Energy Conversion and Management 51 (2010) 2724–2732

Contents lists available at ScienceDirect

Energy Conversion and Management

journal homepage: www.elsevier.com/locate/enconman

Thermodynamic analysis of an integrated solid oxide fuel cell cycle

with a rankine cycle
Masoud Rokni
Technical University of Denmark, Dept. Mechanical Engineering, Thermal Energy System, Building 402, 2800 Kgs, Lyngby, Denmark

a r t i c l e i n f o a b s t r a c t

Article history: Hybrid systems consisting of solid oxide fuel cells (SOFC) on the top of a steam turbine (ST) are investi-
Received 23 September 2009 gated. The plants are fired by natural gas (NG). A desulfurization reactor removes the sulfur content in the
Accepted 1 June 2010 fuel while a pre-reformer breaks down the heavier hydro-carbons. The pre-treated fuel enters then into
the anode side of the SOFC. The remaining fuels after the SOFC stacks enter a burner for further burning.
The off-gases are then used to produce steam for a Rankine cycle in a heat recovery steam generator
Keywords: (HRSG). Different system setups are suggested. Cyclic efficiencies up to 67% are achieved which is consid-
SOFC
erably higher than the conventional combined cycles (CC). Both adiabatic steam reformer (ASR) and cat-
Rankine
Steam plant
alytic partial oxidation (CPO) fuel pre-reformer reactors are considered in this investigation.
Hybrid plant Ó 2010 Elsevier Ltd. All rights reserved.
CPO
ASR

1. Introduction
The solid oxide fuel cell (SOFC) is an electrochemical reactor
currently under development by some companies for power and
heat generation application. Depending on the type of the electro-
lyte they are operating at temperature levels of more than about
750–1000 °C. The lower temperature alternative is now being
developed for market entry during the next decade. Due to mate-
rial complication on the balance of plant (BoP) components many
companies are trying to find new materials for the SOFC cells to de-
crease the operating temperature of the SOFC stacks. Temperatures
of about 650 °C are also mentioned.
The biggest advantage of the SOFC in comparison with other
types of fuel cells may be in its flexibility in using different types
of fuel. However, in planar SOFCs one needs to pre-process most
kind of fuels in order to break down the heavier hydro-carbons
which may otherwise damage the fuel cells. The sulphur content
in the fuels must firstly be removed and then pre-reformed before
entering the anode side of the SOFC. Such fuel pre-processing can
be done in two different catalytic reactors operating at different
temperature levels indicated by reactor manufacturers.
SOFC-based power plants have been studied for a while and
some companies, such as Wärtsilä, are trying to realize such a sys-
tem for combined heat and power (CHP) applications; see e.g. Fon-
tell et al. [1]. The SOFC is also combined with combined cycles (CC)
in the literature to achieve ultra high electrical efficiencies, see e.g.
E-mail address: MR@mek.dtu.dk
Rokni [2] and Riensche et al. [3]. Due to the current operating tem-
perature of the SOFC stacks (more than about 750 °C), hybrid SOFC
and GT (gas turbine) systems have also been studied extensively in
the literature, e.g. in Pålson et al. [4] for CHP and in Proell et al. [5]
with internal biomass gasification. Characterization, quantification
and optimization of hybrid SOFC–GT systems have been studied by
e.g. Subramanyan et al. [6] and Calise et al. [7]. The dynamics and
control concept of a pressurized SOFC–GT hybrid system has also
been studied, such as in Wächter et al. [8]. In Roberts and Brouwer
[9] modeling results are compared with measured data for a
220 kW hybrid SOFC–GT power plant. Details on design, dynamics,
control and startup of such hybrid power plants are studied in Ro-
kni et al. [10]. Part–load characteristics of a SOFC–micro GT were
also studied in Costamagna et al. [11].
While hybrid SOFC–GT plants have been extensively studied by
many researchers, the investigations on combined SOFC and steam
turbine (ST) are very limited see Dunbar et al. [12]. In addition the
SOFC manufacturers are trying to decrease the operating tempera-
ture of the SOFC stacks, and then the combination of SOFC–ST hy-
brid system would be more attractive than the SOFC–GT systems.
By decreasing the operating temperature, the material cost for
the SOFC stacks will be decreased as well as many problems asso-
ciated with the BoP components will be diminished.
In hybrid SOFC–GT plants the SOFC stacks must be pressurized
in an extremely large vessel (depending on the size of the plant
which is usually in mega Watt classes). Such a practical problem
will be diminished in hybrid SOFC–ST systems, because the stacks
will work under atmospheric pressure. Another disadvantage of a
SOFC–GT system is that the startup simplicity of the GT is

0196-8904/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.enconman.2010.06.008
 M. Rokni / Energy Conversion and Management 51 (2010) 2724–2732 2725

diminished since the startup time of a SOFC plant is much longer thermodynamic properties of the fluids involved. In addition, the
than that of a GT plant. components include a number of constitutive equations represent-
Fuel pre-reforming can be done in different reactors such as adi- ing their physical properties, e.g. heat transfer coefficients for heat
abatic steam reformer (ASR) and catalytic partial oxidation (CPO). exchangers and isentropic efficiencies for compressors and
The disadvantage of an ASR reactor is that it needs superheated turbines.
steam during startup (depending on the operating temperature During the development of DNA the four key terms, portability,
of the reactor, i.e. 400 °C) which is an extremely power consuming robustness, efficiency, and flexibility have been kept in mind as the
process. During normal operation steam is available after the an- important features for making a generally applicable tool for en-
ode side of the SOFC stacks, which can be recycled into the system. ergy system studies. In recent years DNA has been introduced into
However, in a CPO reactor additional air is needed for the fuel pre- the education at the Department and has been applied to numer-
reforming process, which in turn increases the plant power con- ous different types of systems. The program is written in FORTRAN.
sumption (compressor work) and thereby the plant efficiency will
be decreased. In this study, ASR reforming process versus CPO 2.1. Modelling of SOFC
reforming process is studied in terms of plant design, plant effi-
ciency and plant output power. The SOFC model used in this investigation is based on the planar
In the Rankine cycle a single pressure level is proposed to study type developed by DTU-Risø and TOPSØE Fuel Cell. The model is
plant design, plant efficiency and plant output power. Further, calibrated against experimental data in the range of 650–800 °C
plants characteristics of such system configurations are also stud- (SOFC operational temperature) as described in Petersen et al.
ied. In addition, the off-gases from the heat recovery steam gener- [15]. The operational voltage (EFC) is found to be
ator (HRSG) are used to preheat the incoming air into the SOFC
EFC ¼ ENernst  DEact  DEohm  DEconc  DEoffset ð1Þ
stacks. The effect of such suggested air pre-heating on plant effi-
ciency and output power is thus studied. All the configurations where ENernst , Eact , Eohm , Econc , Eoffset are the Nernst ideal reversible
studied here are novel in terms of designing new plants with very voltage, activation polarization, ohmic polarization, concentration
high efficiencies. polarization and the offset polarization respectively. The activation
It should be noted that the systems presented here are studied polarization can be evaluated from Butler–Volmer equation (see
thermodynamically and the objective of this study is not to present Keegan et al. [16]). The activation polarization is isolated from other
or discuss associated costs. polarization to determine the charge transfer coefficients as well as
exchange current density from the experiment by curve fitting
technique. It follows,
2. Methodology
RT
DEact ¼
ð0:001698T  1:254ÞF
The results of this paper are obtained using the simulation tool " #
dynamic network analysis (DNA), see Elmegaard and Houbak [13], 1 id
 sin h ð2Þ
which is a simulation tool for energy system analysis. It is the pres- 2ð13:087T  1:096  104 Þ
ent result of an ongoing development at the Department of
Mechanical Engineering, Technical University of Denmark, which where R, T, F and id are the universal gas constant, operating tem-
began with a Master’s Thesis work in Perstrup [14]. Since then perature, Faradays constant and current density respectively. Oh-
the program has been developed to be generally applicable cover- mic polarization depends on the electrical conductivity of the
ing unique features, and hence supplementing other simulation electrodes as well as the ionic conductivity of the electrolyte and
programs. can be described as
Some of the important features are:  
t an t el tca
DEohm ¼ þ þ id ð3Þ
ran rel rca
 Simulation of both steady state (algebraic equations) and
dynamic models (differential equations). where tan = 600 lm, tel = 50 lm and tca = 10 lm are the anode thick-
 Handling of discontinuities in dynamic equations. ness, electrolyte thickness and cathode thickness respectively. ran,
 Use of a sparse-matrix-based simultaneous solver for algebraic rel and rca are the conductivity of anode, electrolyte and cathode
equations. respectively.
 No causality implied on the model input, i.e. no restriction of  
5:760  107 0:117
the choice of inputs and outputs. ran ¼ 105 ; rca ¼ exp  5
ð4Þ
T 8:617  10 T
 Medium compositions can be variables.
 Models of thermodynamic states, transport variables and radia-
tive properties for relevant fluids, e.g. steam, ideal gases and rel ¼ 8:588  108 T 3  1:101  104 T 2 þ 0:04679T  6:54 ð5Þ
refrigerants. Concentration polarization is dominant at high current densi-
 Features for modeling solid fuels of arbitrary components. ties for anode-supported SOFC, wherein insufficient amounts of
reactants will be transported to the electrodes and the voltage will
The component library includes models of, heat exchangers, then reduce significantly. Neglecting the cathode contribution (see
burners, Gasifiers, turbo machinery, dryers and decanters, energy e.g. Kim and Virkar) [17], it can be modeled as
storages, engines, valves, controllers as well as more specialized !  !
components and utility components. The user may also implement pH2 id id
additional components. DEconc ¼ B  ln 1 þ  ln 1  ð6Þ
pH2 O; ias ias
In DNA the physical model is formulated by connecting the rel-
evant component models through nodes and by including operat- where B is the diffusion coefficient and is calibrated against exper-
ing conditions for the complete system. The physical model is imental data which found to be,
converted into a set of mathematical equations to be solved
T
numerically. The mathematical equations include mass and energy B ¼ ð0:008039X 1
H2  0:007272Þ ð7Þ
conservation for all components and nodes, as well as relations for T ref
2726 M. Rokni / Energy Conversion and Management 51 (2010) 2724–2732
In the above equations pH2 and pH2 O are the partial pressures for
the H2 and H2O respectively, while Tref is the reference temperature
(1023 K). The anode limiting current is defined as
2FpH2 Dbin V an
ias ¼
RTt an san
where Van and san are the porosity and tortuosity of the anode and
are the physical characteristics as 30% and 2.5 lm in the experi-
mental setup. The binary diffusion coefficient is given by
  T 1:75 P
ref
Dbin ¼ 4:107  105 X H2 þ 8:704  105
T ref P The partial derivation of this equation with respect to n_ i;out can be
which is also calibrated against the experimental data. Pref is the ref-
erence pressure as 1.013 bar and X H2 is the mass reaction rate of H2.
Finally the current density id is directly proportional to the amount
of reacting hydrogen according to the Faraday’s law:
n_ H2 2F
id ¼
A
where n_ H2 is molar reaction rate of H2. The area A is the physical
property of the cell and is 144 cm2.
The fuel composition leaving the anode is calculated by the
Gibbs minimization method as described in Smith et al. [18]. Equi-
librium at the anode outlet temperature and pressure is assumed
for the following species: H2, CO, CO2, H2O, CH4 and N2. The equi-
librium assumption is fair since the methane content in this study
is low enough. The power production from the SOFC (PSOFC) de-
pends on the amount of chemical energy fed to the anode, the
reversible efficiency (grev), the voltage efficiency (gv) and the fuel
utilization factor (UF). It is defined in mathematical form as
PSOFC ¼ ðLHVH2 n_ H2 ;in þ LHVCO n_ CO;in þ LHVCH4 n_ CH4 ;in Þgrev gv U F
where both gv and UF are the set values while the reversible effi-
ciency is the maximum possible efficiency defined as the relation-
ship between the maximum electrical energy available (change in
Gibbs free energy) and the fuels LHV (Lower Heating Value) as
follows
ðDgf Þfuel
grev ¼
LHVfuel
 
1 tures is assumed to be 780 °C. Therefore, the burned fuel after
ðDgf Þfuel ¼ ðgf ÞH2 O  ðgf ÞH2  ðgf ÞO2 yH2 ;in
2
 
1 ð13Þ
þ ðgf ÞCO2  ðgf ÞCO  ðgf ÞO2 yCO;in
2
 
þ ðgf ÞCO2 þ 2ðgf ÞH2 O  ðgf ÞCH4  2ðgf ÞO2 yCH4 ;in
2.2. Modelling of fuel pre–reforming
The reforming process is assumed to be equilibrium by mini-
mizing the Gibbs free energy. The Gibbs free energy of a gas (as-
sumed to be a mixture of k perfect gases) is given by
X
k
 
G_ ¼ n_ i g 0i þ RT lnðni pÞ
i¼1
where g0, R and T are the specific Gibbs free energy, universal gas
constant and gas temperature respectively. Each atomic element
in the inlet gas is in balance with the outlet gas composition, which
yields the flow of each atom has to be conserved. For N elements
this is expressed as (see, e.g. Elmegaard and Houbak) [19]
X
k X
k steam for operation. Such steam must be supplied to the reformer
n_ i;in Aij ¼ n_ i;out Aij for j ¼ 1; N
i¼1 i¼1
The N elements corresponds to H2, CO2 , H2O and CH4 in such
pre – reforming process. Ai,j is a matrix with information of the
mole j in each mole of i (H2, CO2 , H2O and CH4). The minimization
of Gibbs free energy can be formulated by introducing a Lagrange
ð8Þ multiplier, l, for each of the N constraints obtained in Eq. (15).
After adding the constraints, the expression to be minimized is
then
" !#
X
N X
k
/ ¼ G_ out þ lj ðn_ i;out  n_ i;out ÞAij ð16Þ
j¼1 i¼1
ð9Þ
writes as
@/ g 0i;out Xk
_
¼ þ lnðni;out pout Þ þ lj Aij for i ¼ 1; k ð17Þ
@ ni;out RT j¼1
At the minimum each of these is then zero. The additional equa-
ð10Þ tion to make the system consistent is the summation of molar frac-
tions of the outlet gas to be
X
k
ni;out ¼ 1 ð18Þ
i¼1
3. Plant configurations
The first configuration studied is shown in Fig. 1a. The fuel (NG)
is preheated in a heat exchanger before it is sent to a desulphuri-
zation unit to remove the sulphur content in the NG. This unit is
assumed to be using a catalyst, operating at temperature of
200 °C. Thereafter the heavier carbon contents in the NG are
ð11Þ cracked down in a CPO type pre-reformer catalyst. Before that
the fuel must be preheated again to reach to the operational tem-
perature of the CPO catalyst. The CPO catalyst needs additional air
which is supplied by a small pump as shown in the figure. It is as-
sumed that the supplied NG is pressurized and therefore no pump
is needed for the fuel. The pre-reformed fuel is now sent to the an-
ode side of the SOFC stacks. Due to the exothermic nature of the
CPO catalyst, no pre-heating of the fuel is required. The fuel has
ð12Þ a temperature of about 650 °C before entering the stacks. The oper-
ating temperature of the SOFC stacks as well as outlet tempera-
the stacks has a temperature of about 780 °C which is used to pre-
heat the fuel in the heat exchangers mentioned above. On the other
side, air is compressed in a compressor and then preheated in a
recuperator to about 600 °C before entering the cathode side of
the SOFC stacks. It should be mentioned that these entering tem-
peratures are essential requirements for proper functioning of
SOFC stacks. Otherwise, the stacks will shut down automatically.
Since the fuel in the SOFC stacks will not burn completely the rest
of the fuel together with the air coming out of the cathode side of
the stacks are sent to a burner (catalytic burner) for further burn-
ing. The off-gases from the burner have a high heat quality which
can be used to generate steam in a heat recovery steam generator
(HRSG). In the Rankine cycle, the pressurized water after the feed
ð14Þ pump is heated up to steam in the economizer (ECO), evaporator
(EVA) and super-heater (SUP). The generated steam is then
expanded in a steam turbine to generate electricity. Part of the
expanded steam is then extracted for the deaerator. The expanded
steam after the ST is then cooled down in a condenser (including a
sub-cooler part) before pumping to the deaerator. The water is pre-
heated in the deaerator before being recycled.
As mentioned earlier the ASR reformer needs super heated
ð15Þ externally during startup. However, during normal operation
steam is available after SOFC stacks due to reactions of hydrogen
 M. Rokni / Energy Conversion and Management 51 (2010) 2724–2732 2727

(a)

SOFC Fuel
Prereformer Desuplhurizer

HEX HEX
HEX

Burner

Air

Air

ECO EVA SUP

ST

Condenser
Deaerator

(b)
Recycle
Fuel
SOFC

HEX HEX HEX

HEX

Prereformer Desuplhurizer
Burner

Air

ECO EVA SUP

ST

Condenser
Deaerator

Fig. 1. Combined SOFC–ST cycle plants, (a) with CPO reformer and (b) with ASR reformer. Rankine cycle with a single pressure level.

and oxygen. Therefore, the stream after the anode side of SOFC is
recycled as shown in Fig. 1b. There are three alternatives for such
a recirculation unit, a pump, a turbocharger and an ejector. In a real
plant, due to high temperature of this stream (more than 700 °C)
the cost of a pump will be rather expensive. This is also true for
a turbocharger working at such mass flows and pressures. More-
over, using an ejector brings up problems associated with the size
and dimensioning of the ejector (due to combination of pressure
drop and mass flows), which is out of the scope of this investiga-
tion. In fact, the author is studying the possibilities of including
an ejector into the DNA program which takes into account the size
of ejectors and nuzzle area. Based on these facts and for the sake of
simplicity a pump is used in this investigation. Besides the recycle
part, the plant configuration is the same as in the CPO reformer
case, see Fig. 1b.
The main parameters for the plant are given in Table 1. Number
of SOFC stacks is assumed to be 10,000 and number of cells per
stacks is assumed to be 74. The pressure drops in the cathode
and anode sides of the fuel cell are assumed to be 0.05 respective
0.01 bars. These values are the setting values for the program,
however, pressure drops are a function of channel sizes and mass
flows and the channel geometry is not kwon. Therefore, these
2728 M. Rokni / Energy Conversion and Management 51 (2010) 2724–2732

Table 1 Table 2
Main parameters for design point calculations of Fig. 1. Main calculated parameters for design point calculations of Fig. 1.

Parameter CPO ASR Parameter CPO ASR

Compressor intake temperature 25 °C 25 °C Compressor mass flow 51.4 kg/s 59.0 kg/s
Compressor isentropic efficiency 0.85 0.85 Fuel mass flow 1.32 kg/s 1.30 kg/s
Compressor mechanical efficiency 0.95 0.95 SOFC anode mass flow 1.45 kg/s 1.72 kg/s
SOFC cathode inlet temperature 600 °C 600 °C Burner inlet fuel mass flow 5.17 kg/s 5.13 kg/s
SOFC cathode outlet temperature 780 °C 780 °C Burner inlet air mass flow 47.7 kg/s 55.11 kg/s
SOFC utilization factor 0.80 0.80 HRSG gas side inlet temperature 526.3 °C 455.6 °C
SOFC number of cells 74 74 HRSG gas side outlet temperature 218.2 °C 232.0 °C
SOFC number of stacks 10,000 10,000 HRSG water side inlet temperature 121.1 °C 121.0 °C
SOFC cathode side pressure drop ratio 0.05 bar 0.05 bar ST inlet steam temperature 496.3 °C 425.6 °C
SOFC anode side pressure drop ratio 0.01 bar 0.01 bar ST pressure 69.97 bar 59.97 bar
Heat exchangers pressure drops 0.01 bar 0.01 bar ST inlet mass flow 6.33 kg/s 5.47 kg/s
Fuel inlet temperature 25 °C 25 °C Extraction mass flow 5.40 kg/s 4.64 kg/s
Desulphurizer operation temperature 200 °C 200 °C
SOFC anode inlet temperature 650 °C 650 °C
SOFC anode outlet temperature 780 °C 780 °C
Burner efficiency 0.97 0.97 Table 3
ST isentropic efficiency 0.9 0.9 Net powers and efficiencies of the plants for Fig. 1.
Feed water pump pressure 70 bar 60 bar
Parameter CPO ASR
Extraction pressure 2 bar 2 bar
Extraction steam temperature 120.2 °C 120.2 °C Net power output 38.03 MW 36.72 MW
Generators efficiency 0.97 0.97 Output power from SOFC cycle 31.04 MW 31.23 MW
Output power from ST cycle 6.99 MW 5.49 MW
Thermal efficiency of steam cycle (LHV) 0.382 0.368
Thermal efficiency of SOFC cycle (LHV) 0.514 0.524
values are calculated based on the available data for each channel Thermal efficiency (LHV) 0.630 0.617
mass flow and dimensions. In addition, the calculations show that
the final values in terms of plant power and efficacy do not change
significantly if these values are changed slightly. culated to be about 1% point higher than the SOFC plant with CPO
The pressure drops for all heat exchangers are assumed to pre-reforming. This is due to the fact that CPO pre-reforming needs
0.01 bars. Again these values depend on the heat exchanger types additional air mass flow which must be provided by additional
and channel geometries, but they are reasonable for plate heat ex- compressor flow (or additional compressor) which in turn needs
changer types. The minimum temperature differences at pinch additional electrical power. The cycle efficiencies are defined as
(Tmin) are assumed to be 30 °C, 15 °C and 17 °C for superheat, evap-
orator and economizer respectively. The compressor isentropic effi- gSOFC;a ¼ Net power form SOFC cycle=fuel consumption
ciency and mechanical efficiency are assumed to be 0.85 respective gST ¼ Net power form steam cycle=heat into steam
0.97, while the generator efficiency is assumed to be 0.97. The SOFC ð19Þ
cycle through HRSG
plant provides direct current and must be converted to AC through
a converter. Further, the efficiency of the DC/AC converter is as- gHybrid ¼ Net power form hybrid plant=fuel consumption
sumed to be 100%, due to lack of such component in the DNA pro-
Some of the compressor effect is used to provide the pressure
gram. In reality there would be some losses trough the converter
drops in the heat exchangers of the HRSG. Therefore the compres-
and efficiencies of 97% could be assumed for plants of such sizes
sor effect has been divided into two part; one part which is intent
studied in this investigation. The output power would therefore
for the ST cycle only and the rest which applied to the SOFC cycle
be somewhat lower while the plant efficiency would not change
only.
significantly. Furthermore, the utilization factor for the SOFC cells
It can be shown that the thermal efficiency of the integrated
is assumed to be 0.8. Several calculations have been carried out to
systems can be calculated by
find the optimal extraction pressure as well as the optimum live
steam pressure which are not included here. gth ¼ gSOFC þ gST ð1  gSOFC ÞeHRSG ð20Þ
The main calculated parameters are provided in Table 2. The
main difference in the Rankine cycle from CPO and ASR are the T HRSG;in  T HRSG;out
off-gases temperatures entering the HRSG. The CPO pre-reforming eHRSG ¼ ð21Þ
T HRSG;in  T air
plant provides much higher temperature for the off-gases, which is
due to the fact that the air mass flow to its burner is lower than the where gSOFC and gST are the SOFC cycle efficiency and steam cycle
corresponding one for the ASR pre-reforming plant. Consequently, efficiency respectively. HRSG is the effectiveness of the heat recov-
the feeding temperature for the Rankine cycle will be higher in the ery steam generator, which describes the amount of heat that has
CPO type, which will results in higher net power from the steam been transferred from the SOFC cycle to the steam cycle. These
cycle, see Table 3. equations provide HRSG efficiencies 0.615 and 0.519 for CPO
The plants net powers and thermal efficiencies (based on LHV, respective ARS. Thermal efficiencies can then be calculated as
lower heating value) are shown in Table 3. The net power output 0.628 and 0.615 for CPO respective ASR. These values are in agree-
of the hybrid plant with CPO pre-reforming is calculated to be ment with the numerical results given in Table 3.
about 1 MW higher than the corresponding plant with the ARS In order to optimize the systems presented above, the plants
pre-reforming. The main reason is that the steam cycle in the efficiencies versus live steam pressures and moisture contents are
CPO type plant produces somewhat higher power than the corre- presented in Fig. 2. The results indicate that for the CPO case the
sponding ASR type plant (higher off-gases temperature for the maximum efficiency appears to be at about 45 bar while for the
CPO pre-reforming). As a result the thermal efficiency of the plant ASR case this maxima is around 22 bar. At optimal live steam pres-
with CPO pre-reforming will be higher than the corresponding sure the efficiency of the CPO pre-reformer case is higher than the
plant with ASR pre-reforming. However, the thermal efficiency of corresponding optimal live steam pressure for the ASR pre-refor-
the SOFC plant (not the hybrid one) with ASR pre-reforming is cal- mer case. The moisture content after the steam turbine is an
 M. Rokni / Energy Conversion and Management 51 (2010) 2724–2732 2729

64 20 64 20

63.2 16 63.2 16
Thermal Efficiency [%]

Thermal Efficiency [%]

Moisture content [%}]

Moisture content [%]

62.4 12 62.4 12

61.6 8 61.6 8

60.8 4 60.8 moisture 4

moisture
ηth
ηth

60 0 60 0
0 10 20 30 40 50 60 70 80 0 10 20 30 40 50 60 70 80
Live Steam Pressure [bar] Live Steam Pressure [bar]

(a) (b)
Fig. 2. Thermal efficiency and moisture content of the combined SOFC–ST plants as function of live steam pressure: (a) with CPO reformer and (b) with ASR reformer. Rankine
cycle with single pressure level.

important issue to be considered. Too high level of moisture (more for the ARS case the thermal efficiency is decreased to 0.664 from
than about 16%) may cause blades corrosion located at the last 0.677. However, these efficiencies are still much higher compared
stage, see e.g. Kehlhofer et al. [20]. Further, the results indicate that to the traditional plants for such sizes. Further, the drop in effi-
for the CPO pre-reforming plant the moisture content more than ciency for the ASR pre-reforming case is more sensible than the
16% may occur for pressures above 75 bar, while for the ASR pre- CPO case, the lines are getting wider.
reforming plant this is true for pressures above 48 bar. Although, It can be shown by a simple mathematic that the efficiency of
both are much higher than the optimized pressures mentioned any hybrid plants with hybrid recuperating concept (using dissi-
above. pated heat from the bottoming cycle to the topping cycle) can be
calculated from
3.1. Effects of cathode air pre-heating (hybrid recuperating)  
Q recuperated
gth ¼ 1 þ ðgSOFC þ gST ð1  gSOFC ÞeHRSG Þ ð22Þ
Q in
The energy left after the HRSG in the previous designs are rather
high; see the outlet temperatures of HRSG in both cases. The rea- where Qrecuperated is the heat recuperated to the SOFC cycle and Qin is
son is that decreasing the live steam pressure decreases the de- the heat supplied to the plant through fuel. Note that the ratio
pleted gas temperature form the HRSG, but also the steam cycle Qrecuperated/Qin in this equation is very small and therefore it is sim-
thermal efficiency and thereby the combined plant efficiency. plified. For the CPO pre-reforming case Qrecuperated/Qin and HRSG are
However, at the given pressures above the combined plants effi- calculated as 0.096 and 0.731 respectively. Inserting these values
ciencies are close to their respective optimum and the depleted into the Eq. (22) together with cycle efficiencies provides plant ther-
gas temperatures are high as well. Therefore, this energy can be mal efficiency as 0.680 which is in agreement with the correspond-
used to preheat the cathode air flow and thereby decrease the en- ing calculated value shown in Table 4. For the ASR pre-reforming
ergy from the SOFC cycle which is used for pre-heating. This will in case Qrecuperated/Qin and HRSG are calculated as 0.122 and 0.680
turn increase the energy of the off-gases through the HRSG and respectively. Inserting these values into Eq. (22) gives plant thermal
thereby increase the efficiency of the steam cycle. Fig. 3 shows efficiency as 0.677 which is again in agreement with the corre-
the suggested plants. The SOFC cycle efficiency will not decrease sponding calculated value shown in Table 4.
significantly by such pre-heating. It may be mentioned that such
a pre-heating cannot be done in a combined cycle (gas turbine 3.2. Effects of SOFC fuel utilization factor
on the top of a steam turbine cycle) since the temperature after
the compressor is too high because of compressor pressure ratio, The SOFC fuel utilization factor in the previous calculations was
usually much higher than the HRSG outlet gas temperature (gas set to 0.8. However, the computed results changes if the fuel utili-
side). zation factor varies. Fig. 5a shows the SOFC cycle efficiency as func-
Several calculations for each pre-reformer case are carried out tion of fuel utilization factor for the four cases considered. In this
to find the optimal thermal efficiency of the systems. The optimal figure and from now on the SOFC cycle efficiency is defined as
pressure for the steam turbine is calculated as 70 bar and 60 bar for (due to recuperation),
respective CPO and ASR pre-reforming cases. The most calculated
important results are shown in Table 4. gSOFC;b ¼ Net power form SOFC cycle=ðfuel consumption
It is assumed that the HRSG outlet gas-side temperature is
þ recuperated heatÞ ð23Þ
90 °C. This is due to the fact that the off-gases are pure enough
to allow to recover the energy to such temperature. However, in Without recuperating the SOFC cycle efficiency will be in-
order to see how the thermal efficiencies of the proposed systems creased by increasing cell utilization factor. However, at some
will change if the outlet gas temperatures from the HRSG is in- point the demanding cooling effect will be very significant and
creased, Fig. 4 is presented. As the results show the efficiencies de- the SOFC cycle efficiency starts to decrease. Therefore, the SOFC cy-
creases by increasing the HRSG outlet temperature. At 120 °C the cle efficiency has maxima at 0.78 and 0.76 with CPO respective ASR
thermal efficiency of the CPO case drops to 0.668 from 0.680, while pre-reforming, if no recuperating is used. At utilization factor of
2730 M. Rokni / Energy Conversion and Management 51 (2010) 2724–2732

(a)

SOFC Fuel
Prereformer Desuplhurizer

HEX HEX
HEX HEX

Burner

Air

Air

ECO EVA SUP

ST

Condenser
Deaerator

(b)
Recycle
Fuel
SOFC

HEX HEX HEX

HEX HEX

Prereformer Desuplhurizer
Burner

Air

ECO EVA SUP

ST

Condenser
Deaerator

Fig. 3. Combined SOFC–ST cycle plants, (a) with CPO reformer and (b) with ASR reformer. Rankine cycle with a single pressure level and cathode pre-heating in the SOFC cy.

about 0.81 the SOFC cycle efficiency decreases sharply, which is
more significant in the ASR pre-reforming than the CPO pre-
reforming type. Further, the SOFC cycle efficiency with ASR pre-
reforming is higher than CPO pre-reforming without recuperating.
With hybrid recuperating, SOFC cycle efficiency increases with
decreasing fuel utilization factor. This can mainly be explained
by Eq. (23) in which the recuperated heat to the SOFC cycle in-
creases mach faster than the changes on SOFC power output. (Fuel
consumption remains constant.) To give an example, if utilization
factor is increased from 0.72 to 0.82 then the recuperated heat will
be about four times larger. The SOFC cycle efficiency is lower with
recuperating than without recuperating which is mainly due to
definition used in Eq. (23).
The steam cycle efficiency decreases with increasing fuel utili-
zation as shown in Fig. 5b. This is because less fuel will be burned
in the burner and consequently HRSG inlet temperature decreases.
The steam plant efficiency for CPO case is higher than the
corresponding one with ASR case, if no recuperating is used. This
is because the SOFC cycle efficiency with ASR pre-reforming is
higher than the corresponding one with CPO pre-reforming. If
 M. Rokni / Energy Conversion and Management 51 (2010) 2724–2732 2731

Table 4 0.685
Net powers and efficiencies of the plants for Fig. 3.

Parameter CPO ASR

0.68
Net power output 41.06 MW 40.34 MW
Output power from SOFC cycle 30.92 MW 31.13 MW

Efficiency
Output power from ST cycle 10.14 MW 9.21 MW 0.675
Depleted gas temperature 90 °C 90 °C
Steam Pressure/temperature 70 bar/569.9 °C 60 bar/532.2 °C
Thermal efficiency of steam cycle (LHV) 0.395 0.381 0.67
Moisture content after ST 11.5% 12.9%
Thermal efficiency of SOFC cycle (LHV) 0.512 0.523 ASR
Thermal efficiency (LHV) 0.680 0.677 0.665 CPO

recuperating is used then the steam plant efficiency is similar with 0.66
90 95 100 105 110 115 120
both pre-reforming types.
The hybrid plant efficiency as function of SOFC fuel utilization HRSG outlet T [C]
factor is shown in Fig. 6a. Decreasing the fuel utilization factor
Fig. 4. Thermal efficiencies as function of HRSG outlet gas temperature.
would result in increasing hybrid plant efficiency, for all cases con-
sidered. This is a consequent of what has been explained earlier. pre-reformer without recuperating than the other three configura-
Decreasing fuel cell utilization factor increases ST inlet temper- tions. Further, if the temperature limit for ST is set to 600 °C then
ature significantly. Such an increase is more sever for the CPO the fuel cell utilization factor cannot be below about 0.74, 0.77,

0.6 0.5

0.45

0.5 0.4
Steam cycle efficiency
SOFC cycle efficiency

0.35

0.4 0.3
CPO - no recuperating
0.25
ASR - no recuperating CPO- no recuperating
0.3 CPO - with recuperating 0.2 ASR - no recuperating
ASR - with recuperating CPO - with recuperating
0.15 ASR - with recuperating

0.2 0.1
0.72 0.74 0.76 0.78 0.8 0.82 0.72 0.74 0.76 0.78 0.8 0.82
SOFC fuel utilization factor SOFC fuel utilization factor

(a) (b)
Fig. 5. Cycle efficiencies as function of SOFC utilization factors for four cases considered: (a) SOFC cycle efficiency, (b) steam cycle efficiency.

0.8

1000
CPO without recuperating
0.7
ASR without recuperating
900
CPO with recuperating
Hybrid plant efficiency

0.6 ASR with recuperating

800

0.5 700
T[C]

600
0.4 CPO - no recuperating
ASR - no recuperating 500
CPO - with recuperating
0.3
ASR - with recuperating 400

0.2 300
0.72 0.74 0.76 0.78 0.8 0.82 0.72 0.75 0.78 0.81
SOFC fuel utilization factor SOFC fuel utilization factor

(a) (b)
Fig. 6. (a) Hybrid plant efficiency as function of fuel cell utilization factors and (b) inlet temperature of the high pressure steam turbine as function of fuel cell utilization
factor.
2732 M. Rokni / Energy Conversion and Management 51 (2010) 2724–2732
0.77 and 0.8 for ARS without recuperating, ASR with recuperating,
CPO without recuperating and CPO with recuperating respectively.
4. Conclusions
Hybrid combined SOFC–ST plants are presented and analyzed.
The plants are fired by natural gas and therefore the fuel is desu-
phurized and then pre-reformed before sending to the anode side
of the SOFC. Both CPO and ARS pre-reforming processes are used.
The results indicate that for simple combinations the thermal effi-
ciencies of the system can reach to about 62–63% depending on the
choice of pre-reforming type. The SOFC cycle efficiency in the ASR
performing type is higher than the corresponding SOFC cycle with
CPO pre-reforming type. However, the efficiency of the hybrid
plants with CPO case is larger than the corresponding hybrid plants
with ARS case. This is due to the fact that in the CPO case the tem-
perature of the off-gases entering the HRSG is larger than the cor-
responding temperature in the ARS case. This results in higher
efficiency in the bottoming cycle (steam cycle) and thereby better
efficiency for the hybrid plants.
In another suggested system configurations the off-gases from
the HRSG is recuperated to the SOFC cycle by pre-heating incoming
air flow after the compressor and before entering the cathode side
of the SOFC. In this way the energy of off-gases in the HRSG can be
further used and thereby the electrical efficiencies of the systems
are increased to about 68% for both CPO and ASR pre-reforming
types. Such cathode air pre-heating does not change the SOFC plant
efficiency but increases the steam cycle efficiency significantly due
to better use of the off-gases energy and lowering the stack
temperature.
Decreasing the SOFC fuel utilization factor increases plant elec-
trical efficiency significantly. However, this parameter cannot be
less than a certain value (due to configuration) since the inlet tem-
perature of the steam turbine would be too large.
References
[1] Fontell E, Kivisaari T, Christiansen N, Hansen J, Pålsson J-B. Conceptional study
of a 250 kW planar SOFC system for CHP application. J Power Sources
2004;131:49–56.
[2] Rokni M. Introduction of a fuel cell into combined cycle: a competitive choice
for future cogeneration. ASME Cogen – Turbo IGTI 2003;8:255–61.
[3] Riensche E, Achenbach E, Froning D, Haines MR, Heidug WK, Lokurlu A, et al.
Clean combined-cycle SOFC power plant–cell modeling and process analysis. J
Power Sources 2000;86:404–10.
[4] Pålsson J, Selimovic A, Sjunnesson L. Combined solid oxide fuel cell and gas
turbine system for efficient power and heat generation. J Power Sources
2000;86:442–8.
[5] Proell T, Rauch R, Aichering C, Hofbauer H. Coupling of biomass steam
gasification and an SOFC–GT hybrid system for highly efficient electricity
generation. In: ASME Turbo Expo Proceeding, GT2004-53900; 2004. p. 103–9.
[6] Subramanyan K, Diwekar UM. Characterization and quantification of
uncertainty in solid oxide fuel cell hybrid power plants. J Power Sources
2005;142:103–16.
[7] Calise F, Dentice d’ Accadia M, Vanoli L, von Spakovsky MR. Single-level
optimization of hybrid SOFC–GT power plant. J Power Sources
2006;159:1169–85.
[8] Wächter C, Lunderstädt R, Joos F. Dynamic model of a pressurized SOFC–GT
hybrid power plant for the development of control concept. J Fuel Cell Sci
Technol 2006;3(3):271–9.
[9] Roberts RA, Brouwer J. Dynamic simulation of a pressurized 220 kW solid
oxide fuel-cell–gas-turbine hybrid system: modeled performance compared to
measured Results. J Fuel Cell Sci Technol 2006;3(1):18–25.
[10] Rokni M, Fontell E, Ylijoki Y, Tiihonen O, Hänninen M. Dynamic modeling of
wärtsilä 5kW SOFC system. In: Singhal SC, Mizusaki J, editors. Ninth
International symposium on solid oxide fuel cell (SOFC IX), Quebec; 2005. p.
865–75.
[11] Costamagna P, Magistri L, Massardo AF. Design and part-load performance of a
hybrid system based on a solid oxide fuel cell reactor and a micro gas turbine. J
Power Sources 2001;96:352–68.
[12] Dunbar WR, Lior N, Gaggioli RA. Combining fuel cells with fuel-fired power
plants for improved exergy efficiency. J Energy 1991;16(10):1259–74.
[13] Elmegaard B, Houbak N. DNA – a general energy system simulation tool,
Proceeding of SIMS, Trondheim, Norway; 2005.
[14] Perstrup C. Analysis of power plant installation based on network theory (in
Danish). M.Sc. thesis, Technical University of Denmark, Laboratory of
Energetics, Denmark; 1989.
[15] Petersen TF, Houbak N, Elmegaard B. A zero-dimensional model of a 2nd
generation planar SOFC with calibrated parameters. Int J Thermodyn
2006;9(4):161–8.
[16] Keegan KM, Khaleel M, Chick LA, Recknagle K, Simner SP. Diebler J. Analysis of
a planar solid oxide fuel cell based automotive auxiliary power unit. SAE
technical paper series No. 2002-01-0413; 2002.
[17] Kim JW, Virkar AV. The effect of anode thickness on the performance of anode-
supported solid oxide fuel cell. In: Proceedings of the sixth international
symposium on SOFCs, (SOFC-VI), PV99-19. The Electrochemical Society; 1999.
p. 830–9.
[18] Smith JM, Van Ness HC, Abbott MM. Introduction to chemical engineering
thermodynamics. 7th ed. Boston: McGraw-Hill; 2005.
[19] Elmegaard B. Houbak N. On the implementation of energy simulators with
emphasis on chemical equilibrium gasifier models. In: Proceedings of ECOS’99.
Tokyo; 1999. p. 258–63. ISBN: 4-9980762-0-5
[20] Kehlhofer RH, Warner J, Nielsen H, Bachmann R. Combined–cycle gas steam
turbine power plants. PennWell; 1999. ISBN: 0-87814-736-5.