Small-Scale Biogas-SOFC Plant: Technical Analysis and Assessment

May 16, 2014 - A comprehensive electrochemical model of the solid oxide fuel cell stack based on data from literature is applied in order to account f...
0 downloads 0 Views 2MB Size
Subscriber access provided by RUTGERS UNIVERSITY

Article

SMALL SCALE BIOGAS-SOFC PLANT: TECHNICAL ANALYSIS AND ASSESSMENT OF THE EUROPEAN POTENTIAL Bernhard Tjaden, Marta Gandiglio, Andrea Lanzini, Massimo Santarelli, and Mika Järvinen Energy Fuels, Just Accepted Manuscript • DOI: 10.1021/ef500212j • Publication Date (Web): 16 May 2014 Downloaded from http://pubs.acs.org on May 29, 2014

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Energy & Fuels is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

SMALL SCALE BIOGAS-SOFC PLANT: TECHNICAL ANALYSIS AND ASSESSMENT OF THE EUROPEAN POTENTIAL B. Tjaden*a, M. Gandigliob, A. Lanzinib, M. Santarellib, M. Järvinena a

b

Aalto University, Department of Energy Technology, Espoo - Finland Department of Energy, Politecnico di Torino, Corso Duca degli Abruzzi, 24, 10129, Torino - Italy.

* Corresponding author: Bernhard Tjaden – [email protected]

ABSTRACT This paper investigates simulation results of the thermodynamic performance of a 25 kWel small scale solid oxide fuel cell model fuelled with biogas. Hereby, biogas is produced from a predefined group of substrates, namely livestock effluents, energy crops, agricultural waste and organic waste. For the analysis, average methane (CH4) content is assumed to lie between 50 %mol and 60 %mol as large seasonal and daily variations are observed which are independent of used matter. Main biogas contaminants are sulphur compounds, mostly in the form of H2S. Sulphur leads to fast catalyst deactivation in reformer and fuel cell which is why an effective gas cleaning system is established. For this, ZnO and activated carbon are the most practical gas cleaning solution for small-scale plants. The energy model of the solid oxide fuel cell plant is designed and analysed through detailed energy and mass balance calculations throughout all system components and streams. The above mentioned model is set up in in such a way that different gas reforming options can be analysed and compared with each other in which steam, partial oxidation and auto thermal reforming are included. A comprehensive electrochemical model of the solid oxide fuel cell stack based on data from literature is applied in order to account for polarisation losses under varying operating conditions.1 The system analysis shows, that highest electric efficiency of 56.55% based on lower heating value is achieved under steam reforming. This value lies around 15 percentage points above average electric efficiencies of biogas engines based on lower heating value. Highest total plant efficiency (electric plus thermal) of 74.14% is reached under partial oxidation reforming, as exothermic reforming reactions increase thermal output of the plant. Within the parametric study it is concluded, that due to low electric efficiency and high sensitivity to biogas composition, auto thermal reforming is an unfitting reforming option.

KEYWORDS: SOFC, BIOGAS, GAS REFORMING, CLEAN-UP, CHP.

1. INTRODUCTION Anaerobic digestion (AD) biogas is a widely available energy resource which can be produced from a variety of biological substances which contain carbohydrates, proteins, fats, cellulose and/or hemicellulose.2 It is considered to be carbon neutral and combined with its local accessibility it features an excellent fuel for highly efficient solid oxide fuel cells for decentralized electric power supply. Biogas as fuel for solid oxide fuel cells (SOFCs) has already been treated in several articles by Van Herle et al.1,3 More recently, Papadias et al. performed a detailed analysis of impurities

ACS Paragon Plus Environment

1

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 37

contained in digester and landfill gas combined with a sensitivity analysis of electricity cost of a fuel cell system focusing on establishing a fitting gas cleaning unit.4 In their paper, Trendewicz and Braun also applied a fuel cell for valorising gas from waste water treatment units. Hereby, a thermo-economic analysis was carried out in order to evaluate electric efficiency and cost of electricity as a function of the most important operating parameters (among others: fuel utilisation, operating voltage, extent of internal steam reforming).5 Moreover, Braun provides a thorough techno-economic analysis of different fuel cell system layouts again as a function of operating parameters mentioned before.6 Even though system configurations which are analysed in 7-9 lie out of scope of this article, these papers provide profound fuel cell modelling correlations which are used for verifying purposes in the framework of this project. As gas reformation is inevitable when using a hydrocarbon fuel such as biogas to ensure continuous operation, several recent articles focused on comparing gas reforming options.10,11 In both articles it is concluded, that steam reforming yields highest electric efficiency as no fuel is combusted. Lanzini et. al (2010) investigated the feasibility of direct internal steam-, dry- and POx reforming in anode-supported as well as electrolyte supported cells from both an experimental and energy system analysis standpoint of view, concluding that the highest energy performance is achieved by steam reforming.12 In past years, more attention was paid on investigating cell degradation mechanisms due to biogas impurities.13-16 Looking at biogas, sulphur compounds pose the largest threat to catalysts deployed in reformer and fuel cell leading to catalyst poisoning and cell degradation. Yet, a holistic system modelling approach for small scale biogas fuelled solid oxide fuel cells is rarely presented. This work tries to close this gap by analysing the overall energy conversion system of a small scaled biogas fuelled solid oxide fuel cell. At first, composition and impurity loading of biogas stemming from substrates applicable to small scale systems are determined. This is followed by developing proper gas cleaning and gas reforming steps which aim on providing clean synthetic gas to the fuel cell in order to avoid carbon deposition in the plant. Three reforming options are implemented for comparison purposes: steam, partial oxidation and auto thermal reforming. The fuel cell itself is modelled by implementing thermodynamic and electrochemical correlations taken from literature combined with updated laboratory experimental data. Finally, a sensitivity analysis using a set of decision variables is carried out at the end of project work. The present work has been carried out in the framework of the SOFCOM project, where a small demonstration plant (2 kWel) with a biogas fuelled SOFC power generator will be installed in Turin at the end of 2014.17

1.1 Small size biogas plant: analysis and potential 1.1.1

Biogas substrate types

Three different plant categories are identified and classified as a function of installed power capacity of the biogas user (ICE or SOFC as in this analysis). Connected to each category, substrates considered for biogas production varies: the larger the power rating, the larger the biogas resources which has to be harvested as shown in Table 1. This article focuses on small scale biogas plants and thus, plant size is considered to be below 100 kWel. According to Table 1, the following feedstocks are used to determine biogas composition and type of contamination: •

Livestock effluents: manure from farm animals are used in most agricultural biogas plants; in practise, manure is mixed with straw, bedding material, fodder and other residues from animal husbandry;



Energy crops: such crops are grown to be specifically used for energetic valorisation, whereas anaerobic digestion is one option; energy crops include (among others) cereals, corn and grasses;

ACS Paragon Plus Environment

2

Page 3 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels



Agricultural waste: any kind of biological residue and green waste generated on a farm is considered in this substrate group; more precisely, it includes plant residues, side products of agricultural production processes, saw dust and other wastes;



Organic waste: small municipalities gather and separate waste from restaurants, abattoirs, other small scale businesses and households in order to utilize the organic waste fraction. 1.1.2

Biogas impurities

Depending on the substrate used for biogas production, type and amount of contained impurities vary largely. This is illustrated by Table 2 and Table 3, which list detailed biogas compositions produced from different substrates including waste-water treatment-plants. Concentrations of CH4 and CO2 vary even though the same classification of substrate is used. This is also valid for other biogas constituents such as sulphur and halogen compounds. Table 2 also shows that type and amount of impurities within biogas depends on the originating substrate. Combining data presented in Table 2 with biogas substrate considered in this project, the following conclusions can be drawn: •

Main contaminants in biogas produced from agricultural wastes and biological substrates are sulphur compounds, among which hydrogen sulphide (H2S) is the most dominant one; this conclusion is also validated in other references;3,5,18



In biogas stemming from food and animal waste as well as waste water (Table 2), halogenated compounds are present in very small trace amounts;



Organic silicon compounds are only detected in landfill gas and gas from WWTUs;11,18 although biogas stemming from these sources lies outside the scope of this work, Table 3 lists average values for different siloxane compounds for comparison reasons;

Therefore, the following paragraphs describe origin, concentration levels as well as possible impact of sulphur and halogen compounds on biogas fuelled SOFC plants in which the focus lies on anode side contamination. Sulphur is present in nearly all biological compounds as part of amino acids such as methionine and cysteine.19 In addition, biomass itself is made up by < 2 % (on weight basis of dry and ash free biomass) of sulphur taken up through soil and air.20 During digestion, sulphur is converted into gaseous compounds including H2S, carbonyl sulphide (COS), mercaptans and disulphides among which H2S is the most common one.18 Concentration levels of H2S in biogas along with the overall chemical build-up of biogas vary significantly depending not only on substrate but also on operating conditions. Sklorz et al. (2003) observed the following three correlations of H2S concentration fluctuations in a 45 kWel biogas plant using a gas engine for power generation:21 1.

As soon as the integrated gas engine stops running, H2S concentration in biogas decreases; this is explained by microorganisms or chemical reactions of H2S with galvanized steel tubing which reduce more H2S the slower the gas is flowing and the longer the residence time of the gas inside the digester system;

ACS Paragon Plus Environment

3

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

2.

Page 4 of 37

Due to mechanical stirring starting every 4 hours for twenty minutes, fresh biogas with a higher loading of H2S is released from the digestate; the time for desulphurization reactions to take place is not sufficient to absorb the increasing H2S concentration effectively;

3.

When a new batch of substrate is added to the digester, a continuous increase in H2S concentration is observed due to the injection of fresh, sulphur containing matter;

However, not only H2S is present within biogas but also several other sulphur compounds:22 in this project, kitchen waste is digested in a pilot scale anaerobic digester while gas quality is monitored over a period of 30 days. Aside of H2S, other sulphur compounds such as methanethiol (CH3SH), propanethiol (C3H7SH), butanethiol (C4H9SH) and dimethylsulphide (DMS) are detected which at times even surpass concentration levels of H2S. As a consequence, at least two gas cleaning steps are needed for effective biogas cleaning: one step to remove bulk H2S concentration and a second step to remove remaining sulphur compounds from the biogas as H2S removal systems do not necessarily remove other sulphur compounds.18 Halogens are contained within waste in the form of kitchen salts and polymers (polytetraflouroethylene: PTFE, polyvinylchloride: PVC). As such, these compounds are mostly found in biogas stemming from landfills.11,18 In 4, halogenated compounds measured in landfill gas and biogas produced from anaerobic digestion of sewage sludge are compared with each other. It is shown, that average concentrations in landfill gas are higher than maximum values detected in gas stemming from sewage sludge (compare Table 3). Although sewage sludge lies outside the scope of considered substrates, Table 2 indicates that halogen content in biogas from WWTUs lies in the same range as biogas produced on farms. An explanation for halogen content in biological substances is that small quantities of chlorine are taken up by plants through salts which are washed out of soils. In average, chlorine build-up in plants amounts to < 1 %wtdb.20 As a result, trace amounts of halogens can be detected in biogas produced from substrate considered in this paper. The quantities of halogens reported in literature are below 1 ppmmol as shown in Table 2. 1.1.3

Biogas Fuelled Internal Combustion Engines

Currently, biogas is converted into electricity mostly using internal combustion engines (ICEs). This is why the following section presents an analysis on average electric efficiencies of ICEs. Figure 1 displays the trend of electrical efficiencies of available biogas fuelled engines as a function of electrical power output based on data from engine manufactures. As expected, large size engines can reach high efficiency values of up to 45 % on lower heating value basis which are slightly lower compared to electric efficiencies yielded by fuel cells. On the other side, a rapid decrease of efficiency is observed for power capacities below 500 kWel: small size ICEs present low electrical efficiency values of < 35 % for Pel < 100 kWel which is also shown in 26. It is visible, that for small size plant, a significant increase in electric efficiency can be expected when deploying a fuel cell. Generally ICEs are more robust to biogas contamination which results in lower investment costs for cleaning systems, as summarized in 5.

1.2 Biogas use in high-efficiency electrochemical generators: SOFC case 1.2.1

Effects of biogas on plant components

As explained in the section above, biogas contains several different contaminants which can affect the SOFC and other power plant components. Sulphur contamination of fuel gas results in poisoning of nickel catalysts inside the SOFC anode as well as in the reformer of the power plant. Sasaki et al. (2007) show that the presence of sulphur compounds in solid oxide fuel cells leads to a voltage drop due to the adsorption of sulphur on the triple phase boundary

ACS Paragon Plus Environment

4

Page 5 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

inhibiting anode half-reaction. Already small quantities of 5 pmmmol H2S in the gas stream result in an initial cell voltage drop of 0.14 V at 1,000 °C followed by a stable cell voltage development. In addition, when internal reforming of hydrocarbon fuels is carried out directly on the anode, 5 ppmmol H2S lead to an even larger voltage drop as catalytic reforming reactions of CH4 is inhibited which causes carbon deposition on the electrode. In small quantities, sulphur is adsorbed on the surface of the anode in which case cell voltage can be recovered over time when sulphur impurities are removed from the gas.15 When introducing COS and CH3SH into an SOFC, initial cell voltage drop is of the same magnitude compared to the cell voltage drop caused by H2S contamination. However, after initial voltage drop CH3SH contamination leads to a continuous voltage decrease (not observed with H2S and COS) which still requires further investigations.27,28 Chemical equilibrium calculations show that H2S is the most stable sulphur compound at operating temperature of solid oxide fuel cells. This explains, why initial cell voltage drop caused by different sulphur compounds in the fuel gas are of the same dimension, but not why long-term exposure to mercaptans can be more critical than H2S or COS contamination. Initial cell voltage drop due to sulphur contamination is temperature dependent and decreases with increasing temperature.27 Consequently, it is necessary to remove any kind of sulphur compound from biogas when using it as fuel for SOFCs. However, it has to be pointed out, that extent and nature of cell degradation is depending on applied cell materials and microstructural characteristics and thus, the phenomenon of sulphur degradation is not trivial to generalize. It is thus assumed, that any kind of sulphur compound has to be removed to a level < 1 ppmmol to ensure long term performance of the SOFC. Halogenated compounds lead to corrosion in delicate power plant components and measures have to be taken that concentrations remain low.29 In connection to solid oxide fuel cells, effects of chlorine gas is analysed in several articles: experiments show that fuel gas containing 5 ppmmol of Cl does not cause cell degradation or voltage drops in the SOFC.15,18,30 Błesznowski et al. (2013) investigate the effect of HCl contaminated fuel gas and conclude that at 10 ppmmol, a recoverable voltage drop is identified. Yet, when increasing concentration levels to 1,000 ppmmol, cell voltage starts to decrease continuously at a rate of 9.4 % over 100 h. It is suggested that Cl reacts with Ni in the anode subliming in the form of nickel(II) chloride (NiCl2) at Cl concentrations of > 100 ppmmol.31 Thus, it can be inferred, that the low levels of halogen compounds typically observed in biogas pose no threat to the SOFC system. However, halogen compounds have to be considered when applying adsorptive gas cleaning methods: adsorption efficiency of activated carbon (AC) decreases in the presence of halogenated and aromatic co-vapours. Absolute breakthrough times of various sulphur compounds decrease by up to 14 % when halogen containing co-vapours are added to the gas stream.32 As mentioned previously, siloxanes are not considered to be present in biogas stemming from substrates included in this analysis and are only mentioned in brief: siloxanes impurities lead to SiO2 formation on the anode of the fuel cell which, in turn, causes a continuous degradation of cell voltage and leading to fatal degradation of the cell. Interested readers are referred to.18 1.2.2

Small scale biogas-SOFC plants: possible system designs

To analyse the performance of the biogas fuelled SOFC system for a range of operating scenarios, the system is modelled using the software Aspen Plus® allowing the use of the built-in thermo-physical property library. Hereby, each plant component is defined and implemented in the software according to the plant layout and connected with each

ACS Paragon Plus Environment

5

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 37

other via material streams, which contain all necessary information to resolve the mathematical correlations. Figure 2 shows the power plant layout highlighting the main components: •

Biogas compound mixer: allowing to vary biogas composition;



Biogas cooler: to dry entering biogas;



Reformer: to convert CH4 into CO and H2 using air and/or recirculated anode off-gas as reforming agent;



Solid oxide fuel cell: directly converting chemical energy of fuel gas into electrical energy;



After burner: oxidising unspent fuel;



Heat integration: providing heat for endothermic reactions in reformer and for preheating incoming streams;



Hot water supply: utilisation of remaining heat in exhaust gases for domestic hot water demand;

2. BIOGAS-SOFC PLANTS: TECHNICAL ANALYSIS

2.1 Water and contaminants removal unit It is shown in preceding sections, that integration of biogas fuel with solid oxide fuel cell systems can only be realised via an effective gas cleaning system. Consequently, several cleaning steps are necessary to ensure risk-free operation. The following list gives a summary of gas cleaning stages applied in the system analysis: 1.

In-situ H2S precipitation by injecting 2 %mol to 6 %mol of air directly into the digester; this way, bacteria oxidise H2S into elemental sulphur which is enriched in the digestate lowering hydrogen sulphide concentrations in the gas stream to < 100 ppmmol;29,33

2.

Saturated biogas leaving the digester is dried by condensation using cold water to avoid corrosion in power plant parts and then reheated again with incoming biogas; calculations show, that a relative humidity of approximately 60 % is achieved;

3.

The remaining content of H2S which is not removed by bacterial activity is cleaned by adsorption on ZnO, ideally reaching concentrations levels of < 1 ppmmol;34

4.

Trace impurities such as other sulphur and halogen compounds are removed in a final polishing step by an adsorption bed of activated carbon (AC) aiming for concentrations levels of < 1 ppmmol;32

It is suggested, that the cleaning system comprises of two parallel adsorption vessels for each filter to ensure continuous operation during bed material change or breakdown. After gas cleaning, contamination levels of biogas impurities are considered to be below 1 ppmmol and thus, do not pose any threat to power plant components during system operation. Thus, contaminants are not included in the gas streams of the simulation model but rather pressure drops of the gas cleaning vessels are included. The biogas cleaning system is not included in the Aspen Plus® model since it does not give any sensitive contribution related to the thermodynamics of the system: there are no changes in macro-components composition, temperature, pressure, flow rate. The inlet biogas to the model is supposed to be already clean. Yet, the gas cleaning system is included in form of pressure drops along the cleaning columns.

ACS Paragon Plus Environment

6

Page 7 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

2.2 Reforming options Direct injection of methane into the anode of the fuel cell can lead to carbon deposition and enhance in-plane thermal gradients across the fuel cell membrane. Both effects lead to due to fast endothermic reactions as soon, cell degradation and finally to complete shutdown of the SOFC. Solid carbon formation on the anode reduces mass transport of gas reactants to the triple phase boundary and deactivates catalytic activity of Ni contained in the anode which decreases the performance of electrochemical as well as direct reforming reactions.35 This is why a significant portion of methane has to be reformed into CO and H2 prior to SOFC injection by adding a reforming agent into the reformer. The maximum amount of CH4 that is generally safe to reform directly in Ni anodesupported type stacks is generally below 40 - 50%mol of the overall electrochemical reactive fuel. In this analysis, three reforming options are considered: steam reforming (SR) with simultaneous dry reforming (DR), partial oxidation reforming (POx) and auto thermal reforming (ATR) where steam and dry reforming run in parallel as well. The amount of either reforming agent, which is needed to avoid carbon deposition on the anode, can be estimated by equilibrium calculations as shown in Figure 3. The three options are briefly outlined below. 2.2.1

Steam Reforming (SR)

Under steam reforming, H2O is used as reforming agent to convert CH4 into a H2 and CO rich synthesis gas. For a small scale system, no external supply of any kind of reforming agent is implemented. Consequently, steam is provided to the reformer by anode exhaust gas recirculation from anode exhaust stream back into the reformer. Aside of H2O, CO2 is present in the recirculated gas as well as in the biogas itself. As deployed catalysts favour steam as well as dry reforming, both reforming options are considered to run in parallel under SR. The mass flow of steam is adjusted in such a way, that a steam to carbon ratio (S/C) of two is reached in the reformer. S/C ratio is calculated using the following equation:  ⁄ =

   

.

Eq.1

In which can be defined: S/C

steam to carbon ratio [-]

 

molar flow rate of steam [mol/s]

  

molar flow rate of methane [mol/s]

Despite the carbon dioxide contained in the biogas, the S/C ratio is defined in its general equation as ratio between the methane and the water. Figure 4 shows a ternary diagram of the inlet composition as a function of S/C ratio. It is visible, that temperatures above 675 °C no carbon deposition is caused even at low S/C ratios of one. Thus, the nominal chosen value of two leaned on parameters used in literature such as 36,37 is safe when the reformer temperature lies higher than 700 °C which is always guaranteed in this analysed plant. Steam is injected into the reformer by recirculating anode off-gas back into the reformer to meet the S/R ratio. 2.2.2

Partial Oxidation (POx)

Partial oxidation reforming combusts a part of the fuel by adding air as reforming agent. Due to partial combustion and dilution with N2 contained in air, system efficiency is expected to be lower compared to steam reforming. Yet, POx offers lower system complexity and thus, economic advantages. The lambda air ratio in the reformer λPOx is assumed to be a constant value amounting to 0.25.

ACS Paragon Plus Environment

7

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

 

 = 

,

.

Page 8 of 37

Eq.2

In which can be defined: λPOx

lambda air ratio [-]

 

actual air flow rate [mol/s]

 ,

stoichiometric air flow rate for complete combustion [mol/s]

In order to avoid the carbon formation risk when POx is performed, a ternary diagram is shown in Figure 5 for the inlet gas composition varying the lambda air ratio. With a reduced air flow rate (low oxygen to carbon ratio in the reformer), the risk of solid carbon formation can be reach at temperature lower than 800 °C. On the other side, analysing the chosen nominal conditions of 0.25 the risk is reduced and the safety area is guaranteed at the chosen reformer nominal temperature.36,37 2.2.3

Auto-thermal reforming (ATR)

Auto thermal reforming is a combination of all aforementioned reforming options which means that exothermic partial oxidation reforming drives endothermic steam and dry reforming reactions. In the end, reforming agents are combined in such a way, that overall enthalpy of reaction amounts to zero. Reforming reactions and corresponding enthalpies of reactions at standard conditions are shown in the following equations: 



+ " → 2" + 2





+

→ " + 3





+  " → " + 2

" 5

Δ ℎ' = 247 +,/./0

Eq.3



Δ ℎ2 = 206 +,/./0

Eq.4



Δ ℎ2 = −36 +,/./0

Eq.5



In which can be defined: ∆rh0

specific reaction enthalpy at standard conditions [kJ/mol]

The system analysis compares SR, POx and ATR in which the amount of internal reforming for SR is varied. DR is considered to take place in parallel to SR and ATR due to elevated CO2 content present in the biogas stream as well as in recirculated anode exhaust gas stream and because of applied catalysts. Once the biogas flow rate is determined (DR fixed), infinite possible combinations of S/C and λPOx could be determined in order to have a zero heat duty of the reformer: for this reason in the model the S/C ratio has been fixed to its nominal value. 2.2.4

Technical reformer layout

Depending on the reforming agent, a different reactor design is needed. However, due to the small system scale analysed in this project, only few manufacturers offer adequate solutions. Nevertheless, the following paragraphs summarise solutions on applicable reformer options: •

Steam reforming: feasible in the 450 °C - 900 °C range, (best results are achieved between 550 °C and 800 °C). Because of the endothermic nature of the process, an external heat source is always required when performing SR. On the other side, this process leads to the highest electrical efficiency between the three solutions. A commercial small size steam reformed for fuel cell applications is presented on the

ACS Paragon Plus Environment

8

Page 9 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

website of the following manufacturer: 38. In order to provide the required amount of thermal power, two possible technical solutions are available: o

Combustion heated reformer: the reformer chamber is set inside a combustion chamber in

which exhaust gases (together with NG if needed) are burned with spent air in order to provide the heat required. In 39, the following layout can be found: a first combustion chamber made of metal material for the first combustion stage is connected to a second combustion chamber defining a second combustion stage. The reaction unit comprises also one or more reaction elements housed partly inside the combustion chambers so that the heat generated inside them is transferred by thermal heat conduction to these elements.39 The chosen catalyst is a nickel based which is the most common catalyst suitable for these applications. The chosen system size lies below the one proposed in the work, so that a SR could be considered a viable solution for the analyzed plant. From other patents, a similar layout with packages of tubes is also found.40,41 o

Indirect internal reformer: the reformer is posed in the vicinity of the SOFC in a position where it is able to receive heat radiation from the SOFC. This combination is able to compensate the endothermic heat requirement of the reforming with the surplus of heat generated in the SOFC yielding a reduction in the cathode air flow and blower consumption.



Partial oxidation reformer: POx is an exothermic reaction and thus no heat requirement is needed. Despite this first advantage, the reaction consumes a part of the fuel and thus leads to a reduced electrical efficiency. Some commercial products can be found in the fuel cell market using catalytic POx.42,43



Auto thermal reformer: when a low S/C ratio and a low temperature are coupled together, as seen before in the ternary diagram in Figure 4, the risk of coke formation cannot be avoided. This problem can be overcome by adding air or oxygen to the hydrocarbon/steam fuel mixture. An auto thermal reforming catalyst is a catalyst which promotes a steam reforming as well as partial oxidation reforming. In general high activity nickel reforming catalysts containing 15 %wt - 25%wt nickel on a α-alumina or magnesia doped alumina are used but higher efficiency products with rhodium on a impregnated alumina support are also be found in practice.44

Catalyst materials are different for each reforming solution since a material able to activate the chosen reforming agent is needed. Higher efficient catalyst can be rhodium or platinum based ones, even if this will lead to a high cost increase.40 Catalysts are usually filled in the reformer in pellet or honeycomb form. For start-up purposes, at a first sight for the high efficiency SR an external hydrogen cylinder is necessary to start the endothermic reforming reactions and subsequently start up the SOFC. In recent studies in the field of SOFC start up, more complex but more flexible system layouts were introduced such as in 45. Here, the system is composed of a first steam reformer heated by the SOFC itself which works under nominal operating conditions and a second reformer heated by a NG combustor for the start-up procedure.45 Another option for system start-up is presented in 46 where a first thermally auto-sustained POx phase is followed by a SR for nominal operation. The two stages can be operated as one single reformer with an auto-thermal catalyst or as two separate reformers. This way the system can be started-up in a short time without losing the advantages of SR. In this work, the reforming reactor is modeled using a Gibbs equilibrium reactor. This choice is justified by literature findings which validated reforming models by comparing calculated equilibrium gas compositions with

ACS Paragon Plus Environment

9

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 37

experimental findings.35,36 For this reason the choice of an equilibrium reactor, as done in the presented work, is a good approximation of the real reforming behavior despite the technical and sometimes even complex flow and thermal design arrangements proposed by the manufacturers of fuel processing units for SOFC systems. 1.3.1

Anode off-gas recirculation

For anode recirculation purposes under steam reforming, two possible pathways can be considered: •

Hot recirculation blower: this was historically the preferred solution for molten carbonate fuel cells due to its simple working principle and the availability of materials which can withstand temperatures of up to 700 °C.47,48 For SOFCs systems the temperatures could be higher (up to 900 °C) and thus new expensive and innovative materials are required. Commercial solution can be found on the market which focus on applications in the automotive sector.49,50



Ejector (Venturi device): ejectors are a promising solution for the future of fuel cell plants. They operate without moving parts which results in considerably lower stress. As a result, conventional materials can be applied for which no lubrication is needed which might pose a risk of anode poisoning.51 In the ejector, a primary fluid at high pressure expands in a nozzle and enters a duct at high velocity, where it mixes with another gas coming from a second line. The ejector aims on to maintaining required pressure in the fuel cell while enough exhaust gas is recirculated to obtain the desired S/C ratio. The main disadvantage of the presented solution is related to the poor availability of commercial high temperature solution and the difficulty of responding to load variations in the fuel cell.52,53

2.3 Fuel cell modelling The electrochemical model developed in this work includes activation, diffusion as well as ohmic loss mechanisms. For this, ohmic and activation calculation procedure presented in 1 are extended by a diffusion model taken from.54 The electrochemical model calculates the ASR of the cell as a function of current density in which three loss mechanisms are characterised as follows:55 •

Ohmic losses caused by resistance of migrating electrons and ions in cell material;



Activation losses are connected to overcoming energy barriers of involved reactions;



Diffusion losses arise due to mass transport limitations of reactants;

A detailed derivation and list of correlations used to calculate the voltage drop due to polarization losses is shown in the Supporting Information document accompanying this article. Polarization and power density curves as function of current density under steam reforming are shown in Figure 6. Constant voltage operating strategy has been chosen because it maintains high plant efficiency even during cell degradation: in 56 a comparison between constant voltage and constant current strategies is presented. It is made clear, that constant voltage operating strategy reduces stresses to the stack caused by degradation due to decreasing current in order to ensure chosen operating voltage. As operating current is decreased during cell degradation, efficiency is kept constant. In order to provide constant power output while cell current decreases, the installation of an excess spare capacity is required. Nevertheless, this results in an up to five times longer operating lifetime of the stack when constant voltage is carried out in the system compared to constant current operating strategy

ACS Paragon Plus Environment

10

Page 11 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

Correlation between operating voltage VOp, area specific resistance ASR and current density j is presented in the following equation: 78 = 79:; − (=> ∗ @).

Eq.6

In which can be defined: VOp

operating voltage [V]

VRev

Gibbs voltage of cell [V]

ASR

area specific resistance of the cell [Ω·cm2]

j

current density in the cell [A/cm2]

For calculating ASR and j, Gibbs (thermodynamic) voltage rather than Nernst voltage is used as reversible voltage VRev. The reason for this is that Nernst voltage is depending on the molar fractions of educts and products in the fluid streams entering and exiting the cell. Consequently, Nernst voltage is highest when there is no electrochemical reaction occurring in the cell. However, this only takes place when no current is drawn from the cell and thus when no load is connected to it.1 From a power plant operation point of view, such a scenario is of limited practical use. Gibbs voltage on the other hand provides the maximum (reversible) work that can be extracted from a control volume after certain amount of fuel and oxidant stream have reacted and eventually mixed. Aside of operating voltage, the installed power capacity of the fuel cell is fixed. As a result, current density is varying according to operating conditions and applied reforming option. Once the current density and the ASR have been determined via electrochemical correlations introduced above, the active area of the stack can be calculated from the knowledge of the total faradic current. =B =

CDEEF,GH I

=

JK

NL M I

.

Eq.7

In which can be defined: AAct

active area of stack [cm²]

IFarad,Tot

total faradic current of stack [A]

j

current density in the cell [A/cm2]

WDC

SOFC DC power output [W]

VOp

operating voltage [V]

Eq. 7 indicates, that another fixed parameter in this SOFC model is the power output WDC generated by stack. Together with a fixed operating voltage, WDC is the main input parameter to solve the electrochemical model of the SOFC and calculate ASR and j.

2.4 Heat integration and Balance of Plant As can be seen in the system layout of Figure 2 hot exhaust gases leaving the after burner are used to preheat incoming fluid streams. For implementing heat exchangers in the system, the following hypotheses are applied: •

Counter-flow double-pipe heat exchangers;



No fouling resistance in heat exchangers;



No mixing of hot and cold streams;

ACS Paragon Plus Environment

11

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60



Page 12 of 37

No heat losses to the environment;

In total, five heat exchangers are implemented in the system layout: two in the biogas cooler system, two for preheating fresh biogas and air and a final one for extracting the remaining heat from the exhaust gases for hot water supply (CHP heat-exchanger for cogeneration). For the heat exchangers implemented in the biogas cooler, a simplified sizing approach using average U values for gas to gas and water to gas heat transfer are used: for gas to gas, U value amounts to 25 W/(m²K); for water to gas, U value amounts to 130 W/(m²K). These values are taken from 57 and 58 and are inserted in the following equations to calculate the heat exchanger surface areas: OP = Q ∙ = ∙ ∆TUV . ΔTUV =

Eq.8

XY,Z[Y\,] ^[(Y\,Z[Y,] ) _`

aG,Z bG\,] c

.

Eq.9

(G\,Z bG,] )

In which can be defined: Φef

transferred heat flow [Wth]

U

overall heat transfer coefficient [W/(m²·K)]

A

area of heat exchange [m²]

ΔT_h

logarithmic mean temperature difference [K]

Ti,f , Tj,f

temperature of entering and leaving hot fluid [K]

Ti,k , Tj,k

temperature of entering and leaving cold fluid [K]

For the remaining three heat exchangers, which operate at considerably higher temperature levels compared to the ones utilized in the biogas cooler, a more precise sizing methodology taken from 57 is applied. Remaining Balance of Plant (BoP) components are blowers and power conditioning, for which the following assumptions are applied: •

Internal blower losses are characterized by an isentropic efficiency;



Isentropic efficiency of biogas blower amounts to ηis = 0.5;



Isentropic efficiency of partial oxidation blower amounts to ηis = 0.5;



Isentropic efficiency of air blower amounts to ηis = 0.6;



DC/AC inverter efficiency 0.96;

The power output needed from the blowers is calculated by determining the pressure loss in the system and thus, by calculating the magnitude of overpressure which has to be supplied. Table 4 lists pressure drop assumptions for each component in the system which are summed up to give the total overpressure needed for each fluid line, which are slightly lower compared to values assumed in 5: pressure drop values have been taken from manufacturers in the framework of the SOFCOM European project. The overpressures in the system add up to 0.165 bar and 0.095 bar on the biogas and air side, respectively. The Partial oxidation reformer provides air directly at SOFC operating pressure of 1.065 bar. Hypotheses for isentropic efficiencies of all blowers are stated above in which the efficiency of the air blower is assumed to be higher. The reason for this is that amount of air needed to cool the SOFC exceeds the amount of biogas and air for partial oxidation and thus, a larger blower with higher efficiency is installed. For anode off gas recirculation, a hot recirculation blower is

ACS Paragon Plus Environment

12

Page 13 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

applied for which it is assumed, that the load is negligible compared to the remaining blowers in the system. In addition, it is assumed that blowers are running on direct current. Thus, the AC/DC inverter is implemented after the power demand of the blowers is satisfied. Consequently, net AC electric power output of the system is calculated by first subtracting parasitic loads of balance of plant equipment from the stack DC power output and then applying inverter efficiency.

3. NOMINAL CASE STUDY AND SIMULATION SCENARIOS To validate and characterize the behaviour of the established thermodynamic simulation model, the performance of the system is evaluated first at baseline and then under varying operating conditions. Table 5 lists main operating parameters of the 25 kWel fuel cell system and their baseline values. Table 6 tabulates decision variables and their values for the baseline system as well as range of values for the parametric study. These values are applied for each reforming option and the results are then compared. In all cases, electric power output of the stack, operating temperature and pressure (i.e., atmospheric) are constant. Chosen decision variables are characterised below: •

Steam to carbon ratio and lambda air ratio are the amount of reforming agents for POx and SR to be injected into the reformer in order to convert the methane into synthetic gas.



Internal reforming determines the amount of gas reformed directly on the anode under steam reforming operation. The higher the amount of internal reformation, the smaller the amount of cooling air needed which has positive effects on parasitic power demand.



Operating voltage is the main parameter for running a SOFC plant under constant voltage operation. Varying Vop directly influences amount of produced irreversibilities in the SOFC and thus, stack efficiency.



Fuel utilisation affects productivity of the fuel cell for it indicates how much of the incoming fuel is processed in the fuel cell itself. Thus, FU directly impacts molar flow rate of fuel and the amount heat which is released in the after burner.



Molar fraction of CH4 in biogas underlies significant hourly and seasonal variations for which the flow rate of biogas has to be adapted adequately to supply a constant electric power output.



Molar fraction of N2 in biogas depends on the amount of air injected into the digester to reduce H2S. As a consequence, biogas is diluted with a considerable amount of nitrogen which is compensated by an increase in biogas flow rate. 25 %mol as upper boundary is chosen in order to account for air leakages in the digester or in tubing as well as to account for imprecise air injection for sulphur reduction.

Net electric and total efficiency (electric plus thermal power supply) of the system are the most important thermodynamic parameters when analysing the performance of a SOFC power plant. Both indicators allow a good interpretation of the performance of the system. Under base conditions, all efficiencies shown in this section are based on higher heating value of entering biogas. In addition, total active area of the stack is calculated as well. The reason for this is that overall installed active area of the stack connects thermodynamic parameters with investment costs: the larger the area of the stack, the more expensive the investment costs of the fuel cell. However, it can be inferred by

ACS Paragon Plus Environment

13

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 37

Figure 6 that an increase in power density achieved by an increase in current density and thus, a decrease in installed cell area results in a decrease of electric efficiency of the stack as polarization losses increase. Therefore an optimal techno-economic trade-off must be found.

4. RESULTS

4.1 Analysis of reforming options Table 7 shows the aforementioned performance indicators for each reforming option under baseline simulation conditions. Compared to electric efficiency values from gas engines shown in Figure 1 it is clear that all reforming options in the fuel cell system surpass electric efficiency of ICEs by between five and 12 percentage points. Steam reforming offers the highest electric efficiency of 50.65 % under base case conditions. This is around 10 percentage points higher compared to POx and approximately 12 percentage points higher compared to ATR. On the other hand, total plant efficiency for partial oxidation reforming is highest amounting to 74.14 %. High electric efficiency using SR is explained by the fact that in POx as well as ATR, a portion of the biogas is burned prior to feeding it into the SOFC. However, due to endothermic reforming reactions in the steam reformer, heat output is significantly lower compared to the other two reforming options. Under ATR, plant efficiencies are lower compared to POx, which is due to the large amount of CO2 present in the biogas stream: the large amount of CO2 present in biogas demands a large amount of heat in order to compensate endothermic dry reforming reactions to reach auto thermal conditions. For this, a larger quantity of biogas has to be combusted compared to partial oxidation reforming which reduces electric and total plant efficiency. Hereby, it has to be pointed out that no anode off-gas has to be recirculated as the amount of CO2 in the biogas is large enough for ATR purposes. Looking at the stack active area, it is visible, that under POx the smallest fuel cell area of 9.09 m² is needed. Under steam reforming and auto thermal reforming, needed stack area amounts to 11.35 m² and 10.62 m², respectively. It can be inferred, that the investment cost for a stack with larger active area is more expensive, resulting in lowest investment costs under POx. However, a detailed economic analysis of this project is presented in a following article. In order to have a better understanding of the effects of S/C ratio and λPOx on the system performance, a further sensitivity analysis is carried out varying the ratios for POx and S/C ratio in the ranges of 0.1 - 0.6 for POx and 1 - 4 for S/C ratio. Figure 7 shows that varying S/C ratio the electrical efficiency only increases slightly by approximately 0.2 % points. In relation to analysing the sensitivity of the system to changes in λPox, Figure 8 shows that increasing the air flow rate into the reformer results in an increase in thermal power released from the reformer and an increase in hot water production. On the other hand, a drop in electrical efficiency is the direct consequence due to the combustion of biogas in the reformer. A final analysis related to the reforming modelling with the Aspen Plus® software is related to the choice of the reactor model for the partial oxidation simulation: in the analysed model the reformer can be modelled either through a stoichiometric reactor in which the POx reaction is predefined or through a Gibbs equilibrium reactor. Figure 9 shows results of the reformer outlet composition with the two different reactors varying the air ratio. Under base case conditions (λPox = 0.25) a relative small difference between the two systems is observed, while at the boundaries of the sensitivity analysis, differences are becoming more significant which can be explained by different reactions simulated in the Gibbs reactor model.

4.2 Operating voltage sensitivity analysis

ACS Paragon Plus Environment

14

Page 15 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

Behaviour of performance indicators in the parametric study is qualitatively similar throughout each reforming option. Yet, quantitative differences are observed. Decreasing operating voltage from 0.8 V to 0.7 V goes along with a decrease in electric efficiency and an increase in thermal efficiency under each reforming option. When operating voltage of the SOFC is increased to 0.85 V, the contrary is the case: lower over-potentials generate less waste heat caused by irreversibilities in the stack which decreases demand of cooling air, thus providing higher electric efficiencies as shown in Table 8. Consistently, the thermal power output of the plant decreases leading to lower thermal efficiencies. Yet, all three reforming options show, that total plant efficiency for higher operating voltages increases as the amount of higher electric efficiency compensates decreasing thermal efficiency. Simulation results for varying VOp are presented in Figure 10, which shows the increase and decrease in percentage points compared to results under base case conditions. It is visible, that steam reforming has higher sensitivity towards varying operating voltage compared to partial oxidation and auto thermal reforming. It is assumed, that due to higher reforming efficiency of SR, response to voltage changes are of higher sensitivity. Development of electric efficiency of POx and ATR as function of varying VOp is of comparable extent.

4.3 Fuel utilisation sensitivity analysis Manner and extent of variation in performance indicators under changing fuel utilisation is comparable with simulation results under varying operating voltage: with higher fuel utilisation, less fuel is needed for the same electric power output increasing ηel. However, thermal power output of the plant and thus, thermal efficiency decreases. In this case, the slope of decreasing thermal plant efficiency is more negative compared to the slope of increasing electric efficiency. Thus, total plant efficiency decreases. The reason for this is that, with higher FU, less fuel is burned in the after burner and therefore, less heat is provided by the system. Results for varying FU are presented in Figure 11 and Table 9. When comparing Figure 10 with Figure 11 it is noticed, that ηel under SR shows highest sensitivity towards varying fuel utilisation. Likewise, POx and ATR show similar results under changing FU as under changing VOp.

4.4 Internal reforming analysis Internal steam reforming (Int. Ref.) has positive effects on ηtot and, of lesser extent, positive effects on ηel. Endothermic reforming reactions taking place in the anode decrease the amount of cooling air needed in the system. Also, as less heat is needed in the external reformer, a larger amount of thermal power can be provided in the form of hot water. Internal reforming is the only parameter, which increases electric as well as thermal efficiency which is visible in Figure 12.

4.5 Analysis of different biogas inlet mixtures Varying biogas composition does not have a large effect on electric as well as total efficiency compared to results in previous sections. Although the amount of overall biogas required must increase when its CH4 content is lower, the demand of cooling air is reduced. A larger quantity of CO2 and/or N2 in the biogas in fact has a cooling effect on the stack. Consequently, the higher electric demand for the biogas blower is partly compensated by the smaller demand of cooling air (Figure 13 and Figure 14). Connected to this, one peculiarity has to be pointed out: when increasing CH4 fraction under ATR, increase of electric and total efficiency is of a larger extent compared to the other two reforming options. The reason for this is, that with lower amount of CO2 in the biogas stream, less dry reforming reactions are taking place which in turn means, that a smaller amount of biogas has to be combusted to reach auto thermal conditions. As a result, a smaller amount of air is needed in the reformer. These interrelations have a large effect on the sensitivity of electric and total efficiency under ATR. In addition, ATR also shows highest sensitivity towards N2 dilution as

ACS Paragon Plus Environment

15

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 37

illustrated in Figure 14. From a practical point of view, lower sensitivity to varying biogas composition is advantageous as chemical build-up of biogas underlies significant, seasonal as well as hourly variations, as outlined in sections above. As a result, to ensure stable power output and plant efficiency, steam reforming and partial oxidation reforming are more suitable for this area of application. Coupled with lowest electric efficiency and larger active area demand for the fuel cell stack compared to POx, auto thermal reforming seems to be an unfitting choice for converting biogas into synthetic gas for small scale solid oxide fuel cell plants. The presented work will contain only sensitivity analyses and not optimizations: even if the Aspen Plus® software has the tool for running optimization, they have not been included in the paper. In fact, looking at all the sensitivity graphs and tables, trends are monotonic increasing/decreasing and thus the optimal value is set on one of the sensitivity boundaries. Future work will then be related to multi-objective optimization in order to find the best set of parameters to achieve a maximum efficiency value.

5. CONCLUSIONS This article investigated the potential of biogas as fuel for small scaled solid oxide fuel cell plants. For this, a simulation model of a 25 kWel solid oxide fuel cell system was established using thermodynamic and electrochemical correlations from literature. The model was used to analyse thermodynamic behaviour of the system under varying operating conditions in which three reforming options were compared, namely: steam, partial oxidation and auto thermal reforming. Analysing the results of the parametric study showed, that steam reforming offered highest electric efficiency throughout all analysed cases: under base case conditions, electric efficiency of steam reforming amounted to 50.65 % based on higher heating value. Partial oxidation and auto thermal reforming lie approximately 10 and 12 percentage points below that value, respectively. In addition, active stack area under base conditions was investigated in which partial oxidation reforming featured smallest active stack area with approximately 9.1 m². Steam reforming and auto thermal reforming surpass this value by approximately 2.3 m² and 1.5 m² respectively. Due to lowest active area needed for the fuel cell stack compared to the other reforming options, system costs under partial oxidation reforming was expected to be lowest. Among all reforming options, auto thermal reforming showed highest sensitivity to varying biogas composition. Thus, auto thermal reforming was interpreted as unsuitable choice for biogas reforming in solid oxide fuel cell plants. As future work, the thermodynamic system analysis of this article is extended by an economic analysis by implementing cost functions for each component and deploying a comprehensive financial structure. Results of the economic analysis will be presented in a different article following shortly. Acknowledgements This work has been funded by the European Union (FCH JU Project 278798 SOFCOM, www.sofcom.eu).

ACS Paragon Plus Environment

16

Page 17 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

NOMENCLATURE AAct

active area of stack [cm²]

ASR

Area Specific Resistance of the cell [Ω·cm2]

ATR

Auto-thermal Reforming

DR

Dry Reforming

F

Faraday constant [C/mol]

IFarad,Tot

total faradic current of stack [A]

j

current density in the cell [A/cm2]

 ,

stoichiometric air flow rate for complete combustion [mol/s]

 

actual air flow rate [mol/s]

  

molar flow rate of methane [mol/s]

 

molar flow rate of steam [mol/s]

pAn

Pressure in anode [bar]

POx

Partial Oxidation

R

Universal gas constant [J/(mol·K)]

S/C

steam to carbon ratio [-]

SOFC

Solid Oxide Fuel Cell

SOFCOM

Solid Oxide Fuel Cell Operation and Management

SR

Steam Reforming

T

Temperature [K]

TAn

Temperature in anode [K]

Ti,f , Tj,f

temperature of entering and leaving hot fluid [K]

Ti,k , Tj,k

temperature of entering and leaving cold fluid [K]

U

overall heat transfer coefficient [W/(m²·K)]

VOp

Operating voltage [V]

VRev

Gibbs voltage of cell [V]

WDC

SOFC DC power output [W]

xa, xb

Molar fraction [-]

∆rh0

specific reaction enthalpy at standard conditions [kJ/mol]

ΔT_h

logarithmic mean temperature difference [K]

λPOx

lambda air ratio [-]

Φef

transferred heat flow [Wth]

ACS Paragon Plus Environment

17

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 37

REFERENCES (1)

Van Herle, J., Maréchal, F., Leuenberger, S. and Favrat, D. (2003), “Energy balance model of a SOFC cogenerator operated with biogas”, Scientific Advances in Fuel Cell Systems, Vol. 118 1–2, pp. 375–383.

(2)

Deublein, D. and Steinhauser, A. (2008), Biogas from waste and renewable resources: An introduction, Wiley-VCH, Weinheim.

(3)

Van Herle, J., Maréchal, F., Leuenberger, S., Membrez, Y., Bucheli, O. and Favrat, D. (2004), “Process flow model of solid oxide fuel cell system supplied with sewage biogas”, Selected papers presented at the Eighth Grove Fuel Cell Symposium, Vol. 131 1–2, pp. 127–141.

(4)

Papadias, D., Ahmed, S. and Kumar, R. (2012), “Fuel quality issues with biogas energy – An economic analysis for a stationary fuel cell system”, Integration and Energy System Engineering, European Symposium on Computer-Aided Process Engineering 2011, Vol. 44 No. 1, pp. 257–277.

(5)

Trendewicz, A. and Braun, R. (2013), “Techno-economic analysis of solid oxide fuel cell-based combined heat and power systems for biogas utilization at wastewater treatment facilities”, J. Power Sources, Vol. 233 No. 0, pp. 380–393.

(6)

Braun, R. (2010), “Techno-Economic Optimal Design of Solid Oxide Fuel Cell Systems for MicroCombined Heat and Power Applications in the U.S”, J. of Fuel Cell Science and Technology, Vol. 7 No. 3, p. 31018.

(7)

Gandiglio, M., Lanzini, A., Leone, P., Santarelli, M. and Borchiellini, R. (2013), “Thermoeconomic analysis of large solid oxide fuel cell plants: Atmospheric vs. pressurized performance”, Energy, Vol. 55 No. 0, pp. 142–155.

(8)

Zhao, Y., Sadhukhan, J., Lanzini, A., Brandon, N. and Shah, N. (2011), “Optimal integration strategies for a syngas fuelled SOFC and gas turbine hybrid”, J. of Power Sources, Vol. 196 No. 22, pp. 9516–9527.

(9)

Zhang, X., Guo, J. and Chen, J. (2012), “Thermodynamic Modeling and Optimum Design Strategy of a Generic Solid Oxide Fuel Cell-Based Hybrid System”, Energy Fuels, Vol. 26 No. 8, pp. 5177–5185.

(10)

Liso, V., Olesen, A., Nielsen, M. and Kær, S. (2011), “Performance comparison between partial oxidation and methane steam reforming processes for solid oxide fuel cell (SOFC) micro combined heat and power (CHP) system”, Energy, Vol. 36 No. 7, pp. 4216–4226.

ACS Paragon Plus Environment

18

Page 19 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

(11)

Pfeifer, T., Nousch, L., Lieftink, D. and Modena, S. (2013), “System design and process layout for a SOFC micro-CHP unit with reduced operating temperatures”, European Fuel Cell 2011, Vol. 38 No. 1, pp. 431– 439.

(12)

Lanzini, A., Leone, P., (2010), “Experimental investigation of direct internal reforming of biogas in solid oxide fuel cells”, Int. J. of Hydrogen Energy, Vol. 35, No. 6, pp. 2463-2476.

(13)

Wolak, F. (2012), Fuel Cell Power Plants: Biofuel Case Study - Tulare, CA, Golden. Available at http://www1.eere.energy.gov/hydrogenandfuelcells/wkshp_biogas_fuel_cells.html.

(14)

Rasi, S., Läntelä, J. and Rintala, J. (2011), “Trace compounds affecting biogas energy utilisation – A review”, Energy Convers. and Manage., Vol. 52 No. 12, pp. 3369–3375.

(15)

Sasaki, K., Adachi, S., Haga, K., Uchikawa, M., Yamamoto, J., Iyoshi, A., Chou, J., Shiratori, Y. and Ito, K. (2007), “Fuel Impurity Tolerance of Solid Oxide Fuel Cells”, ECS Trans., Vol. 7 No. 1, pp. 1675–1683.

(16)

Norheim, A., Wærnhus, I., Broström, M., Hustad, J. and Vik, A. (2007), “Experimental Studies on the Influence of H2S on Solid Oxide Fuel Cell Performance at 800 °C”, Energy Fuels, Vol. 21 No. 2, pp. 1098–1101.

(17)

“SOFCOM Research Project. SOFC CCHP with poly-fuel: Operation and Maintenance”, available at: http://areeweb.polito.it/ricerca/sofcom/en/ (accessed 13 May 2014).

(18)

Arnold, M. (2009), Reduction and monitoring of biogas trace compounds, VTT Technical Research Centre of Finland, Espoo, VTT Research Notes 2496.

(19)

Du, W. and Parker, W. (2012), “Modeling volatile organic sulphur compounds in mesophilic and thermophilic anaerobic digestion of methionine”, Water Res., Vol. 46 No. 2, pp. 539–546.

(20)

Vassilev, S., Baxter, D., Andersen, L. and Vassileva, C. (2010), “An overview of the chemical composition of biomass”, Fuel, Vol. 89 No. 5, pp. 913–933.

(21)

Sklorz, M., Schnelle-Kreis, J., Gottlieb, A., Kühner, N. and Schmid, B. (2003), Untersuchungen zum Einsatz von Oxidationskatalysatoren an landwirtschaftlichen Biogas-Verbrennungsmotoren, Bayerisches Institut für Angewandte Umweltforschung und -technik, Augsburg.

(22)

Papurello, D., Soukoulis, C., Schuhfried, E., Cappellin, L., Gasperi, F., Silvestri, S., Santarelli, M. and Biasioli, F. (2012), “Monitoring of volatile compound emissions during dry anaerobic digestion of the

ACS Paragon Plus Environment

19

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 37

Organic Fraction of Municipal Solid Waste by Proton Transfer Reaction Time-of-Flight Mass Spectrometry”, Bioresour. Technol., Vol. 126 No. 0, pp. 254–265. (23)

SEVA Energie AG, available at: http://www.seva.de/index.php/en/component/content/article/7 (accessed 13 May 2014).

(24)

2G Energy AG, available at: http://www.2-g.de/biogas-module/ (accessed 13 May 2014).

(25)

General Electric, available at: https://www.ge-distributedpower.com/products/power-generation (accessed 13 May 2014).

(26)

Kukuk, J. (2011), “BHKW-Kenndaten 2011”, ASUE Arbeitsgemeinschaft für sparsamen und umweltfreundlichen Energieverbrauch e.V., available at: http://asue.de/cms/upload/broschueren/2011/bhkw-kenndaten/asue-bhkw-kenndaten-0311.pdf (accessed 10 April 2013).

(27)

Haga, K., Adachi, S., Shiratori, Y., Ito, K. and Sasaki, K. (2008), “Poisoning of SOFC anodes by various fuel impurities”, Solid State Ionics 16: Proceedings of the 16th International Conference on Solid State Ionics (SSI-16), Part II, Vol. 179 27–32, pp. 1427–1431.

(28)

Sasaki, K. (2008), “Thermochemical Stability of Sulphur Compounds in Fuel Cell Gases Related to Fuel Impurity Poisoning”, J. of Fuel Cell Sci. and Technol., Vol. 5 No. 3, p. 31212.

(29)

Ryckebosch, E., Drouillon, M. and Vervaeren, H. (2011), “Techniques for transformation of biogas to biomethane”, Biomass Bioenergy, Vol. 35 No. 5, pp. 1633–1645.

(30)

Sasaki, K., Haga, K., Yoshizumi, T., Minematsu, D., Yuki, E., Liu, R., Uryu, C., Oshima, T., Ogura, T., Shiratori, Y., Ito, K., Koyama, M. and Yokomoto, K. (2011), “Chemical durability of Solid Oxide Fuel Cells: Influence of impurities on long-term performance”, J. of Power Sources, Vol. 196 No. 22, pp. 9130– 9140.

(31)

Błesznowski, M., Jewulski, J. and Zieleniak, A. (2013), “Determination of H2S and HCl concentration limits in the fuel for anode supported SOFC operation”, Cent. Eur. J. of Chem., Vol. 11 No. 6, pp. 960-967.

(32)

Papurello, D., Schuhfried, E., Lanzini, A., Romano, A., Cappellin, L., Märk, T., Silvestri, S. and Biasioli, F. (2014), “Influence of co-vapors on biogas filtration for fuel cells monitored with PTR-MS (Proton Transfer Reaction-Mass Spectrometry)”, Fuel Process. Technol., Vol. 118 No. 0, pp. 133–140.

ACS Paragon Plus Environment

20

Page 21 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

(33)

Krich, K., Augenstein, D., Batmale, J., Benemann, J., Rutledge, B. and Salour, D. (2005), Biomethane from Dairy Waste: A Sourcebook for the Production and Use of Renewable Natural Gas in California, Western United Dairymen, California

(34)

Anttila, P., Kaisalo, N., Suomalainen, M., Simell, P. and Rautanen, M. (2013), Fuel Cleaning Guidelines for Implementation of Proofs-of-Concepts, SOFC CCHP with Poly-Fuel: Operation and Maintenance, VTT Technical Research Centre of Finland, Espoo, SOFCOM Project no. 278798.

(35)

Lanzini, A., Leone, P., Guerra, C., Smeacetto, F., Brandon, N. and Santarelli, M. (2013), “Durability of anode supported Solid Oxides Fuel Cells (SOFC) under direct dry-reforming of methane”, Chem. Eng. J., Vol. 220 No. 0, pp. 254–263.

(36)

Galvagno, A., Chiodo, V., Urbani, F. and Freni, F. (2013), “Biogas as hydrogen source for fuel cell applications”, Int. J. of Hydrogen Energy, Vol. 38 No. 10, pp. 3913–3920.

(37)

Chiodo, V., Urbani, F., Galvagno, A., Mondello, N. and Freni, S. (2012), “Analysis of biogas reforming process for molten carbonate fuel cells”, J. Power Sources, Vol. 206 No. 0, pp. 215–221.

(38)

Canta Forda Computer Laboratory, available at: http://www.cfcl.com/ (accessed 13 May 2014).

(39)

Antonini, M., Saldivia, A., Saracco, G., Farina, C., Solaro, S., Mori, M. and Anastasi, M. (2009), Endothermic reaction unit and steam reforming device comprising this reaction unit. EP2519342 A1.

(40)

Mizuno, Y. and Hatada, S. (2008), Solid oxide fuel cell and reformer. EP2009726 A1.

(41)

Sammes, N. (2002), Integrated solid oxide fuel cell and reformer. US6492050 B1.

(42)

Hexis Galileo, available at: http://www.hexis.com/ (accessed 13 May 2014).

(43)

SOFC Power Stack Box, available at: http://www.sofcpower.com/ (accessed 13 May 2014).

(44)

Setzer, H., Lesieur, R., Karavolis, S. and Wnuck, W. (1984), Autothermal reforming catalyst and process. US4473543.

(45)

Sugiura, Y. (2013), Indirect internal reforming solid oxide fuel cell system. US20130011758 A1.

(46)

Anzai, I. (2014), Method for starting-up solid oxide fuel cell system. US8623563 B2.

(47)

AVL Fuel Cell Engineering, available at: https://www.avl.com/home (accessed 13 May 2013).

ACS Paragon Plus Environment

21

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

(48)

Page 22 of 37

Grillo, O., Magistri, L. and Massardo, A. (2003), “Hybrid systems for distributed power generation based on pressurisation and heat recovering of an existing 100 kW molten carbonate fuel cell”, J. Power Sources, Vol. 115 No. 2, pp. 252–267.

(49)

Massardo, A. and Bosio, B. (2001), “Assessment of Molten Carbonate Fuel Cell Models and Integration With Gas and Steam Cycles”, J. Eng. Gas Turbines Power, Vol. 124 No. 1, pp. 103–109.

(50)

Chick, L., Sprenkle, V., Powell, M., Meinhardt, K. and Whyatt, G. (2013), Solid oxide fuel cell steam reforming power system. US8394544 B2.

(51)

Marsano, F., Magistri, L. and Massardo, A. (2004), “Ejector performance influence on a solid oxide fuel cell anodic recirculation system”, J. Power Sources, Vol. 129 No. 2, pp. 216–228.

(52)

Hakala, T. (2012), Method and arrangement for controlling anode recirculation. EP2494641 A1

(53)

Hakala, T. (2013), Method and arrangement for controlling anode recirculation for fuel cells using a steam jet ejector. US8580443 B2.

(54)

Wuillemin, Z. (2009), “Experimental and modeling investigations on local performance and local degradation in solid oxide fuel cells”, Doctoral dissertation, Laboratoire d'énergétique industrielle, École polytechnique fédérale de Lausanne, Lausanne.

(55)

Kendall, K. and Singhal, S. (2003), High-temperature solid oxide fuel cells: Fundamentals, design and applications, Elsevier Advanced Technology, Oxford.

(56)

Thijssen, J. and Surdoval, W. (2009), Stack Operating Strategies for Central Station SOFC, 2009 Fuel Cell Seminar, Palm Springs.

(57)

Çengel, Y. (2008), Introduction to thermodynamics and heat transfer, 2nd ed., McGraw-Hill, Dubuque.

(58)

Kakaç, S. and Liu, H. (1998), Heat exchangers: Selection, rating, and thermal design, CRC Press, Boca Raton.

ACS Paragon Plus Environment

22

Page 23 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

LIST OF FIGURES

Figure 1: Biogas engines efficiency map23,24,25................................................................................................ 28 Figure 2: Small scale biogas fed SOFC system layout. ................................................................................... 29 Figure 3: Ternary C-H-O diagram for biogas reforming options (equilibrium calculations for 600 °C to 800 °C). ........................................................................................................................................................... 30 Figure 4: Ternary C-H-O diagram for the steam reformer inlet composition varying the S/C ratio ............... 31 Figure 5: Ternary C-H-O diagram for the POx reformer inlet composition varying the λPox.......................... 32 Figure 6: Polarization (×) and power density (•) curve for steam reforming in anode-supported cell operating at 800 °C .......................................................................................................................................................... 33 Figure 7: Electrical Efficiency ηel for varying S/C ratio.................................................................................. 33 Figure 8: Reformer heat duty QRef and molar flow rate of air nair for Pox for varying λPOx ............................ 34 Figure 9: Reformate outlet gas composition for different reactor models ....................................................... 34 Figure 10: Change in ηel as function of varying VOp........................................................................................ 35 Figure 11: Change in ηel as function of varying FU ........................................................................................ 35 Figure 12: Change in ηel (blue) and ηtot (red) as function of a varying degree of direct internal reforming within the SOFC .............................................................................................................................................. 36 Figure 13: Change in ηel (blue) and ηtot (red) under varying xCH4.................................................................... 37 Figure 14: Change in ηel (blue) and ηtot (red) under varying xN2 ..................................................................... 37

ACS Paragon Plus Environment

23

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 24 of 37

LIST OF TABLES

Table 1: Classification of biogas substrate according to power plant capacity ............................................... 25 Table 2: Biogas composition for different biogas plant types13 ...................................................................... 25 Table 3: Average and maximum values of main contaminants in biogas from WWTU4 ............................... 25 Table 4: Pressure Drops in System Components............................................................................................. 25 Table 5: Baseline case study main parameters ................................................................................................ 26 Table 6: Range of decision variables for parametric study ............................................................................. 26 Table 7: Baseline Simulation Results for SR, POx and ATR.......................................................................... 26 Table 8: ηel and ηtot under varying VOp ............................................................................................................ 27 Table 9: ηel and ηtot under varying FU ............................................................................................................. 27

ACS Paragon Plus Environment

24

Page 25 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

Table 1: Classification of biogas substrate according to power plant capacity Plant Size

Installed Power

S

Small Size

10 - 100 kWel

M

Medium Size

< 1 MWel

L

Large Size

1 - 10 MWel

Biogas Substrate • • • • • • • • •

Livestock effluents Energy crops Agricultural waste Organic waste Livestock effluents + energy crops +agricultural waste Agro-industrial waste Small waste water treatment units Large scale waste water treatment units Landfills

Table 2: Biogas composition for different biogas plant types13 Biogas Composition

Natural gas Waste water

Food waste

Animal waste

Landfill

80-100

50-60

50-70

45-60

40-55

Carbon dioxide (% vol.)