- Email: [email protected]
Analysis of solid oxide fuel cell system concepts with anode recycling
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 8 ( 2 0 1 3 ) 6 8 0 9 e6 8 2 0
Available online at www.sciencedirect.com
journal homepage: www.elsevier.com/locate/he
Analysis of solid oxide fuel cell system concepts with anode recycling Roland Peters a,*, Robert Deja a, Ludger Blum a, Jari Pennanen b, Jari Kiviaho b, Tuomas Hakala c a
Forschungszentrum Ju¨lich GmbH, Wilhelm-Johnen-Straße, 52425 Ju¨lich, Germany VTT, Technical Research Centre of Finland, Biologinkuja 5, FIN-02044 Espoo, Finland c Wa¨rtsila¨ Finland Oy, Tekniikantie 12, FIN-02150 Espoo, Finland b
article info
abstract
Article history:
The main drivers for anode recirculation are the increased fuel efficiency and the inde-
Received 26 October 2012
pendence of the external water supply for the fuel pre-reforming process. Different sys-
Received in revised form
tem flow-schemes have been defined and a set of parameters have been elaborated as
1 March 2013
basis for various calculations. Taking into account the combinations of layout, cell type,
Accepted 21 March 2013
fuel utilization, fuel, and recycling ratio the total number of cases modeled was about 220.
Available online 19 April 2013
All calculated SOFC systems are on a high level of electrical net efficiency in the range of 50e66%. The electrical and thermal efficiencies are mainly influenced by the fuel utili-
Keywords:
zation. The layout itself, the choice of fuel gas or the type of cell have minor effects on the
Fuel cells
system efficiency.
SOFC systems
Copyright ª 2013, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved.
Anode recycling Efficiencies
1.
Introduction
A research consortium is formed for Development of Anode Sub-System for small-scale and large-scale SOFC systems. The consortium comprises several European companies: VTT Technical Research Centre of Finland, HTceramix SA, EBZ Entwicklungs- und Vertriebsgesellschaft Brennstoffzelle mbH, Wa¨rtsila¨ Finland Oy, Hexis AG and Forschungszentrum Ju¨lich GmbH. The high temperature fuel cell technologies have potential for high electrical efficiency up to 60% [1], and for total efficiency higher than 90% [2,3] based on the lower heating value. SOFC has the added benefit of offering commercial applications from 1 kW residential to several MW stationary units with high fuel flexibility. Whilst much effort is devoted to cell and stack issues, less attention has been paid to the
components and sub-systems required for an operational system. The consortium is focused on the development of fuel and water management for SOFC systems. The fuel management, and especially recirculation, is a key question in achieving high electrical efficiency and rejecting external water supply. The recirculation increases the fuel utilization rate and can provide the water needed for the reforming of fuels [4].
2.
Scientific approach
SOFC applications are working in a wide range of power. Small-scale SOFC systems are often addressed to electrical power of 1e5 kW. Currently a lot of development activities are carried out at this power level [1e3,5,6]. For Large-scale SOFC
* Corresponding author. Tel.: þ49 2461 614664; fax: þ49 2461 616695. E-mail address: [email protected] (Ro. Peters). 0360-3199/$ e see front matter Copyright ª 2013, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.ijhydene.2013.03.110
6810
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 8 ( 2 0 1 3 ) 6 8 0 9 e6 8 2 0
Abbreviations AC DC LHV SOFC O/C UF UFsystem RR NG BG ESC ASC P U R I
alternating current direct current lower heating value solid oxide fuel cell oxygen to carbon ratio stack fuel utilization system fuel utilization recycling ratio natural gas biogas electrolyte supported cells anode supported cells power voltage resistance current
systems with a power range of some 10 kW up to several hundred kW the situation is different. Especially for planar cells currently only a few development activities are pursued in this power level [7,8]. Within these developments planar electrolyte supported cells (ESC) as well as anode supported cells (ASC) are used. Concerning the anode subsystems several publications have been presented previously. A detailed analysis of hydrogen and methane fueled 5 kW planar SOFC system with anode gas recycling of 60% using a fuel driven ejector is discussed in Ref. [9]. The transient behavior of a 300 kW hybrid system based on the coupling of a recuperated micro-gas turbine with a tubular SOFC using a fuel driven ejector for the anode gas recycling is discussed in Ref. [10]. An analysis of an ethanol fueled SOFC systems with anode gas recycling ratios between 40 and 70% is mentioned in Ref. [11]. Experimental and simulation results of a tubular 5 kW SOFC system having an ejector driven recycling loop using ethanol with 52% recycling are discussed in Ref. [12]. Results of a methane fueled 2 kW SOFC system with anode gas recycling ratios between 83 and 90% in which the anode gas recirculation loop is driven by a blower operating below 200 C are presented in Ref. [13]. Experimental results of a 300 W SOFC system using propane with external reforming and a recycling ratio of about 35% are presented in Ref. [14]. Within the presented work different concepts of anode recycling loops are investigated concerning complexity and efficiency. The feasibility of the different recycling solutions for anode subsystems is governed by several general aspects: the choice of fuel, the reforming technology, the performance of the SOFC stack, the carbon formation limits and the accumulated pressure losses of all system components. The anode subsystems concepts of main interest are: No anode off-gas recycling, with condensing of water (Type A) Blower-based approach, with or without condensing of water (Type B) Ejector-based approach with or without condensing of water (Type C)
i N A T p r cp k Dp Q_ _ m hisen hmech hel hinv hth htot
current density number area temperature pressure density heat capacity heat capacity ratio pressure drop heat flux mass flow isentropic efficiency mechanical efficiency electrical efficiency inverter efficiency thermal efficiency total efficiency
For small-scale systems an electrical net power output of 3 kW was set and the system uses stacks with ESC. For largescale systems an electrical net power output of 250 kW was fixed and stacks operating with ASC are used. The development of SOFC systems for stationary application is mainly focusing on natural gas or other methane rich gases like biogas as fuel. Therefore four types of gases were defined for the calculation (Table 1).
2.1.
Component and system modeling
The stack model considers electrochemical and chemical reactions needed to determine gas compositions, voltage, power and mass flows. The model converts methane and higher hydrocarbons into a hydrogen rich gas using steam. CH4 þ H2 O4CO þ 3H2
(1)
C2 H6 þ 2H2 O42CO þ 5H2
(2)
C3 H8 þ 3H2 O43CO þ 7H2
(3)
Hydrogen is directly converted in the electrochemical reaction. Carbon monoxide is converted into hydrogen in the shift reaction. CO þ H2 O4CO2 þ H2
(4)
Table 1 e Definition of fuel gases. Fuel type
Natural gas 1 (NG 1) Natural gas 2 (NG 2) Biogas 1 (BG 1) Biogas 2 (NG 2)
Fuel composition/mol-% CH4
C2H6
C3H8
N2
CO2
98.0 85.0 65.0 40.0
0.8 6.0 e e
0.2 3.0 e e
0.9 5.0 e 30.0
0.1 1.0 35.0 30.0
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 8 ( 2 0 1 3 ) 6 8 0 9 e6 8 2 0
The formation of solid carbon is calculated by using the Boudouard reaction. 2CO4CO2 þ C
(5)
It is assumed that the reactions above will always result in equilibrium at reaction temperature and pressure. The DC power of the stack is calculated using Eq. (6): PDC Stack ¼ Ucell $Ncell $Icell
(6)
The Cell voltage is determined by using the Nernst voltage considering the cell resistance: Ucell ¼ ðUNernst Rcell $Icell Þ
(7)
with: Icell ¼ i$Acell
(8)
The fuel mass flow is calculated by Faraday’s law, taken into account current and gas composition. The required air mass flow is calculated by the energy balance. Like the stack, the reformer converts methane and higher hydrocarbons into a hydrogen rich gas. In case of a heated reformer the chemical equilibrium composition is calculated for a given reaction temperature and pressure. The reformate outlet temperature is equal to the reaction temperature. The outlet temperature of the heating gas is calculated by the energy balance of the reformer. For an adiabatic reformer the reaction temperature is equal to the outlet temperature of the reformer which again results from the energy balance. The power consumption of the blower is calculated considering mass flow, gas composition, pressure difference and temperatures. The theoretical outlet temperature is calculated using Eq. (9): Tout ideal ¼ Tin
k1 pout k pin
(9)
The isentropic efficiency is defined in Eq. (10): hisen ¼
cpout ideal $Tout ideal cpin $Tin cpout real $Tout real cpin $Tin
(10)
Assuming a constant heat capacity the outlet temperature of the blower is calculated using Eq. (11): Tout real ¼
Tout ideal Tin þ Tin hisen
_ m$cp$ðT out real Tin Þ hmech
_ pout pin m$ r$hmech
hel ¼
PAC net _ fuel in $LHVfuel in m
(14)
The electric net AC output is the electric DC power output of the SOFC stack multiplied with the inverter efficiency of which the power consumption of rotary system components (e.g. blower, pump) is then subtracted. Other additional power consumption is not considered. - Electric net AC power: PAC net ¼ PDC Stack $hinv
X
PBlower; Pump
(15)
The thermal efficiency is determined relating the heat output given to the heat sink to the reaction enthalpy of the fuel entering the system. The heat taken out of the system strongly depends on the temperature of the flue gas leaving the heat sink and that is again depending on the application. However, within all calculations the flue gas outlet temperature is set to 40 C. - Thermal efficiency: hth ¼
Q_ heat sink _ fuel in $LHVfuel in m
(16)
- Total efficiency: htot ¼ hel þ hth
(12)
The power consumption of the pump is calculated with Eq. (13): PPump ¼
the latent heat coming from condensation of steam or evaporation of water, including superheating of steam. In case of condensation a saturated gas flow at outlet temperature is assumed and the amount of condensed water is calculated. In the off-gas burner a known air and fuel flow react with each other. The burnable components are converted to steam and carbon dioxide. The composition of the flue gas is determined by the chemical equilibrium at outlet conditions. The heat released by the reaction increases the temperature of the flue gas. The different efficiencies are defined in Eq. (14)e(17). They are always related to the lower heating value (LHV). The electrical efficiency can be determined by relating the electric net AC output to the reaction enthalpy of the fuel entering the system. - Electrical efficiency:
(17)
(11)
The power consumption of the blower is calculated using Eq. (12): PBlower ¼
6811
(13)
For the heat exchanger calculation two mass flows and three temperatures must be known or must be determined at a preceding component. The fourth temperature is calculated based on the energy balance. The model considers also
The total efficiency is the sum of electrical and thermal efficiency. The detailed definitions for these efficiencies are described in detail in Refs. [15,16]. The electrical efficiency is influenced by the power consumption of the rotary equipment. This effect increases with flow rates and pressure drop the blower has to overcome. The system and component models are calculated based on energy balance and do not consider any dimension, like channel geometry, diameter and lengths of pipes etc. To implement the pressure drop caused by the components a design pressure drop for a design mass flow was set for the air side. It was assumed that 80% of the pressure drop is caused by components with a linear dependency of the flow rate (laminar flow, e.g. stack, reformer etc.) and 20% by a quadratic dependency (turbulent flow, e.g. pipes, branches etc.). The flow pass of the
6812
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 8 ( 2 0 1 3 ) 6 8 0 9 e6 8 2 0
Fig. 1 e Pressure drop characteristic for the fuel and air side.
fuel side contains mainly components with small flow channels and mainly laminar flow regime. Therefore 100% linear dependency of the flow rate was chosen. Design mass flow, pressure drop and pressure drop characteristic are depicted in Fig. 1. To achieve realistic and comparable results further parameters had to be set for the calculation, as presented in Table 2.
The calculations were performed under steady-state conditions without heat losses using two software codes. CycleTempo is a flow-sheet program for thermodynamical modeling and optimization of energy conversion systems developed by Delft University of Technology [17]. ProSofc is an in-house developed MATLAB/SIMULINK-based code by VTT Technical Research Centre of Finland which provides application oriented model libraries for steady-state fuel cell
Table 2 e Definition of system operation parameters.
System nominal power output Cell type Stack operation temperature Active cell area Cells per system Current density at rated conditions ASR at stack operation temperature Stack inlet temperature fuel Stack outlet temperature fuel Stack inlet temperature air Stack outlet temperature air Stack temperature System inlet temperature fuel System inlet temperature air System outlet temperature flue gas Fuel utilization stack Minimum air utilization stack Recycling ratio of anode off-gas Pre-reformer type Pre-reformer equilibrium and outlet temperature Efficiency of blower and pump Efficiency of DC/AC inverter
Small-scale systems
Large-scale systems
3 kW Electrolyte supported (ESC) 850 C 100 cm2 180 250 mA cm2 400 mOhm cm2 750 C 850 C 750 C 850 C 850 C
250 kW Anode supported (ASC) 750 C 360 cm2 2500 400 mA cm2 250 mOhm cm2 650 C 725 C 650 C 725 C 725 C
25 C 25 C 40 C 60/70/80% 50% 50/60/70% Steam reformer (adiabatic or heated), at biogas operation no pre-reforming Heated pre-reformer ¼ 500 C adiabatic pre-reformer ¼ according to energy balance Isentropic efficiency ¼ 60% mechanical efficiency ¼ 50% 95%
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 8 ( 2 0 1 3 ) 6 8 0 9 e6 8 2 0
systems [18]. The equations to calculate the ejector parameter (e.g. pressure drop through the ejector, entrainment ratio) are taken from Ref. [19].
3.
System layout and parameters
To study the behavior of the different anode gas subsystems, the consortium agreed on six different simplified system flowschemes. One with water condensation and without anode gas recirculation (Type A1, reference case), three concepts with a blower based anode gas recirculation loop (Type B) and two cases with an ejector based anode gas recirculation loop (Type C). All systems are generally able to run in a water autarkic mode which means a sufficient amount of water is recycled with or without condensation of water.
3.1.
System layout A 1
Layout A1 (Fig. 2) works without an anode gas recycling loop. A heat exchanger on the fuel side is not needed, because the heated reformer is able to heat up the fuel mixture to the stack inlet temperature. The steam supply for the reforming process depends on the amount of condensed water. In this layout the system fuel utilization is limited to the stack fuel utilization.
3.2.
6813
System layout B 1
Layout B1 (Fig. 3) has an anode recycling loop. The anode off gas is cooled in two steps by the incoming fuel gas mixture. The recirculation blower has to handle inlet temperatures of the recycling gas up to 600 C. The fuel mixture leaving the adiabatic pre-reformer has to be heated up to the stack inlet temperature with an additional heat exchanger. Due to the low inlet temperature of air and fuel a catalytic off-gas burner has to be used.
3.3.
System layout B 2
Layout B2 (Fig. 4) has an anode recycling loop. The anode off gas is cooled to 200 C in two steps by the incoming fuel gas mixture and by the incoming air. The anode off gas is mixed in front of the blower with the fuel entering the system and therefore the recirculation blower always works at inlet temperatures below 200 C.
3.4.
System layout B 3
Layout B3 (Fig. 5) is based on layout B2. Additionally the anode off gas is cooled by water, to condense steam in the anode off gas. The recirculation blower works with inlet temperatures
Fig. 2 e Flow scheme of system layout A 1.
Fig. 3 e Flow scheme of system layout B 1.
6814
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 8 ( 2 0 1 3 ) 6 8 0 9 e6 8 2 0
Fig. 4 e Flow scheme of system layout B 2.
below 60 C. The steam supply for the reforming process depends on the amount of water condensed out of the anode off gas.
3.5.
For the six system layouts described above a calculation matrix was defined. To keep the effort for calculation on a manageable level a selection of possible variants was made (Table 3).
System layout C 1
Layout C1 (Fig. 6) is a variant of layout B1. Instead of the blower a steam driven ejector is used. Like layout B3 the anode off gas is cooled by water below 60 C. This water is pumped to a steam generator and this steam is used as driving force for the ejector.
4.
With the operation parameters and the calculation matrix described above in total 220 different system variants were calculated.
4.1. 3.6.
Results and discussion
Efficiencies of layouts
System layout C 2
Layout C2 (Fig. 7) is also a variant of layout B1. Instead of the steam driven ejector a fuel driven ejector is used. The fuel blower works with ambient temperature at the blower inlet. There is no condensation in the anode gas loop needed.
As an example the results of layout B2 with NG 1 as fuel (Fig. 8) are shown and discussed in detail. The calculated electrical efficiencies are on the desired high level for all cases. For small-scale systems the electrical efficiency increases with fuel utilization in the stack (UF) and recycling ratio (RR) except
Fig. 5 e Flow scheme of system layout B 3.
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 8 ( 2 0 1 3 ) 6 8 0 9 e6 8 2 0
6815
Fig. 6 e Flow scheme of system layout C 1.
for UF 80% and RR 70%. The positive effect of rising UF and RR on the electrical efficiency is related to the decreasing fuel mass flow entering the system for the same power production, due to the higher fuel utilization in the system. System fuel utilization [16]: UFsystem ¼
UF 1 RR$ð1 UFÞ
(18)
Vice versa the air flow increases due to reduced methane flow entering the system and the reduced cooling effect of the internal reforming process. This effect is enhanced by the higher heat production due to the decreasing cell voltage. The drop of the cell voltage is related to a lower Nernst voltage caused by the dilution of the fuel. Therefore the efficiency increases less intensively for higher UF and RR or may even decrease, as for UF 80% and RR 70%. For large-scale systems, this effect is more pronounced. For UF 60% the electrical efficiency increases for all RR. For UF 70
and 80% the electrical efficiency always decreases with RR. In addition to the above-discussed effects the chosen cathode temperature differences (100 C for small-scale and 75 C for large-scale systems) increase the air flow to the detriment of the large-scale systems. Finally the increased power consumption of the air blower due to the higher design pressure drop has additionally a negative influence on the electrical efficiency. The pressure drop and air factors are shown in Fig. 9. A part of blower power consumption leads to an increase of the gas temperature and this can be recovered in the heat sink. This leads to a direct shift from electrical efficiency to thermal efficiency. The other part is lost for the system. Due to the declining fuel flow less water vapor is in the flue gas. According to the fixed flue gas temperature at the heat sink outlet less water is condensed and so less heat is recovered. Both effects lead to a lower thermal efficiency and a lower total efficiency.
Fig. 7 e Flow scheme of system layout C 2.
6816
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 8 ( 2 0 1 3 ) 6 8 0 9 e6 8 2 0
Table 3 e Calculation matrix.
Lay-out A 1 No anode gas recycling Lay-out B 1 Fuel blower (hot) Lay-out B 2 Fuel blower (<200 C) Lay-out B 3 Fuel blower (<60 C) Lay-out C 1 Ejector, steam driven Lay-out C 2 Ejector, fuel driven
Fuel type
Power
Cell type
Water condens.
Calculated variants
NG 1, NG 2
3 kW, 250 kW
ESC, ASC
e
80%
Heated
Yes
4
NG 1, NG 2, BG1, BG2 NG 1, NG 2, BG1, BG2 NG 1, NG 2
3 kW, 250 kW
ESC, ASC
50/60/70%
60/70/80%
Not heated
No
54
3 kW, 250 kW
ESC, ASC
50/60/70%
60/70/80%
Heated
No
72
3 kW
ESC
50/60/70%
60/70/80%
Heated
Yes
18
NG 1, NG 2, BG 1, BG 2 NG 1, NG 2
250 kW
ASC
50/60/70%
60/70/80%
Not heated
Yes
36
3 kW, 250 kW
ESC, ASC
50/60/70%
60/70/80%
Not heated
No
36
For RR 50% and UF 60% the total efficiency reaches values higher than one. The efficiencies within these calculations are related to the lower heating value, which means the latent heat of the steam in the flue gas is not considered in the energy input into the system. If a significant amount of water is condensed, it is possible to get values higher than one. For layout B2 using NG1 the optimal operation range to reach high electrical efficiencies is depicted in Fig. 10. Independent of the system scale for RR 50 and 60% with all UF carbon formation in the pre-reformer occurs and only RR 70% can be chosen as a suitable recycling ratio. With RR 70% for small-scale systems the efficiency increases always with UF and electrical efficiencies higher than 60% are achieved for UF 70 and 80%. For large-scale systems an electrical efficiency of nearly 60% is reached for UF 60%. As discussed above this is because of the higher power consumption of the blower and the higher design pressure drop. Therefore systems with low parasitic power consumption should operate with UF between 70 and 80% and systems with high power consumption between 60 and 70%. Te results of all calculations are presented in Figs. 11 and 12. The Layouts A 1, B 2, B 3 and C 2 are calculated with CycleTempo and B 1 and C 1 are done with ProSofc.
Recycling ratio
Fuel utilization
Pre-reformer
In Figs. 11 and 12 the results of the layout types B and C are arranged in the following way. For each fuel and layout there are nine efficiency points. The first three points from the left are for UF 60%, the following three for UF 70% and the last three points for UF 80%. The points within one UF are RR 50%, 60% and 70%, from left to right. The electrical and thermal efficiencies are mainly influenced by the fuel utilization and recycling ratio as already described in detail for layout B 2 with NG 1. The systems operating with biogas will not have a significant decrease of electrical efficiency at high UF and RR because they are working without a pre-reformer. That is possible because the two biogases contain only traces of higher hydrocarbons. Therefore the methane content at the inlet of the SOFC is higher and the air flow rate will not increase that much as for natural gas systems. Also Layouts with unheated reformer (B 1, C 1 and C 2) are reacting less sensitive on high UF and RR. The equilibrium temperature of the adiabatic reformer is lower (typical between 450 and 460 C) than for the heated reformers (500 C). This has an influence on the methane concentration entering the stack and also on the air flow. Systems using ESC cells are generally reacting less sensitive on the fuel utilization and recycling rates due to the chosen operation parameters.
Fig. 8 e Efficiencies of layout B 2 with NG 1.
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 8 ( 2 0 1 3 ) 6 8 0 9 e6 8 2 0
Fig. 9 e Pressure drop and air factor of layout B 2 with NG 1.
Fig. 10 e Optimal operation range for layout B2 with NG1.
Fig. 11 e Results of layout A 1, B 2, B 3 and C 2 calculated with CycleTempo.
6817
6818
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 8 ( 2 0 1 3 ) 6 8 0 9 e6 8 2 0
Fig. 12 e Results of layout B 1 and C 1 calculated with ProSofc.
4.2.
Feasibility of layouts
The consortium agreed on some restriction of operation parameters, in order to keep the calculated systems on a level, which can be transferred to a technical feasible system. Therefore the following knock-out criteria were defined: The carbon formation based on thermodynamical equilibrium in pre-reformer should be avoided. The terminal temperature difference for gasegas heat exchangers should be at least 50 C and for heated reformers at least 100 C, in order to keep the size and cost of the components in a feasible range. The outlet temperature of the burner should be below 1000 C due to the requirements for the burner material. The air utilization in the stack should be lower than 50%, to avoid a poor cell performance. The knock-out criteria described above were applied to all calculated systems and the results are discussed below. A marked amount of the calculated systems do not fulfill one or more knock-out criteria. The general effects leading to a feasible system are discussed in detail for the layout B 2.
From the thermodynamical point of view often carbon formation according to Eq. (5) occurs in the pre-reformer. Fig. 13 shows the dependency of the carbon formation process from UF, RR and O/C-ratio for an equilibrium temperature of 500 C. With higher UF and RR the risk of carbon formation is reduced. NG 2 needs higher O/C-ratios than NG 1 because of the concentration of higher hydrocarbons. Beyond an O/Cratio of 2.3 no carbon formation is taking place. This point can also be considered as a suggestive criterion, as technology may allow successful, carbon-free operation of the prereformer even below the thermodynamical carbon formation limit as discussed in Ref. [20]. The terminal temperature difference criterion for the air heaters (minimum 50 C) is infringed 14 times always with UF 80% and RR 70%. Due to the higher air flow rates and the lower off-gas burner outlet temperature the terminal temperature difference becomes smaller. The terminal temperature difference criterion for the fuel heaters and pre-reformer is not touched. The maximum outlet temperature criterion of the off-gas burner is infringed 16 times, at UF (60%) and RR (50, 60, 70%). Due to the high methane concentration in the fuel gas
Fig. 13 e Carbon formation in pre-reformer for layout B 2.
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 8 ( 2 0 1 3 ) 6 8 0 9 e6 8 2 0
entering the stack and the cooling effect of the internal reforming process in the stack the air flow rate is very low. At the same time UF is low, that means a high amount of unconverted fuel is entering the off-gas burner. This combination leads to a high off-gas burner outlet temperature. The air utilization criterion for the stack (max. 50%) is infringed six times, always with UF 60% and RR 50%. Also here the air flow rate is very low due to the cooling effect of the internal reforming.
5.
Conclusions
To study the influence of the fuel management on the electrical efficiency and the water supply for the reforming process in total 220 systems calculations were carried out. All calculated systems are on a high level of efficiency: Electrical efficiency :
0:50 to 0:66
Thermal efficiency :
0:26 to 0:49
Total efficiency :
0:76 to 1:06
Systems without anode gas loop have up to 16% less electrical efficiency compared to anode off-gas recycling systems depending on the operation parameters. The electrical and thermal efficiencies are mainly influenced by fuel utilization and anode off-gas recirculation ratio. For systems operating with natural gas the results reveal that for higher values of UF and RR the electrical efficiency can also decrease. This is related to the higher air flow rates and the resulting power consumption of the blower. On the other hand high UF and RR are needed to avoid carbon formation in the reformer. That means an optimal operation range to avoid carbon formation in the pre-reformer and to get a high electrical efficiency requires RR of 70% and UF between 60 and 80%, depending on the parasitic power consumption. The layout itself, the choice of fuel gas or the type of cell has a minor effect on system efficiency. That means other criteria are important to choose the most promising approach, like number of components, complexity of system and so on. Layout A 1 without anode off-gas recirculation has the lowest complexity, but also the lowest electrical efficiencies. Layouts B 1, B 2 and B 3 are all using fuel blowers for the gas recycling loop, resulting in blower operating temperatures up to 600 C (B 1). Layouts C 1 and C 2 are using ejectors in the recycling loop. These components can easily work at high temperatures, but they are not easily to control, especially at part load operation. The fuel blowers and ejectors mentioned above are not available on the shelf and have to be developed. Therefore further more detailed calculations have to be done, with focus on part-load operation and controllability of the systems to elaborate a complete set of design and operation parameters.
Acknowledgment The authors would like to express their thanks to all partners who have contributed to the study. Acknowledgment goes
6819
to the European Commission for co-financing the project ASSENT under the contract number 244821.
references
[1] Fo¨ger K. Commercialisation of CFCL’s residential power Station BluGen. Proceedings, European Fuel Cell Forum Lucerne 2010;2:22e9. [2] Mai A, Sfeir J, Schuler A. Status of SOFC stack and system development at Hexis. Fuel Cell Seminar 2007;12(1). CD 000453. [3] Schuler A, Nerlich V, Doerk T, Mai A. Galileo 1000 N e status of development and operation experiences. Proceedings of 9th European Solid Oxide Fuel Cell Forum 2010;2:98. [4] European Commission. http://cordis.europa.eu/projects/rcn/ 94269_en.html. April 2012. [5] Bertoldi M, Bucheli O, Moden S, Larrain D, Ravagni A. Manufacturing and market-oriented development of SOFC generators at SOFCpower SpA. ECS Transactions 2011;35(1):127e38. [6] Schimanke D, Posdziech O, Mai BE, Kluge S, Strohbach T, Wunderlich C. Demonstration of a highly efficient SOFC system with combined partial oxidation and steam reforming. ECS Transactions 2011;35(1):231e42. [7] Bloomenergy e data-sheet ES-5700 system. http://www. bloomenergy.com/fuel-cell/es-5700-data-sheet; April 2012. [8] Fontell E, Kivisaari T, Christiansen N, Hansen JB, Pa˚lsson J. Conceptual study of a 250 kW planar SOFC system for CHP application. Journal of Power Sources 2004;131:49e56. [9] Kazempoor P, Dorer V, Ommi F. Evaluation of hydrogen and methane-fuelled solid oxide fuel cell systems for residential applications: system design alternative and parameter study. International Journal of Hydrogen Energy 2009;34:8630e44. [10] Ferrari ML, Traverso A, Magistri L, Massardo AF. Influence of the anodic recirculation transient behaviour on the SOFC hybrid system performance. Journal of Power Sources 2005;149:22e32. [11] Saebea D, Patcharavorachot Y, Arpornwichanop A. Analysis of an ethanol-fuelled solid oxide fuel cell system using partial anode exhaust gas recirculation. Journal of Power Sources 2012;208:120e30. [12] Lanzini A, Santarelli M, Orsello G. Residential solid oxide fuel cell generator fuelled by ethanol: cell, stack and system modelling with a preliminary experiment. Fuel Cells 2010;4:654e75. [13] Powell M, Meinhardt K, Sprenkle V, Chick L, McVay G. Demonstration of a highly efficient solid oxide fuel cell power system using adiabatic steam reforming and anode gas recirculation. Journal of Power Sources 2012;205:377e84. [14] Dietrich R-U, Oelze J, Lindermeir A, Spitta C, Steffen M, Ku¨ster T, et al. Efficiency gain of solid oxide fuel cell systems by using anode offgas recycle eResults for a small scale propane driven unit. Journal of Power Sources 2011;196:7152e60. [15] Blum L, Deja R, Peters R, Stolten D. Comparison of efficiencies of low, mean and high temperature fuel cell Systems. International Journal of Hydrogen Energy 2011;36:11056e67. [16] Blum L, Riensche E. Systems. In: Juergen Garche, Chris Dyer, Patrick Moseley, Zempachi Ogumi, David Rand, Bruno Scrosati, editors. Encyclopedia of electrochemical power sources, vol. 3. Amsterdam: Elsevier; 2009. p. 99e119. [17] Hemmes K, Patil A, Woudstra N. Flexible coproduction of hydrogen and power using internal reforming solid oxide fuel cells system. Journal of Fuel Cell Science and Technology 2008;5:041010e1.
6820
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 8 ( 2 0 1 3 ) 6 8 0 9 e6 8 2 0
[18] ProSofc is a MATLAB/SIMULINK-based simulation tool for stationary fuel cell system simulations. The code is developed by VTT, Technical Research Centre of Finland. [19] Zhu Y, Cai W, Li Y, Wen C. Anode gas recirculation behavior of a fuel ejector in hybrid solid oxide fuel cell systems:
performance evaluation in three operational modes. Journal of Power Sources 2008;185:1122e30. [20] Halinen M, Thomann O, Kiviaho J. Effect of anode off-gas recycling on reforming of natural gas for solid oxide fuel cell systems. Fuel Cells 2012;5:754e60.
Available online at www.sciencedirect.com
journal homepage: www.elsevier.com/locate/he
Analysis of solid oxide fuel cell system concepts with anode recycling Roland Peters a,*, Robert Deja a, Ludger Blum a, Jari Pennanen b, Jari Kiviaho b, Tuomas Hakala c a
Forschungszentrum Ju¨lich GmbH, Wilhelm-Johnen-Straße, 52425 Ju¨lich, Germany VTT, Technical Research Centre of Finland, Biologinkuja 5, FIN-02044 Espoo, Finland c Wa¨rtsila¨ Finland Oy, Tekniikantie 12, FIN-02150 Espoo, Finland b
article info
abstract
Article history:
The main drivers for anode recirculation are the increased fuel efficiency and the inde-
Received 26 October 2012
pendence of the external water supply for the fuel pre-reforming process. Different sys-
Received in revised form
tem flow-schemes have been defined and a set of parameters have been elaborated as
1 March 2013
basis for various calculations. Taking into account the combinations of layout, cell type,
Accepted 21 March 2013
fuel utilization, fuel, and recycling ratio the total number of cases modeled was about 220.
Available online 19 April 2013
All calculated SOFC systems are on a high level of electrical net efficiency in the range of 50e66%. The electrical and thermal efficiencies are mainly influenced by the fuel utili-
Keywords:
zation. The layout itself, the choice of fuel gas or the type of cell have minor effects on the
Fuel cells
system efficiency.
SOFC systems
Copyright ª 2013, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved.
Anode recycling Efficiencies
1.
Introduction
A research consortium is formed for Development of Anode Sub-System for small-scale and large-scale SOFC systems. The consortium comprises several European companies: VTT Technical Research Centre of Finland, HTceramix SA, EBZ Entwicklungs- und Vertriebsgesellschaft Brennstoffzelle mbH, Wa¨rtsila¨ Finland Oy, Hexis AG and Forschungszentrum Ju¨lich GmbH. The high temperature fuel cell technologies have potential for high electrical efficiency up to 60% [1], and for total efficiency higher than 90% [2,3] based on the lower heating value. SOFC has the added benefit of offering commercial applications from 1 kW residential to several MW stationary units with high fuel flexibility. Whilst much effort is devoted to cell and stack issues, less attention has been paid to the
components and sub-systems required for an operational system. The consortium is focused on the development of fuel and water management for SOFC systems. The fuel management, and especially recirculation, is a key question in achieving high electrical efficiency and rejecting external water supply. The recirculation increases the fuel utilization rate and can provide the water needed for the reforming of fuels [4].
2.
Scientific approach
SOFC applications are working in a wide range of power. Small-scale SOFC systems are often addressed to electrical power of 1e5 kW. Currently a lot of development activities are carried out at this power level [1e3,5,6]. For Large-scale SOFC
* Corresponding author. Tel.: þ49 2461 614664; fax: þ49 2461 616695. E-mail address: [email protected] (Ro. Peters). 0360-3199/$ e see front matter Copyright ª 2013, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.ijhydene.2013.03.110
6810
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 8 ( 2 0 1 3 ) 6 8 0 9 e6 8 2 0
Abbreviations AC DC LHV SOFC O/C UF UFsystem RR NG BG ESC ASC P U R I
alternating current direct current lower heating value solid oxide fuel cell oxygen to carbon ratio stack fuel utilization system fuel utilization recycling ratio natural gas biogas electrolyte supported cells anode supported cells power voltage resistance current
systems with a power range of some 10 kW up to several hundred kW the situation is different. Especially for planar cells currently only a few development activities are pursued in this power level [7,8]. Within these developments planar electrolyte supported cells (ESC) as well as anode supported cells (ASC) are used. Concerning the anode subsystems several publications have been presented previously. A detailed analysis of hydrogen and methane fueled 5 kW planar SOFC system with anode gas recycling of 60% using a fuel driven ejector is discussed in Ref. [9]. The transient behavior of a 300 kW hybrid system based on the coupling of a recuperated micro-gas turbine with a tubular SOFC using a fuel driven ejector for the anode gas recycling is discussed in Ref. [10]. An analysis of an ethanol fueled SOFC systems with anode gas recycling ratios between 40 and 70% is mentioned in Ref. [11]. Experimental and simulation results of a tubular 5 kW SOFC system having an ejector driven recycling loop using ethanol with 52% recycling are discussed in Ref. [12]. Results of a methane fueled 2 kW SOFC system with anode gas recycling ratios between 83 and 90% in which the anode gas recirculation loop is driven by a blower operating below 200 C are presented in Ref. [13]. Experimental results of a 300 W SOFC system using propane with external reforming and a recycling ratio of about 35% are presented in Ref. [14]. Within the presented work different concepts of anode recycling loops are investigated concerning complexity and efficiency. The feasibility of the different recycling solutions for anode subsystems is governed by several general aspects: the choice of fuel, the reforming technology, the performance of the SOFC stack, the carbon formation limits and the accumulated pressure losses of all system components. The anode subsystems concepts of main interest are: No anode off-gas recycling, with condensing of water (Type A) Blower-based approach, with or without condensing of water (Type B) Ejector-based approach with or without condensing of water (Type C)
i N A T p r cp k Dp Q_ _ m hisen hmech hel hinv hth htot
current density number area temperature pressure density heat capacity heat capacity ratio pressure drop heat flux mass flow isentropic efficiency mechanical efficiency electrical efficiency inverter efficiency thermal efficiency total efficiency
For small-scale systems an electrical net power output of 3 kW was set and the system uses stacks with ESC. For largescale systems an electrical net power output of 250 kW was fixed and stacks operating with ASC are used. The development of SOFC systems for stationary application is mainly focusing on natural gas or other methane rich gases like biogas as fuel. Therefore four types of gases were defined for the calculation (Table 1).
2.1.
Component and system modeling
The stack model considers electrochemical and chemical reactions needed to determine gas compositions, voltage, power and mass flows. The model converts methane and higher hydrocarbons into a hydrogen rich gas using steam. CH4 þ H2 O4CO þ 3H2
(1)
C2 H6 þ 2H2 O42CO þ 5H2
(2)
C3 H8 þ 3H2 O43CO þ 7H2
(3)
Hydrogen is directly converted in the electrochemical reaction. Carbon monoxide is converted into hydrogen in the shift reaction. CO þ H2 O4CO2 þ H2
(4)
Table 1 e Definition of fuel gases. Fuel type
Natural gas 1 (NG 1) Natural gas 2 (NG 2) Biogas 1 (BG 1) Biogas 2 (NG 2)
Fuel composition/mol-% CH4
C2H6
C3H8
N2
CO2
98.0 85.0 65.0 40.0
0.8 6.0 e e
0.2 3.0 e e
0.9 5.0 e 30.0
0.1 1.0 35.0 30.0
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 8 ( 2 0 1 3 ) 6 8 0 9 e6 8 2 0
The formation of solid carbon is calculated by using the Boudouard reaction. 2CO4CO2 þ C
(5)
It is assumed that the reactions above will always result in equilibrium at reaction temperature and pressure. The DC power of the stack is calculated using Eq. (6): PDC Stack ¼ Ucell $Ncell $Icell
(6)
The Cell voltage is determined by using the Nernst voltage considering the cell resistance: Ucell ¼ ðUNernst Rcell $Icell Þ
(7)
with: Icell ¼ i$Acell
(8)
The fuel mass flow is calculated by Faraday’s law, taken into account current and gas composition. The required air mass flow is calculated by the energy balance. Like the stack, the reformer converts methane and higher hydrocarbons into a hydrogen rich gas. In case of a heated reformer the chemical equilibrium composition is calculated for a given reaction temperature and pressure. The reformate outlet temperature is equal to the reaction temperature. The outlet temperature of the heating gas is calculated by the energy balance of the reformer. For an adiabatic reformer the reaction temperature is equal to the outlet temperature of the reformer which again results from the energy balance. The power consumption of the blower is calculated considering mass flow, gas composition, pressure difference and temperatures. The theoretical outlet temperature is calculated using Eq. (9): Tout ideal ¼ Tin
k1 pout k pin
(9)
The isentropic efficiency is defined in Eq. (10): hisen ¼
cpout ideal $Tout ideal cpin $Tin cpout real $Tout real cpin $Tin
(10)
Assuming a constant heat capacity the outlet temperature of the blower is calculated using Eq. (11): Tout real ¼
Tout ideal Tin þ Tin hisen
_ m$cp$ðT out real Tin Þ hmech
_ pout pin m$ r$hmech
hel ¼
PAC net _ fuel in $LHVfuel in m
(14)
The electric net AC output is the electric DC power output of the SOFC stack multiplied with the inverter efficiency of which the power consumption of rotary system components (e.g. blower, pump) is then subtracted. Other additional power consumption is not considered. - Electric net AC power: PAC net ¼ PDC Stack $hinv
X
PBlower; Pump
(15)
The thermal efficiency is determined relating the heat output given to the heat sink to the reaction enthalpy of the fuel entering the system. The heat taken out of the system strongly depends on the temperature of the flue gas leaving the heat sink and that is again depending on the application. However, within all calculations the flue gas outlet temperature is set to 40 C. - Thermal efficiency: hth ¼
Q_ heat sink _ fuel in $LHVfuel in m
(16)
- Total efficiency: htot ¼ hel þ hth
(12)
The power consumption of the pump is calculated with Eq. (13): PPump ¼
the latent heat coming from condensation of steam or evaporation of water, including superheating of steam. In case of condensation a saturated gas flow at outlet temperature is assumed and the amount of condensed water is calculated. In the off-gas burner a known air and fuel flow react with each other. The burnable components are converted to steam and carbon dioxide. The composition of the flue gas is determined by the chemical equilibrium at outlet conditions. The heat released by the reaction increases the temperature of the flue gas. The different efficiencies are defined in Eq. (14)e(17). They are always related to the lower heating value (LHV). The electrical efficiency can be determined by relating the electric net AC output to the reaction enthalpy of the fuel entering the system. - Electrical efficiency:
(17)
(11)
The power consumption of the blower is calculated using Eq. (12): PBlower ¼
6811
(13)
For the heat exchanger calculation two mass flows and three temperatures must be known or must be determined at a preceding component. The fourth temperature is calculated based on the energy balance. The model considers also
The total efficiency is the sum of electrical and thermal efficiency. The detailed definitions for these efficiencies are described in detail in Refs. [15,16]. The electrical efficiency is influenced by the power consumption of the rotary equipment. This effect increases with flow rates and pressure drop the blower has to overcome. The system and component models are calculated based on energy balance and do not consider any dimension, like channel geometry, diameter and lengths of pipes etc. To implement the pressure drop caused by the components a design pressure drop for a design mass flow was set for the air side. It was assumed that 80% of the pressure drop is caused by components with a linear dependency of the flow rate (laminar flow, e.g. stack, reformer etc.) and 20% by a quadratic dependency (turbulent flow, e.g. pipes, branches etc.). The flow pass of the
6812
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 8 ( 2 0 1 3 ) 6 8 0 9 e6 8 2 0
Fig. 1 e Pressure drop characteristic for the fuel and air side.
fuel side contains mainly components with small flow channels and mainly laminar flow regime. Therefore 100% linear dependency of the flow rate was chosen. Design mass flow, pressure drop and pressure drop characteristic are depicted in Fig. 1. To achieve realistic and comparable results further parameters had to be set for the calculation, as presented in Table 2.
The calculations were performed under steady-state conditions without heat losses using two software codes. CycleTempo is a flow-sheet program for thermodynamical modeling and optimization of energy conversion systems developed by Delft University of Technology [17]. ProSofc is an in-house developed MATLAB/SIMULINK-based code by VTT Technical Research Centre of Finland which provides application oriented model libraries for steady-state fuel cell
Table 2 e Definition of system operation parameters.
System nominal power output Cell type Stack operation temperature Active cell area Cells per system Current density at rated conditions ASR at stack operation temperature Stack inlet temperature fuel Stack outlet temperature fuel Stack inlet temperature air Stack outlet temperature air Stack temperature System inlet temperature fuel System inlet temperature air System outlet temperature flue gas Fuel utilization stack Minimum air utilization stack Recycling ratio of anode off-gas Pre-reformer type Pre-reformer equilibrium and outlet temperature Efficiency of blower and pump Efficiency of DC/AC inverter
Small-scale systems
Large-scale systems
3 kW Electrolyte supported (ESC) 850 C 100 cm2 180 250 mA cm2 400 mOhm cm2 750 C 850 C 750 C 850 C 850 C
250 kW Anode supported (ASC) 750 C 360 cm2 2500 400 mA cm2 250 mOhm cm2 650 C 725 C 650 C 725 C 725 C
25 C 25 C 40 C 60/70/80% 50% 50/60/70% Steam reformer (adiabatic or heated), at biogas operation no pre-reforming Heated pre-reformer ¼ 500 C adiabatic pre-reformer ¼ according to energy balance Isentropic efficiency ¼ 60% mechanical efficiency ¼ 50% 95%
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 8 ( 2 0 1 3 ) 6 8 0 9 e6 8 2 0
systems [18]. The equations to calculate the ejector parameter (e.g. pressure drop through the ejector, entrainment ratio) are taken from Ref. [19].
3.
System layout and parameters
To study the behavior of the different anode gas subsystems, the consortium agreed on six different simplified system flowschemes. One with water condensation and without anode gas recirculation (Type A1, reference case), three concepts with a blower based anode gas recirculation loop (Type B) and two cases with an ejector based anode gas recirculation loop (Type C). All systems are generally able to run in a water autarkic mode which means a sufficient amount of water is recycled with or without condensation of water.
3.1.
System layout A 1
Layout A1 (Fig. 2) works without an anode gas recycling loop. A heat exchanger on the fuel side is not needed, because the heated reformer is able to heat up the fuel mixture to the stack inlet temperature. The steam supply for the reforming process depends on the amount of condensed water. In this layout the system fuel utilization is limited to the stack fuel utilization.
3.2.
6813
System layout B 1
Layout B1 (Fig. 3) has an anode recycling loop. The anode off gas is cooled in two steps by the incoming fuel gas mixture. The recirculation blower has to handle inlet temperatures of the recycling gas up to 600 C. The fuel mixture leaving the adiabatic pre-reformer has to be heated up to the stack inlet temperature with an additional heat exchanger. Due to the low inlet temperature of air and fuel a catalytic off-gas burner has to be used.
3.3.
System layout B 2
Layout B2 (Fig. 4) has an anode recycling loop. The anode off gas is cooled to 200 C in two steps by the incoming fuel gas mixture and by the incoming air. The anode off gas is mixed in front of the blower with the fuel entering the system and therefore the recirculation blower always works at inlet temperatures below 200 C.
3.4.
System layout B 3
Layout B3 (Fig. 5) is based on layout B2. Additionally the anode off gas is cooled by water, to condense steam in the anode off gas. The recirculation blower works with inlet temperatures
Fig. 2 e Flow scheme of system layout A 1.
Fig. 3 e Flow scheme of system layout B 1.
6814
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 8 ( 2 0 1 3 ) 6 8 0 9 e6 8 2 0
Fig. 4 e Flow scheme of system layout B 2.
below 60 C. The steam supply for the reforming process depends on the amount of water condensed out of the anode off gas.
3.5.
For the six system layouts described above a calculation matrix was defined. To keep the effort for calculation on a manageable level a selection of possible variants was made (Table 3).
System layout C 1
Layout C1 (Fig. 6) is a variant of layout B1. Instead of the blower a steam driven ejector is used. Like layout B3 the anode off gas is cooled by water below 60 C. This water is pumped to a steam generator and this steam is used as driving force for the ejector.
4.
With the operation parameters and the calculation matrix described above in total 220 different system variants were calculated.
4.1. 3.6.
Results and discussion
Efficiencies of layouts
System layout C 2
Layout C2 (Fig. 7) is also a variant of layout B1. Instead of the steam driven ejector a fuel driven ejector is used. The fuel blower works with ambient temperature at the blower inlet. There is no condensation in the anode gas loop needed.
As an example the results of layout B2 with NG 1 as fuel (Fig. 8) are shown and discussed in detail. The calculated electrical efficiencies are on the desired high level for all cases. For small-scale systems the electrical efficiency increases with fuel utilization in the stack (UF) and recycling ratio (RR) except
Fig. 5 e Flow scheme of system layout B 3.
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 8 ( 2 0 1 3 ) 6 8 0 9 e6 8 2 0
6815
Fig. 6 e Flow scheme of system layout C 1.
for UF 80% and RR 70%. The positive effect of rising UF and RR on the electrical efficiency is related to the decreasing fuel mass flow entering the system for the same power production, due to the higher fuel utilization in the system. System fuel utilization [16]: UFsystem ¼
UF 1 RR$ð1 UFÞ
(18)
Vice versa the air flow increases due to reduced methane flow entering the system and the reduced cooling effect of the internal reforming process. This effect is enhanced by the higher heat production due to the decreasing cell voltage. The drop of the cell voltage is related to a lower Nernst voltage caused by the dilution of the fuel. Therefore the efficiency increases less intensively for higher UF and RR or may even decrease, as for UF 80% and RR 70%. For large-scale systems, this effect is more pronounced. For UF 60% the electrical efficiency increases for all RR. For UF 70
and 80% the electrical efficiency always decreases with RR. In addition to the above-discussed effects the chosen cathode temperature differences (100 C for small-scale and 75 C for large-scale systems) increase the air flow to the detriment of the large-scale systems. Finally the increased power consumption of the air blower due to the higher design pressure drop has additionally a negative influence on the electrical efficiency. The pressure drop and air factors are shown in Fig. 9. A part of blower power consumption leads to an increase of the gas temperature and this can be recovered in the heat sink. This leads to a direct shift from electrical efficiency to thermal efficiency. The other part is lost for the system. Due to the declining fuel flow less water vapor is in the flue gas. According to the fixed flue gas temperature at the heat sink outlet less water is condensed and so less heat is recovered. Both effects lead to a lower thermal efficiency and a lower total efficiency.
Fig. 7 e Flow scheme of system layout C 2.
6816
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 8 ( 2 0 1 3 ) 6 8 0 9 e6 8 2 0
Table 3 e Calculation matrix.
Lay-out A 1 No anode gas recycling Lay-out B 1 Fuel blower (hot) Lay-out B 2 Fuel blower (<200 C) Lay-out B 3 Fuel blower (<60 C) Lay-out C 1 Ejector, steam driven Lay-out C 2 Ejector, fuel driven
Fuel type
Power
Cell type
Water condens.
Calculated variants
NG 1, NG 2
3 kW, 250 kW
ESC, ASC
e
80%
Heated
Yes
4
NG 1, NG 2, BG1, BG2 NG 1, NG 2, BG1, BG2 NG 1, NG 2
3 kW, 250 kW
ESC, ASC
50/60/70%
60/70/80%
Not heated
No
54
3 kW, 250 kW
ESC, ASC
50/60/70%
60/70/80%
Heated
No
72
3 kW
ESC
50/60/70%
60/70/80%
Heated
Yes
18
NG 1, NG 2, BG 1, BG 2 NG 1, NG 2
250 kW
ASC
50/60/70%
60/70/80%
Not heated
Yes
36
3 kW, 250 kW
ESC, ASC
50/60/70%
60/70/80%
Not heated
No
36
For RR 50% and UF 60% the total efficiency reaches values higher than one. The efficiencies within these calculations are related to the lower heating value, which means the latent heat of the steam in the flue gas is not considered in the energy input into the system. If a significant amount of water is condensed, it is possible to get values higher than one. For layout B2 using NG1 the optimal operation range to reach high electrical efficiencies is depicted in Fig. 10. Independent of the system scale for RR 50 and 60% with all UF carbon formation in the pre-reformer occurs and only RR 70% can be chosen as a suitable recycling ratio. With RR 70% for small-scale systems the efficiency increases always with UF and electrical efficiencies higher than 60% are achieved for UF 70 and 80%. For large-scale systems an electrical efficiency of nearly 60% is reached for UF 60%. As discussed above this is because of the higher power consumption of the blower and the higher design pressure drop. Therefore systems with low parasitic power consumption should operate with UF between 70 and 80% and systems with high power consumption between 60 and 70%. Te results of all calculations are presented in Figs. 11 and 12. The Layouts A 1, B 2, B 3 and C 2 are calculated with CycleTempo and B 1 and C 1 are done with ProSofc.
Recycling ratio
Fuel utilization
Pre-reformer
In Figs. 11 and 12 the results of the layout types B and C are arranged in the following way. For each fuel and layout there are nine efficiency points. The first three points from the left are for UF 60%, the following three for UF 70% and the last three points for UF 80%. The points within one UF are RR 50%, 60% and 70%, from left to right. The electrical and thermal efficiencies are mainly influenced by the fuel utilization and recycling ratio as already described in detail for layout B 2 with NG 1. The systems operating with biogas will not have a significant decrease of electrical efficiency at high UF and RR because they are working without a pre-reformer. That is possible because the two biogases contain only traces of higher hydrocarbons. Therefore the methane content at the inlet of the SOFC is higher and the air flow rate will not increase that much as for natural gas systems. Also Layouts with unheated reformer (B 1, C 1 and C 2) are reacting less sensitive on high UF and RR. The equilibrium temperature of the adiabatic reformer is lower (typical between 450 and 460 C) than for the heated reformers (500 C). This has an influence on the methane concentration entering the stack and also on the air flow. Systems using ESC cells are generally reacting less sensitive on the fuel utilization and recycling rates due to the chosen operation parameters.
Fig. 8 e Efficiencies of layout B 2 with NG 1.
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 8 ( 2 0 1 3 ) 6 8 0 9 e6 8 2 0
Fig. 9 e Pressure drop and air factor of layout B 2 with NG 1.
Fig. 10 e Optimal operation range for layout B2 with NG1.
Fig. 11 e Results of layout A 1, B 2, B 3 and C 2 calculated with CycleTempo.
6817
6818
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 8 ( 2 0 1 3 ) 6 8 0 9 e6 8 2 0
Fig. 12 e Results of layout B 1 and C 1 calculated with ProSofc.
4.2.
Feasibility of layouts
The consortium agreed on some restriction of operation parameters, in order to keep the calculated systems on a level, which can be transferred to a technical feasible system. Therefore the following knock-out criteria were defined: The carbon formation based on thermodynamical equilibrium in pre-reformer should be avoided. The terminal temperature difference for gasegas heat exchangers should be at least 50 C and for heated reformers at least 100 C, in order to keep the size and cost of the components in a feasible range. The outlet temperature of the burner should be below 1000 C due to the requirements for the burner material. The air utilization in the stack should be lower than 50%, to avoid a poor cell performance. The knock-out criteria described above were applied to all calculated systems and the results are discussed below. A marked amount of the calculated systems do not fulfill one or more knock-out criteria. The general effects leading to a feasible system are discussed in detail for the layout B 2.
From the thermodynamical point of view often carbon formation according to Eq. (5) occurs in the pre-reformer. Fig. 13 shows the dependency of the carbon formation process from UF, RR and O/C-ratio for an equilibrium temperature of 500 C. With higher UF and RR the risk of carbon formation is reduced. NG 2 needs higher O/C-ratios than NG 1 because of the concentration of higher hydrocarbons. Beyond an O/Cratio of 2.3 no carbon formation is taking place. This point can also be considered as a suggestive criterion, as technology may allow successful, carbon-free operation of the prereformer even below the thermodynamical carbon formation limit as discussed in Ref. [20]. The terminal temperature difference criterion for the air heaters (minimum 50 C) is infringed 14 times always with UF 80% and RR 70%. Due to the higher air flow rates and the lower off-gas burner outlet temperature the terminal temperature difference becomes smaller. The terminal temperature difference criterion for the fuel heaters and pre-reformer is not touched. The maximum outlet temperature criterion of the off-gas burner is infringed 16 times, at UF (60%) and RR (50, 60, 70%). Due to the high methane concentration in the fuel gas
Fig. 13 e Carbon formation in pre-reformer for layout B 2.
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 8 ( 2 0 1 3 ) 6 8 0 9 e6 8 2 0
entering the stack and the cooling effect of the internal reforming process in the stack the air flow rate is very low. At the same time UF is low, that means a high amount of unconverted fuel is entering the off-gas burner. This combination leads to a high off-gas burner outlet temperature. The air utilization criterion for the stack (max. 50%) is infringed six times, always with UF 60% and RR 50%. Also here the air flow rate is very low due to the cooling effect of the internal reforming.
5.
Conclusions
To study the influence of the fuel management on the electrical efficiency and the water supply for the reforming process in total 220 systems calculations were carried out. All calculated systems are on a high level of efficiency: Electrical efficiency :
0:50 to 0:66
Thermal efficiency :
0:26 to 0:49
Total efficiency :
0:76 to 1:06
Systems without anode gas loop have up to 16% less electrical efficiency compared to anode off-gas recycling systems depending on the operation parameters. The electrical and thermal efficiencies are mainly influenced by fuel utilization and anode off-gas recirculation ratio. For systems operating with natural gas the results reveal that for higher values of UF and RR the electrical efficiency can also decrease. This is related to the higher air flow rates and the resulting power consumption of the blower. On the other hand high UF and RR are needed to avoid carbon formation in the reformer. That means an optimal operation range to avoid carbon formation in the pre-reformer and to get a high electrical efficiency requires RR of 70% and UF between 60 and 80%, depending on the parasitic power consumption. The layout itself, the choice of fuel gas or the type of cell has a minor effect on system efficiency. That means other criteria are important to choose the most promising approach, like number of components, complexity of system and so on. Layout A 1 without anode off-gas recirculation has the lowest complexity, but also the lowest electrical efficiencies. Layouts B 1, B 2 and B 3 are all using fuel blowers for the gas recycling loop, resulting in blower operating temperatures up to 600 C (B 1). Layouts C 1 and C 2 are using ejectors in the recycling loop. These components can easily work at high temperatures, but they are not easily to control, especially at part load operation. The fuel blowers and ejectors mentioned above are not available on the shelf and have to be developed. Therefore further more detailed calculations have to be done, with focus on part-load operation and controllability of the systems to elaborate a complete set of design and operation parameters.
Acknowledgment The authors would like to express their thanks to all partners who have contributed to the study. Acknowledgment goes
6819
to the European Commission for co-financing the project ASSENT under the contract number 244821.
references
[1] Fo¨ger K. Commercialisation of CFCL’s residential power Station BluGen. Proceedings, European Fuel Cell Forum Lucerne 2010;2:22e9. [2] Mai A, Sfeir J, Schuler A. Status of SOFC stack and system development at Hexis. Fuel Cell Seminar 2007;12(1). CD 000453. [3] Schuler A, Nerlich V, Doerk T, Mai A. Galileo 1000 N e status of development and operation experiences. Proceedings of 9th European Solid Oxide Fuel Cell Forum 2010;2:98. [4] European Commission. http://cordis.europa.eu/projects/rcn/ 94269_en.html. April 2012. [5] Bertoldi M, Bucheli O, Moden S, Larrain D, Ravagni A. Manufacturing and market-oriented development of SOFC generators at SOFCpower SpA. ECS Transactions 2011;35(1):127e38. [6] Schimanke D, Posdziech O, Mai BE, Kluge S, Strohbach T, Wunderlich C. Demonstration of a highly efficient SOFC system with combined partial oxidation and steam reforming. ECS Transactions 2011;35(1):231e42. [7] Bloomenergy e data-sheet ES-5700 system. http://www. bloomenergy.com/fuel-cell/es-5700-data-sheet; April 2012. [8] Fontell E, Kivisaari T, Christiansen N, Hansen JB, Pa˚lsson J. Conceptual study of a 250 kW planar SOFC system for CHP application. Journal of Power Sources 2004;131:49e56. [9] Kazempoor P, Dorer V, Ommi F. Evaluation of hydrogen and methane-fuelled solid oxide fuel cell systems for residential applications: system design alternative and parameter study. International Journal of Hydrogen Energy 2009;34:8630e44. [10] Ferrari ML, Traverso A, Magistri L, Massardo AF. Influence of the anodic recirculation transient behaviour on the SOFC hybrid system performance. Journal of Power Sources 2005;149:22e32. [11] Saebea D, Patcharavorachot Y, Arpornwichanop A. Analysis of an ethanol-fuelled solid oxide fuel cell system using partial anode exhaust gas recirculation. Journal of Power Sources 2012;208:120e30. [12] Lanzini A, Santarelli M, Orsello G. Residential solid oxide fuel cell generator fuelled by ethanol: cell, stack and system modelling with a preliminary experiment. Fuel Cells 2010;4:654e75. [13] Powell M, Meinhardt K, Sprenkle V, Chick L, McVay G. Demonstration of a highly efficient solid oxide fuel cell power system using adiabatic steam reforming and anode gas recirculation. Journal of Power Sources 2012;205:377e84. [14] Dietrich R-U, Oelze J, Lindermeir A, Spitta C, Steffen M, Ku¨ster T, et al. Efficiency gain of solid oxide fuel cell systems by using anode offgas recycle eResults for a small scale propane driven unit. Journal of Power Sources 2011;196:7152e60. [15] Blum L, Deja R, Peters R, Stolten D. Comparison of efficiencies of low, mean and high temperature fuel cell Systems. International Journal of Hydrogen Energy 2011;36:11056e67. [16] Blum L, Riensche E. Systems. In: Juergen Garche, Chris Dyer, Patrick Moseley, Zempachi Ogumi, David Rand, Bruno Scrosati, editors. Encyclopedia of electrochemical power sources, vol. 3. Amsterdam: Elsevier; 2009. p. 99e119. [17] Hemmes K, Patil A, Woudstra N. Flexible coproduction of hydrogen and power using internal reforming solid oxide fuel cells system. Journal of Fuel Cell Science and Technology 2008;5:041010e1.
6820
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 8 ( 2 0 1 3 ) 6 8 0 9 e6 8 2 0
[18] ProSofc is a MATLAB/SIMULINK-based simulation tool for stationary fuel cell system simulations. The code is developed by VTT, Technical Research Centre of Finland. [19] Zhu Y, Cai W, Li Y, Wen C. Anode gas recirculation behavior of a fuel ejector in hybrid solid oxide fuel cell systems:
performance evaluation in three operational modes. Journal of Power Sources 2008;185:1122e30. [20] Halinen M, Thomann O, Kiviaho J. Effect of anode off-gas recycling on reforming of natural gas for solid oxide fuel cell systems. Fuel Cells 2012;5:754e60.