160.0
11
160.0
12
160.0
16
85.3
17
820.0
18
698.3
19
597.2
20
592.5
21
469.4
22
313.6
23
10.0
25
200.0
27
60.0
30
60.0
33
90.8
75%-load operation
1
10.0
7
677.2
8
250.0
9
323.9
10
160.0
11
160.0
12
160.0
16
88.3
17
718.0
18
678.2
19
593.0
20
589.0
21
422.6
22
316.2
23
10.0
25
200.0
27
60.0
30
60.0
32
160.0
33
64.0
50%-load operation
1
10.0
7
588.3
8
250.0
9
290.2
10
160.0
11
160.0
12
160.0
16
60.6
17
610.0
18
588.6
19
514.5
20
511.0
21
328.9
22
95.3
23
10.0
25
200.0
27
60.0
30
60.0
32
160.0
33
94.6
25%-load operation
1
10.0
7
577.6
8
250.0
9
285.5
10
160.0
11
160.0
12
160.0
16
56.7
17
585.0
18
577.6
_
mkg=s
xCH4
xCO
xCO2
xO2
xN2
xair
xH2 O
xH2
5.530E-05
2.587E-04
2.587E-04
3.259E-04
3.259E-04
1.523E-04
4.569E-05
4.768E-03
4.841E-03
4.841E-03
4.841E-03
4.841E-03
4.841E-03
4.841E-03
4.871E-03
2.034E-04
6.000E-04
4.000E-03
1.695E-03
1.000
0.031
0.031
0.026
0.026
0.043
0.043
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.007
0.000
0.048
0.048
0.003
0.003
0.005
0.005
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.001
0.000
0.088
0.088
0.112
0.112
0.187
0.187
0.000
0.006
0.006
0.006
0.006
0.006
0.006
0.000
0.000
0.000
0.000
0.031
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.195
0.195
0.195
0.195
0.195
0.195
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.777
0.777
0.777
0.777
0.777
0.777
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
1.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.809
0.000
0.336
0.336
0.402
0.402
0.000
0.000
0.000
0.022
0.022
0.022
0.022
0.022
0.022
1.000
1.000
1.000
1.000
0.152
0.000
0.498
0.498
0.457
0.457
0.764
0.764
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
3.440E-05
1.609E-04
1.609E-04
2.047E-04
2.047E-04
1.044E-04
3.132E-05
3.538E-03
3.589E-03
3.589E-03
3.589E-03
3.589E-03
3.589E-03
3.589E-03
1.170E-03
1.265E-04
4.000E-04
2.000E-03
1.224E-03
1.224E-03
1.000
0.007
0.007
0.006
0.006
0.009
0.009
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.001
0.001
0.000
0.076
0.076
0.006
0.006
0.010
0.010
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.002
0.002
0.000
0.077
0.077
0.122
0.122
0.190
0.190
0.000
0.005
0.005
0.005
0.005
0.005
0.005
0.000
0.000
0.000
0.000
0.030
0.030
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.197
0.197
0.197
0.197
0.197
0.197
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.778
0.778
0.778
0.778
0.778
0.778
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
1.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.813
0.813
0.000
0.306
0.306
0.356
0.356
0.000
0.000
0.000
0.020
0.020
0.020
0.020
0.020
0.020
1.000
1.000
1.000
1.000
0.153
0.153
0.000
0.534
0.534
0.509
0.509
0.791
0.791
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
2.730E-05
1.277E-04
1.277E-04
1.602E-04
1.602E-04
7.176E-05
2.153E-05
3.314E-03
3.345E-03
3.345E-03
3.345E-03
3.345E-03
3.345E-03
3.345E-03
4.733E-03
1.004E-04
2.600E-04
1.800E-03
7.782E-04
7.782E-04
1.000
0.041
0.041
0.035
0.035
0.060
0.060
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.010
0.010
0.000
0.041
0.041
0.003
0.003
0.004
0.004
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.001
0.001
0.000
0.088
0.088
0.106
0.106
0.184
0.184
0.000
0.004
0.004
0.004
0.004
0.004
0.004
0.000
0.000
0.000
0.000
0.031
0.031
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.199
0.199
0.199
0.199
0.199
0.199
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.781
0.781
0.781
0.781
0.781
0.781
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
1.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.807
0.807
0.000
0.355
0.355
0.423
0.423
0.000
0.000
0.000
0.015
0.015
0.015
0.015
0.015
0.015
1.000
1.000
1.000
1.000
0.152
0.152
0.000
0.475
0.475
0.433
0.433
0.751
0.751
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
1.350E-05
6.314E-05
6.314E-05
7.904E-05
7.904E-05
3.447E-05
1.034E-05
1.708E-03
1.722E-03
1.722E-03
1.000
0.048
0.048
0.041
0.041
0.072
0.072
0.000
0.000
0.000
0.000
0.036
0.036
0.002
0.002
0.004
0.004
0.000
0.000
0.000
0.000
0.088
0.088
0.103
0.103
0.183
0.183
0.000
0.004
0.004
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.200
0.200
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.782
0.782
0.000
0.000
0.000
0.000
0.000
0.000
0.000
1.000
0.000
0.000
0.000
0.367
0.367
0.436
0.436
0.000
0.000
0.000
0.014
0.014
0.000
0.461
0.461
0.418
0.418
0.742
0.742
0.000
0.000
0.000
APPENDIX
11
1000
Table 3 (continued )
Node
T( C)
_
mkg=s
xCH4
xCO
xCO2
xO2
xN2
xair
xH2 O
xH2
19
20
21
22
23
25
27
30
32
33
506.2
502.9
148.3
93.3
10.0
200.0
60.0
60.0
160.0
90.9
1.722E-03
1.722E-03
1.722E-03
1.722E-03
4.666E-03
4.964E-05
1.300E-04
9.000E-04
3.673E-04
3.673E-04
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.012
0.012
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.001
0.001
0.004
0.004
0.004
0.004
0.000
0.000
0.000
0.000
0.031
0.031
0.200
0.200
0.200
0.200
0.000
0.000
0.000
0.000
0.000
0.000
0.782
0.782
0.782
0.782
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.805
0.805
0.014
0.014
0.014
0.014
1.000
1.000
1.000
1.000
0.152
0.152
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
Table 4
Overall power-to-heat performance for the adopted load prole.
12
Load
_
W
electrical net
Q_ HEx cogen:
Total cogeneration
Heat load fullled (%)
Winter, mean
Winter, max
Summer, mean
Summer, max
Spring, mean
Spring, max
540
950
380
650
460
920
840
1560
580
1000
700
1470
47.2
51.8
193.0
64.1
76.9
54.6
characteristics, etc. The results for all the nodes, shown in Fig. 1, are
given with the following parameters: temperature, mass ow rate,
and mole fractions. In Table 3, the results for full-load, 75%, 50% and
25%-load operation are given. The fuel cell stack operates at
a constant temperature of 160 C to avoid fast degradation of the
membrane, if operated at higher temperatures [14,15]. Also the
temperature is kept constant, because rapid and frequent variations
of the operating temperature would contribute to faster deterioration of the membrane [5,14]. The corresponding fuel cell operating pressure is kept at a value slightly above atmospheric, since an
increase of pressure would not contribute signicantly in the
system efciency, and also the degradation of the membrane would
be faster [14,15].
The current density variation for the given loads and the HTPEMFC stack efciency (LHV) variation for the given loads are given
in Fig. 5. The current density is varied from 0.19 A/cm2 at full-load
linearly up to 0.04 A/cm2 at 25% load. The efciency follows an
analogous path, increasing at part-load to 45.5%.
The net electrical power output at full-load and part-loads vs.
the corresponding cogeneration heat loads is shown in Fig. 6. The
systems useful heat capacity was analyzed in Section 2. For this
load prole the net electrical power output, along with the corresponding total cogeneration heat output are given in Table 4. In the
APPENDIX
6. Conclusions
Table 5
Component geometries for the proposed micro-CHP system.
Geometric variable
description
ncells
DSMR
i
DSMR
o
Number of cells
SMR inlet
tube diameter
SMR outlet
tube diameter
Value
Geometric variable
description
Value
176
0.1016 m
LSMR
DWGS
i
0.8 m
0.13 m
0.1322 m
LWGS
0.45 m
same table, the percentage of heat output fullling the aforementioned load prole at every time segment is also given. As expected,
the combined heat load is satised better in the summer segment.
It should be noted that there would still be a decit of heat at some
instances in the summer if no heat storage was provided. In the case
of the other two time segments, winter and spring, an auxiliary
burner will be required at some instances, even if a heat storage
system is used. More specically for the winter segment, the heat
load can be fullled at around 50% in most instances. The system
should avoid operation at part-loads below 20% because then
higher selected nominal cell voltage does not necessarily mean
higher operating efciency; at very low current densities where
parasitic losses (including gas permeation through the polymer
membrane) may not be negligible [16].
The mass ow variation of hydrogen, methane, and carbon
monoxide throughout the proposed micro-CHP system is shown in
Fig. 7. Mole fractions have not been used (although it could be done
with mole fractions at dry basis), because it would appear as if the H2
content has decreased at the exit of the WGS reactor. This is because
of the water addition at the reactors inlet, which alters the total
mass ow rate (and therefore the mole fractions) exiting the WGS
reactor. By observation the SMR reactor is very efcient, converting
hydrogen to 2.01E-5 kg/s for a methane input of 5.52E-5 kg/s. This
suggests that the current conguration is very efcient and appropriate for the application under study, provided this can be proved
with experimental calculations in the future. The CO content is
reduced from 2.71E-5 kg/s, at the SMR outlet, to 0.21E-5 kg/s. The
HT-PEMFC stack is assumed to withstand CO-contents in the inlet
reforming fuel of up to 10,000 ppm. The CO-content was maintained
within this limit at full- and part-load. At full-load it was calculated
to be 5480 ppm (see Table 3). This result veries that the current
design of the fuel processing system, although simple, it can be
adequate for the needs of the fuel cell stack, in terms of CO-content
removal.
The HT-PEMFC stack, SMR reactor and the WGS reactor geometries for the proposed micro-CHP system are given in Table 5. The
UA-values for the heat exchangers used in the proposed micro-CHP
system are given in Table 6. These component values satisfy
a compact, yet efcient micro-CHP system, as required for this
application [4].
Table 6
UA-values [W/(m2- C)] for the heat exchangers used in the proposed micro-CHP
system.
Geometric variable
description
UAI
UAII
UAIII
UASG,sup
UASG,ev
Natural gas
pre-heater
SMR/WGS cooler
Value
Geometric variable
description
0.0538784 UAIV
0.720725
UAV
WGS/HT-PEMFC
0.548661
cooler
Superheater
0.107073
(steam generator)
Evaporator
0.843147
(steam generator)
UAVI
UASG,ec
1st Cogeneration
heat exchanger
2nd Cogeneration
heat exchanger
3rd Cogeneration
heat exchanger
Economizer
(steam generator)
1001
Value
1.41641
2.36245
1.39233
0.0896212
APPENDIX
13
1002
14
[10] Kim K. Dynamic proton exchange membrane fuel cell system synthesis/design
and operation/control optimization under uncertainty. Blacksburg, VA: PhD
Dissertation, Virginia Tech; 2008.
[11] Li C, Finlayson BA. Heat transfer in packed bedsea reevaluation. Chemical
Engineering Science 1977;32:1055e66.
[12] Keiski RL, Salmi T, Niemisto P, Ainassaari J, Pohjola VJ. Stationary and transient
kinetics of the high temperature water-gas shift reaction. Applied Catalysis A:
General 1996;137(2):349e70.
[13] Davies J, Lihou D. Optimal design of methane steam reformer. Chemical and
Process Engineering 1971;52:71e80.
[14] Korsgaard AR, Refshauge R, Nielsen MP, Bang M, Kr SK. Experimental
characterization and modeling of commercial polybenzimidazole-based MEA
performance. Journal of Power Sources 2006;162(1):239e45.
[15] Bchi FN, Inaba M, Schmidt TJ. Polymer electrolyte fuel cell durability. New
York, NY, USA: Springer; 2009.
[16] Barbir F. PEM fuel cells: theory and practice. Elsevier; 2005.
[17] Peacock A, Newborough M. Impact of micro-CHP systems on domestic
sector CO emissions. Applied Thermal Engineering 2005;25(17e18):
2653e76.
Nomenclature
Acell: Fuel cell active area (cm2)
Cp,s: Specic heat of shell-side gas in SMR (kJ/(kmol.K))
dp: Catalyst diameter, SMR reactor (m)
dHr,1: Enthalpy of reaction (kJ/mol)
Di: Inlet diameter (m)
Do: Outlet diameter (m)
Deq: Equivalent diameter (m)
F: Faraday constant (C/mol)
Gs: Mass ux (kg/(m2.s))
hs: Shell-side heat transfer coefcient (W/m2.K)
hw: Tube-side heat transfer coefcient (W/m2.K)
i: Current density (A/cm2)
Istack: Fuel cell stack current (A)
i0: Exchange current density (A/cm2)
keh: H2 electro-oxidation rate constant (A/cm2)
ki: Thermal conductivity of species i (W/K.m)
KT: Equilibrium constant
KWGS: Kinetic power law t parameter (m^1.92/kmol^-0.36.kg.s)
LHVi: Lower heating value of species i (J/kg)
_ i : Mass ow rate of species i (kg/s)
m
ncells: Number of cells in the fuel cell stack
n_ CO : CO molar feed rate (kmol/h)
Pstack: Fuel cell stack electrical power output (W)
Q_ i : Heat transfer rate of component i (W)
rCO: CO reaction rate (kmol/kg.s)
R: Ideal gas constant (J/K.mol)
Rdiff: Diffusion resistance (Ohm.cm2)
Rohmic: Ohmic resistance (Ohm.cm2)
Rep: Reynolds number
Tcell: Fuel cell operating temperature (oC)
Ti: Temperature of species i (oC)
Vcell: Total cell voltage (V)
Vstack: Fuel cell stack voltage (V)
V0: Open circuit voltage (V)
us: Supercial velocity (m/s)
Wcell: Power density (W/m2)
_
W
electrical : Electrical power output (W)
z: Distance along the reactor length (m)
ai: Charge transfer coefcient
ha: Anode overpotential (V)
hc: Cathode overpotential (V)
hohmic: Ohmic losses (V)
hi: Efciency of component/system i
qH2 : H2 adsorption/desorption constant
lair: Air stoichiometry, cathode side
mi: Viscosity of species i (kg/s.m)
rg: Density of the reformate gas (kg/m3)
rwiremesh: Density of the wiremesh catalytic material (kg/m3)
u: WGS reactor cross-sectional area (m)
Subscripts/superscripts
cogen.: Cogeneration
HEx: Heat exchanger
in: Input stream
out: Output stream
ref: Reformate gas
s: Flue gas
thermal: Thermal power
APPENDIX
Journal Article II
APPENDIX
15
16
APPENDIX
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 6 ( 2 0 1 1 ) 5 0 1 0 e5 0 2 0
Available at www.sciencedirect.com
article info
abstract
Article history:
17 January 2011
systems, since it can operate at higher temperature than Nafion-based fuel cells, and
therefore can reach higher cogeneration efficiencies. The proposed system can provide
electric power, hot water, and space heating for a typical Danish single-family household.
Keywords:
PBI
modeled and coupled to the fuel cell stack subsystem. The micro-CHP systems synthesis/
HT-PEMFC system
design and operational pattern is analyzed by means of a parametric study. The parametric
Micro-CHP
Residential system
maximizing total system efficiency, without violating the requirements of the system. Four
Parametric study
1.
Introduction
* Corresponding author.
E-mail address: aar@et.aau.dk (A. Arsalis).
0360-3199/$ e see front matter Copyright 2011, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved.
doi:10.1016/j.ijhydene.2011.01.121
APPENDIX
17
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 6 ( 2 0 1 1 ) 5 0 1 0 e5 0 2 0
1.1.
Several research groups investigated the modeling and operation of fuel cell-based micro-CHP systems throughout the
recent years. Georgopoulos [7] studied the design, synthesis
and operational optimization of a PEMFC-based micro-CHP
system using a decomposition methodology, while considering different system configurations. Ferguson and Ugursal
[8] studied operating strategies for different PEMFC sizes and
their effect on the performance of a cogeneration system,
while demonstrating that these are critical factors affecting
the performance of these systems. Obara [9] studied the
heating network of residential fuel cell-based micro-CHP
systems showing how the hot-water piping network can be
optimized. Godat and Marechal [10] applied process integration techniques on a PEMFC-based micro-CHP to identify its
optimal operating conditions along with the fuel cells optimal
process structure. Braun [11] studied a SOFC-based residential
system in terms of fuel supply and concluded that the efficiency performance advantages of methane-fueled SOFC
systems compared to hydrogen-based SOFC systems can be as
high as 6%.
Finally, Korsgaard et al. [4,5] investigated the modeling
prospects of a higher temperature PEMFC-based micro-CHP
system. The system was modeled in MATLAB Simulink to
enable a dynamic modeling of the system with the application
of different control strategies. Based on the promising results
extracted from this latter research study, the current research
work investigates the modeling of such a high temperature
PEMFC-based micro-CHP system to investigate in greater
detail and precision the design/synthesis of key system
components, such as the reactors, heat exchangers and
combustor. Further on, the modeling of a novel plate heat
18
5011
2.
System layout
Q_ HEx cogen
_ CH4 ;in LHVCH4
m
(1)
_ el
W
net
_
mCH4 ;in LHVCH4
(2)
cogen
3.
_ el Q_ HEx
W
net
cogen
_ CH4 ;in LHVCH4
m
(3)
System modeling
EES was selected as the modeling tool for this research study
because it includes many built-in mathematical and thermophysical property functions, while selected input variables
APPENDIX
5012
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 6 ( 2 0 1 1 ) 5 0 1 0 e5 0 2 0
APPENDIX
19
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 6 ( 2 0 1 1 ) 5 0 1 0 e5 0 2 0
3.1.
The fuel cell stack model is based on [4,12,13]. The fuel cell stack
is sized in a design point that will fulfill the required electrical
power output of the micro-CHP system by adjusting the number
of fuel cells in the stack and the current density at the nominal
load. The model considers only the reaction of hydrogen with
oxygen, while all other species are considered inactive.
The ohmic and diffusion resistances are based on the
following linear regressions with experimental values found
in [12],
Rohmic 0:0001667Tcell 0:2289
(4)
(5)
hc
RTcell
i
1
sinh
aanode F
2keh qH2
(6)
RTcell
i0 i
i
Rdiff
ln
4acathode F
i0
lair 1
(7)
5013
nie nio
2
X
ni;j xj
i1
(8)
(9)
_i
m
in
X
_i
m
(10)
out
Hi in WFC;el:
Hi out
(11)
The inputs are the enthalpies of all the flows into the fuel
cell stack, while the outputs are the electric power produced
and enthalpies of all the flows out of the fuel cell.
3.2.
20
APPENDIX
(12)
5014
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 6 ( 2 0 1 1 ) 5 0 1 0 e5 0 2 0
DH298 206:2kJ=mol
DH298 41:1kJ=mol
(13)
(14)
3.3.
The inlet composition for the WGS reactor is the sum of the exit
composition of the SMR reactor and the injected water. It has
been proved experimentally [14] that an addition of water in the
WGS reactor can increase the carbon monoxide conversion as
a result of the higher S/C, despite the higher space velocity.
Therefore this method will allow a smaller reactor area, which
is favored for a micro-CHP system design. Thus, when the WGS
reactor is operated at an elevated temperature, will allow size
reduction. Moreover, lesser time and energy demand for startup are required [14]. The CO content should be reduced to an
acceptable level (for the fuel cell stack) of 0.1%e0.2%. The
kinetic constant is based on a power law relationship and
calculated at steady state. The kinetic model of the WGS reactor
is described in detail in [16]. The water gas shift reaction (Eq.
(12)) is moderately exothermic and its thermodynamic equilibrium is calculated according to the following expression,
4400
4:063
KT exp
T
(15)
3.5.
Combustor
Depleted fuel (hydrogen-rich gas) from the fuel cell and additional natural gas from the line are combusted with air in the
catalytic combustor to fulfill the SMR reaction and the thermal
load residential demand. The adiabatic flame temperature of
the combustor unit is limited by the combustor material. The
amount of fuel input to the system is a model input, while the
air is regulated according to the fuel input with a fixed air
stoichiometry. The effect of NOx-content has not been investigated in the current model.
The mass flow rate of the flue gas must be high enough to
fulfill the required thermal load demand. Since the depleted
fuel from the fuel cell stack anode is not sufficient, more fuel is
added to the combustor from the fuel line throughout a bypass valve. The air used in the combustor is twice preheated
by the two plate heat exchangers situated before and after the
WGS reactor. The relevant stoichiometric combustion reactions are given below:
H2 0:5O2 3:76N2 /H2 O 0:53:76N2
(19)
(20)
(21)
4.
Ea
KWGS k0 exp
RT
(16)
The proposed micro-CHP system must be validated to investigate possible modeling errors and other sources of
(17)
System validation
3.4.
(18)
UAI
UAII
UAIII
UASG;sup
UASG;ev
UAIV
UAV
UAVI
UASG;ec
APPENDIX
Value
0.0538784
0.720725
0.548661
0.107073
0.843147
1.41641
2.36245
1.39233
0.0896212
21
5015
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 6 ( 2 0 1 1 ) 5 0 1 0 e5 0 2 0
Results
_ net
Electrical power output, W
Korsgaard, 2008
Proposed system
Korsgaard, 2008
Proposed system
22
Combined cogeneration
system efficiency
Uncertainty
(%)
5.42
0.00
0.23
0.02
0.00
Korsgaard, 2008
Proposed system
0.88
0.75
14.92
0.23
0.00
4.67
5.
Sensitivity analysis
0.63
3.10
1.63
0.11
0.00
0.00
0.39
25.27
7.32
7.58
1.51
26.98
0.00
Variable
Uncertainty
75:32 6:74
Value
ncell
Acell
T6
T8
LSMR
wSMR
LWGS
dWGS
185
45.16
126
250
0.49
0.09
0.45
0.13
APPENDIX
5016
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 6 ( 2 0 1 1 ) 5 0 1 0 e5 0 2 0
Initial Value
Range
160
150e180
900
700e1000
1.3
4.0
1.3e1.6
2.5e4.5
determined as a function of the uncertainties of each variableUX . By assuming the individual calculations are uncorrelated and random (neglecting covariance), the uncertainty
of the efficiency can be expressed as,
v
u 2
uX vhi
U2X
Uhi t
vXi
i
where hi f X1 ; X2 ; .; XN
(22)
Objective functions
S=C
lH2
Tcell
Tcomb
hsys cogen
hel FC
hel net
hther cogen
_ el
W
FC
2.5
3
3.5
3.75
4
4.25
4.5
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
1.3
1.3
1.3
1.3
1.3
1.3
1.3
1.3
1.35
1.4
1.45
1.5
1.55
1.6
1.3
1.3
1.3
1.3
1.3
1.3
1.3
1.3
1.3
1.3
1.3
1.3
1.3
1.3
433
433
433
433
433
433
433
433
433
433
433
433
433
433
423
428
433
438
443
448
453
433
433
433
433
433
433
433
1173
1173
1173
1173
1173
1173
1173
1173
1173
1173
1173
1173
1173
1173
1173
1173
1173
1173
1173
1173
1173
973
1023
1073
1123
1173
1223
1273
0.7999
0.8166
0.8256
0.8285
0.8308
0.8327
0.8522
0.8308
0.8497
0.8704
0.8932
0.9182
0.9455
0.9755
0.7163
0.8206
0.8308
0.8366
0.8406
0.8436
0.8459
0.7663
0.7974
0.8208
0.8301
0.8308
0.8299
0.829
0.3444
0.3628
0.372
0.375
0.3774
0.3793
0.3765
0.3774
0.366
0.3552
0.345
0.3354
0.3264
0.3178
0.2263
0.364
0.3774
0.3851
0.3904
0.3944
0.3974
0.4048
0.3948
0.3854
0.3798
0.3774
0.3761
0.375
0.2455
0.2621
0.271
0.274
0.2762
0.2781
0.2923
0.2762
0.2765
0.2774
0.2788
0.2809
0.2836
0.2869
0.1618
0.266
0.2762
0.282
0.2861
0.289
0.2913
0.2123
0.2431
0.2663
0.2755
0.2762
0.2754
0.2744
0.5545
0.5545
0.5546
0.5546
0.5546
0.5546
0.5599
0.5546
0.5732
0.5931
0.6144
0.6373
0.662
0.6886
0.5546
0.5546
0.5546
0.5546
0.5546
0.5546
0.5546
0.5541
0.5543
0.5545
0.5546
0.5546
0.5546
0.5546
933.1
994
1026
1037
1046
1052
1094
1046
1014
983.9
955.8
929.2
904.1
880.3
626.9
1008
1046
1067
1082
1092
1101
812.5
924.8
1009
1043
1046
1042
1039
APPENDIX
_ el
W
net
897.8
958.7
991.1
1002
1010
1017
1059
1010
978.4
948.6
920.5
893.9
868.8
845
591.6
972.8
1010
1031
1046
1057
1065
777.1
889.4
974
1007
1010
1007
1004
23
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 6 ( 2 0 1 1 ) 5 0 1 0 e5 0 2 0
required for the fuel cell stack (similarly for the combustor).
The most interesting parameter in this case with little effect
on system performance is the temperature of the methanesteam mixture SMR input. Again a positive effect is observed,
this time on the overall system thermal efficiency; since a low
temperature for the methane-steam mixture is sufficient,
meaning a minimum amount of heat from the flue gas will be
extracted for heating it prior to entrance to the SMR reactor. In
turn the flue gas heat will be utilized in the cogeneration heat
exchanger IV and V, producing a higher thermal cogeneration
output. This will increase the thermal power cogeneration
output of the system, which in turn will increase the overall
system efficiency.
6.
basis for the system. Finally, the SMR and WGS reactors inlet
reformate mixture temperatures were held constant to
preserve the validity of the experimental results. A description
of the fixed parameters along with their values is given in
Table 4.
The S=C is varied from 2.5 to 4.5 in 0.5 increments. Values
below 2.5 are not included in order to avoid problems of
carbon deposition and coke formation on the anode of the fuel
cell stack as reported in [14,19]. On the other hand, values
above 4.5 are not included, since a high S=C favors hydrogen
production but overall process economics favor a low steamto-carbon ratio, because exit gases include steam, which is not
utilized (condensed in the water knockout stage) [14]. Fuel cell
stacks are not usually operated with a lH2 value below 1.25,
when hydrogen-rich reformate gas is used as fuel. In addition,
since a reasonable amount of fuel is required for the
combustor, then extremely low values are not appropriate
[20]. As an upper limit, the lH2 must remain lower than 1.7 to
avoid the risk of fuel depletion in the fuel cell stack [4,11]. The
Tcell is varied from 150 C to 180 C in 5 C increments. A value
beyond 180 C is not used because it would exceed the operating limit of the fuel cell stack. The Tcell is dictated by the
materials set and cell design, which sets the conduction
properties and durability of the fuel cell tri-layer [6]. The Tcomb
is limited to a maximum value of 1273 K for two reasons: (1) to
minimize heat losses in the combustor, (2) a higher temperature could cause severe damages to the SMR reactor, since it
has not been tested experimentally for higher temperatures.
24
5017
APPENDIX
5018
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 6 ( 2 0 1 1 ) 5 0 1 0 e5 0 2 0
APPENDIX
25
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 6 ( 2 0 1 1 ) 5 0 1 0 e5 0 2 0
5019
Fig. 7 e Variation of H2, CH4, CO2, H2O and CO throughout the SMR and WGS reactors.
7.
Conclusions
26
the one operating at an average load (as required by the residential load demand). A very significant, and still unexplored,
area of optimization is the geometry of the system. The
parametric study suggested the SMR and WGS reactors must
be sized and simulated in greater detail, in order to achieve the
geometry which will provide the highest amount of H2
conversion. This will lead to an even higher combined
cogeneration system efficiency and therefore lower the fuel
consumption.
Finally, EES was chosen as the modeling tool for this
research study for the reasons explained in the introduction.
The modeling was completed within a small amount of time,
taking advantage of the nature and the buildin capabilities of
the software (such as parametric tables, graphs, NewtoneRaphson method of solving algebraic equations, and
thermophysical properties).
Acknowledgments
The authors would like to acknowledge the support of Danfoss
and Dantherm Power throughout the realization of this
research study.
references
APPENDIX
5020
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 6 ( 2 0 1 1 ) 5 0 1 0 e5 0 2 0
[4] Korsgaard AR, Nielsen MP, Kr SK. Part one: a novel model of
HTPEM-based micro-combined heat and power fuel cell
system. International Journal of Hydrogen Energy 2008;33(7):
1909e20.
[5] Korsgaard AR, Nielsen MP, Kr SK. Part two: control of
a novel HTPEM-based micro combined heat and power fuel
cell system. International Journal of Hydrogen Energy 2008;
33(7):1921e31.
[6] Buchi FN, Inaba M, Schmidt TJ. Polymer electrolyte fuel cell
durability. New York, NY, USA: Springer; 2009.
[7] Georgopoulos N. Application of a Decomposition Strategy to
the Optimal Synthesis/Design and Operation of a Fuel Cell
Based Total Energy System, M.S. Thesis, 2002, Virginia Tech,
Blacksburg, VA, USA.
[8] Ferguson A, Ugursal VI. Fuel cell modelling for building
cogeneration applications. Journal of Power Sources 2004;
137:30e42.
[9] Obara S. Dynamic characteristics of a PEM fuel cell system
for individual houses. International Journal of Energy
Research 2006;30(15):1278e94.
[10] Godat J, Marechal F. Optimization of a fuel cell system using
process integration techniques. Journal of Power Sources
2003;118:411e23.
[11] Braun R. Optimal Design and Operation of Solid Oxide Fuel
Cell Systems for Small-scale Stationary Applications,
Doctoral Dissertation, 2002, University of
WisconsineMadison, USA.
[12] Korsgaard AR, Nielsen MP, Bang M, Kr SK. Modeling of CO
influence in PBI electrolyte PEM fuel cells. In: Proceedings of
the FUELCELL; 2006 [Irvine, CA, USA].
[13] Korsgaard AR, Refshauge R, Nielsen MP, Bang M, Kr SK.
Experimental characterization and modeling of commercial
polybenzimidazole-based MEA performance. Journal of
Power Sources 2006;162(1):239e45.
[14] Kolb G. Fuel processing for fuel cells. Weinheim, Germany:
Wiley-VCH; 2008.
[15] Ahlstrom-Silversand AF, Odenbrand CUI. Thermally
sprayed wire-mesh catalysts for the purification of flue
gases from small-scale combustion of bio-fuel catalyst
preparation and activity studies. Applied Catalysis A 1997;
153:177.
[16] Keiski RL, Salmi T, Niemisto P, Ainassaari J, Pohjola VJ.
Stationary and transient kinetics of the high temperature
water-gas shift reaction. Applied Catalysis A: General 1996;
137(2):349e70.
[17] Davies J, Lihou D. Optimal design of methane steam
reformer. Chemical and Process Engineering 1971;52:
71e80.
[18] F-Chart Software, 2010, EES Manual, www.fchart.com, v8.
590 edition.
[19] Larminie J, Dicks A. Fuel cell systems explained. Chichester,
UK: John Wiley & Sons; 2003.
[20] Barbir F. PEM fuel cells: theory and practice. Elsevier; 2005.
[21] Rasmussen PL, Nielsen MP, Kr SK, Andreasen SJ.
Experimental study and modeling of degradation
Nomenclature
dHr;1 : Enthalpy of reaction (kJ/mol)
Ea : Activation energy (kJ/mol)
F: Faraday constant (C/mol)
i: Current density (A/cm2)
i0 : Exchange current density (A/cm2)
keh : Electrooxidation rate of hydrogen (A/cm2)
k0 : Pre-exponential factor
KT : Equilibrium constant
KWGS : Kinetic power law fit parameter (m^1.92/kmol^-0.36.kg.s)
LHVi : Lower heating value of species i (J/kg)
_ i : Mass flow rate of species i (kg/s)
m
n_ CO : CO molar feed rate (kmol/h)
Q_ i : Heat transfer rate of component i (W)
rCO : CO reaction rate (kmol/kg.s)
R: Ideal gas constant (J/K.mol)
Rdiff : Diffusion resistance (Ohm.cm2)
Rohmic : Ohmic resistance (Ohm.cm2)
S=C: Steam-to-carbon ratio
Tcell : Fuel cell operating temperature ( C)
Tcomb : Combustor output temperature ( C)
Ti : Temperature of species or state point i ( C)
Vcell : Total cell voltage (V)
V0 : Open circuit voltage (V)
us : Superficial velocity (m/s)
_ i : Work rate of component/subsystem i (W)
W
ai : Charge transfer coefficient
ha : Anode overpotential (V)
hc : Cathode overpotential (V)
hohmic : Ohmic losses (V)
hi : Efficiency of component/system i
qH2 : Surface coverage of hydrogen
lair : Cathode air stoichiometric ratio
lH2 : Anode stoichiometric ratio
rg : Density of the reformate gas (kg/m3)
rwiremesh : Density of the wire mesh catalytic material (kg/m3)
u: Water gas shift reactor cross-sectional area (m)
Subscripts/superscripts
cogen: Cogeneration
el:
Electrical power
FC:
Fuel cell subsystem
HEx: Heat exchanger
in:
Input stream
net:
Net amount
out:
Output stream
sys:
System
ther: Thermal power
APPENDIX
27
28
APPENDIX
APPENDIX
29
30
APPENDIX
1. Introduction
The application of fuel cell technology in residential micro-CHP (Combined Heat
and Power) systems has been gaining an increased amount of interest during the
recent years, which is mainly due to their promising efficiency performance. Many
advantages can be listed for these systems, as well for the fuel cell itself and the
system in general. The main advantages of fuel cells include, among others, high
electrical efficiencies and low greenhouse gas emissions [1-4]. In addition to these advantages, the micro-CHP technology can provide increased cogeneration efficiencies,
due to the thermal efficiency utilization. Also on-site power-and-heat production
eliminates transmission and distribution losses, which are inevitable for centralized
systems [5-7]. Another important aspect of fuel cell-based micro-CHP technology is
the fact that fuel cell systems can operate efficiently at part-load operation, which
provides greater flexibility in adopting an operational strategy. Further on, fuel cells
possess the capability of responding rapidly to load changes. As a consequence of
high efficiency, fuel cell systems can offer the potential of reduced operational costs,
which has been becoming extremely influential in the recent years due to the rapidly
increasing cost of fossil fuels.
In the current research study, different operational strategies are examined to
investigate the response of the system to changing loads, primarily with respect to
efficiency. The overall goal of this research work is to investigate the application of
conventional operational strategies such as heat-led and electricity-led operation [6],
Received by the editors 2011.
Key words and phrases. PBI; HT-PEMFC system; micro-CHP; residential system; operational
strategy; fuel cell system.
1
APPENDIX
31
A. ARSALIS ET AL.
and then formulate an improved operational strategy, which provides a higher average net electrical (and total) efficiency for the system. In this manner a comparison
can be made with the conventional operational strategies, drawing assumptions for
discussion and analysis. Also a thermal storage tank is modeled and coupled to the
system to increase its operational flexibility. Finally, an investigation of the heat
losses exhibited by the thermal storage tank is carried out throughout the whole
annual operational load profile. This is done to monitor the periods exhibiting the
highest losses. The applied load profile is based on averaged consumption data for
a typical single-family household in Denmark [5,6].
2. Literature Review
Residential micro-CHP systems can be categorized in four main types, based
on their respective prime mover technology: Stirling engines (SE), internal combustion engines (ICE), proton exchange membrane fuel cells (PEMFC), and solid
oxide fuel cells (SOFC). SE- and ICE-based micro-CHP systems are considered
to be the most mature technologies, since they are commercially available from a
number of manufacturers [8-11]. The operating principle of the SE is based on an
external combustion engine with an internal piston being driven via a temperature
difference between the ends of a cylinder. SE residential systems typically have
low electrical efficiencies ranging from 0.07 to 0.15 (based on LHV) [9,12], while
their respective overall efficiencies are comparable to that of a condensing boiler. In
terms of operational behavior, they have rapid startup times and quick load changing responses [12]. The operating principle of an ICE micro-CHP system is based
on combustion occurring inside a cylinder driving a piston to create mechanical
work, which is converted to electrical energy in a generator. Electrical efficiencies
are typically at around 0.25 (based on LHV), with similar operational behavior as
SE-based systems [12,13].
Fuel cell-based micro-CHP systems is the least mature technology, since a number of technical and economic issues must be resolved before its potential can be
realized. Nevertheless, it is the most promising technology because it has the potential of achieving high electrical and overall efficiencies [2,5,12,14]. Startup times
are rather slow for the SOFC type [15,16], while for PEMFCs are significantly
faster. On the other hand, low-temperature PEMFCs (Nafion) have a very low
heat-to-power ratio. They are also very sensitive when operated with reformate
fuel, requiring extensive carbon monoxide reduction. A relatively new technology
is HT-PEMFCs, which operate at temperatures at around 160o C and they are relatively tolerant to carbon monoxide poisoning [5,17,18]. Fuel cell systems are able
to respond to load changes rapidly, although a regular load changing operational
pattern will deteriorate the fuel cell membrane, and therefore reduce its lifetime.
Electrical efficiencies can reach 0.40 (based on LHV), while the overall efficiency is
as high as the ICE and SE types at nominal load. In addition, the fuel cell electrical
efficiency is significantly higher at part-load, because of the higher cell voltage at
lower loads [17].
2.1. Market penetration projections. Micro-CHP technology has the potential to reduce transmission and distribution losses that typically occur when a large
remotely located power station transmits electricity to a household [5]. This suggests that onsite produced and consumed electricity will have negligible losses,
while imported/exported electricity will still have the usual losses. Therefore an
effort should be made to choose an operational pattern which will minimize grid
interaction. According to the Danish Energy Association [19], the fleet of central
power plants in Denmark is relatively old, and the newest power plant was built in
2002. This fact in connection to the new initiative to promote the fuel cell-based
32
APPENDIX
APPENDIX
33
A. ARSALIS ET AL.
34
APPENDIX
APPENDIX
35
A. ARSALIS ET AL.
36
APPENDIX
Finally grid interaction should be minimized not only because of technical reasons
(transmission losses, etc.), but also because of economic reasons (e.g., taxation
imposed for purchasing/selling electricity). Ideally, electricity should be imported
and exported according to a combination of fuel prices, electricity import/export
prices, and their interaction with the electrical/overall efficiency profiles [6]. Based
on the above, some general preliminary assumptions can be made on the nature of
this operational strategy: it is expected that the system will follow an electricity-led
operational strategy in the winter months, while the system will follow a heat-led
strategy in the summer months. It is also possible to shut down the system in some
instances during the summer months.
4. System layout and modeling
4.1. HT-PEMFC-based micro-CHP system design, layout and modeling.
The developed model is based on the PBI (Polybenzimidazole) HT-PEMFC technology and is described in detail in [23]. It is noted that the system should be
considered coupled with a grid network to import/export electricity as needed in
case a purely electricity-led operation is not followed. Equivalently, a thermal storage tank and an auxiliary burner are coupled to the system to satisfy the heating
load demand in case a purely heat-led strategy is not followed. The operating principle of the system, shown in Fig. 3, is the same as in [23], with the exception of
recovering the thermal energy of the flue gas utilizing a thermal storage tank, instead of heat exchangers. The fuel processing subsystem consists of a desulfurizer,
steam methane reforming (SMR) reactor, water gas shift (WGS) reactor, combustor and water-knock out device. Syngas is directed to the fuel cell stack anode. The
combustor outlet flue gas is used in the SMR reactor, the fuel preheater and the
steam generator. The fuel cell stack exhaust mixture is used to preheat the natural
gas and subsequently it is mixed with the combustor outlet flue gas for heat recovery
in the thermal storage tank. The system, modeled in the commercially available
software EES (Engineering Equation Solver), includes 27 state points. Figure 3
indicates the temperature and mass flow rate for every node at 1 kWe net power
output and current density 0.2 A/cm2 . The net electrical efficiency and the total
system efficiency at these conditions are 0.291 and 0.829, respectively. It should
be noted that the operating pressure of the system is assumed to be slightly above
atmospheric conditions.
4.2. HT-PEMFC. The HT-PEMFC model considers only the reaction of hydrogen with oxygen, while all other species are considered inactive. The ohmic and
diffusion resistances are based on linear regressions evaluated experimentally in [24],
(1)
(2)
where Tcell is the fuel cell operating temperature (K). The anode and cathode
overpotentials are given by the following expressions, respectively,
i
RTcell
1
sinh
(3)
a =
anode F
2keh H2
(4)
c =
RTcell
ln
4cathode F
i0 + i
i0
+ Rdif f
i
air 1
where i is the charge transfer coefficient, air is the cathode air stoichiometric
ratio, keh is the electro-oxidation rate of hydrogen, H2 is the surface coverage of
APPENDIX
37
A. ARSALIS ET AL.
ohmic = iRohmic
Vcell = V0 a c ohmic
CH4 +H2 O
CO + 3H2
(8)
CO+H2 O
CO2 +H2
0 = 206.2kJ/mol
h
rxn
0 = 41.1kJ/mol
h
rxn
Heat transfer rates are evaluated in each element as a single total heat transmission
coefficient. The reactor is discretized into 14 individual cells on either side of the
heat exchanger. The number of cells was increased until a sufficient grid consistency
was obtained. The WGS reactor receives the cooled reformate gas and its purpose
is to reduce the CO content to a fuel cell tolerable percentage, which is up to 3%
(per mole basis). The thermodynamic equilibrium constant of the WGS reaction
38
APPENDIX
where TW GS is the average operating temperature of the WGS reactor (K). The
carbon monoxide conversion is given by:
dxCO
rCO
(10)
=w
dz
n CO
where is the cross-sectional area of the WGS reactor (m2 ), w is the density of
the wire mesh catalytic material (kg/m3 ), rCO is the carbon monoxide reaction
rate (kmol/kg-s), and n CO is the molar flow rate of carbon monoxide (kmol/s).
4.4. Thermal storage tank. The model for the thermal storage tank assumes
simple heat transfer balances of the three working mediums (flue gas, tank water,
cogeneration water), while heat losses are accounted only to occur from the top
surface of the tank as in [5,26]. It is assumed that auxiliary heating is available
(although not modeled in detail), in case the heating load demand exceeds the
provided heat transfer from the flue gas. The auxiliary heating can be provided
by means of an auxiliary burner or an electric circuit inside the tank as in [26].
Therefore the tank temperature can be formulated as a function of the heat transfer
rates and the tank capacity:
Z
1
Q f luegas + Q aux Q heatingdemand Q loss dt
(11)
Ttank =
mtank cp,tw
where Ttank is the tank temperature, mtank is the mass of the tank and cp,tw is the
average heat capacity of the water in the tank. Finally the auxiliary heat transfer
rate input is defined as,
(12)
Q aux = m
tank cp,tw (60 Ttank )
suggesting that the auxiliary heater will be utilized when the tank temperature
drops below 60o C. The set point temperature for the tank is 60o C, while the
temperatures for the incoming and outgoing residential water streams are 15 and
55o C, respectively.
4.5. General remarks. The proposed system is assumed to be able to fulfill any
heat- or electrical-load from low to peak values, based on the adopted Danish load
profile as explained above. Also the supply/demand matching performance of the
micro-CHP system will be affected by the transient characteristics of the micro-CHP
system (start/stop, operational strategy selection). The current system simulation
assumes a perfect startup/shutdown characteristic since a detailed pattern of the
start/stop behavior is not available. However it should be noted that a purely
electricity-led operational strategy suggests the system will never be shut down
since the electrical load demand is never zero. A shutdown/startup is needed in
the case of adopting a purely heat-led strategy, and also for a combined heat- and
electricity-led operational strategy. Finally, the analysis assumes a household with
an area of 130 m2 , inhabited by four persons. The thermal storage tank size is
assumed to have a volume of 250 L as defined in [6].
5. Results and Discussion
5.1. Application and analysis of conventional operational strategies. Before an improved strategy can be formulated, the two conventional operational
strategies are applied to investigate trends on the behavior of the system, which
can be used later on the improved strategy. Table 1 shows the calculated results
for the minimum, maximum, and average values for net electrical power output,
APPENDIX
39
10
A. ARSALIS ET AL.
min
Wel,net
Welavg
, net
max
Wel,net
Q hmin
Q havg
Q hmax
min
sys
avg
sys
max
sys
elmin
elavg
elmax
thmin
thavg
thmax
Electricity-led
162
567
2211
267
926
5056
0.536
0.850
0.891
0.163
0.326
0.345
0.373
0.524
0.546
Heat-led
816
2453
1439
6186
0.478
0.784
0.891
0.134
0.291
0.345
0.345
0.493
0.546
total cogeneration heat transfer rate, total efficiency, thermal efficiency, and net
electrical efficiencies. The net electrical efficiency (net,el ), the cogeneration thermal efficiency (uh ), and the total system efficiency (sys ) are used to monitor the
performance of the proposed micro-CHP system. The net electrical efficiency is
el ) divided by the chemical energy input to the
defined as the net power output (W
system (m
CH4,in LHVCH4 ). The net power output is the remaining power generated
by the fuel cell stack after parasitic losses from balance-of-plant components have
been subtracted. The cogeneration thermal efficiency is defined as the recovered
useful heat (Q uh ) divided by the chemical energy input to the system. The total
system efficiency is the sum of the net power output and the recovered useful heat
divided by the chemical energy input to the system.
el
W
m
CH4,in LHVCH4
(13)
net,el =
(14)
uh =
Q uh
m
CH4,in LHVCH4
(15)
sys =
el + Q uh
W
m
CH4,in LHVCH4
The power and heat outputs suggest the system will be required to operate over a
large range of operation. Although fuel cell systems are capable of achieving high
efficiencies over a large range of operation, the system has to be constrained to
avoid operation at low efficiency regimes. Low net electrical efficiencies are caused
by high current densities. For example, if the fuel cell stack follows a purely heat-led
operational strategy, and since heat and electricity production are interrelated, during time periods of high heat demand the high current density will cause a decrease
in fuel cell stack efficiency. The three aforementioned representative days are used
with the applied operational strategies and the results are shown in Fig. 4. From
an observation on these figures it can be concluded that the relation between electricity and heat production is quite linear. Also it can be concluded that adoption
of a heat-led strategy will require frequent utilization of the grid (import/export),
while a thermal storage tank/external burner will be needed when an electricity-led
operation is adopted.
5.2. Formulation and application of an improved strategy. The application
of the improved strategy is formulated on the basis of the following assumptions:
The annual load profile is split into 35 time periods, to locate the periods which require electricity-led operation, heat-led operation or system
shutdown.
The net electrical power output of the micro-CHP system is constrained
for a maximum value of 1 kWe. This is done to avoid operation at low
efficiency regimes.
40
APPENDIX
11
APPENDIX
41
12
A. ARSALIS ET AL.
Figure 5. Annual variation of electricity (production and demand) and heat (production, demand and auxiliary).
The annual variation of imported and exported electricity for the improved operational strategy is shown in Fig. 6. A negative value for the import/export curve
indicates electricity is exported during that averaged time period, while a positive
value indicates import of electricity. The goal of keeping these values at a minimum
range is primarily fulfilled for the export part, where a significant amount has to
be imported during the shutdown of the system. It should be noted that small
amounts of electricity import, primarily during the winter season, are caused due
to the constraining of the maximum power output of the system during electricityled operation. If a purely electricity-led operation is followed during the summer
period, where the heating load demand is rather limited, then there will be a need
for heat dumping, since the TST will be unable to fully store the incoming heat.
If the heat demand vs. heat production is calculated for every averaged time period, it is observed that during the periods 10 to 25, there is an excess of heat
production as compared to heat demand. Therefore the total heat dumping value
for electricity-led operation is the sum of the differences between heat recovered
by the micro-CHP system and heat demanded by the household for the aforementioned periods. The rate of heat dumping for electricity-led operation is illustrated
in Fig. 7.
5.3. Thermal storage tank heat losses. The thermal storage tank can exhibit
variable heat losses depending on the operational profile of the system, the external
heater and the heat demand. During heat-led operation, where all the heat demand
is supplied by the micro-CHP system, heat losses increase rapidly. The reason is
due to the fact that the micro-CHP system has to supply all the required heat
demand, while the external heater is not utilized. In other words, the thermal
42
APPENDIX
13
APPENDIX
43
14
A. ARSALIS ET AL.
Operational strategy
Variable
Description (unit)
Electricity-led
Heat-led
Improved
ELHV ,in
15630
26475
13805
Eel , prod
4984
7164
4533
Eel ,imp
753
1127
Eel,exp
2933
676
Eheat , prod
8136
12639
7334
Eheat , aux
6502
5305
Eheat ,dump
1999
el ,net
31.9
27.1
32.8
th
39.3
47.7
53.1
sys
71.2
74.8
85.9
Table 2, which summarizes the effect of the applied strategies on system performance. Utilizing the heat-led operation requires almost twice fuel input to the
micro-CHP system (compared to the improved strategy), which produces an increased amount of electricity overall. Nevertheless electricity has to be imported
at periods of low heat demand, such as the summer season. The electricity-led
operation requires significantly less fuel than the heat-led operation, but a high
amount of heat dumping is required. The merits of the improved operation are
observed in all categories, although a significant amount of imported electricity is
required due to the system shutdown in the summer period and the constraining
of the system at 1 kWe. The most important parameter of the study is the improvement of the average total system efficiency, which is 85.9% for the improved
operational strategy, while the respective total system efficiencies for electricity-led
and heat-led operations are 71.2 and 74.8%, respectively. It is also important to
observe that the thermal efficiency, and subsequently the total system efficiency, for
the electricity-led strategy is significantly low due to the need for heat dumping.
6. Conclusions
A proposed residential micro-CHP system, based on PBI fuel cell technology, is
analyzed in terms of operational performance. The applied load profile consists of
the electrical and heat demands for a single-family household in Denmark. The
results indicate that the proposed micro-CHP system has the potential to maintain high efficiencies throughout the annual load profile, if an improved strategy
is applied. The improved parameters include the net electrical efficiency, thermal
(cogeneration) efficiency, and in effect the total system efficiency. Also, heat dumping is avoided by use of a thermal storage tank. The study also indicates that the
system has to be coupled to the grid network and an external burner, to prohibit
system operation at low efficiency regimes, and within critical values of current density. The coupling of the system to the grid also allows the system to be shutdown
44
APPENDIX
15
in periods of low demand, such as the summer period. This means that only a single
turn on/off is required, which in effect can assist in maintaining a high lifetime for
the fuel cell stack, and also minimize losses caused by frequent starts/stops.
The average net electrical efficiency of the system is 32.8%, but it is expected that
higher efficiencies will be achieved, when the experimental SMR reactor, utilized in
this study, has been optimized. The average total system efficiency is 85.9%, which
verifies the ability of such a system to achieve high total system efficiencies. Nevertheless, a market penetration breakthrough of the proposed system into the area
of residential prime movers will only be justified when it has met a threshold point,
in terms of total life cycle cost. In comparison to non-fuel cell-based micro-CHP
systems, such as SE- and ICE-based systems, the proposed system has significantly
higher net electrical efficiencies. In addition, the system allows lower heat-to-power
ratios, which aid the system to avoid high thermal surpluses throughout annual
operation. The improved operation also targets into minimizing grid interaction,
which can be justified by economic reasons, taxation of purchased/sold electricity,
and technical reasons such as distribution losses. Although the exported electricity
has been reduced to 676 kWh, the imported electricity remains significantly high
at 1,127 kWh. This is caused by the inevitable constraining of the net electrical
power output of the system and the system shutdown.
Acknowledgment
The authors would like to acknowledge the support of Danfoss and Dantherm
Power throughout the realization of this research study.
References
[1] R. OHayre, W. Colella, S.-W. Cha, F.B. Prinz, Fuel Cell Fundamentals, 2nd
ed., Wiley, Hoboken, New Jersey, 2009.
[2] L. Barelli, G. Bidini, F. Gallorini, A. Ottaviano, An energeticexergetic comparison between PEMFC and SOFCbased microCHP systems, International
Journal of Hydrogen Energy. 36 (2011) 32063214.
[3] L. Barelli, G. Bidini, F. Gallorini, A. Ottaviano, Dynamic analysis of PEMFCbased CHP systems for domestic application, Applied Energy. 91 (2012) 13-28.
[4] V. Dorer, A. Weber, Energy and CO2 emissions performance assessment of
residential micro-cogeneration systems with dynamic whole-building simulation
programs, Energy Conversion and Management. 50 (2009) 648-657.
[5] A.R. Korsgaard, M.P. Nielsen, S.K. Kr, Part one: A novel model of HTPEMbased micro-combined heat and power fuel cell system, International Journal of
Hydrogen Energy. 33 (2008) 1921-1931.
[6] A.R. Korsgaard, M.P. Nielsen, S.K. Kr, Part two: Control of a novel HTPEMbased micro combined heat and power fuel cell system, International Journal of
Hydrogen Energy. 33 (2008) 1921-1931.
[7] B.V. Mathiesen, H. Lund, Comparative analyses of seven technologies to facilitate the integration of fluctuating renewable energy sources, Engineering and
Technology. 3 (2009) 190- 204.
[8] M. De Paepe, P. DHerdt, D. Mertens, Micro-CHP systems for residential applications, Energy Conversion and Management. 47 (2006) 3435-3446.
[9] B. Thomas, Benchmark testing of Micro-CHP units, Applied Thermal Engineering. 28 (2008) 2049-2054.
[10] H. Onovwiona, V. Ismetugursal, A. Fung, Modeling of internal combustion engine based cogeneration systems for residential applications, Applied Thermal
Engineering. 27 (2007) 848-861.
APPENDIX
45
16
A. ARSALIS ET AL.
[11] K. Alanne, N. S
oderholm, K. Siren, I. Beausoleil-Morrison, Techno-economic
assessment and optimization of Stirling engine micro-cogeneration systems in
residential buildings, Energy Conversion and Management. 51 (2010) 26352646.
[12] M. Dentice dAccadia, M. Sasso, S. Sibilio, L. Vanoli, Micro-combined heat
and power in residential and light commercial applications, Applied Thermal
Engineering. 23 (2003) 1247-1259.
[13] F. Caresana, C. Brandoni, P. Feliciotti, C.M. Bartolini, Energy and economic
analysis of an ICE-based variable speed-operated micro-cogenerator, Applied
Energy. 88 (2011) 659-671.
[14] A. Arsalis, M.P. Nielsen, S.K. Kr, Modeling and off-design performance of
a 1 kWe HT-PEMFC (high temperature-proton exchange membrane fuel cell)based residential micro-CHP (combined-heat-and-power) system for Danish singlefamily households, Energy. 36 (2011) 5010-5020.
[15] R.J. Braun, Techno-economic optimal design of solid oxide fuel cell systems for
micro-combined heat and power applications in the U.S., Journal of Fuel Cell
Science and Technology. 7 (2010).
[16] T. Wakui, R. Yokoyama, K.-ichi Shimizu, Suitable operational strategy for
power interchange operation using multiple residential SOFC (solid oxide fuel
cell) cogeneration systems, Energy. 35 (2010) 740-750.
[17] J. Zhang, Z. Xie, J. Zhang, Y. Tang, C. Song, T. Navessin, et al., High temperature PEM fuel cells, Journal of Power Sources. 160 (2006) 872-891.
[18] S.J. Andreasen, S.K. Kr, Modelling and evaluation of heating strategies for
high temperature polymer electrolyte membrane fuel cell stacks, International
Journal of Hydrogen Energy. 33 (2008) 4655-4664.
[19] Danish Energy Association, Danish Electricity Supply08 - Statistical Survey,
2009.
[20] M. Houwing, A. Ajah, P. Heijnen, I. Bouwmans, P. Herder, Uncertainties in the
design and operation of distributed energy resources: The case of micro-CHP
systems, Energy. 33 (2008) 1518-1536.
[21] A. Ferguson, Fuel cell modelling for building cogeneration applications, Journal
of Power Sources. 137 (2004) 30-42.
[22] F. Barbir, PEM Fuel Cells: Theory and Practice, Elsevier, 2005.
[23] A. Arsalis, M.P. Nielsen, S.K. Kr, Modeling and parametric study of a 1 kWe
HT-PEMFC-based residential micro-CHP system, International Journal of Hydrogen Energy. 36 (2011) 5010-5020.
[24] A.R. Korsgaard, R. Refshauge, M.P. Nielsen, M. Bang, S.K. Kr, Experimental
characterization and modeling of commercial polybenzimidazole-based MEA
performance, Journal of Power Sources. 162 (2006) 239-245.
[25] R.L. Keiski, T. Salmi, P. Niemisto, J. Ainassaari, V.J. Pohjola, Stationary and
transient kinetics of the high temperature water-gas shift reaction, Applied
Catalysis A: General. 137 (1996) 349-370.
[26] D.S. Salcines, C.R. Estebanez, V.C. Herrero, Simulation of a solar domestic
water heating system with different collector efficiencies and different volumen
storage tanks, in: International Conference on Renewable Energies and Power
Quality, Barcelona, 2004.
Department of Energy Technology, Aalborg University, Pontoppidanstrde 101,
9220, Aalborg ., Denmark
E-mail address: aar@et.aau.dk
46
APPENDIX
Journal Article IV
APPENDIX
47
48
APPENDIX
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 7 ( 2 0 1 2 ) 2 4 7 0 e2 4 8 1
article info
abstract
Article history:
algorithm (GA) optimization strategy. The proposed system consists of a fuel cell stack,
11 October 2011
steam methane reformer (SMR) reactor, water gas shift (WGS) reactor, heat exchangers,
and other balance-of-plant (BOP) components. The objective function of the single-
objective optimization strategy is the net electrical efficiency of the micro-CHP system.
The implemented optimization procedure attempts to maximize the objective function by
Keywords:
variation of nine decision variables. The value of the objective function for the optimum
PBI
design configuration is significantly higher than the initial one, with a 20.7% increase.
HT-PEMFC
Copyright 2011, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights
Micro-CHP
reserved.
Residential system
Optimization
Fuel cell system
1.
Introduction
* Corresponding author.
E-mail address: aar@et.aau.dk (A. Arsalis).
0360-3199/$ e see front matter Copyright 2011, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved.
doi:10.1016/j.ijhydene.2011.10.081
APPENDIX
49
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 7 ( 2 0 1 2 ) 2 4 7 0 e2 4 8 1
2.
System layout
50
2471
APPENDIX
2472
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 7 ( 2 0 1 2 ) 2 4 7 0 e2 4 8 1
RTcell
i
1
sinh
aanode F
2keh qH2
(3)
hc
RTcell
i0 i
i
Rdiff
ln
4acathode F
i0
lair 1
(4)
Time Segment
Winter
Summer
Spring
ha
540
1450
330
950
1930
1080
380
70
230
650
140
1420
460
600
310
920
1070
1620
3.
3.1.
The fuel cell stack model is based on the modeling assumptions previously published by some of the authors in [4,10].
The model considers only the reaction of hydrogen with
oxygen, while all other species are considered inactive. The
ohmic and diffusion resistances are based on linear regressions evaluated experimentally as follows,
Rohmic 0:0001667Tcell 0:2289
(1)
(2)
(6)
3.2.
(5)
SMR reactor
KSMR
1=2
1=2
7=2
2
yCO;o
yCO2 ;o
yH2 ;o
peq
3=2
p0
yCH4 ;o yH2 O;o
(7)
LSMR
n_ CH4 ;i
ntubes rB A
XCH4 ;eq
dXCH4
rCH4
(8)
APPENDIX
51
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 7 ( 2 0 1 2 ) 2 4 7 0 e2 4 8 1
2473
3.3.
WGS reactor
52
Description (unit)
Value
160
45.16
267
0.13
1385
1200
KT exp
4400
4:063
T
(9)
(10)
(11)
(12)
APPENDIX
2474
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 7 ( 2 0 1 2 ) 2 4 7 0 e2 4 8 1
Table 3 e Initial, minimum, maximum and optimum values of the decision variables used in the optimization procedure.
Variable
S/C
lH2
nSMR
tubes
LSMR
reactor
WGS
Lreactor
TComb
fg;out
TSMR
ref;in
TWGS
ref;in
TFP
fg;in
4.
Description (unit)
Initial
Min
Max
Optimum
1st simulation
Optimum
2nd simulation
Steam-to-carbon ratio
Hydrogen stoichiometry
Number of tubes in the SMR
SMR reactor length (m)
WGS reactor length (m)
Combustor output temperature (K)
SMR reformate inlet temperature (K)
WGS reformate inlet temperature (K)
Fuel preheater flue gas inlet temperature (K)
4.000
1.500
275
1.800
0.350
1173
400
499
670
2.500
1.500
275
1.800
0.350
1050
399
470
550
4.000
1.550
300
2.000
0.500
1174
530
500
670
2.908
1.509
297
1.994
0.465
1095
530
474
620
3.002
1.502
298
1.996
0.441
1073
530
472
614
(13)
_ el;net
W
_
mCH4 ;in LHVCH4
(14)
APPENDIX
53
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 7 ( 2 0 1 2 ) 2 4 7 0 e2 4 8 1
2475
Fig. 4 e Behavioral pattern of the steam-to-carbon ratio at the design point throughout the optimization procedure.
Fig. 5 e Behavioral pattern of the hydrogen stoichiometry at the design point throughout the optimization procedure.
54
APPENDIX
2476
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 7 ( 2 0 1 2 ) 2 4 7 0 e2 4 8 1
Fig. 6 e Behavioral pattern of the number of tubes in the SMR reactor at the design point throughout the optimization
procedure.
values for the SMR and WGS reactor parameters are selected
from [7] and [16], respectively.
5.
Fig. 7 e Behavioral pattern of the SMR reactor length at the design point throughout the optimization procedure.
APPENDIX
55
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 7 ( 2 0 1 2 ) 2 4 7 0 e2 4 8 1
2477
Fig. 8 e Behavioral pattern of the WGS reactor length at the design point throughout the optimization procedure.
Fig. 9 e Behavioral pattern of the combustor output temperature at the design point throughout the optimization procedure.
56
APPENDIX
2478
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 7 ( 2 0 1 2 ) 2 4 7 0 e2 4 8 1
Fig. 10 e Behavioral pattern of the SMR reformate inlet temperature at the design point throughout the optimization
procedure.
the latter, is the behavior of the combustor output temperature value (shown in Fig. 9), which is interdependent to the
SMR reactor length. Obviously when a larger reactor area is
available, more flue gas can be utilized by the endothermic
reaction in the SMR reactor. Another interrelated parameter is
the SMR reformate inlet temperature (shown in Fig. 10). Its
value tends to maximize, because a higher temperature in the
inlet of the SMR reactor will convert hydrogen more effectively in a restricted reactor area. The initial value was very
low, suggesting that this was another source of low efficiency
in the initial configuration. In Fig. 11, the optimum WGS
reformate inlet temperature value deviates significantly from
the initial one, tending to minimize, suggesting the WGS
reaction is more efficient at lower values. Finally, Fig. 12
illustrates the behavioral pattern of the fuel preheater flue
gas inlet temperature. Interestingly the optimum value
Fig. 11 e Behavioral pattern of the WGS reformate inlet temperature at the design point throughout the optimization
procedure.
APPENDIX
57
Table 4 e Values of the thermodynamic properties for all state points for the initial and optimum configurations.
Ti (K)
Node
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
298
298
298
420
420
420
845
499
566
433
433
433
298
298
468
1014
1173
1016
666
298
487
298
298
298
420
432
564
670
298
298
298
530
530
530
942
472
561
433
433
433
298
298
458
890
1073
909
605
298
464
298
298
298
530
432
553
614
initial
2.833E-05
2.834E-06
2.550E-05
2.550E-05
2.550E-05
1.400E-04
1.400E-04
1.400E-04
1.400E-04
1.400E-04
6.698E-05
2.679E-05
2.928E-04
2.928E-04
1.182E-03
1.182E-03
1.500E-03
1.500E-03
1.500E-03
1.125E-03
1.500E-03
1.500E-03
1.145E-04
1.145E-04
1.145E-04
1.182E-03
1.182E-03
1.500E-03
yCH4
yCO
yCO2
yH2
yH2 O
yN2
yO2
optimum initial optimum initial optimum initial optimum initial optimum initial optimum initial optimum initial optimum
2.550E-05
9.501E-10
2.550E-05
2.550E-05
2.550E-05
1.114E-04
1.114E-04
1.114E-04
1.114E-04
1.114E-04
7.644E-05
3.840E-05
1.038E-04
1.038E-04
1.408E-03
1.408E-03
1.550E-03
1.550E-03
1.550E-03
1.349E-03
1.550E-03
1.550E-03
8.595E-05
8.595E-05
8.595E-05
1.408E-03
1.408E-03
1.550E-03
1.0000
1.0000
1.0000
1.0000
1.0000
0.2000
0.0411
0.0411
0.0411
0.0411
0.0678
0.0678
0.0000
0.0000
0.0060
0.0060
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0060
0.0060
0.0000
1.0000
1.0000
1.0000
1.0000
1.0000
0.2499
0.0140
0.0140
0.0140
0.0140
0.0177
0.0177
0.0000
0.0000
0.0013
0.0013
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0013
0.0013
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0568
0.0568
0.0031
0.0031
0.0051
0.0051
0.0000
0.0000
0.0005
0.0005
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0005
0.0005
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0787
0.0787
0.0099
0.0099
0.0126
0.0126
0.0000
0.0000
0.0009
0.0009
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0009
0.0009
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0568
0.0568
0.1104
0.1104
0.1823
0.1823
0.0000
0.0000
0.0162
0.0162
0.0328
0.0328
0.0328
0.0000
0.0328
0.0328
0.0000
0.0000
0.0000
0.0162
0.0162
0.0328
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0787
0.0787
0.1474
0.1474
0.1864
0.1864
0.0000
0.0000
0.0136
0.0136
0.0284
0.0284
0.0284
0.0000
0.0284
0.0284
0.0000
0.0000
0.0000
0.0136
0.0136
0.0284
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.3973
0.3973
0.4510
0.4510
0.7447
0.7447
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.5506
0.5506
0.6193
0.6193
0.7834
0.7834
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.8000
0.4481
0.4481
0.3944
0.3944
0.0000
0.0000
0.0000
0.0000
0.0995
0.0995
0.1300
0.1300
0.1300
0.0000
0.1300
0.1300
1.0000
1.0000
1.0000
0.0995
0.0995
0.1300
0.0000
0.0000
0.0000
0.0000
0.0000
0.7501
0.2781
0.2781
0.2094
0.2094
0.0000
0.0000
0.0000
0.0000
0.0973
0.0973
0.1424
0.1424
0.1424
0.0000
0.1424
0.1424
1.0000
1.0000
1.0000
0.0973
0.0973
0.1424
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.7900
0.7900
0.6935
0.6935
0.6928
0.6928
0.6928
0.7900
0.6928
0.6928
0.0000
0.0000
0.0000
0.6935
0.6935
0.6928
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.7900
0.7900
0.7006
0.7006
0.6797
0.6797
0.6797
0.7900
0.6797
0.6797
0.0000
0.0000
0.0000
0.7006
0.7006
0.6797
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.2100
0.2100
0.1844
0.1844
0.1444
0.1444
0.1444
0.2100
0.1444
0.1444
0.0000
0.0000
0.0000
0.1844
0.1844
0.1444
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.2100
0.2100
0.1862
0.1862
0.1495
0.1495
0.1495
0.2100
0.1495
0.1495
0.0000
0.0000
0.0000
0.1862
0.1862
0.1495
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 7 ( 2 0 1 2 ) 2 4 7 0 e2 4 8 1
initial optimum
_ i kg=s
m
2479
58
APPENDIX
2480
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 7 ( 2 0 1 2 ) 2 4 7 0 e2 4 8 1
Fig. 12 e Behavioral pattern of the fuel preheater flue gas inlet temperature at the design point throughout the optimization
procedure.
6.
Conclusions
A GA-based optimization strategy was applied to a HT-PEMFCbased micro-CHP system. The value of the net electrical efficiency (objective function) obtained is 0.412, while the values
of the cogeneration thermal efficiency and total system efficiency are 0.497 and 0.91, respectively. The results of this
research study indicate that the proposed micro-CHP system
must focus on all system components, paying special attention at their geometric dimensions and operational variables.
The applied global optimization strategy successfully
managed to locate a system configuration with a significantly
higher net electrical efficiency than the one obtained initially.
The choice on the values of the independent variables is still
within the feasible limits of operating the system. The theoretical high tolerance of PBI technology on carbon monoxidecontent in the reformate fuel fed to the fuel stack anode,
allows the use of a low steam-to-carbon ratio as indicated by
the results. A more in-depth study for the water knockout/
condenser stage might be necessary in future studies to
minimize the losses occurred in that stage.
Finally, the applied global optimization strategy avoids the
typical trap of designing a fuel cell-based system with a high
APPENDIX
59
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 7 ( 2 0 1 2 ) 2 4 7 0 e2 4 8 1
fuel cell stack efficiency, but low net electrical efficiency. This
is accomplished by taking into account the geometric and
operational parameters of the fuel processing components,
such as the SMR and WGS reactors. A rigorous variation of
these parameters shifts the results toward the maximization
of the net electrical efficiency. Although this study has not
included cost-related parameters, such as the unit cost of fuel,
it should be noted that a high net electrical efficiency has a high
impact factor on the reduction of the system operational cost.
Acknowledgments
The authors would like to acknowledge the support of Danfoss
and Dantherm Power throughout the realization of this
research study.
references
[1] Hawkes AD, Leach MA. The capacity credit of microcombined heat and power. Energy Policy 2008;36(4):1457e69.
[2] Hawkes AD, Leach MA. Modelling high level system design
and unit commitment for a microgrid. Applied Energy 2009;
86(7e8):1253e65.
[3] Hawkes AD, Leach MA. Cost-effective operating strategy for
residential micro-combined heat and power. Energy 2007;
32(5):711e23.
[4] Korsgaard AR, Nielsen MP, Kr SK. Part one: a novel model of
HTPEM-based micro-combined heat and power fuel cell
system. International Journal of Hydrogen Energy 2008;33(7):
1921e31.
[5] Peacock A, Newborough M. Impact of micro-CHP systems on
domestic sector co emissions. Applied Thermal Engineering
2005;25(17e18):2653e76.
[6] Gunes MB, Ellis MW. Evaluation of energy, environmental,
and economic characteristics of fuel cell combined heat and
power systems for residential applications. Journal of Energy
Resources Technology 2003;125(3):208.
[7] Georgopoulos NG. Application of a decomposition strategy to
the optimal synthesis/design and operation of a fuel cell
based total energy system. Virginia Tech; 2002.
60
2481
APPENDIX
Journal Article V
APPENDIX
61
62
APPENDIX
1. Introduction
Combined-heat-and-power (CHP) is a well-known technology for large-scale stationary plants. These systems are designed to convert the chemical energy in a fuel
into both electrical power and useful heat. During recent years, another application
of the CHP technology has evolved, which includes smaller scale applications such
as a residential micro-CHP system. This system can provide electricity and heat
(hot water and space heating) for a detached single-family household. Systems
operating on natural gas must be integrated with a fuel processing subsystem, to
allow conversion of natural gas to hydrogen. Heat exchangers are also used for the
thermal management of the system, while balance-of-plant components perform
various necessary tasks, such as air compressing or water pumping. The thermal
management of the system includes heating and cooling processes (e.g. preheating
of the steam-methane mixture, or cooling of the reformate mixture prior to the
water gas shift reaction), and also heat recovery to fulfill a residential load profile
(e.g. space heating). Finally, a thermal storage tank can be coupled to the system
to provide greater operational flexibility during transient heating load demands.
Pinch analysis and process integration is a method used to analyze the heating
and cooling processes of a system and then assist in identifying energy saving opportunities, through a more efficient heat exchange network (HEN) design [1-10].
An optimum HEN design will allow cost reductions, both capital and operating
ones [1,11,12]. The objective is to assist process engineers to identify and model
HEN configurations which are initially considered unknown. The technique can
be summarized in two main steps: the targeting phase, aiming at identifying the
Received by the editors 2011.
Key words and phrases. PBI; heat exchange network; micro-CHP; optimization; process integration; fuel cell system.
1
APPENDIX
63
A. ARSALIS ET AL.
64
APPENDIX
APPENDIX
65
A. ARSALIS ET AL.
CH4 + H2 O
CO + 3H2
(2)
CH4 + 2H2 O
CO2 + 4H2
(3)
2CH4 + 3H2 O
CO + CO2 + 7H2
(5)
LSM R
n CH4 ,i
=
ntubes B A
Z
0
dXCH4
SM R
rCH
4
where n CH4 ,i is the inlet molar flow rate of methane (kmol/s), ntubes is the number of tubes, B is the catalyst bulk density (kg/m3 ), A is the cross-sectional area
SM R
of a single tube (m2 ), XCH4 is the kinetic methane conversion, and rCH
is the
4
demethanation reaction rate (kmol/kg-h), which is given by the following expression,
EA
SM R
SM R
(6)
rCH
=
k
exp
pCH4
0
4
SM R
RTavg
where k0SM R is the Arrhenius frequency factor (kmol/kg-h), EA is the Arrhenius
SM R
is the average reformate gas temperature (K),
activation energy (kJ/kmol), Tavg
R is the ideal gas constant (kJ/kmol-K), and pCH4 partial pressure of methane
(bar).
3.2. Water gas shift reactor. The WGS reactor receives the reformate gas from
the SMR reactor and its main purpose is to reduce the carbon monoxide content,
while simultaneously achieving a slight increase in the hydrogen-content. The carbon monoxide content should be reduced at a level below the maximum acceptable
for the fuel cell stack, which is 5% according to the fuel cell stack manufacturer
[18]. The kinetic constant is based on a power law relationship and calculated at
steady-state. The thermodynamic equilibrium of the slightly exothermic water-gas
shift reaction is calculated as follows [19],
4400
(7)
KT = exp
4.063
TW GS
where KT is the equilibrium constant given in [20], and TW GS is the temperature
of the WGS reactor. The kinetic power law fit parameters are obtained from [19],
EA
W GS
(8)
KW GS = k0
exp
RT
where k0W GS is the pre-exponential factor. The carbon monoxide conversion is given
by,
dxCO
rCO
(9)
= w
dz
n CO
where is the cross-sectional area of the WGS reactor (m), w is the density of the
wire mesh catalytic material (kg/m3 ), rCO is the carbon monoxide reaction rate
66
APPENDIX
(kmol/kg-s), and n CO is the molar flow rate of carbon monoxide (kmol/s). The
temperature gradient is defined as,
dT
w
(10)
=
dz
g cp,mass us (rCO dHr,1 )
where g is the density of the reformate gas (kg/m3 ), us is the superficial velocity
(m/s), and dHr,1 is the enthalpy of reaction (kJ/mol).
3.3. Fuel cell stack. The fuel cell operating temperature is an important factor in
regards of efficiency and fuel cell membrane degradation. Increased operating temperatures aid in the reduction of cooling requirements, simplify water management
and minimize contamination problems. An HT-PEMFC utilizes a PBI (Polybenzimidazole) membrane, operating at temperatures between 140 and 180o C [21].
Therefore this fuel cell type can be coupled effectively in a micro-CHP system, because waste heat can be recovered, and lower quality reformed hydrogen may be
used as fuel [22]. The model considers only the reaction of hydrogen with oxygen,
while all other species are considered inactive. The ohmic and diffusion resistances
are based on linear regressions evaluated experimentally in [23],
(11)
(12)
where Tcell is the fuel cell operating temperature. The anode and cathode overpotentials are given by the following expressions, respectively,
i
RTcell
sinh1
(13)
a =
anode F
2keh H2
i0 + i
RTcell
i
ln
(14)
c =
+ Rdif f
4cathode F
i0
air 1
where i is the charge transfer coefficient, air is the cathode air stoichiometric
ratio, keh is the electro-oxidation rate of hydrogen, H2 is the surface coverage of
hydrogen, i is the current density, and i0 is the exchange current density. The
ohmic losses are given by,
(15)
ohmic = iRohmic
Vcell = V0 a c ohmic
3.4. Heat losses. Component heat losses are calculated on the basis of each components dimensions and the insulations material properties and thickness. Heat
exchanger and piping losses are neglected since the initial HEN configuration is
not known in advance. Therefore component heat losses are only considered for the
SMR reactor, fuel cell stack and combustor. A thermal conductivity of k = 0.04 W/(m-K)
and a convective heat transfer coefficient of h = 6 W/(m2 -K) are used, as found
in the literature [24]. The total heat losses for each component are given by the
following relation,
Tc Tadj
Tc T
+
(17)
Q loss = Q loss,ca + Q loss,cc = Lca
Lcc
1
kAca + hAca
kAcc
where Q loss,ca and Q loss,cc are the component-to-atmosphere and component-tocomponent heat loss rates (W), respectively. Tc is the average temperature of the
component (K), Tadj is the temperature of the adjacent component (K), T is the
atmospheric temperature (K), Lca is the thickness of the outer insulation (m), Lcc
APPENDIX
67
A. ARSALIS ET AL.
Description (unit)
Value
Tcell
Acell
ncells
267
diWGS
0.13
Density of the wire mesh catalytic material of the WGS reactor (kg/m3)
1385
1200
160
45.16
The HEN is designed and optimized by use of pinch analysis techniques. Basic
pinch analysis and process integration leads to a HEN with a MER. This in turn
initiates the basis for an optimum HEN design and integration with the microCHP system. Since residential micro-CHP systems based on fuel cell technology
are highly integrated [3], an effective process design methodology must be applied to
allow the system to operate in an efficient and realistic manner. Such a methodology
can be summarized in the following steps:
(1) Development and optimization of a computational simulation model that
allows the calculation of the energy requirements, which can be defined as
a list of hot and cold streams;
(2) Removal of all heat exchangers from the initial configuration and application of pinch analysis to evaluate possible HEN design opportunities [1];
(3) Reevaluation of the HEN design by means of a process change optimization strategy and modification of the stream data set to minimize utility
requirements [1,2];
(4) Optimization of the HEN, with respect to its total annual cost, by use of a
simultaneous optimization algorithm as found in the literature [12,16,25].
A systematic approach of this methodology allows evaluation of different HEN
configurations, which may be integrated to the micro-CHP system as shown in
the next section. Before the pinch analysis is applied, the fixed values for the
simulation model must be defined. These values, shown in Table 1, are kept fixed
due to operational limitations imposed either by the manufacturer or defined in the
literature.
4.1. Pinch analysis. After the stream data set (including temperatures, flow rates
and heat loads) has been extracted from the system simulation model, and a minimum temperature difference Tmin is selected, the energy targets and pinch temperature can be calculated. It is predicted that the pinch analysis will yield towards
68
APPENDIX
a threshold problem with a single pinch. Therefore only a cooling utility will be
required. The integration model considers cooling water as a cold utility to close
the system balance. The design can be determined by applying the plus-minus
principle (also known as the ticking-off rule), and by identifying essential matches
at the non-utility end [1,2]. The plus-minus principle suggests that a process change
will reduce the utility targets if: (a) above the pinch, the total hot stream heat load
is increased, and/or the total cold stream load is decreased, (b) below the pinch,
the total hot stream heat load is decreased, and/or the total cold stream load is
increased. Obviously some changes are not feasible due to component/process constraints, as explained in Section 3. Prior to HEN design, some processes may allow
modifications which can lead in minimization of the utilities consumption.
APPENDIX
69
A. ARSALIS ET AL.
Table 2. Initial process stream data used in the heat integration analysis.
#
Description
Stream
Tsupply ( K )
Ttarget ( K )
H ( kW )
Hot
906.7
523.2
0.26740
Hot
601.8
433.2
0.11390
Hot
1078.2
298.2
6.03600
Cold
432.0
970.6
4.00000
Cold
298.2
599.2
0.04540
373.2
599.2
0.08869
373.2
374.2
0.44400
298.2
373.2
0.06183
6a
6b
Cold
6c
70
APPENDIX
Figure 2. Hot and cold composite curves extracted from the initial process stream data set.
more detailed observation of the composite curves shows that apart from the two
dominant cold (fuel cell stack exhaust preheating) and hot (SMR reactor flue gas)
streams, which are interdependent, the steam generation (shown in the lower left
cold composite curve) does not cause any pinch matching problems. This suggests
that no modification, such as local minimum temperature difference contribution, is
required. Finally it should be noted that some streams are very small compared to
others (e.g. #4 vs. #5), but still they cannot be removed due to their importance
in the unimpeded operation of the system.
5.1.2. Process change and system optimization. As indicated above process change
must be applied in the early stages of the pinch analysis, since the plus-minus
principle is only useful for first-stage screening [1]. In the system under study it
is vital to apply the principle simultaneously with system optimization to ensure
the feasibility of the modified processes. Specifically for the system under study,
and since this is a threshold problem, two modifications below the pinch can be
attempted: (i) decrease the total hot stream load and (ii) increase the total cold
stream load. The streams allowing modification are #3, #5, and #6 (shown in
Table 2). Streams #5 and #6, correspond to the inlet temperature of the SMR
reactor. The WGS reactor, combustor and fuel cell stack inlet/outlet temperatures
cannot be modified, because they are either fixed or dependent. The temperature
for stream #3 cannot be modified but its heat load can be reduced by reduction of
its corresponding flow rate.
The simulation model of the micro-CHP system is optimized using the genetic
algorithm method built-in in EES. The objective function is the maximization of
the net electrical efficiency of the system (el,net ) at nominal load. Four independent design parameters are selected and varied during the system optimization
process, and they are shown in Table 3, with their range of variation. The selection of the four independent variables should not be considered accidental, because
the optimization is concentrated around the SMR reactor parameters, which is an
endothermic reaction and can subsequently affect the heating utilization related to
the process change described above. The obtained value of the objective function
APPENDIX
71
10
A. ARSALIS ET AL.
Description (unit)
Initial
Min
Max
Optimum
T6
599
400
600
587
T11
626
500
700
597
m 17
6.82E-3
6.50E-3
6.87E-3
6.55E-3
S C
3.19
2.50
4.00
2.83
Table 4. Optimum process stream data used in the heat integration analysis.
#
Tsupply ( K )
Ttarget ( K )
H ( kW )
924.1
523.2
0.26220
603.3
433.2
0.10790
1054.4
298.2
5.59900
432.0
1021.8
3.86200
298.2
587.2
0.04325
6a
373.2
587.2
0.07464
6b
373.2
374.2
0.39500
6c
298.2
373.2
0.05500
is 35.16% for the design parameter values shown in Table 3. Therefore the optimization resulted in a significant increase of the objective function, compared to a
value of 31.86% prior to optimization.
5.1.3. Pinch analysis based on the optimum stream data set. Pinch analysis is repeated for the new stream data set obtained by the optimization strategy. Table 4
shows a list of the optimum process stream data set, which was calculated by the
optimization process. The corresponding composite curves are shown in Fig. 3.
An observation of the figure shows that the process change was successful, since
it accomplished minimization of the cold utility requirement (1.54 kW). Figure 4
shows the grand composite curve (GCC), which illustrates the net heat flow against
the shifted temperature. This graph represents the difference between the available
heat from the hot streams and the heat required by the cold streams, relative to
the pinch, at a given shifted temperature. For a threshold problem, like the one
under study, the net heat flow at the top (heating utility requirement) is zero, while
the corresponding one at the bottom (cooling utility requirement) is 1.54 kW. The
pinch temperature is 771o C, since it is the point where the net heat flow is 0 and
the GCC touches the axis.
In a threshold problem it is necessary to distinguish its characteristics. In the
analysis under study, the closest temperature approach between the hot and cold
composites is at the non-utility end and the curves diverge away from this point, as
shown in Fig. 2-4. In this case, design can be started from the non-utility end, using
72
APPENDIX
11
Figure 3. Hot and cold composite curves extracted from the optimum process stream data set.
the pinch design rules. In this case the network design is relatively slack [1], and a
great many designs are possible as the thermodynamic constraint of the pinch does
not apply. Another aspect that can be studied is whether the endothermic SMR
reactor has been placed appropriately in terms of heat integration. Pinch analysis
can suggest minor refinements allowing improved process integration, although this
would suggest changing the reaction conditions, which may alter the overall system
performance. The possible extent of heat integration of the reactor with the rest of
the process mainly depends on the reaction temperature. The appropriate placement principle [1] suggests that for endothermic reactions, the heat should ideally
come below the pinch so that it can be driven by waste heat from the process.
This is followed for the system under study, since the pinch temperature (771o C)
is higher than the reformate outlet temperature (651o C).
The system performance is then tested at part-load operation to examine the
behavior of the proposed system at off-design conditions. Figure 5 shows the variation of efficiency at different net electrical power outputs. As the load decreases,
the net electrical efficiency increases linearly, while the thermal efficiency (and total
efficiency) decreases initially linearly, while at very low loads the decrease is higher.
This is because of the effect of the heat losses, which are temperature and area
dependent. Since component temperatures and areas remain constant at all loads,
heat losses have a more drastic effect at lower operating loads. Therefore, system
operation at low, and especially critical, loads may not be favored at periods of high
demand on cogeneration heat. The greater loss is due to the SMR reactor, which
although it is heavily insulated, it contributes 50% of the losses. This is because of
the high temperature flows (reformate & flue gas) occurring in the SMR reactor.
5.2. Simultaneous HEN optimization. The simultaneous optimization strategy analyzed in the previous section is applied to the HEN to obtain an optimized
configuration. The MINLP model with two stages is solved in GAMS, with the
DICOPT++ solver [27]. The compact nature of the micro-CHP system does not
favor the use of stream splitting, and therefore this is restricted in the code. In
addition to the potentially more complex operation, stream splitting would require
APPENDIX
73
12
A. ARSALIS ET AL.
an additional investment cost for a control valve [12]. As in the initial configuration, the pattern of the composite curves, shown in Fig. 6, is identical to the
one shown in Fig. 3. This means that the HEN configuration still requires only
cooling (threshold problem), and therefore maintains its MER nature. This can be
explained by the significantly higher cost of running a hot utility, as compared to
the cooling one [12,16,25].
The grid diagram representation, shown in Fig. 7, shows the heat transfer operation of the optimum HEN. Hot streams are at the top running left to right,
while cold streams are located at the bottom running right to left. Heat exchange
matches are represented by vertical lines joining two arrows on the two streams
being matched. An exchanger using cold utility is represented by an arrow with
a W. The coupling of the optimized HEN configuration with the micro-CHP
system is shown in Fig. 8. The new configuration is significantly different than the
original one (shown in Fig. 1), due to the cost reduction requirements imposed in
the objective function [16]. Therefore, the optimum configuration encompasses the
need for heat exchangers with a cheaper structure, such as a minimum heat transfer
74
APPENDIX
13
APPENDIX
75
14
A. ARSALIS ET AL.
Figure 7. HEN configuration after the application of the simultaneous HEN optimization.
76
APPENDIX
15
Description (unit)
Value
Q util
N HEx
AHEx
0.13
Autil
0.05
0.18
AHEx ,tot
1.53
8
Cutil
Cinv
8070
8147
CHEN
77
6. Conclusions
An optimization methodology, based on process integration techniques, is applied
on an HT-PEMFC-based micro-CHP simulation model to improve its performance,
in terms of net electrical efficiency and total HEN cost. This is accomplished by
solving two objective functions: (1) maximization of the net electrical efficiency
of the system, (2) minimization of the total HEN cost. The first optimization
strategy includes a pinch analysis and process change procedure. Pinch analysis is
applied on the initial set of stream data, and then by use of the plus-minus principle
(process change), the system is optimized with a genetic algorithm optimization
methodology. This approach is diverse, because it allows studying and evaluation
of the optimum system parameters (e.g. steam-to-carbon ratio), based on the
knowledge input suggested by the pinch analysis technique. The proposed system
exhibits high electrical efficiencies, which vary from 35.2 to 43.6% (based on LHV).
The total system efficiencies vary from 65.2 to 91.1%. The system total efficiencies
are significantly decreased at low operating loads, because of the higher impact of
heat losses at those loads.
In the second optimization strategy the optimum stream data set, calculated
in the previous optimization strategy, is used as the input in an MINLP problem
formulation. This optimization attempts to minimize the total annual HEN cost,
by suggesting a HEN configuration fulfilling the targets with a reduced cost. This
is successfully accomplished, and most importantly not in the expense of the net
electrical efficiency, since the energy input is not forced to change. Therefore the
heat integration is accomplished requiring only cold utilities (threshold problem).
A HEN, significantly different than the initial one, is generated by the optimization
strategy, suggesting the effectiveness of the optimization algorithm. The minimized
total annual cost of the HEN is $8,147/year. The calculation is rapid because of the
nature of the MINLP formulation, which consists of a nonlinear objective function,
but with linear constraints.
APPENDIX
77
16
A. ARSALIS ET AL.
Acknowledgment
The authors would like to acknowledge the support of Danfoss and Dantherm
Power throughout the realization of this research study.
References
[1] I.C. Kemp, Pinch Analysis and Process Integration: A User Guide on Process
Integration for the Efficient Use of Energy, 2nd ed., Butterworth-Heinemann,
2007.
[2] R. Smith, Chemical Process Design and Integration, Wiley, Chichester, UK,
2005.
[3] J. Godat, F. Marechal, Optimization of a fuel cell system using process integration techniques, Journal of Power Sources. 118 (2003) 411-423.
[4] R. OHayre, W. Colella, S.-W. Cha, F.B. Prinz, Fuel Cell Fundamentals, 2nd
ed., Wiley, 2009.
[5] J. Larminie, A.L. Dicks, Fuel Cell Systems Explained, 2nd ed., Wiley, Chichester, UK, 2003.
[6] F. Marechal, F. Palazzi, J. Godat, D. Favrat, Thermo-Economic Modelling and
Optimisation of Fuel Cell Systems, Fuel Cells. 5 (2005) 5-24.
[7] B. Linnhoff, J.R. Flower, Synthesis of heat exchanger networks: I. Systematic
generation of energy optimal networks, AIChE Journal. 24 (1978) 633-642.
[8] B. Linnhoff, J.R. Flower, Synthesis of heat exchanger networks: II. Evolutionary
Generation of Networks with Various Criteria of Optimality, AIChE Journal.
24 (1978) 642-654.
[9] B. Linnhoff, E. Hindmarch, The pinch design method for heat exchanger networks, Chemical Engineering Science. 38 (1983) 745-763.
[10] N.C. Monanteras, C.A. Frangopoulos, Towards synthesis optimization of a fuelcell based plant, Energy Conversion and Management. 40 (1999) 1733-1742.
[11] C. Wallmark, P. Alvfors, Design of stationary PEFC system configurations to
meet heat and power demands, Journal of Power Sources. 106 (2002) 83-92.
[12] L.T. Biegler, I.E. Grossmann, A.W. Westerberg, Systematic Methods of Chemical Process Design, 1st ed., Prentice Hall, 1997.
[13] V. Verda, F. Nicolin, Thermodynamic and economic optimization of a MCFCbased hybrid system for the combined production of electricity and hydrogen,
International Journal of Hydrogen Energy. 35 (2010) 794-806.
[14] F. Palazzi, N. Autissier, F. Marechal, D. Favrat, A methodology for thermoeconomic modeling and optimization of solid oxide fuel cell systems, Applied
Thermal Engineering. 27 (2007) 2703-2712.
[15] N. Autissier, F. Palazzi, F. Marechal, J. van Herle, D. Favrat, Thermo-Economic
Optimization of a Solid Oxide Fuel Cell, Gas Turbine Hybrid System, Journal
of Fuel Cell Science and Technology. 4 (2007) 123.
[16] T.F. Yee, I.E. Grossmann, Simultaneous optimization models for heat integrationII. Heat exchanger network synthesis, Computers & Chemical Engineering. 14
(1990) 1165-1184.
[17] N.G. Georgopoulos, Application of a Decomposition Strategy to the Optimal
Synthesis/Design and Operation of a Fuel Cell Based Total Energy System,
Virginia Tech, 2002.
[18] Serenergy, Available at: http://www.serenergy.dk/ [Accessed February 2, 2011]
[19] R.L. Keiski, T. Salmi, P. Niemisto, J. Ainassaari, V.J. Pohjola, Stationary and
transient kinetics of the high temperature water-gas shift reaction, Applied
Catalysis A: General. 137 (1996) 349-370.
[20] J. Davies, D. Lihou, Optimal design of methane steam reformer, Chem. Proc.
Eng. 52 (1971) 71-80.
78
APPENDIX
17
[21] D.F. Cheddie, N. Munroe, Mathematical model of a PEMFC using a PBI membrane, Energy Conversion and Management. 47 (2006) 1490-1504.
[22] J. Zhang, Z. Xie, J. Zhang, Y. Tang, C. Song, T. Navessin, et al., High temperature PEM fuel cells, Journal of Power Sources. 160 (2006) 872-891.
[23] A.R. Korsgaard, R. Refshauge, M.P. Nielsen, M. Bang, S.K. Kr, Experimental
characterization and modeling of commercial polybenzimidazole-based MEA
performance, Journal of Power Sources. 162 (2006) 239-245.
[24] K. Kim, M.R. von Spakovsky, M. Wang, D.J. Nelson, A hybrid multi-level
optimization approach for the dynamic synthesis/design and operation/control
under uncertainty of a fuel cell system, Energy. (2010).
[25] J. Ponce-Ortega, A. Jimenez-Gutierrez, I.E. Grossmann, Optimal synthesis of
heat exchanger networks involving isothermal process streams, Computers $
Chemical Engineering. 32 (2008) 1918-1942.
[26] S. Sieniutycz, J. Jezowski, Energy Optimization in Process Systems, 1st ed.,
Elsevier, Amsterdam, 2009.
[27] J. Viswanathan, I.E. Grossmann, A combined penalty function and outer approximation method for MINLP optimization, Computers & Chemical Engineering. 14 (1990) 769-782.
Department of Energy Technology, Aalborg University, Pontoppidanstrde 101,
9220, Aalborg ., Denmark
E-mail address: aar@et.aau.dk
APPENDIX
79