Ceramic Fuel Cells Ltd: Residential Micro-CHP System Demonstrations

Proceedings of

12th European SOFC & SOE Forum 2016

Chapter 04 - Session A13 Development of systems and balance of plant components

Edited by Prof. Nigel Brandon (Chair) Dr. Richard Dawson Dr. Zeynep Kurban Dr. Farid Tariq

Dr. Antonio Bertei Dr. Kristina Kareh Dr. Mardit Matian Dr. Enrique Ruiz Trejo

Dr. Paul Boldrin Dr. Jung-Sik Kim Dr. Paul Shearing Dr. Vladimir Yufit

Co-Edited by Olivier Bucheli

Gabriela Geisser

Fiona Moore

Dr. Michael Spirig

Copyright © European Fuel Cell Forum AG These proceedings must not be made available for sharing through any open electronic means.

ISBN 978-3-905592-21-4

www.EFCF.com/Lib

th

12 European SOFC & SOE Forum

www.EFCF.com/Lib ISBN 978-3-905592-21-4

5 - 8 July 2016, Lucerne/Switzerland

Chapter 04 - Session A13 Development of systems and balance of plant components

Content

Page A13 - ..

A1301 (Published in EFCF Special Issue Series, www.EFCF.com/LIB)......................... 5 Development of highly efficient SOFC power generating system using fuel concentration recovery process Kazuo Nakamura, Takahiro Ide, Shumpei Taku, Tatsuya Nakajima, Marie Shirai, Tatsuki Dohkoh, Takao Kume, Yoichi Ikeda, Takaaki Somekawa, Takuto Kushi, Kei Ogasawara, Kenjiro Fujita Tokyo Gas Co., Ltd., Fundamental Technology Dept.; 1-7-7, Suehiro-cho, Tsurumi-ku, Yokohama 230-0045 / Japan

A1302 (Published in EFCF Special Issue Series, www.EFCF.com/LIB)......................... 6 Prognostics-oriented simulation of an MSR fuel processor for SOFCs Federico Pugliese (1), Andrea Trucco (2), Paola Costamagna (1) (1) Department of Civil, Chemical and Environmental Engineering (DICCA) (2) Department of Electrical, Electronics and Telecommunications Engineering (DITEN) University of Genoa Via Opera Pia 15, 16145 Genoa, Italy.

A1303 .................................................................................................................................... 7 A Planar Steam Reformer Designed for 60,000 h Operation Yves De Vos (1), Jean-Paul Janssens (1) (1) Bosal ECS NV Dellestraat 20, B-3560 Lummen/Belgium

A1304 ................................................................................................................................... 12 Proof of concept for solid oxide electrolysis systems DI Richard Schauperl, Bsc Beppino Defner, Bsc Dominik Dunst, DI Jürgen Rechberger AVL List GmbH Hans-List-Platz 1 A-8020 Graz, Austria

A1305 ................................................................................................................................... 20 SchIBZ – application of large diesel fuelled SOFC systems for seagoing vessels and decentralized onshore applications Keno Leites thyssenkrupp Marine Systems GmbH Hermann-Blohm-Str. 3, 20457 Hamburg, Germany

A1306 (Published in EFCF Special Issue Series, www.EFCF.com/LIB)....................... 32 Development of a SOFC/Battery-Hybrid System for Distributed Power Generation in India Thomas Pfeifer, Mathias Hartmann, Markus Barthel, Jens Baade, Ralf Näke, Christian Dosch Fraunhofer IKTS Winterbergstraße 28 D-01277 Dresden / Germany

A1307 .................................................................................................................................. 33 Sulfur Tolerant WGS-Catalysts Thorsten Dickel (1), André Weber (1) Michael Scharrer (2), Claus Peter Kluge (2) (1) Institute for Applied Materials (IAM-WET), Karlsruhe Institute of Technology (KIT), Adenauerring 20b, D-76131 Karlsruhe/Germany (2) CeramTec GmbH CeramTec-Weg 1 D-95615 Marktredwitz

Development of systems & balance of plant components Chapter 04 - Session A13 - 2/58

th

12 European SOFC & SOE Forum

www.EFCF.com/Lib ISBN 978-3-905592-21-4

5 - 8 July 2016, Lucerne/Switzerland

A1309 (Abstract only, published elsewhere) .................................................................. 42 Control strategy for a SOFC gas turbine hybrid power plant Moritz Henke (1), Mike Steilen (1), Ralf Näke (2), Marc Heddrich (1), K. Andreas Friedrich (1) (1) German Aerospace Center (DLR), Pfaffenwaldring 38-40, 70569 Stuttgart, Germany (2) Fraunhofer IKTS, Winterbergstraße 28, 01277 Dresden, Germany

A1312 (Abstract only, published elsewhere) .................................................................. 43 rSOC plant concept for renewable energy storage Matthias Frank (1), Roland Peters (1), Van Nhu Nguyen (1), Robert Deja (1), Ludger Blum (1), Detlef Stolten (1,2) (1) Juelich Research Center IEK-3: Electrochemical Process Engineering 52428 Juelich/Germany (2) RWTH Aachen University Lehrstuhl für Brennstoffzellen, Fakultät für Maschinenwesen 52072 Aachen/Germany

A1313 ................................................................................................................................... 44 Investigation of a novel catalytic partial oxidation and pre-reforming radial reactor of a micro-CHP SOFC-system with anode off-gas recycle Timo Bosch (1), Maxime Carré (1), Angelika Heinzel (2), Michael Steffen (2), François Lapicque (3) (1) Robert Bosch GmbH Robert-Bosch-Campus 1, DE-71272 Renningen (2) Zentrum für BrennstoffzellenTechnik GmbH Carl-Benz-Strasse 201, DE-47057 Duisburg (3) Laboratoire Réactions et Génie des Procédés, CNRS-Univ. Lorraine 1 rue Grandville, FR-54000 Nancy

A1315 (Abstract only) ........................................................................................................ 54 Performance evaluation of solid oxide carbon fuel cells operating on steam gasified carbon fuels Tak-Hyoung Lim(*), Jong-Won Lee, Seung-Bok Lee, Seok-Joo Park, Rak-Hyun Song Fuel Cell Research Laboratory, Korea Institute of Energy Research (KIER) 152 Gajeong-ro, Yuseong-gu, Daejeon 34129, Korea

A1316 (Abstract only, published elsewhere) .................................................................. 55 Methane Steam Reforming Reaction over Ni/CeO2-ZrO2 Catalysts Loaded on Metallic Monolith Jong Dae Lee (1) (1) Department of Chemical Engineering, Chungbuk National University. 1 Chungdea-ro, Seowon-gu Cheong-ju, Chungbuk 28644, Korea

A1317 (Will be published elsewhere) ............................................................................... 56 System validation tests for a SOFC power system at INER Shih-Kun Lo*, Wen-Tang Hong, Hsueh-I Tan, Huan-Chan Ting, Ting-Wei Liu and Ruey-Yi Lee Institute of Nuclear Energy Research No. 1000 Wenhua Road, Longtan District Taoyuan City / Taiwan (R.O.C.)

A1319 (Abstract only, published elsewhere) .................................................................. 57 A Global Reaction Model of Carbon Gasification with K2CO3 in the External Anode Media of a DCFC Shinae Song, Jun Ho Yu, Kyungtae Kang, Jun Young Hwang Korea Institute of Industrial Technology Ansan, 426-173, South Korea

Development of systems & balance of plant components Chapter 04 - Session A13 - 3/58

th

12 European SOFC & SOE Forum

www.EFCF.com/Lib ISBN 978-3-905592-21-4

5 - 8 July 2016, Lucerne/Switzerland

A1320 (Abstract only, published elsewhere) .................................................................. 58 Experimental study on the fuel ejector for solid oxide fuel cell system Kanghun Lee (1), Sanggyu Kang (1, 2), Youngduk Lee (1), Kook-Young Ahn (1,2) (1) Korea Institute of Machinery and Materials (KIMM); Gajeongbukro 156; Daejeon/Republic of Korea (2) University of Science and Technology (UST), Gajeongro 217; Daejeon/Republic of Korea

Development of systems & balance of plant components Chapter 04 - Session A13 - 4/58

th

12 European SOFC & SOE Forum

www.EFCF.com/Lib ISBN 978-3-905592-21-4

5 - 8 July 2016, Lucerne/Switzerland

A1301 (Published in EFCF Special Issue Series, www.EFCF.com/LIB) Development of highly efficient SOFC power generating system using fuel concentration recovery process Kazuo Nakamura, Takahiro Ide, Shumpei Taku, Tatsuya Nakajima, Marie Shirai, Tatsuki Dohkoh, Takao Kume, Yoichi Ikeda, Takaaki Somekawa, Takuto Kushi, Kei Ogasawara, Kenjiro Fujita Tokyo Gas Co., Ltd., Fundamental Technology Dept.; 1-7-7, Suehiro-cho, Tsurumi-ku, Yokohama 230-0045 / Japan Tel.: +81-45-500-8772 Fax: +81-45-500-8790 [email protected]

Abstract Although a large number of residential fuel cell systems have been installed, the market for business-use fuel cell systems is at an initial stage in Japan. We believe that the realization of highly efficient power generation is the key issue in creating the market for business-use fuel cell systems. In view of the above, we aim to develop a highly efficient power generation system using the solid oxide fuel cell (SOFC). In order to realize the SOFC module with highly efficient power generation, we have laid out a two-stage SOFC stack configuration with a fuel concentration recovery process between the stacks. The fuel concentration recovery process is designed to remove 96% of the H2O content in the anode off-gas from the first SOFC stack. Since the fuel utilization rate of the second SOFC stack using the fuel from the process can be raised to about the same level as that of the first SOFC stack (for example 70%), a total fuel utilization rate above 91% can be achieved. Therefore, the module can generate power with high electric efficiency using this configuration. In order to demonstrate highly efficient power generation, the SOFC module using the configuration was manufactured and operated. The power generation test was carried out successfully, and thermally self-sustainable operation was confirmed. The total output power was DC 2.27 kW and the power generation-end efficiency was DC 69.2% (lower heating value, LHV) at the total fuel utilization rate of 86.3%. Taking inverter loss (5%) and auxiliary devices loss (6%) into consideration, the AC electrical efficiency was estimated to be 61.8% (LHV). We have established the method of achieving highly efficient power generation using the SOFC module with the two-stage SOFC stacks and the fuel regeneration process. In order to realize even higher power generation efficiency, it is required to remove the CO2 content by the fuel regeneration process and to also prevent heat from escaping outside the system. In the future, we aim to develop SOFC systems with high power generation efficiency above AC 65% (LHV) by improving the SOFC module and integrating it into the system. Remark: The full paper is published in EFCF Special Issue Series (www.EFCF.com/LIB, SI EFCF 2016) in Journal "FUEL CELLS - From Fundamentals to Systems".

Development of systems & balance of plant components Chapter 04 - Session A13 - 5/58

th

12 European SOFC & SOE Forum

www.EFCF.com/Lib ISBN 978-3-905592-21-4

5 - 8 July 2016, Lucerne/Switzerland

A1302 (Published in EFCF Special Issue Series, www.EFCF.com/LIB) Prognostics-oriented simulation of an MSR fuel processor for SOFCs Federico Pugliese (1), Andrea Trucco (2), Paola Costamagna (1) (1) Department of Civil, Chemical and Environmental Engineering (DICCA) (2) Department of Electrical, Electronics and Telecommunications Engineering (DITEN) University of Genoa Via Opera Pia 15, 16145 Genoa, Italy. Tel.: +39-010-353-2922 Fax: +39-010-353-2586 [email protected]

Abstract In solid oxide fuel cell (SOFC) plants, failure of the methane steam reforming (MSR) fuel processor can result in increased levels of methane being fed into the fuel cell stack, with possible consequent damage. In view of this, diagnostics and prognostics of the MSR reactor is of utmost importance. The development of methods for early prediction and detection of faults in chemical reactors is based on numerical tools for the steady-state and transient simulation. In the present work, we investigate in detail the problem of carbon deposition in an MSR reactor for application in SOFC power plants, through a first-principle model based on microscopic mass balances embedding a local chemical kinetics. The partial differential and algebraic equations (PDAEs) are integrated numerically using a finite element method, implemented through COMSOL Multiphysics. The results allow to identify the areas where carbon deposition is expected to occur, and show that, even with a steam-tocarbon (S/C) ratio of 3, carbon deposition can occur in some specific operating conditions.

Remark: The full paper is published in EFCF Special Issue Series (www.EFCF.com/LIB, SI EFCF 2016) in Journal "FUEL CELLS - From Fundamentals to Systems".

Development of systems & balance of plant components Chapter 04 - Session A13 - 6/58

th

12 European SOFC & SOE Forum

www.EFCF.com/Lib ISBN 978-3-905592-21-4

5 - 8 July 2016, Lucerne/Switzerland

A1303 A Planar Steam Reformer Designed for 60,000 h Operation Yves De Vos (1), Jean-Paul Janssens (1) (1) Bosal ECS NV Dellestraat 20, B-3560 Lummen/Belgium Tel.: +32-13-530-971 Fax: +32-13-531-411 [email protected]

Abstract A planar steam reformer is designed for meeting 60,000 h lifetime. The component is designed as a plate heat exchanger, whereby the reaction heat for the steam reforming is extracted from the hot cathode flow through thin, catalytically coated heat exchanging foils. The surface wall reactions were modeled in a periodic CFD domain, consisting of a coated foil, and periodic half anode and cathode channels on the opposing sides of the foil. The flow parameters, heat exchange and wall surface reactions were solved by the CFD. The catalyst aging by deactivation was determined by reactivity measurements on washcoat powder. The washcoat mass loss by flaking was obtained using SEM/EDX. Aging was simulated in CFD by partly deactivating the reactions at wall temperatures > 800°C. The reaction zone and the temperature profile shifted as a result. Washcoat redundancy was validated by calculating the overall performance. Redundancy parameters were drafted, so that the designed component proved unaffected by catalyst aging. The measured performance of new and aged reformers was in line with the calculations. Oxide scale growth, and scale flaking was determined by post-mortem analysis at 20,000 h. A stable Cr oxide scale was measured, 20 - 40 m thick. Locally, Cr / Fe oxide scale was present, resulting in flaking, and reduction of the bulk plate thickness. Iron oxides contributed for 40 to 85% of the collected flake mass, as determined using x-ray diffraction (XRD). The Cr evaporation was modeled by CFD at the cathode path, using rate parameters as measured on steel sample plates in an inert reactor. Optimization for both Cr evaporation and scale flaking was achieved by reducing the hotspots, and optimizing the material grade and component cost.

Fig 1: Contour plot of the gas composition in the planar steam reformer.

Development of systems & balance of plant components Chapter 04 - Session A13 - 7/58

th

12 European SOFC & SOE Forum

www.EFCF.com/Lib ISBN 978-3-905592-21-4

5 - 8 July 2016, Lucerne/Switzerland

Introduction The reformer is designed as a plate heat exchanger, with two flow paths, separated by heat exchanging plates. The anode gas flows between these plates, which are catalytically coated (Ni / Rh coating). The steam reforming occurs in this flow path, and the endothermic reaction heat is delivered by hot cathode gas, passing at the other side of the heat plates. The architecture of the heat exchanging plate is identical to standard heat exchangers. Hence, by using the same material, wall thickness and assembly process, one can reduce R&D and manufacturing costs. The coating has been validated on reformers for the chemical process industry, at 850°C nominal operating temperature. The risk for local hot spots above 850°C is avoided by 1D and 3D simulations. In this design phase, the local heat exchange and chemical reactions are modeled, so that the wall temperature and gas flow conditions can be solved. The lifetime of the component is however dependent on the chemical stability of the steel heat exchanging plates, as the application exposes the alloy to high humidity, high temperature, and possibly high carbon activity. Post mortem analysis on austenitic and Nibased alloys is required to determine if the component meets its required lifetime.

1. Scientific Approach The pre-reforming is modeled in CFD, using literature-based conversion rate at equilibrium [1]. The conversion rate is thereby calculated using the wall temperature and partial pressure of CO, H2, CH4 and H2O. The conversion rate deviates from the equilibrium value, depending on the Gas Hourly Space Velocity of the reactor. Data by the coating supplier were used to model the conversion in each cell of the CFD domain. The domain was meshed so that the gas resided on average 1 ms in the cells of the CFD domain. The calculated conversion was 14 % higher than the measured data on an actual reformer. The activation energy and pre-exponential factor were tuned, leading to new CFD calculations.

2. Experiments/Calculations/Simulations 2.1. Pre-reforming and temperature profile 1D simulations were executed, splitting up the reformer in 135 points, each representing 2.45 cm² surface area, and 0.245 cm³ reactor volume. The heat exchange was solved using validated data from previous experiments, on the same plate design, albeit without chemical reactions. The reaction heat was determined by calculating the reforming rate, using the procedure as described in §1. The conversion rate and heat exchange was converged in each point, leading to a calculated temperature profile and gas composition. Fig 2 shows the thermal profile, as obtained from a 1D analysis. The reformer uses coflow arrangement, as this leads to low temperature gradient over the length of the plate, and thus reproducible stable operating points. The inlet of both flows is at position 0 and the outlet at 135 mm. The pre-reforming rate is close to equilibrium at the outlet temperature of 480°C.

Development of systems & balance of plant components Chapter 04 - Session A13 - 8/58

th

12 European SOFC & SOE Forum

www.EFCF.com/Lib ISBN 978-3-905592-21-4

5 - 8 July 2016, Lucerne/Switzerland

Fig 2: Thermal profile over the length of the reformer.

2.2. Chemical stability of the alloy The used material is austenitic steel, containing 22% Cr. Cr is expected to form a dense Cr2O3 scale, which inhibits further corrosion. The stability of the scale was validated by post mortem SEM-EDX on three units: -

Steam reformer, 800 h of operation, XH2O = .04, T in = 815°C, Tout = 750°C Anode preheater, 500 h of operation, XH2O = .05, T in = 750°C, Tout = 300°C Cathode preheater, 23.000 h of operation, XH2O = .015, T in = 920°C, Tout = 230°C

Fig 3: SE (left) and BSE (right) image of a delaminated washcoat area The steam reformer application used porous Al2O3 washcoat, which was locally lost due to flaking. Fig 3 shows an area that lost its washcoat. The washcoat (Al2O3) appears white on the SE image, while the exposed steel surface shows a layered structure. The deepest Development of systems & balance of plant components Chapter 04 - Session A13 - 9/58

th

12 European SOFC & SOE Forum

www.EFCF.com/Lib ISBN 978-3-905592-21-4

5 - 8 July 2016, Lucerne/Switzerland

area appears grey in the BSE image, indicating a zone with lower average atomic number. The EDX spectrum shows predominantly Fe and O in this area, as compared to Cr and O in the darker zone, and Al / O in the washcoat area. This shows that the flaking is mainly initiated in the oxide scale of the alloy, causing loss of the washcoat covering the scale. The loss was <2% of the total area at 800h. Extrapolation to 60.000 h of required lifetime suggests that this loss rate should be improved by at least a factor 15. The anode preheater showed multiple small pits (<1 m diameter) on the anode flowpath, reminiscent of metal dusting. The pits were observed on heat exchanging plates exposed to CO and CO2 containing atmosphere, at temperatures in the 500 – 700°C range. Carbon deposits were observed in some areas of the plates (darker shades on the BSE image of Fig 4). We concluded that the alloy was susceptible to metal dusting, although it was not proven that Cr and Fe carbides were formed. However, literature shows that the selected austenitic material has only medium metal dusting resistance, as compared to Ni-based alloys Inconel 693 and 696. Furthermore, the abovementioned temperature range and gas composition match the conditions for metal dusting.

Fig 4: Pits visible in SE and BSE imagery, on the side as exposed to high C-activity.

Fig 5: Delamination of uncoated heat exchanger as exposed to high H2O partial pressure. The cathode preheater showed local loss of the scale, and local presence of Fe oxides (light area in the BSE image of Fig 5, mark d). Bubble-like structures (mark a) were observed, consisting of Fe and Cr oxides. The presence of Fe in the oxide scale indicates runaway corrosion, resulting in accelerated material loss. Development of systems & balance of plant components Chapter 04 - Session A13 - 10/58

th

12 European SOFC & SOE Forum

www.EFCF.com/Lib ISBN 978-3-905592-21-4

5 - 8 July 2016, Lucerne/Switzerland

3. Results The analysis results of §2.2. show that the austenitic alloy is not suited for meeting 60.000 h operation. Therefore, Ni-based alloys were analyzed. These alloys bring a significant cost penalty, which can be mitigated when the wall thickness is reduced to 120 – 160 m. Forming of these thin-walled foils is challenging, and requires high elongation at break. Good resistance to metal dusting is obtained by elevated Al or Si in the alloy composition, which adversely affects the elongation at break. Furthermore, thin foils are typically hot rerolled, starting from 1-2 mm coil material. The rerolling reduces the elongation at break. A compromise Ni-based material was selected, meeting the requirements on cost, manufacturability and chemical stability. Accelerated material testing showed stable oxide scale, without indications for scale loss during the test time (50 and 750 h). Fig 6 shows the scale on the tested material after 50 h exposure, while Fig 7 shows that the scale contains no Fe.

Fig 6: Se (left) and BSE (right) SEM/EDX image of the selected Ni-based alloy, 50 h exposure.

Fig 7: Spectrum of the selected Ni-based alloy, 50 h exposure.

References [1]

Mbodji, M. Steam methane reforming reaction process intensification by using a millistructured reactor: Experimental setup and model validation for global kinetic reaction rate estimation. Paper, Nancy, France: Chem. Eng. J., 2012.

Development of systems & balance of plant components Chapter 04 - Session A13 - 11/58

th

12 European SOFC & SOE Forum

www.EFCF.com/Lib ISBN 978-3-905592-21-4

5 - 8 July 2016, Lucerne/Switzerland

A1304 Proof of concept for solid oxide electrolysis systems DI Richard Schauperl, Bsc Beppino Defner, Bsc Dominik Dunst, DI Jürgen Rechberger AVL List GmbH Hans-List-Platz 1 A-8020 Graz, Austria Tel.: +43-316-787-2168 [email protected]

Abstract Today’s energy supply systems are not very suitable for highly stochastic energy production from renewable sources (wind, PV). The availability of wind and solar power is not sufficiently predictable and the storage of this excess energy is not possible in quantities based on today’s technologies. Solid oxide electrolysis offer an efficient and scalable energy storage technology with the potential to contribute in finding solutions for the typical issues of renewable electricity production and in developing carbon neutral fuels. A third opportunity is to use the products, including O2, for further chemical synthesis, for example in the pharmaceutical or plastic industry. Electrical energy can be stored in chemical energy by producing molecular hydrogen or syngas. This happens via electrolysis of steam or co-electrolysis of steam and carbon dioxide. These energy carriers can either be used as a buffer for fluctuating energy production, or used as transport fuels. The synthetic fuels are potentially carbon neutral, when the electricity comes from renewable energy production. Within the project “HydroCell” various advantageous system concepts for hydrogen production, including all components which are necessary to operate the electrolysis stacks, were identified. Furthermore, several electricity storage technologies where taken into comparison, with the goal to investigate the high temperature electrolysis technology’s potential for the energy sector. Two system concepts where identified, reaching electrical to chemical energy conversion efficiencies up to 79%. (1) Based on these results, two system concepts were developed, downscaled into “Proof-of-Concept” systems and analyzed on a test rig. Furthermore, suitable operating strategies were developed for an efficient and safe operation. The presentation will explain the theoretical background and will show the Proof-of-Concept system design and measurements performed in the project “HydroCell”. ___________ (1) Electrical conversion efficiency is defined by:

m … Mass of H2 output [kg] Pel … Electrical power [W] LHV … Lower heating value oh H2

Development of systems & balance of plant components Chapter 04 - Session A13 - 12/58

th

12 European SOFC & SOE Forum

www.EFCF.com/Lib ISBN 978-3-905592-21-4

5 - 8 July 2016, Lucerne/Switzerland

Introduction With the fast developments in the share of renewable electricity production in the energy mix, the issue of storing surplus electricity in periods of high production has gained a new dimension. Often conventional grid control mechanisms are overloaded, with the result being that power plants have to shut down more frequently. However, the concept idea of Power-to-Gas (PtG) promises to be a possible solution for handling huge amounts of electrical energy, when surplus electricity is used for electrolysis of water and/or coelectrolysis of water and CO2 with a product of H2, and/or CH4. (Figure 1)

Power demand

Power grid

e-

e-

SOEC electrolysis H2 storage H2

NG gas grid

CHP (Fuel Cell)

H2 O2

CO

Methanation CH4 storage

CH4

Figure 1: Connecting power to gas grid using SOEC/SOFC With conventional Alkaline Electrolysis (AEL) and Polymer Electrolyte Membrane Electrolysis (PEMEL) keeping high efficiencies for high capacity ranges while being cost effective is currently not foreseeable. Steam electrolysers with a ceramic membrane as an oxygen ion conductor (HTEL / SOEC / SOEL), operate at temperatures between 700 – 1000 °C. Thus, a major part of the required reaction energy can be supplied in form of heat, which shifts the thermodynamically equilibrium of splitting water further towards the elements of hydrogen and oxygen, which results in increasing electrical conversion efficiencies, while keeping high hydrogen production rates.

Development of systems & balance of plant components Chapter 04 - Session A13 - 13/58

th

12 European SOFC & SOE Forum

www.EFCF.com/Lib ISBN 978-3-905592-21-4

5 - 8 July 2016, Lucerne/Switzerland

The project “HydroCell” was an Austrian funded project by FFG, with a consortium coordinated by AVL List GmbH with the partners Plansee SE, Fraunhofer-Institut für Keramische Technologien und Systeme (Fraunhofer IKTS), Montanuniversität Leoben (MUL - Leoben University of Natural Resources and Applied Life Sciences) and Fraunhofer Austria Research GmbH. The aims of the project were:  Identifying, developing and assembling key components for 2x30 cells SOEC stacks  Developing, simulating and optimizing complete system concepts  Developing and testing Proof-of-Concept applications on the test rig  Developing a 3D Computational Fluid Dynamics (CFD) simulation platform for reliability analysis  Identifying requirements for an industrial scaled SOEC plant  Investigating application scenarios  Identifying conceptions for H2O/CO2 Co-Electrolysis with system-implemented downstream methanation The focus of this presentation is on the system conception part, comparing the simulation results with the Proof-of-Concept measurements at the test rig and finally conclusions of the results.

1. Scientific Approach Based on a performance map of 2x30cell stacks of Scandia doped Zirconia electroceramic and chromium based interconnects (CFY), an AVL simulation platform was used to identify, test and optimize various system concepts. These contained all necessary components to operate a SOEC stack (including heaters, vaporizers, compressors, heatexchanger, etc.). Based on the results, specifications for testing the 2x30cell stacks in hardware were developed and two most promising system conceptions were down scaled to Proof-of-Concept designs, manufactured, operated and characterized on a test rig at AVL.

2. Experiments/Calculations/Simulations Three system concepts were elaborated for the target product of hydrogen. These concepts were optimized in system simulations, aiming to identify the requirements for efficient hardware applications. The concepts are briefly described in the following. All system designs have included a partly recirculation of the fuel gas mass-flow following the stack’s outlet back to its inlet. With this concept it is possible to regulate a reduced atmosphere of H2O in H2 between 50 to 80mol%, in order to keep a reduced atmosphere on gas side of the SOEC stack. In the system conceptions 1 and 2 a part of the produced hydrogen is recirculated and reacts in oxidation catalysts with oxygen enriched air coming from the stacks air side outlet. Thereby, the produced heat energy will increase the temperature of the inlet streams. Concept number 2 differs from the first one in an additional mass recirculation at the air-path of the stack. The motivation was to further increase the sustainment of heat energy in the system. (Figure 2)

Development of systems & balance of plant components Chapter 04 - Session A13 - 14/58

th

12 European SOFC & SOE Forum

www.EFCF.com/Lib ISBN 978-3-905592-21-4

5 - 8 July 2016, Lucerne/Switzerland

Figure 2: Simplified flowchart illsutrating system conceptions 1 & 2 The third system concept is designed without recirculation of the produced hydrogen for heat recovery. Instead, the heat content of the outlet streams gets optimal recycled using heat exchangers, while the lack of heat energy gets provided using electrical gas heaters. (Figure 3)

Figure 3: Simplified flowchart illsutrating system conception 3 The presented systems in figures 2 and 3 were simulated in terms of stability, calibration and change of parameters, such as steam conversion rate (SC) and external heat supply. Based on the simulation results, concepts number 1 (Hex + hydrogen recirculation) and 3 (HEX + electrical heaters) were down scaled to Proof-of-Concept systems (PoC 2 and PoC 1 in the following) realized in hardware at a test rig at AVL. According to that the systems were reduced to the most essential components (stack module, vaporizer, heat Development of systems & balance of plant components Chapter 04 - Session A13 - 15/58

th

12 European SOFC & SOE Forum

www.EFCF.com/Lib ISBN 978-3-905592-21-4

5 - 8 July 2016, Lucerne/Switzerland

exchangers, gas heaters and oxidation catalysts). The media supply and hydrogen recirculation cycle was performed by the test rig (Figure 4 and 5). The core of the PoC systems was a module of 2x30 cells stacks (developed by IKTS and Plansee), which demonstrated a total electrical power consumption of about 5kW el at highest operation points. The system was tested based on previously developed operational strategies. This included heat up and cool down strategies as well a hot stand by and optimized hydrogen production modes.

Figure 4: Simplified flowchart of the Proof-of-Concept installation.

Figure 5: Proof-of-Concept in hardware at the AVL test rig. [1] = Hotbox containing the stack, [2] = gas heating module (electrical and/or with burners and HEX), [3] = Vaporizer, [4] = media supply, (5) = oxidation catalysts, (6) = heat exchanger With PoC1, using the electrical gas heaters only, a total of 64 points of operation were measured. The parameters varied were: Steam Conversion rate (SC) between 70%, 80% and 90%, temperature at the stacks air outlet between 750°C, 780°C, 800°C and 820°C, as well as the current consumptions between 30A to 50A in 5A steps. Analyzing these data resulted in a comprehensive picture of the stack- and system performance. Development of systems & balance of plant components Chapter 04 - Session A13 - 16/58

th

12 European SOFC & SOE Forum

www.EFCF.com/Lib ISBN 978-3-905592-21-4

5 - 8 July 2016, Lucerne/Switzerland

PoC2 was generally based on PoC1. Two oxidation catalysts were added to regulate the input streams’ temperature. Instead of measuring again all 64 points of operation, the most promising operation points were selected. Thus, the stack temperature was varied between 750°C, 780°C and 800°C and the current between 40A, 50A and 60A, while the steam conversion was fixed at SC=80±2%.

3. Conceptual and Experimental Results The simulation results established concepts 1 and 2 as most promising with the highest system efficiencies, while concept number 3 is the simplest and most stable in terms of calibration. Concept number 2 demonstrates slightly higher results for the electrical efficiency, but showed to be far more instable, caused by the additional feedback at the air path, which makes reaching steady points of operation more difficult (Figure 6).

Figure 6: Simulation results – electrical conversion efficiency over stack outlet temperature comparing system conceptions 1,2 and 3 All simulations suggested an ideal operation stack temperature between 780°C to 800°C. The additional heat input at higher temperatures is unproportional higher compared to the increase in hydrogen production. Therefore, an increase in hydrogen recirculation result in concepts 1 and 2 at temperatures above 800°C. Proof-of-Concept measurement results showed best efficiencies at higher steam conversion rates. A stable operation exceeding a steam conversion rate of 85% could not be demonstrated. A long term operation at steam conversions >85% is not recommended and may damage the stack. In terms of system performance, optimal temperatures were measured between 750°C and 780°C (Figure 7). In comparison, stack tests performed in furnace environment at IKTS, showed highest efficiencies at operation temperatures around 800°C.

Development of systems & balance of plant components Chapter 04 - Session A13 - 17/58

th

12 European SOFC & SOE Forum

www.EFCF.com/Lib ISBN 978-3-905592-21-4

5 - 8 July 2016, Lucerne/Switzerland

Figure 7: PoC2 measurement results – electrical conversion efficiency over steam conversion rate The overall Proof-of-Concept measuring results meets the simulation results satisfactory. With PoC2 an electrical system efficiency of 79% was reached at a stack outlet temperature of T=780°C, current density of i = 0,47A/cm2 (I=60A) and a steam conversion of SC=80%. Additional simulations including an optimal heat recovery at the system resulted in a potential for an electrical system efficiency up to 83%. (Figure 8)

Figure 8: Results overview comparing electrical conversion efficiency in simulation and measurements

Development of systems & balance of plant components Chapter 04 - Session A13 - 18/58

th

12 European SOFC & SOE Forum

www.EFCF.com/Lib ISBN 978-3-905592-21-4

5 - 8 July 2016, Lucerne/Switzerland

The significant difference between simulated and measured electrical conversion efficiency for PoC1 is mainly due to heat losses in the system, cased by the gas heater. When optimizing the heat recovery and thus minimizing heat energy losses, an electrical system efficiency potential of 70,5% is theoretically possible.

4. Conclusion and Outlook The overall project results are very satisfying. Estimated electrical system efficiencies of up to about 80% were demonstrated at the test rig. These results shows also that hightemperature electrolysis has significant advantages compared to conventional techniques (PEM ,alkaline). The “HYDROCELL” project has generated substantial knowledge and experience for the operation of SOEC stacks/moduls and first SOEC Proof of concept systems. Further developments shall increase performance improvements, especially when stack performance and heat sustainment were optimized. Also reverse applications like SOFC/SOEC systems are investigated in current research projects.

Development of systems & balance of plant components Chapter 04 - Session A13 - 19/58

th

12 European SOFC & SOE Forum

www.EFCF.com/Lib ISBN 978-3-905592-21-4

5 - 8 July 2016, Lucerne/Switzerland

A1305 SchIBZ – application of large diesel fuelled SOFC systems for seagoing vessels and decentralized onshore applications Keno Leites thyssenkrupp Marine Systems GmbH Hermann-Blohm-Str. 3, 20457 Hamburg, Germany Tel.: +49-431-700-1466 Fax: +49-431-700-16001466 [email protected]

Abstract Under the project name SchIBZ thyssenkrupp Marine Systems and 6 partners from industry and science developed a fuel cell system for seagoing vessels. The unique feature of this system is the use of low sulphur diesel oil as fuel. The system is based on solid oxide fuel cells coupled with a unique reforming unit for the diesel fuel and connected with an energy buffer. The components are modular to realize power outputs roughly between 50 and 500 kW per system. The advantages of the system are a high electrical efficiency, around 50%, very low gaseous emissions without exhaust gas treatment, low heat radiation and noise, very low maintenance due to few active components, possibility for heat recovery for further energy efficiency, high intrinsic redundancy and the potential to reduce the power installed on board. Additionally the availability of energy supply can be increased by decentralized installation of the units on board of oceangoing ships. Furthermore, the system offers advantages for transportable, remote power supply when installed in container. The consortium is right now in the phase of the construction of a 50 kW demonstrator which is going to be installed on-board a merchant vessel for several months for sea trials in 2016. It is planned to offer the system commercially after that successful test. Further development activities will comprise adaption to other fuels, improvements at the electrical side and scaling. This paper will present the results of tests as well as an outlook for further development of the technology and application.

Development of systems & balance of plant components Chapter 04 - Session A13 - 20/58

th

12 European SOFC & SOE Forum

www.EFCF.com/Lib ISBN 978-3-905592-21-4

5 - 8 July 2016, Lucerne/Switzerland

Introduction The global shipping is faced with ever more stringent emission regulations. While short sea shipping is evaluating battery electric solutions, ocean going vessels have, from today’s perspective, to further use hydrocarbon based fuels, to be able to store the necessary energy content in an acceptable volume. Still a favourite fuel is diesel oil, since the handling is very well known and the intrinsic safety is high. To reduce emissions from the usage in internal combustion engines, the engines have to be equipped with a number of auxiliary systems. These add not only space and weight, but also complexity and maintenance requirements. An elegant alternative solution would be a fuel cell system, if able to use diesel oil. Therefore thyssenkrupp Marine Systems decided to investigate the possibilities and develop a system, if it seems feasible. After rating the features of all configurations the combination of low sulphur diesel oil and high temperature fuel cells promised to be the best solution, although one with considerable development needs. To execute this development thyssenkrupp Marine Systems sought for partners with the respective know how. Finally the consortium consists of thyssenkrupp Marine Systems, DNV GL, sunfire, Oel-Waerme-Institut, Motion Control & Power Electronics, Leibniz-University Hannover and the ship owner Braren. Additionally funding by the German government was applied for under the NIP. The project is part of a so called lighthouse initiative, named e4ships. Actually, a 50 kWe demonstrator is under construction and shall be set to work on board a merchant ship this summer.

1. Development Approach For the development of the system a number of boundary conditions were chosen:  Use of proven components as far as possible  Development of a form factor for scaling from 50 to 500 kW approx.  Preparation for the external forces on board a seagoing ship  Matching to consumer network  Fully automated operation The development was structured in certain phases:  Selection of the process design  Experimental prove of the reforming process  System test in a small scale around 10 kW e  Construction of a large demonstrator for seaborne tests 1.1 Design of the fuel process Since fuel cells cannot convert liquid fuels directly, these have to be converted into a suitable fuel gas, by means of a reformer. The reformate gas from hydrocarbon fuels contains besides hydrogen also methane, carbon oxides and water vapour. Depending on the fuel cell type different mixtures of these can be facilitated. The composition of the reformate gas can be adjusted by certain process parameters and steps, which will not be described in detail in this paper.

Development of systems & balance of plant components Chapter 04 - Session A13 - 21/58

th

12 European SOFC & SOE Forum

www.EFCF.com/Lib ISBN 978-3-905592-21-4

5 - 8 July 2016, Lucerne/Switzerland

From the variety of fuel cell types currently on the market the two most suitable for ship borne use were selected for a comparison of process designs. The fuel cell types in question are:  PEMFC - polymer electrolyte membrane fuel cell  SOFC – solid oxide fuel cell With these two fuel cell types a study was performed, to design the necessary processes for an operation on diesel fuel. 1.1 (a) PEMFC The fuel gas quality requirements of the PEMFC are very stringent:  >70% H2 (depending on manufacturer)  no hydrocarbons  virtually no trace gases (e.g. CO)  no sulphur

Figure 1: PEMFC fuelled by road diesel [1] To use diesel oil as fuel a sophisticated gas conditioning and cleaning process is necessary (Figure 1). Therefore the system design incorporates a desulphurisation stage for the fuel gas. Even ultra-low sulphur fuel with 10 ppmS will poison the PEMFC in hours. A save value is below 50 ppb (gas concentration) [OWI]. An important factor in this system is the need for a dedicated burner to heat the reformer. This burner can make use of the anode off-gas but needs additional fuel depending on the system operation point. The added fuel reduces the overall efficiency of the system down to a range of 35% since it is not converted in electrical power. For comparison: the electrical efficiency of diesel engine driven generators reaches 40%. The process design shown in Figure 1 is one proposal which was developed in the project. Depending on the specific fuel cell and the catalysts used others are possible. Today simulation tools are Development of systems & balance of plant components Chapter 04 - Session A13 - 22/58

th

12 European SOFC & SOE Forum

www.EFCF.com/Lib ISBN 978-3-905592-21-4

5 - 8 July 2016, Lucerne/Switzerland

available which can be used to predict a suitable system layout with respect to the materials used. 1.1 (b) SOFC A more simple system can be achieved by using solid oxide fuel cells The fuel processing is different to the PEMFC. Due to the high internal temperature of SOFC of around 800 C there is enough excess heat to promote a reforming process with a high level of hydrogen production. Although the anode material is less sensitive to sulphur as in a PEMFC desulphurisation enhances performance and lifetime. This is realised with a nickel based reformer material, which serves as sulphur trap. [2] To make best use of the thermal energy content of the off-gas it is used to pre-heat the process gas and then oxidised in a catalytic burner for other pre-heating purposes. The advantage, that no surplus fuel is needed for the reforming allows high rates of system efficiency. Values of around 50% have been measured. Variations are possible, e.g. the use of the off-gas burner to heat the reformer by fuel injection in the start-up phase will be investigated. The process air is heated by the cathode off-gas to keep thermal energy inside the process. In comparison to PEMFC, MEAs of SOFCs are small with a stack size limited to around 5 kW e as of today. This requires an advanced design and thermal integration of the numerous stacks in order to attain a useful output, however SOFC offer better possibilities for varying the power output and an inherently higher system efficiency than the other types.

Figure 2: SOFC fuelled by road diesel [1] The system shown in Figure 2 is a basis for further improvements, especially in the areas of reformer and stack configuration potential to gain up to 10% more efficiency is expected. The adaption to other hydrocarbon fuels is possible with acceptable efforts. 1.2 Design of the electrical process Since SOFC are not capable of rapid load changes in so called isle networks, energy buffers are needed to cope with the transient load requirements. Based on the knowledge Development of systems & balance of plant components Chapter 04 - Session A13 - 23/58

th

12 European SOFC & SOE Forum

www.EFCF.com/Lib ISBN 978-3-905592-21-4

5 - 8 July 2016, Lucerne/Switzerland

about the possible load change rates from small scale tests, a calculation method was developed to size the energy buffer such that it suits the actual network load behaviour. The idea behind this concept is that the fuel cell provides an average current over some time and the energy buffer takes care of higher or lower loads by additionally feeding the network or receiving the surplus power from the fuel cell, as shown in Figure 3. The size of the fuel cell is selected according to the maximum condition in the load balance. For the selection of an energy buffer two values have to be determined:  the amount of energy which is contained in the integral between the load curves of the network and the fuel cells  the maximum short term peak currents of consumers like pumps or compressors or short circuit failures There is no general relation between fuel cell power and energy buffer size. In principal it can be said, that the energy buffer can be smaller the larger the network of a vessel is, because a large number of consumers and generators averages each other somewhat statistically.

Figure 3: Detail from the load graph of the test vessel MS FORESTER Fuel cells as well as the most energy buffers are DC sources. It is useful to connect them via a DC rail and then convert the total current to AC for the board network, see Figure 4. For this concept every power source is coupled with a DC/DC converter to provide a fixed voltage level at the DC rail.

Development of systems & balance of plant components Chapter 04 - Session A13 - 24/58

th

12 European SOFC & SOE Forum

www.EFCF.com/Lib ISBN 978-3-905592-21-4

5 - 8 July 2016, Lucerne/Switzerland

Figure 4: Basic concept of electric coupling of fuel cells and energy buffers ©Imtech Marine Germany GmbH

2. Experiments 2.1 Experimental development of the process For the design and construction of a demonstration plant detailed knowledge of the components and process parameters is necessary. The first important component is the reforming catalyst. This material is sensitive to operating temperatures and the activation procedure [2]. In course of the project SchIBZ the best operation conditions are evaluated and proven by a 3000 h test. The reformate gas composition is of a very high quality and fits the requirements of the SOFC, see Figure 7. In the following a test plant with reformer and 10 kW SOFC was built as next step, see Figure 6. The test operation for more than 1000 h showed very good values for performance, efficiency and degradation. Based on these experiences the final design of the demonstration plant is developed with support of a simulation model [4].

Figure 5: test rig for basic catalyst evaluation ©OWI GmbH

Figure 6: Test stand with 10 kW SOFC system for diesel fuel oil ©OWI GmbH

Development of systems & balance of plant components Chapter 04 - Session A13 - 25/58

th

12 European SOFC & SOE Forum

www.EFCF.com/Lib ISBN 978-3-905592-21-4

5 - 8 July 2016, Lucerne/Switzerland

Figure 7: Reformate gas composition from adiabatic reforming [3] 2.2 Testing of the components In order to achieve unit ratings of several 100 kW a basic fuel cell module is defined and developed. It was decided that this module shall have dimensions which allow it to be moved on board along the normal passageways if needed and stacked 2 or 3 rows high by 2 columns wide. Under this restriction a module with dimensions of 800 by 800 by 2700 mm was the result. An output of about 50 kW will be available with the actual generation of fuel cells. Special consideration was given to the even distribution of the process gases over the several stacks inside the module. Finally the insulation was designed to achieve a surface temperature below 60 C and equal temperatures even for the stacks at the ends of the module.

Figure 8: basic module in test booth ©sunfire GmbH To ensure the suitability of the fuel cell modules for ship borne use, some physical tests are made. Every component on board a ship must stand certain inclination angles and accelerations. The test conditions are defined by the classes. Two tests were performed:  Inclination test with 22,5° (Figure 9)  Acceleration test with DNV GL vibration profile and single shocks (Figure 10)

Development of systems & balance of plant components Chapter 04 - Session A13 - 26/58

th

12 European SOFC & SOE Forum

www.EFCF.com/Lib ISBN 978-3-905592-21-4

5 - 8 July 2016, Lucerne/Switzerland

Both tests are performed with cold and hot stacks to verify that operating and nonoperating systems can stand the forces acting on a vessel in rough seas. The continuous vibrations have accelerations up to 2g and the shocks up to 6g. In each test no failures were observed but the necessity of a better fixing of the insulation. The rigid perlite plates chafed at each other with the standard fixation, which was designed for stationary use. After a modification no wear is observed.

Figure 9: Inclination test at 22,5° ©sunfire GmbH

Figure 10: Vibration and shock test ©thyssenkrupp Marine Systems GmbH

For the connection of the fuel cells to the board network galvanically separated DC/DC converters, based on commercial models, are developed by the project partner M&P. The expected fuel efficiency is based on the individual test results estimated for the complete system. Figure 11 shows the efficiency of the fuel cell system compared with a typical diesel engine genset over the load range. It can be seen that the efficiency advantage of the fuel cell system is roughly between 25 and 50%.

Figure 11: expected fuel cell efficiency (blue) and diesel genset efficiency (orange), ©sunfire GmbH, thyssenkrupp Marine Systems GmbH 2.3 The demonstration plant To verify the seaworthiness of the power supply system a 50 kW demonstrator was designed and will be commissioned this summer. A failure mode and effect analysis (FMEA) was performed to ensure that it fulfils the class requirements of DNV GL for main Development of systems & balance of plant components Chapter 04 - Session A13 - 27/58

th

12 European SOFC & SOE Forum

www.EFCF.com/Lib ISBN 978-3-905592-21-4

5 - 8 July 2016, Lucerne/Switzerland

power supplies on merchant ships. Based on this the approval process is started to gain an approval for experimental use. This will be the basis for a type approval for commercial applications later. Accompanying work is done to support the development of fuel cell safety regulations (IGF code) from the International Maritime Organisation. Especially the safety analyses are provided for the discussion in the respective working group [5]. To improve the availability of fuel cell power hot gas valves are installed in the fuel gas supply lines to the fuel cell modules. These valves allow cutting off one module from the system in case of a failure. Therefore only the power of one basic module is lost instead of several 100 kW. This feature is hardly possible with combustion engines. The core of the demonstration plant is a 40” container, carrying in two spaces the process components (fuel cell modules, reformer module) and the electronic components (cabinets with power electronics, automation and monitoring systems, the energy buffer and the generator switchboard for the connection to the board network), see Figure 12. The container is necessary to mount the test system on deck of the testing vessel as additional power supply to the network. The volume of the container could accommodate up to 200 kW SOFC power plus the necessary electronic systems.

Figure 12: 50 kW e SOFC demonstration power supply system for operation on ULSD ©thyssenkrupp Marine Systems GmbH The complete test set up is shown in Figure 13 and the location on board MS FORESTER of Reederei Braren in Figure 14. The demonstration plant is a complete power station which can operate as standalone unit. For this reason it comprises additionally to the power supply system an auxiliary container with ventilation, air filtration and safety systems like firefighting and a tank container. The ventilation is dimensioned according to the requirements from the IMO IGF code for spaces containing gas fuelled machinery. These spaces must be either inherently safe or equipped with sufficient ventilation to prevent an explosive atmosphere under any circumstances (ESD concept) with underpressure against surrounding spaces. These facilities are normally part of the vessels infrastructure.

Development of systems & balance of plant components Chapter 04 - Session A13 - 28/58

th

12 European SOFC & SOE Forum

www.EFCF.com/Lib ISBN 978-3-905592-21-4

5 - 8 July 2016, Lucerne/Switzerland

Figure 13: standalone power supply system for technology demonstration on board a ship

Figure 14: artists’ impression of the installation of the demonstration plant at MS FORESTER

Development of systems & balance of plant components Chapter 04 - Session A13 - 29/58

The vessel has a power consumption between 100 and 200 kW. The supply is shared between the on board generator and the fuel cell system such, that the short term load variations are covered by the energy buffers and the base load according to Figure 3 is covered by both aggregates proportionally. In this configuration the fuel cell supplies between 25 and 50% of the power consumption, which has not been demonstrated before. The control and monitoring system is designed for a fully automatic operation of the plant with just an observation monitor and restricted user interface for the crew, inside the machinery control room.

3. Results The results achieved during the development stages underline the assumption, that fuel cells are a viable alternative to conventional power supply for ocean going ships. Beneath the lower emissions and improved maintenance efforts they offer the possibility for the design of a distributed electrical network, which will increase redundancy significantly.

Figure 15: schematic concept for a distributed installation of fuel cell systems in a passenger vessel ©thyssenkrupp Marine Systems GmbH In long term perspective, the application of fuel cell systems will result in the (re)introduction of DC network. This opens opportunities for further efficiency and cost improvements. In addition to that the selected form factors of the components are well suited for an installation in standard containers. Based on the presented demonstration plant also applications onshore, for temporary or permanent base load supply of facilities without stable network connections and the need for liquid fuels, are a promising field. By interconnection of the DC terminals powers in the MW range are possible.

www.EFCF.com/Lib

th

12 European SOFC & SOE Forum

www.EFCF.com/Lib ISBN 978-3-905592-21-4

5 - 8 July 2016, Lucerne/Switzerland

References [1] [2] [3] [4] [5]

K. Leites, S. Krummrich, P. Nehter, A. Bauschulte, N. Kleinohl, W. Kühnau, S. Claussen (2013), ‘SchIBZ – Application of solid oxide fuel cells for oceangoing ships’, 5th international conference on Fundamentals and Developments of Fuel Cells Pedro Nehter, Nils Kleinohl, Ansgar Bauschulte, Keno Leites (2014), ‘Diesel based SOFC APU for marine applications‘,11th European SOFC and SOE Forum 2014 Nils Kleinohl (2015), ‘SOFC fed by pre-reformer’, ASTM Materials: Performance and Characterization, Special Issue on CURRENT STATUS AND FUTURE ADVANCES IN FUEL CELL TECHNOLOGY Michael Dragon (2014), ‘Simulation und exergetische Verlustanalyse eines SOFCBrennstoffzellensystems mit Anodengasrezirkulation‘, Berichte aus dem Institut für Thermodynamik, Leibniz-Universität-Hannover IMO SUB-COMMITTEE ON CARRIAGE OF CARGOES AND CONTAINERS (2015), Report of the working group‘, CCC 2/WP.3

Abbreviations AC Braren DC DNV GL HT-PEM IGF IMO LUH M&P MCFC MEA OWI PEMFC SchIBZ SOFC sunfire

alternating current Rörd Braren GmbH direct current Det Norske Veritas Germanischer Lloyd High Temperature Polymer Electrolyte Membrane Fuel Cell International Code for the use of gas as fuel International Maritime Organisation Leibniz-University Hannover motion control and power electronics GmbH Molten Carbonate Fuel Cell Membrane Assembly Oel-Waerme-Institut GmbH Polymer Electrolyte Membrane Fuel Cell SchiffsIntegration BrennstoffZelle Solid Oxide Fuel Cell sunfire GmbH

Acknowledgments The consortium of the project SchIBZ would like to thank the Federal Ministry of Transport and Digital Infrastructure (BMVI) and the National Organisation for Hydrogen and Fuel Cell Technology (NOW) for the continuing support under the National Innovation Program for Hydrogen and Fuel Cells (NIP).

Development of systems & balance of plant components Chapter 04 - Session A13 - 31/58

th

12 European SOFC & SOE Forum

www.EFCF.com/Lib ISBN 978-3-905592-21-4

5 - 8 July 2016, Lucerne/Switzerland

A1306 (Published in EFCF Special Issue Series, www.EFCF.com/LIB) Development of a SOFC/Battery-Hybrid System for Distributed Power Generation in India Thomas Pfeifer, Mathias Hartmann, Markus Barthel, Jens Baade, Ralf Näke, Christian Dosch Fraunhofer IKTS Winterbergstraße 28 D-01277 Dresden / Germany Tel.: +49-351-2553-7822 Fax: +49-351-2554-302 [email protected]

Abstract In recent years India faces demanding challenges in covering an aggressively increasing electricity consumption through economic growth and progressive consumer requirements. Renewable sources and small distributed power generators have been identified as one of the options to establish a diversified power supply infrastructure. The present situation offers promising opportunities for fuel cell systems, as “grid-parity” is not the common measure of competitiveness, but rather the installation speed and availability of reliable power sources. Contracted by the company h2e Power Systems Pvt. Ltd. based in Pune, India, Fraunhofer IKTS has developed a 1 kW(el) SOFC power generator during a three-year system engineering and technology transfer project. The fuel cell system is based on the CFY stack technology by Plansee SE and IKTS, incorporating state-of-the-art ESC with Scandia-doped Zirconia electrolytes. CFY-stacks have proven to be robust and reliable, showing power degradation rates below 0.6 % per 1.000 hours during endurance operation over 20.000 hours and a cyclization capability of more than 120 near-system cycles under full RedOx-conditions. For the SOFC power generator a CFY stack is integrated with a pre-reformer, a tail-gas oxidizer and heat exchangers into a HotBox-module following a novel concept for leastspace-demanding reactor integration and flow distribution. Aside from compactness, a simple and robust, yet highly efficient system concept was set as the primary development goal for the project. To meet these requirements, two major design decisions have been introduced in the process layout, i.e. a rated fuel utilization in the stack of 85 % as well as a POX-air pre-heater for reducing the reformer air flow to lowest possible values. This approach leads to a water-less SOFC system with a net electrical efficiency above 40 %. In 2015, two Proof-of-Concept (PoC) prototype systems were commissioned and tested at IKTS. One of the PoC-prototypes was installed later at the customer’s laboratory in Pune, India, for test and demonstration purposes. In Project Phase II, three improved prototype systems were built at IKTS and shipped to India in May 2016 for initial demonstration projects and field trials. At the same time, the technology transfer to the customer was initiated, in order to enable for a local manufacturing and deployment of SOFC systems in India. By completion of the prototype delivery and technology transfer, the initial development project was successfully finished. Various follow-up activities are currently under negotiation between h2e Power Systems and Fraunhofer IKTS. Remark: The full paper is published in EFCF Special Issue Series (www.EFCF.com/LIB, SI EFCF 2016) in Journal "FUEL CELLS - From Fundamentals to Systems". Development of systems & balance of plant components Chapter 04 - Session A13 - 32/58

th

12 European SOFC & SOE Forum

www.EFCF.com/Lib ISBN 978-3-905592-21-4

5 - 8 July 2016, Lucerne/Switzerland

A1307 Sulfur Tolerant WGS-Catalysts Thorsten Dickel (1), André Weber (1) Michael Scharrer (2), Claus Peter Kluge (2) (1) Institute for Applied Materials (IAM-WET), Karlsruhe Institute of Technology (KIT), Adenauerring 20b, D-76131 Karlsruhe/Germany Tel.: +49-721-608-47494 Fax: +49-721-608-47492 [email protected]

(2) CeramTec GmbH CeramTec-Weg 1 D-95615 Marktredwitz

Abstract The sulfur content in fuels as reformed natural gas or diesel results in a severe power loss of state of the art anode supported SOFCs. Previous studies showed that the performance loss is related to a sulfur poisoning of the Ni/YSZ-cermets resulting in a deactivation of (i) the catalytic watergas-shift-reaction (WGS) at the nickel surfaces and (ii) the electrooxidation of hydrogen at the triple phase boundaries (TPB). In a first step towards a sulfur tolerant SOFC different ceria and nickel/ceria catalysts were investigated with respect to sulfur poisoning of the WGS-reaction. The catalysts were applied onto dense zirconia substrates and tested in a single cell test bench that enables in-operando gas analysis along the gas channel. A model fuel consisting of 50% CO and 50% H2O was applied. To study sulfur poisoning 2 ppm of H2S was added. The experiments revealed that pure ceria exhibits a low catalytic activity but a good sulfur tolerance. Nickel showed a significantly higher initial catalytic activity but strong poisoning effects. In case of Ni/ceria cermets exhibiting an appropriate microstructure and layer thickness high catalytic activity and excellent sulfur tolerance can be achieved. The results indicate that the sulfur tolerance is increasing with the density of TPBs between ceria, nickel and the fuel. The conversion rate and its stability in sulfur containing fuels increases with the thickness of the catalyst layer. The applicability of such Ni/ceria catalyst layer was validated by performance tests of state of the art anode supported cells operated with a sulfur containing diesel reformate. Due to the additional hydrogen generated by the WGSreaction in this sulfur tolerant catalyst layer the cell performance was increased by 32%.

Development of systems & balance of plant components Chapter 04 - Session A13 - 33/58

th

12 European SOFC & SOE Forum

www.EFCF.com/Lib ISBN 978-3-905592-21-4

5 - 8 July 2016, Lucerne/Switzerland

Introduction One of the major problems of SOFCs working under real conditions are the sulfur compounds contained in almost all available fuels such as natural gas, diesel reformat or gas from hydrothermal gasification of biomass [1],[2]. Depending on the fuel composition, which usually contains different amounts of hydrogen, water vapor, carbon monoxide, carbon dioxide, methane and nitrogen, the performance loss due to sulfur poisoning can be severe. Impedance analysis of anode supported SOFCs revealed that the polarization resistance of the Ni/YSZ anodes increased by a factor up to 5 after adding small amounts of H2S (below 1 ppm). In addition the CO conversion by the watergas-shift-reaction (WGS) was deactivated [3],[4]. Considering a diesel reformate with CO to H2 ratio of approx. 1, the deactivation of the WGS results in a fuel- and thus efficiency-loss of 50% if CO is no longer converted in the stack. In these studies it was shown that the losses are mainly due to a reduced activity of the nickel acting as a catalyst for the WGS as well as an electrocatalyst for the electrooxidation of hydrogen at the three phase boundary (TPB). These two effects can be explained by the occupation of the nickel surfaces by sulfur. In figure 1 a reaction scheme for reformate operation and the effects of sulfur poisoning is illustrated.

figure 1: paths of gas diffusion, ionic transport and electrical conduction in a SOFC-Anode, sulfur occupies the nickel surface and inhibits the shift-reaction on the Ni-surfaces and the electro-oxidation at the TPBs Sulfur is known to occupy metallic surfaces such as Nickel reversibly, depending on the concentration of hydrogen sulfide in the gas phase and the temperature as shown in equation (1) which was proposed by Hansen [5],

with the sulfur coverage on the Ni surface S. The sulfur chemisorbs at the free nickel surface as shown in equation (2), where Ni* stands for a free Ni-site at the surface. Ceria as well as gadolinium-doped ceria (GDC) as mixed ionic electronic conductors (MIEC) in reducing atmospheres allow different reaction mechanisms (figure 2), which can help to improve the performance of the Nickel-ceria-catalyst in sulfur poisoning conditions. Development of systems & balance of plant components Chapter 04 - Session A13 - 34/58

th

12 European SOFC & SOE Forum

www.EFCF.com/Lib ISBN 978-3-905592-21-4

5 - 8 July 2016, Lucerne/Switzerland

figure 2: two reaction mechanisms that can be promoted by a Nickel-GDC-catalyst: a) the oxidation of absorbed sulfur b) the oxidation of CO at the TPB

The first mechanism includes the WGS via the TPB: water dissociates at the GDC surface, where it forms oxygen ions and hydrogen. The ions react at the TPB with the carbon monoxide adsorbed on the nickel surface to form carbon dioxide. In the second mechanism chemisorbed sulfur reacts with oxygen ions at the TPB to form sulfur dioxide. These two mechanisms show that particularly a high number of TBPs can help to improve the catalytic activity under sulfur poisoning conditions. Thus Ni/ceria catalysts are well established in the field of catalysis [6]. In order to increase the amount of hydrogen in sulfur containing reformates such a WGS-catalyst can be added on the anode backbone or in the gas channel. The advantages of Ni/ceria anodes with respect to sulfur poisoning have already been shown several times. In a number of papers an acceptable to excellent stability of the performance of cells exhibiting a Ni/ceria anode has been proven. Mostly these experiments were carried out on electrolyte supported cells operated at temperatures of 850 °C and above [7],[8]. Further on in these publications the effect of catalyst (steam reforming, WGS) and electrocatalyst (anode polarization resistance) poisoning is summarized in an overall performance loss. In this contribution our work towards a sulfur tolerant intermediate temperature (750 °C) SOFC, namely the integration of sulfur-tolerant, ceria based WGS-catalysts is shown (sulfur tolerant anodes will be addressed in reference [11]). Next to model structures revealing the catalytic properties of nickel, ceria and nickel/ceria systems, performance and stability of advanced catalyst layers is presented.

Experimental A number of different model structures (MS) and Ni/ceria catalyst layers (CL) were prepared by applying the catalyst materials onto an 8YSZ-substrate. The active area that is in contact with the gas was 40 x 40 mm² for all samples. - MS1: - MS2: -

Ni-sheet with planar surface screen printed GDC layer on 8YSZ-substrate, sintered at 1650 °C (dense layer) to obtain a dense layer with a smooth surface (see fig. 4) MS3: screen printed GDC layer on 8YSZ-substrate, “sintered” only in operando during heating up to 750 °C and operation (highly porous) MS4: very thin Ni/GDC layer on 8YSZ-substrate, unsintered MS5: very thin GDC-layer on Ni-coated 8YSZ-substrate, unsintered CL1…6: Ni/ceria layers differing in thickness

Development of systems & balance of plant components Chapter 04 - Session A13 - 35/58

th

12 European SOFC & SOE Forum

www.EFCF.com/Lib ISBN 978-3-905592-21-4

5 - 8 July 2016, Lucerne/Switzerland

To test the performance of the abovementioned samples in terms of the WGS-reaction and the impact of sulfur poisoning, they were mounted in a specially designed single cell housing located in the furnace of a test bench [9],[10]. The housing enables a gas extraction at five points along the gas channel. The extracted gas was analyzed by a Varian CP 4900 micro GC to evaluate the gas conversion along the gas channel. All experiments were executed at 750 °C applying an inlet gas composition of 50 % H2O and 50 % CO at a flow rate of 0.6 slm (vg = 0.72 m/s). For the sulfur poisoning 2 ppm of H2S was added.

figure 3: simplified sketch of the single cell housing and the sample located in the furnace of a test bench

Results As previously discussed, the material combination of nickel and ceria is a promising catalyst to promote the WGS reaction. In a first step the two materials were examined separately (MS1, MS2) and in combination in thin layers (MS3 … 5) to avoid additional transport losses due to gas diffusion in the layer. SEM pictures of these samples are shown in figure 4. The CO conversion rate (100% conversion at equilibrium conditions) before and after poisoning with 2 ppm H2S is displayed in figure 5. The comparison of the dense structures MS1 and MS2 reveals that the catalytic activity of nickel is very high before H2S-poisoning whereas GDC shows a quite low conversion rate. After adding 2 ppm of H2S to the fuel a relative drop in conversion rate of 85% at the Ni(MS1) and the 36% at the GDC-surface is observed. It should be noted that the comparison of the SEM-images of MS1 and MS2 (fig. 4) after the test reveals a much rougher surface and thus a larger surface area for Ni (MS1). The relative drop in conversion rate of the porous GDC-layer (sample MS3, 38%) is quite similar to that of the dense GDC sample. In case of the samples MS4 and MS5 the aim was to achieve a small number of TPBs between Ni, GDC and the fuel. In case of MS4 the very thin Ni/GDC layer printed onto the 8YSZ-substrate hardly exhibits any TPBs as Ni and GDC particles are distributed over the substrate surface. In case of MS5 the GDC-particles are in contact with the Ni thin film sputtered on the 8YSZ substrate. Thus a limited number of contact points will definitely be there. The comparison of MS1 and MS5 shows a smaller relative drop in conversion rate for MS5 (79%) vs. MS1 (85%).

Development of systems & balance of plant components Chapter 04 - Session A13 - 36/58

th

12 European SOFC & SOE Forum

www.EFCF.com/Lib ISBN 978-3-905592-21-4

5 - 8 July 2016, Lucerne/Switzerland

A significant improvement with respect to the drop in conversion rate due to sulfur poisoning can be achieved with the Ni/ceria catalyst layers (CL1 and 6). The Ni/ceriacermets show a much higher initial conversion rate which only decreases slightly after sulfur is added to the fuel.

MS1

MS2

MS3

MS4

MS5

figure 4: SEM pictures of the different model structures taken after sulfur poisoning experiment at 750 °C; MS1: solid nickel surface; MS2: sintered ceria surface (1650 °C); MS3 not sintered porous GDC layer; MS4: single layer Ni/GDC catalyst; MS5: thin layer of ceria screen-printed on a sputtered nickel surface

Development of systems & balance of plant components Chapter 04 - Session A13 - 37/58

th

12 European SOFC & SOE Forum

www.EFCF.com/Lib ISBN 978-3-905592-21-4

5 - 8 July 2016, Lucerne/Switzerland

conversion rate

100% 90%

unpoisioned

80%

poisioned

70% 60% 50% 40% 30% 20% 10% 0%

figure 5: conversion rates of the different model structures and two Ni/ceria catalyst layers, the conversion is displayed in relation to the chemical equilibrium

figure 6: durability experiment of Ni/ceria catalyst layers differing in thickness In fig. 6 the molar fraction of CO2 resulting from the WGS reaction along the gas channel is displayed for six Ni/ceria catalyst layers differing in thickness. Before sulfur poisoning all samples show a CO2-content of approx. 25%. At t = 0 an amount of 2 ppm H2S was added to the fuel. The conversion rate drops quickly, which is expected to be caused by a rapid occupation of the Ni-surface by sulfur. The drop increases with decreasing layer thickness. Due to the high number of TPBs in these layers a high conversion rate is maintained.

Development of systems & balance of plant components Chapter 04 - Session A13 - 38/58

th

12 European SOFC & SOE Forum

www.EFCF.com/Lib ISBN 978-3-905592-21-4

5 - 8 July 2016, Lucerne/Switzerland

After the initial drop a second slow degradation mechanism is observed. The degradation rate is decreasing with increasing layer thickness. Microstructure analysis revealed that this degradation can most likely be attributed to Ni-agglomeration and thus a reduction in TPB-density of the layers. The related degradation rate decreases with time, a durability test over 750 h revealed that in the timeframe 100 h to 750 h the decrease in CO2concentration was only 2.2%.

Concept validation in a single cell test To validate the applicability of the Ni/ceria catalyst layer, single cell tests with and without the examined WGS-catalyst were performed. Anode supported cells fabricated by CeramTec, exhibiting a Ni/8YSZ anode functional layer and substrate, were tested in the abovementioned single cell test bench. The tests were performed with a simulated diesel reformate composed of H2, H2O, CO, CO2 and N2. For the sulfur poisoning experiment 2 ppm of H2S were added. In fig. 7 current-voltage-characteristics of the cells are shown. In case of the H2S-free reformate (dotted lines) the cell without catalyst shows a slightly higher performance and fuel utilization (f.u.) at 0.7 V. This has to be attributed to the additional gas diffusion polarization in the catalyst layer.

Figure 7: performance of anode supported cells with and without WGS-catalyst without and with 2 ppm of H2S in the reformate fuel at an operating temperature of 750°C In case of 2 ppm H2S in the reformate a significant decrease in performance is observed for both cells. The cell without catalyst shows the lowest performance. The slope of the CV-characteristic indicates a strong gas diffusion or conversion at current densities exceeding 0.2 A/cm². At 0.7 V the fuel utilization is only 38% if hydrogen and carbon monoxide are considered as oxidizable species. If we are only considering the supplied hydrogen as the fuel and assume, that CO is not converted at all, the fuel utilization f.u.(H2,supplied) would be 83%. Thus, it is comprehensible that a gas conversion limitation occurs at higher current densities. In case of the cell with Ni/ceria catalyst no diffusion or conversion limitation is visible. At 0.7 V a fuel utilization of 51% is achieved, which would corresponds to a f.u.(H 2,supplied) of 110%. As a f.u. exceeding 100% is impossible it is obvious that carbon monoxide is

Development of systems & balance of plant components Chapter 04 - Session A13 - 39/58

th

12 European SOFC & SOE Forum

www.EFCF.com/Lib ISBN 978-3-905592-21-4

5 - 8 July 2016, Lucerne/Switzerland

converted by the additional WGS catalyst, maintaining a sufficient amount of hydrogen in the fuel gas. To support this result, gas analysis was performed at P1 to P5 along the gas channel. The measured conversion of carbon monoxide is displayed in fig. 8. In case of a H 2S-free reformate close to 70% of the CO is converted at a fuel utilization of 70% for both cells. In case of 2 ppm of H2S in the fuel less than 10% of the supplied carbon monoxide is converted without catalyst, whereas with catalyst close to 50% of CO is converted at 51% fuel utilization.

Figure 8: conversion of carbon monoxide along the gas channel of the tested cells at an operating temperature of 750°C Considering the abovementioned conditions, the addition of the sulfur tolerant Ni/ceria catalyst at the anode results in a conversion of CO by the WGS reaction and translates into an additional power output of 32%. Nevertheless a decrease in performance is observed in comparison to the cell tests in a H2S-free reformate. This decrease has to be attributed to the poisoning of the TPB in the anode functional layer. To achieve a further performance improvement, the integration of a sulfur tolerant anode functional layer is required.

Conclusions Ni/ceria is an excellent WGS catalyst for sulfur containing fuels that is able to convert carbon monoxide and steam into hydrogen and carbon dioxide even at intermediate operating temperatures of 750 °C. The results on model structures indicate that a high density of TPBs results in an improved sulfur tolerance. The investigated Ni/ceria catalyst layers show a high sulfur tolerance and a good durability. A successful integration of the catalyst into a single cell is demonstrated. The additional layer enables the conversion of the carbon monoxide in the reformate and thus increases the cell performance by 32% relative to the sulfur-poisoned cell without catalyst.

Development of systems & balance of plant components Chapter 04 - Session A13 - 40/58

th

12 European SOFC & SOE Forum

www.EFCF.com/Lib ISBN 978-3-905592-21-4

5 - 8 July 2016, Lucerne/Switzerland

Acknowledgements Funding by the Bundesministerium für Wirtschaft und Technologie (BMWi 03ET2048, 03ET6056B and 03ET6056E) is gratefully acknowledged.

References [1] Y. Matsuzaki and I. Yasuda, "The poisoning effect of sulfur-containing impurity gas on a SOFC anode Part I. Dependence on temperature, time, and impurity concentration", Solid State Ionics 132, pp. 261-269 (2000). [2] K. Föger and K. Ahmed, "Catalysis in high-temperature fuel cells", Journal of Physical Chemistry B 109, pp. 2149-2154 (2005). [3] A. Kromp, S. Dierickx, A. Leonide, A. Weber and E. Ivers-Tiffée, "Electrochemical Analysis of Sulfur-Poisoning in Anode Supported SOFCs Fuelled with a Model Reformate", J. Electrochem. Soc. 159, p. B597-B601 (2012). [4] A. Weber, S. Dierickx, A. Kromp and E. Ivers-Tiffée, "Sulfur Poisoning of AnodeSupported SOFCs under Reformate Operation", Fuel Cells 13, pp. 487-493 (2013). [5] J. B. Hansen, "Correlating Sulfur Poisoning of SOFC Nickel Anodesby a Temkin Isotherm", Electrochemical & Solid-State Letters 11, p. B178-B180 (2008). [6] U. Hennings, Sulfur-tolerant Natural Gas Reforming for Fuel-cell Applications, KIT Scientific Publishing (2010). [7] J. P. Ouweltjes, P. V. Aravind, N. Woudstra and G. Rietveld, "Biosyngas utilization in solid oxide fuel cells with Ni/GDC anodes", Journal of Fuel Cell Science and Technology 3, pp. 495-498 (2006). [8] E. Batawi, U. Weissen, A. Schuler, M. Keller and C. Voisard, "Cell manufacturing processes at Sulzer Hexis", in Proceedings of the 7th International Symposium on Solid Oxide Fuel Cells VII, pp. 140-147 (2001). [9] D. Fouquet, D. Klotz, E. Dannhäuser, A. C. Müller, A. Weber and E. Ivers-Tiffée, "SOFC single cell test setup for the use of various hydrocarbons", in S. C. Singhal and M. Dokiya (Eds.), Proceedings of the Eigth International Symposium on Solid Oxide Fuel Cells (SOFC-VIII), pp. 1167-1169 (2003). [10] H. Timmermann, W. Sawady, R. Reimert and E. Ivers-Tiffée, "Kinetics of (reversible) internal reforming of methane in solid oxide fuel cells under stationary and APU conditions", J. Power Sources 195, pp. 214-222 (2010). [11] see B0507, this proceeding

Development of systems & balance of plant components Chapter 04 - Session A13 - 41/58

th

12 European SOFC & SOE Forum

www.EFCF.com/Lib ISBN 978-3-905592-21-4

5 - 8 July 2016, Lucerne/Switzerland

A1309 (Abstract only, published elsewhere) Control strategy for a SOFC gas turbine hybrid power plant Moritz Henke (1), Mike Steilen (1), Ralf Näke (2), Marc Heddrich (1), K. Andreas Friedrich (1) (1) German Aerospace Center (DLR), Pfaffenwaldring 38-40, 70569 Stuttgart, Germany (2) Fraunhofer IKTS, Winterbergstraße 28, 01277 Dresden, Germany Tel.: +49-711-6862-795 [email protected]

Abstract Today, gas steam combined cycle plants are the most efficient power plants converting chemical energy into electrical energy with installed powers of usually several hundred MW. They are technologically mature reaching electrical efficiencies of 60 % (based on the lower heating value). Hybrid power plants consisting of solid oxide fuel cells (SOFC) and a gas turbine (GT) can reach higher electrical efficiencies and can be built at lesser installed power. The general concept of a SOFC/GT hybrid power plant is to use the hot SOFC exhaust gases to drive a gas turbine. High electrical efficiencies are achieved if SOFC and gas turbine match well. In the past few years a hybrid power plant with an electrical power output of 30 kW has been designed and is currently under construction at DLR. One challenge concerning the operation of the hybrid power plant is the control of the hybrid system. Stand-alone gas turbines can react comparatively fast to load changes and quickly reach a new stationary operating point. SOFC system temperatures react much slower due to their large thermal capacity. The operating strategy of the hybrid system needs to ensure a safe and reliable operation and allow for high electrical efficiency in a wide power range. Furthermore, dynamic operation like start-up, load changes, shut-down and emergency procedures are also considered. This work will give an overview of the control concept and show how the requirements are met. Remark: Only the abstract is available, because the authors chose to publish elsewhere. Please see Presentations on www.EFCF.com/LIB or contact the authors directly.

Development of systems & balance of plant components Chapter 04 - Session A13 - 42/58

th

12 European SOFC & SOE Forum

www.EFCF.com/Lib ISBN 978-3-905592-21-4

5 - 8 July 2016, Lucerne/Switzerland

A1312 (Abstract only, published elsewhere) rSOC plant concept for renewable energy storage Matthias Frank (1), Roland Peters (1), Van Nhu Nguyen (1), Robert Deja (1), Ludger Blum (1), Detlef Stolten (1,2) (1) Juelich Research Center IEK-3: Electrochemical Process Engineering 52428 Juelich/Germany (2) RWTH Aachen University Lehrstuhl für Brennstoffzellen, Fakultät für Maschinenwesen 52072 Aachen/Germany Tel.: +49-2461-614394 Fax: +49-2461-616695 [email protected]

Abstract Since 2015, a reversible solid oxide cells plant (rSOC) in the kW-class has been developed at Forschungszentrum Jülich. Based on steam, hydrogen and air; the rSOC plant is environmentally friendly. One of the major benefits of an rSOC plant is that it can be operated in either electrolysis (SOEC) or fuel cell mode (SOFC). This is ideal system to deal with the time discrepancy occurring between energy demand and supply in most renewable energy resources (wind, solar…). At times of high energy surplus, the plant can run in SOEC mode, thereby electrolyzing water into a storable gaseous fuel, H 2. At a later time of energy demand, the rSOC plant can be run in fuel cell (SOFC) mode, producing electricity from the hydrogen. The plant requires a single stack which can be used for both modes. Based on data of previous solid oxide cells stacks carried out at Jülich, a model of an rSOC plant was developed. A balance of plant able to supply the stack with gas compositions required for the two modes was established. Importantly the saturated steam needed for SOEC mode is generated inside the rSOC plant. Different plant concepts were examined and compared, especially in order to increase the overall efficiency of the plant. Concerning heat management, in-depth analysis and optimization of waste heat recovery was carried out. Off-gas recycling was also implemented both in SOFC and SOEC modes. In SOFC mode, anode off-gas recirculation enables the system to reach higher fuel utilizations than of the fuel cell stack alone. In SOEC mode, hydrogen recirculation makes it possible to limit the use of the gas tank to the start-up phase only. The final concept will be discussed. Benchmark data of the developed rSOC plant, as well as the process flow sheet, will be presented. Additionally, simulation results of the rSOC plant will be shown. Remark: Only the abstract is available, because the authors chose to publish elsewhere. Please see Presentations on www.EFCF.com/LIB or contact the authors directly.

Development of systems & balance of plant components Chapter 04 - Session A13 - 43/58

th

12 European SOFC & SOE Forum

www.EFCF.com/Lib ISBN 978-3-905592-21-4

5 - 8 July 2016, Lucerne/Switzerland

A1313 Investigation of a novel catalytic partial oxidation and pre-reforming radial reactor of a micro-CHP SOFCsystem with anode off-gas recycle Timo Bosch (1), Maxime Carré (1), Angelika Heinzel (2), Michael Steffen (2), François Lapicque (3) (1) Robert Bosch GmbH Robert-Bosch-Campus 1, DE-71272 Renningen (2) Zentrum für BrennstoffzellenTechnik GmbH Carl-Benz-Strasse 201, DE-47057 Duisburg (3) Laboratoire Réactions et Génie des Procédés, CNRS-Univ. Lorraine 1 rue Grandville, FR-54000 Nancy Tel.: +49-711-811-27652 Fax: +49-711-811-5193290 [email protected]

Abstract A new radial reactor design using a precious metal catalyst coated wire mesh has been developed. This reactor has been tested standalone by emulating the total microcombined heat and power (micro-CHP) solid oxide fuel cell (SOFC) system (Pel,AC= 1000 W) interfaces by an inlet gas conditioning system and external reactor heating representing the thermal boundary conditions. During startup the total system runs on catalytic partial oxidation (CPOX) mode with internal electric heating at an oxygen to carbon ratio (O/C) of 1.2 and on pre-reforming during SOFC nominal operation (O/C= 2.2) overlaid by anode off-gas recycle in both cases. This reactor is investigated for several operation points by means of nondispersive infrared (NDIR) for CO, CO2 and CH4, a thermal conductivity detector (TCD) for H2, a paramagnetic sensor for O2 and a dew point mirror for H2O, radial and axial temperature distributions and pressure losses.

Fig. 1: Reactor test rig (left); radial reactor 3D CAD image (middle) [1]; installed wiremesh catalyst (right)

Development of systems & balance of plant components Chapter 04 - Session A13 - 44/58

th

12 European SOFC & SOE Forum

www.EFCF.com/Lib ISBN 978-3-905592-21-4

5 - 8 July 2016, Lucerne/Switzerland

Introduction A solid oxide fuel cell (SOFC) is a galvanic element which can convert chemical energy directly into electricity [2]. Hydrogen and carbon monoxide, fed to the SOFC anode, can be converted as well as methane if having sufficient oxygen e.g. in terms of steam at the anode inlet [3]. Fuel not converted within the cells can be reused partially by anode off-gas recycle. Recycling can be done with a recycle blower. The remainder is burnt providing heat so that it is a micro combined heat and power system (micro-CHP). Higher hydrocarbons fed cause carbon deposits on nickel-based anodes which results in their deactivation [3]. Sulfur which enters the anode poisons them due to chemisorption and lowers the cell performance [3]. Diatomic oxygen present at the anode causes nickel oxidation as third deterioration mechanism at elevated temperatures [4]. Sulfur and higher hydrocarbons are constituents of natural gas (NG) [5]. Gas processing is the generation of a gas stream with gas compounds processible by the cells [6]. If fueling NG, SOFC upstream gas processing is necessary therefore. This function is realized by the gas processor (GP). Vaporizers, heat exchanger, reactor, condenser, water treatment and desulfurizer can make up a GP. Reforming of the fuel is done within the reactor. Four main reactions for the production of a synthesis gas rich in hydrogen and carbon monoxide are discussed in literature [5]. Catalytic partial oxidation (CPOX) is substoichiometric exothermal oxidation of hydrocarbon fuel by diatomic oxygen. For catalytic steam reforming (CSR) steam is used instead of diatomic oxygen making this reaction endothermal. The resulting reformer product gas has a higher heating value than the inlet gas. Combining CPOX and CSR is called catalytic oxidative steam reforming (COSR). Catalytic carbon dioxide dry reforming (CCDR) uses carbon dioxide instead of diatomic oxygen or steam. Pre-reforming is conversion of higher hydrocarbons instead of maximizing the hydrogen and carbon monoxide amount [7]. Several side reactions exist [8]. Some of them lead to undesired carbon deposits lowering the reforming performance or even damage the catalyst particles. Thermodynamic equilibrium calculations allow to determine in which conditions carbon deposits are formed [3]. Important criteria are the oxygen to carbon ratio ϕ and the temperature. Low ϕ and temperature shift the equilibrium to the formation of solid carbon. Reforming can be done by multi-channel reactors with well-defined geometry like monoliths or microchannel reactors [9]. They consist of a ceramic or metallic geometric substrate where the catalyst can be deposited. In contrast, fixed-bed reactors consist of a tube or vessel randomly filled with catalyst particles like pellets, pills, spheres or extrudates. Flow direction of gas can be either in axial or radial direction through the reactor [10]. Wire meshs as reactor internals instead of particles provide an ordered structure like multi-channel reactors [9]. Nickel (Ni) or precious metals are often used as active component of catalyst materials for the reforming reactions mentioned [11,12]. Ni is lower in cost but has some drawbacks compared with precious metals [13]. Whisker carbon, the most detrimental form of carbon deposits, forms on Ni. Moreover the sintering temperature of Ni is lower than the one of precious metals, which limits the process temperatures [13]. In addition a pre-reduction of Ni surface is required before use and it is air sensitive and pyrophoric if it is reduced. The facts mentioned make precious metals easier to handle, especially during SOFC-system startup and shutdown [14,15].

Development of systems & balance of plant components Chapter 04 - Session A13 - 45/58

th

12 European SOFC & SOE Forum

www.EFCF.com/Lib ISBN 978-3-905592-21-4

5 - 8 July 2016, Lucerne/Switzerland

In this study the radial reactor prototype of a micro-CHP SOFC-system with anode off-gas recycle was developed cooperatively and a test station was developed to investigate the radial reactor standalone for several operation points, including the system startup. It will be characterized emulating SOFC-system boundaries. A first result presented in this work is the pressure loss characteristic.

1. Scientific Approach Applying construction methodology, several SOFC-system startup concepts were elaborated and evaluated. Deduced from this, a GP startup procedure and reactor for the most promising concept was developed. To study the standalone reactor’s behavior at different operation points including the startup as it would operate within the total SOFC-system with anode off-gas recycle, the emulation of the reactor inlet gas composition, its temperature and the thermal boundaries is necessary. A simplified SOFC-system model was established to determine the inlet gas composition. Based on this, a reactor test rig was developed and built to characterize the reactor as standalone component.

2. Experimental section 2.1 SOFC-system startup The GP startup procedure consists of four phases. Startup cold is the first phase where the GP is heated by e.g. the reactor internal electric heaters and the air enclosed in the anode GP is recirculated. The reactor acts as an air preheater. The next phase is startup warm. NG and air are supplied to the anode GP with ϕ = 1.2, providing a reducing atmosphere to protect the Ni anode from oxidation. At first the reaction mode is CPOX. Due to the implemented anode off-gas recirculation, the reaction mode shifts to a combination of COSR and CCDR. The air supply is the only oxygen source if stack leakage is neglected. Until start of electric energy production with the stack, startup warm is kept. The phase when electricity is produced and air is still fed to the anode GP is defined as hybrid phase. In hybrid phase, ϕ of the reactor is ramped up to its nominal operation of 2.2 by the electrochemical reaction. Oxygen ions diffusing through the electrolyte to the cathode provide a second oxygen source. Subsequently the stack current and stack fuel utilization are ramped up to eliminate anode GP air supply by keeping ϕ at its nominal value. The reaction mode without external air supply is a combination of CSR and CCSR. Feeding only NG to the anode GP the fourth phase defined as power ramp-up starts. At this stage the electric power output is increased until nominal power is reached. This startup procedure is the baseline for the development of the test rig. 2.2 Test rig The test rig comprises the reactor which will be tested and the gas analysis as shown in Fig. 2. Seven mass flow controllers (MFCs) of the EL-FLOW series of Bronkhorst for the gases and one CORI-FLOW for the deionized (DI) water of Bronkhorst are implemented at the test rig for dosing the species according to the operation point. Besides DI water with a full scale (FS) of 1000 g h-1, methane (FS of 4 slpm), carbon monoxide (FS of 2.2 slpm), hydrogen (FS of 6.2 slpm), carbon dioxide (FS of 6.5 slpm), nitrogen (FS of 6 slpm), and air Development of systems & balance of plant components Chapter 04 - Session A13 - 46/58

th

12 European SOFC & SOE Forum

www.EFCF.com/Lib ISBN 978-3-905592-21-4

5 - 8 July 2016, Lucerne/Switzerland

(FS of 2.5 slpm and 35 slpm) can be dosed. This implies that NG consists of pure methane. Media treatment (MT) with the aid of filters is carried out upstream the controlled valves (VCs). The VCs are part of the safety concept. A controlled evaporator and mixer (CEM) of Bronkhorst is used to mix DI water with gases and evaporate it simultaneously. A downstream superheater is used to further heat the gas to the desired reactor inlet temperature with sensor tag T412 (c.f. Fig. 4). Two air paths are realized as shown in Fig. 2. The MFC with higher FS is connected to the CEM. Air supply is permitted as long as the surface temperature of the superheater (sensor tag T411’) is below the ignition temperature of the present fuel-air-mixture. Reaching a defined temperature limit, the air supply is suppressed by the VC of this air path. This air path is used for pressure loss, heat up during startup cold and ignition tests. The second air path is required for startup warm and hybrid tests and is connected close to the reactor inlet.

Fig. 2: Simplified piping and instrumentation diagram of the test rig with the subsystems test specimen and gas analytics with controlled evaporator and mixer (CEM), deionized (DI) water, dew point (DP) mirror, mass flow controller (MFC), media treatment (MT), pressure sensor (P), temperature sensor (T). The subsystem gas analysis contains several heated hoses to suppress water condensation upstream of the dew point (DP) mirror. An EdgeTech DP mirror measuring system with heated cabinet and a measuring range of -35 – 95 °C with a measurement accuracy of ±0.2 °C is part of the measurement equipment. For calculating the molar fraction of water, the pressure downstream of the DP mirror is measured with a type S-10 pressure transmitter from Wika (P001). It has an absolute pressure range of 0 – 1600 mbar and 0.2% error at full scale. A sample gas cooler type EGK1/2 of Bühler is used for dehumidifying the gas stream. It cools the gas stream down to 5 °C. This corresponds to a vapor pressure of 8.718 mbar. The subsequent gas analysis is done with an online gas analyzer XEGP of Emerson. It includes measuring cells for five gas species with two measuring ranges each. Diatomic oxygen concentration is measured by a paramagnetic cell. The two measuring ranges are 0 – 5 vol% and 5 – 25 vol% with a measuring error of ±0.05 vol% for the first range and ±0.25 vol% for the second one. Hydrogen is measured with a thermal conductivity (TCD) sensor with a first range of 0 – 20 vol% and a second of 20 – 100 vol%. Errors are ±0.2 vol% and 1 vol% respectively. Methane, carbon monoxide and carbon dioxide are measured by non-dispersive infrared (NDIR) sensors. All of them have a first range of 0 – 10 vol% and a second of 10 – 100 vol% with respective errors of ±0.1 vol% and 1 vol%. Temperature measurement is Development of systems & balance of plant components Chapter 04 - Session A13 - 47/58

th

12 European SOFC & SOE Forum

www.EFCF.com/Lib ISBN 978-3-905592-21-4

5 - 8 July 2016, Lucerne/Switzerland

carried out with thermocouples type K class 1, wired with class 1 thermocouple wire and an Agilent multiplexer type 34970A is used to measure the occurring thermovoltages. The error chain results in ±4 K. P700 and P701 are measurement positions equipped with a WIKA differential pressure sensor type DPT-10 (c.f. Fig. 4). It has a measurement range of 0 – 10 mbar and an accuracy of 0.15% of the span. The differential pressure sensor for reactor internal measurements is not shown. It is a model 267 MR differential pressure transducer of Setra with a range of 0 – 25 Pa and an accuracy of ±1% FS. The 267 MR is only usable for dry and non-flammable gases during preliminary tests in contrast to the DPT-10. 2.3 Reactor prototype The development of the reactor prototype was done in cooperation with the Swedish company Catator AB. Manufacturing of the prototype including casing and wire mesh catalyst (c.f. Fig. 1, right) was done by the Swedish partner [16]. The ratio of reactor height H and outer diameter Do is smaller than 0.09 (Fig. 4). The inner reactor volume of the hollow cylinder shape is 1.04 l. Each catalyst package consists of several wire gauzes in a row. The orientation of the wire mesh is orthogonal to the radial gas flow direction. The ratio of wire mesh number of the inner to the outer package is smaller than 0.24. The volume occupied by the inner package is smaller than 0.06 l and the one of the outer package small than 0.3 l [16,18,19].

Fig 3: Schematic top view of the internal reactor component arrangement and measurement positions [18,19]. The two catalyst packages have different precious metal compositions instead of using one catalyst material. The first catalyst features superior methane oxidation properties. Development of systems & balance of plant components Chapter 04 - Session A13 - 48/58

th

12 European SOFC & SOE Forum

www.EFCF.com/Lib ISBN 978-3-905592-21-4

5 - 8 July 2016, Lucerne/Switzerland

The reactor is made of AISI 310 and coated with an oxygen conductive ceramic layer to prevent carbon deposits [16,17].

Fig. 4: Schematic sectional view alongside the connecting line of process gas inlet and outlet [18,19]. 20 measurement positions on five radii are established, as indicated in Fig. 3 and Fig. 1 (middle), having access to the reactor inside from above. For position 11 the gas inlet line is used. For the remaining 19 positions, additional holes are machined to provide access. In all positions thermocouples type K are installed centered in z direction. Four measurement positions can be used for parallel measurement of the pressure or taking gas samples during reforming operation. The reactor is embedded in insulation and electric heaters to emulate the GP integration implying heat loss compensation (c.f. Fig. 1, left). The upper and bottom annulus surface (referred to z-direction) of the reactor are heated by heating sleeves with a maximum electric power of approximately 340 W and an insulation thickness of 5 cm each. The inner lateral surface is equipped with a third electric heater. Its maximum electric power is approximately 90 W. The outer lateral surface is insulated with a 5 cm thick insulating medium.

3. Results Pressure loss measurements were carried out for several measurement pairings with the differential pressure sensors of Setra and WIKA. The supply of unheated pressurized air with a temperature of 22 °C and a dew point temperature of -40 °C is done with the air MFC having 35 slpm FS. The MFC set point was increased in steps of 2 slpm from 0 to 34 slpm every 1.5 min. Subsequent decrease to 20, 10, 4 and 0 slpm was carried out in 1.5 min steps to check sensor drifts. No static pressure difference at this procedure could be measured at the inner plenum (positions 41, 21, 31) with regard to the accuracy of the Setra differential pressure sensor. This coincides with calculations prior to the measurements for the volume flow at the inner plenum. In this region pressure losses are almost factor 100 lower compared with the one across the catalyst packages. Therefore homogeneous area specific gas flow in radial direction is assumed. No static pressure difference is measurable between positions 45 and 25. 15 to 45 or 15 to 25 shows the designated slight decrease in static pressure. Here the increase in dynamic pressure contributes to the decrease in static one. Static differential pressure measurements in the catalyst packages were done at y-axis, since dynamic pressure rates are most likely compensated there and a value for the pressure loss is obtained. Fig. 5 shows the experimental value measured between positions 41 and 45, averaged over 60 s for one MFC setpoint, and the simulated values. Simulations were Development of systems & balance of plant components Chapter 04 - Session A13 - 49/58

th

12 European SOFC & SOE Forum

www.EFCF.com/Lib ISBN 978-3-905592-21-4

5 - 8 July 2016, Lucerne/Switzerland

done according to Ergun’s law as described in [20] and Armour and Cannon as depicted in [21]. Porosity and specific surface area were calculated according to [22]. 9 experiment

8

simulation according to Ergun

7

simulation according to Armour and Cannon

ΔP [Pa]

6 5 4 3

2 1 0 0

5

10

15

20

25

30

35

volume flow [slpm]

Fig. 5: Measured and simulated pressure loss at the catalyst packages [20,21,22]. Experimental data are related to the measurement points 41 and 45. Fig. 5 shows that both simulations correlate with the measurements. To generalize pressure loss at the catalyst packages the pressure loss coefficient ζ1 is referred to a modified Reynolds number [23] in Fig. 6 at the superficial surface of the first catalyst package with

and Here represents the superficial velocity in front of the first wire mesh, surface area, the density of the gas and the kinematic viscosity.

the specific

100'000 experiment

90'000

simulation according to Ergun

80'000

simulation according to Armour and Cannon

70'000 60'000

ζ1 50'000 40'000 30'000 20'000

10'000 0

0.0

0.2

0.4

0.6

0.8

1.0

1.2

Re1

Fig. 6: Correlation of catalyst pressure loss coefficient ζ1 and modified Reynolds number Re1 [20,21,22]. Experimental data are related to the measurement points 41 and 45. In a further experiment the pressure loss was measured between P700 and P701 with the WIKA differential pressure sensor. Since the inner diameter is the same for both measurement points, no dynamic pressure correction has to be done. The difference in Development of systems & balance of plant components Chapter 04 - Session A13 - 50/58

th

12 European SOFC & SOE Forum

www.EFCF.com/Lib ISBN 978-3-905592-21-4

5 - 8 July 2016, Lucerne/Switzerland

static pressure delivers directly the pressure loss of the complete test specimen. Again, the pressure loss coefficient ζ2 is referred to the Reynolds number Re2. Re2 is defined for the internal flow at the tube were the static pressure is measured [20]. Fig. 7 shows the correlation for both, the simulated and measured pressure loss. 7

experiment 6

simulation

5

4

ζ2 3 2 1 0 0

1'000

2'000

3'000

4'000

5'000

6'000

Re2

Fig. 7: Correlation of test specimen pressure loss coefficient ζ2 and Reynolds number Re2 [20,22]. Experimental data are related to the measurement points P700 and P701. Simulations are based on empirical data and consider the inlet tubing, the catalyst package modeled according to Ergun’s law, and the outlet tube. The reactor plenums are neglected since their losses are comparatively small. The inlet tubing is modeled as a series of two tubes (Fig. 1 (middle)) with their different cross sections and their sudden channel widening [20]. The outlet tubing, consisting of the same cross sections as the inlet but in another sequence, is modeled as a series of two tubes, pipe inlet section and sudden reduction of cross section [20]. The tube lengths of inlet and outlet differ. Model and experimental values correlate especially for larger Reynolds numbers. The dip in the simulated values at approximatively 2200 to 2600 Re2 is related to the change from laminar to turbulent flow. The relative pressure loss distribution is represented in Fig. 8. For higher volume flows the catalyst packages contribute to less than 10% of the total losses. The air leaving the reactor is directed into an exhaust system. In order to see whether changes in the exhaust path cause differential pressure changes, manual manipulation of the exhaust path flow was done observing the differential pressure measurement. Manipulation causes a short rise or fall of the signal. Within a few seconds the baseline value of before manipulation is reached again. The sample gas pump belonging to the gas analysis line causes no perturbations either.

Development of systems & balance of plant components Chapter 04 - Session A13 - 51/58

th

12 European SOFC & SOE Forum

www.EFCF.com/Lib ISBN 978-3-905592-21-4

5 - 8 July 2016, Lucerne/Switzerland

1.0 0.9 0.8

relative ΔP

0.7 0.6 0.5 0.4

inlet + outlet + catalyst packages inlet + outlet

0.3 0.2

inlet

0.1

0.0 0.06 0.12 0.18 0.24 0.29 0.35 0.41 0.47 0.53 0.59 0.65 0.71 0.76 0.82 0.88 0.94 1.00

relative volume flow

Fig. 8: Simulation based relative pressure loss of the reactor test specimen [20,22].

4. Conclusion In this paper a novel radial reactor and the developed test rig to study this reactor standalone at different operation is presented. Pressure loss measurements are shown for the catalyst packages and for the complete test specimen, both correlating with carried out simulations. For higher volume flow the pressure loss across the catalyst packages contributes to less than 10% of the pressure loss of the total test specimen. Experiments for different operation during system startup are ongoing.

Acknowledgement Thanks go to the company Robert Bosch GmbH for its funding as well as to the company Catator AB for the cooperation in reactor development.

References [1] [2] [3] [4] [5] [6]

N. u. (2015): 3D CAD model of the reactor test specimen. Internal data. Catator AB. Kurzweil, Peter (2013): Brennstoffzellentechnik. Grundlagen, Komponenten, Systeme, Anwendungen. 2., überarb. und aktualisierte Aufl. Wiesbaden: Springer Vieweg (SpringerLink : Bücher). Hansen, John Bøgild (2011): Direct Reforming Fuel Cells. In: Fuel Cells: Technologies for Fuel Processing: Elsevier, pp. 409–450. Halinen, M.; Thomann, O.; Kiviaho, J. (2014): Experimental study of SOFC system heat-up without safety gases. In: International Journal of Hydrogen Energy 39 (1), pp. 552–561. DOI: 10.1016/j.ijhydene.2013.10.043. Speight, James G. (2011): Fuels for Fuel Cells. In: Fuel Cells: Technologies for Fuel Processing: Elsevier, pp. 29–48. Shekhawat, Dushyant; Berry, David A.; Spivey, James J. (2011): Introduction to Fuel Processing. In: Fuel Cells: Technologies for Fuel Processing: Elsevier, pp. 1-9.

Development of systems & balance of plant components Chapter 04 - Session A13 - 52/58

th

12 European SOFC & SOE Forum

[7] [8]

[9] [10] [11] [12] [13] [14]

[15] [16] [17] [18] [19] [20] [21]

[22] [23]

www.EFCF.com/Lib ISBN 978-3-905592-21-4

5 - 8 July 2016, Lucerne/Switzerland

Ahmed, Khaliq; Föger, Karl (2010): Fuel Processing for High-Temperature HighEfficiency Fuel Cells. In Ind. Eng. Chem. Res. 49 (16), pp. 7239–7256. DOI: 10.1021/ie100778g. Aasberg-Petersen, K.; Dybkjær, I.; Ovesen, C. V.; Schjødt, N. C.; Sehested, J.; Thomsen, S. G. (2011): Natural gas to synthesis gas – Catalysts and catalytic processes. In Journal of Natural Gas Science and Engineering 3 (2), pp. 423–459. DOI: 10.1016/j.jngse.2011.03.004. Ilsen Önsan, Z.; Avci, Ahmet K. (2011): Reactor Design for Fuel Processing. In: Fuel Cells: Technologies for Fuel Processing: Elsevier, pp. 451–516. Li, Jeff C. H. (2000): Radial-Flow Packed-Bed Reactors. In: Ullmann's Encyclopedia of Industrial Chemistry. Weinheim, Germany: Wiley-VCH Verlag GmbH & Co. KGaA. Spencer, M. S. (1989): Fundamental Principles. In Martyn V. Twigg (Ed.): Catalyst handbook. 2. ed. London: Wolfe, pp. 17–84. Smith, Mark W.; Shekhawat, Dushyant (2011): Catalytic Partial Oxidation. In: Fuel Cells: Technologies for Fuel Processing: Elsevier, pp. 73–128. Haynes, Daniel J.; Shekhawat, Dushyant (2011): Oxidative Steam Reforming. In: Fuel Cells: Technologies for Fuel Processing: Elsevier, pp. 129–190. Adachi, H.; Ahmed, S.; Lee, S.H.D.; Papadias, D.; Ahluwalia, R. K.; Bendert, J. C. et al. (2009): A natural-gas fuel processor for a residential fuel cell system. In: Journal of Power Sources 188 (1), pp. 244–255. DOI: 10.1016/j.jpowsour.2008.11.097. Farrauto, Robert J. (2014): New catalysts and reactor designs for the hydrogen economy. In: Chemical Engineering Journal 238, pp. 172–177. DOI: 10.1016/j.cej.2013.07.004. Silversand, Fredrik A. (2015): Delivery documentation and relevant information. Internal document. Catator AB. Silversand, Fredrik A. (2014): DESIGN STUDY. Reactors for hot desulphurization and catalytic partial oxidation/pre reforming. Internal report. Catator AB. N. u. (2015): Test specimen manufacturing drawing. Reactor. Internal document. Catator AB. N. u. (2015): Test specimen manufacturing drawing. Top lid. Internal document. Catator AB. VDI-Wärmeatlas (2013). 11., bearb. und erw. Aufl. Berlin, Heidelberg: Springer (VDI-Buch). Kołodziej, Andrzej; Jaroszyński, Mieczysław; Janus, Bożena; Kleszcz, Tadeusz; Łojewska, Joanna; Łojewski, Tomasz (2009): AN EXPERIMENTAL STUDY OF THE PRESSURE DROP IN FLUID FLOWS THROUGH WIRE GAUZES. In: Chemical Engineering Communications 196 (8), pp. 932–949. DOI: 10.1080/00986440902743851. Zhao, Zenghui; Peles, Yoav; Jensen, Michael K. (2013): Properties of plain weave metallic wire mesh screens. In: International Journal of Heat and Mass Transfer 57 (2), pp. 690–697. DOI: 10.1016/j.ijheatmasstransfer.2012.10.055. Richardson, J. F.; Harker, J. H.; Backhurst, J.R (2006, c2002): Coulson and Richardson's chemical engineering. Particle technology and separation processes. 5th ed. reprint. Amsterdam, NE: Butterworth/Heinemann (Chemical engineering series, v.2).

Development of systems & balance of plant components Chapter 04 - Session A13 - 53/58

th

12 European SOFC & SOE Forum

www.EFCF.com/Lib ISBN 978-3-905592-21-4

5 - 8 July 2016, Lucerne/Switzerland

A1315 (Abstract only) Performance evaluation of solid oxide carbon fuel cells operating on steam gasified carbon fuels Tak-Hyoung Lim(*), Jong-Won Lee, Seung-Bok Lee, Seok-Joo Park, Rak-Hyun Song Fuel Cell Research Laboratory, Korea Institute of Energy Research (KIER) 152 Gajeong-ro, Yuseong-gu, Daejeon 34129, Korea Tel.: +82-42-860-3608 Fax: +82-42-860-3297 [email protected]

Abstract We investigated the operating characteristics of solid oxide carbon fuel cells (SO-CFCs) integrated with a steam gasifier that used carbonaceous fuels, including activated carbon and biomass driven charcoal. Steam gasification was carried out in a specially designed gasifier, which was directly integrated with a solid-oxide based carbon fuel cell. We studied the effect of gasification temperature, steam flow rate and catalyst addition on the electrochemical performance of SO-CFC, and the results showed that among the three tested fuels, activated carbon with a K2CO3 catalyst performed the best. At 850°C, the maximum power density 108mW/cm2, 161mW/cm2 and 181mW/cm2 was achieved when the SO-CFC operated on activated carbon, biomass driven charcoal and activated carbon with a K2CO3 catalyst, respectively. The SO-CFC operated continuously for 100h and it showed relatively stable performance. This study suggests that by using a catalytic steam gasifier integrated with the SO-CFC, the solid carbon fuel resources can be used for power generation with higher efficiency and minimal carbon footprint.

Remark: Only the abstract was available at the time of completion. Please see Presentations on www.EFCF.com/LIB or contact the authors directly.

Development of systems & balance of plant components Chapter 04 - Session A13 - 54/58

th

12 European SOFC & SOE Forum

www.EFCF.com/Lib ISBN 978-3-905592-21-4

5 - 8 July 2016, Lucerne/Switzerland

A1316 (Abstract only, published elsewhere) Methane Steam Reforming Reaction over Ni/CeO2-ZrO2 Catalysts Loaded on Metallic Monolith Jong Dae Lee (1) (1) Department of Chemical Engineering, Chungbuk National University. 1 Chungdea-ro, Seowon-gu Cheong-ju, Chungbuk 28644, Korea Tel.: +82-43-261-2375 Fax: +82-43-269-2370 [email protected]

Abstract In recent years, the over consumption of fossil fuels leads to critical environmental problems and arises a great concern on energy security. Great research effort has been focused on the production of hydrogen and the fuel cell systems. Hydrogen has been proposed as a clean and renewable energy. Among the hydrocarbon fuels, methane is a commercial gas that is easily transported and stored. Some typical fuel reforming technologies are steam reforming, partial oxidation, autothermal reforming and CO2 reforming etc. In general, steam reforming has the advantage of producing a higher H 2 concentration than catalytic partial oxidation. Currently, steam reforming process by precious metal catalysts (e.g., Ru, Pd, and Pt) has generally been used to convert CH4 to H2. In this study, the catalytic behaviors of Ni Ni/CexZr1-xO2 loaded on the metallic monolith were investigated for the steam reforming reaction of CH4. Ni, Pd and Ru were loaded on the Al2O3-MgO supports by the impregnation method after dissolving in 5M-HNO3 and then these catalysts were thermally treated at 800 ℃ for 2h. Metallic monolith with diameter of 2.5 cm and height 2 cm was prepared by winding a combination of flat plate and flexural plate of 50 μm thickness. Before loading the catalyst to metallic monolith, alumina sol was coated on the surface of metallic monolith for improvement of catalyst adhesion, and pre-heated at 900 ℃. The catalyst slurrys were washcoated on the metallic monolith of honeycomb structure that has excellent heat conductivity. Prepared supports and catalysts were analyzed by XRD, SEM and BET. The effect of Ni content on the Ni/Ce0.80Zr0.20O2 catalysts was also investigated and the catalyst loaded with 15wt% Ni had the highest activity for the steam reforming reaction. Also, the effect of temperature, GHSV and H2O/CH4 ratio, was investigated to find optimum operating conditions for each processes. As GHSV decreased and H 2O/CH4 ratio increased, CH4 conversion and H2 yield were increased. Among the catalysts, the Ni(15wt%)/Ce0.80Zr0.20O2 and Ni(15wt%)-Ru(0.5wt%)/Ce0.80Zr0.20O2 catalysts showed high CH4 conversion at 800℃ for the steam reforming reaction. The optimum operating conditions of both catalysts were GHSV under 10,000h-1 and H2O/CH4 ratio over 3 at 800℃. Catalytic activity of Ni(15wt%)-Ru(0.5wt%)/Ce0.80Zr0.20O2 loaded on the metal monolith was tested at 800℃ for 10 h and the activity of the catalyst remained stable in steam reforming reaction for mono and double layer metallic monolith catalysts. Remark: Only the abstract is available, because the authors chose to publish elsewhere. Please see Presentations on www.EFCF.com/LIB or contact the authors directly.

Development of systems & balance of plant components Chapter 04 - Session A13 - 55/58

th

12 European SOFC & SOE Forum

www.EFCF.com/Lib ISBN 978-3-905592-21-4

5 - 8 July 2016, Lucerne/Switzerland

A1317 (Abstract only, published elsewhere) System validation tests for a SOFC power system at INER Shih-Kun Lo*, Wen-Tang Hong, Hsueh-I Tan, Huan-Chan Ting, Ting-Wei Liu and Ruey-Yi Lee Institute of Nuclear Energy Research No. 1000 Wenhua Road, Longtan District Taoyuan City / Taiwan (R.O.C.) *Tel.: +886-3-471-1400 Ext. 6787 Fax: +886-3-471-3980 [email protected]

Abstract This research presents the results of system validation tests for a SOFC power system. In the study, the system was heated up without electric device, i.e., the fuel providing the required thermal energy through an integrated balance of plant (BOP). The ex-situ experiments, without a SOFC stack installed in the system, were first conducted to investigate the operability of a BOP apparatus. It was found that the BOP possessed high conversion efficiencies for both steam reforming and water gas shift reactions. The total fuel concentration of hydrogen and carbon monoxide from the reformer was 91.2 %. The system validation tests showed that, with the natural gas as fuel, the output power from the stack reached to 1060 W, while the fuel utilization efficiency and electrical efficiency were 67.16 % and 45.0 %, respectively. A steady 600-hour system operation test was carried out at an average system temperature of 694 oC. Of which, a 36-cell stack was employed for the test. Meanwhile, the current, voltage and output power were 26 A, 32.3 V and 840 W, respectively, and its electrical efficiency was 33.4 %.

Remark: Only the abstract is available, because the authors chose to publish elsewhere. Please see Presentations on www.EFCF.com/LIB or contact the authors directly.

Development of systems & balance of plant components Chapter 04 - Session A13 - 56/58

th

12 European SOFC & SOE Forum

www.EFCF.com/Lib ISBN 978-3-905592-21-4

5 - 8 July 2016, Lucerne/Switzerland

A1319 (Abstract only, published elsewhere) A Global Reaction Model of Carbon Gasification with K2CO3 in the External Anode Media of a DCFC Shinae Song, Jun Ho Yu, Kyungtae Kang, Jun Young Hwang Korea Institute of Industrial Technology Ansan, 426-173, South Korea Tel.: +82-31-8040-6434 Fax: +82-31-8040-6430 [email protected]

Abstract The present study was conducted to develop a practical reaction model for high temperature gasification based on the mechanism of the key elementary reactions, considering its applications to the external anode media of a direct carbon fuel cell (DCFC). The characteristics of gasification reactions were experimentally investigated for carbon black-K2CO3 mixtures in carbon dioxide ambient atmosphere at high temperatures up to 900 oC. Changes in the exit gas composition were monitored during the heating process (Fig. 1). Based on the experimental observations, a simplified reaction model for a global gasification reaction was suggested in the form of a linear combination of the Boudouard reaction, carbonate-catalysed reactions, and metal-catalysed reactions. The correlation between the equilibrium concentrations of carbonates and oxides in the mixture media was also given, where the ratio of the carbonate concentration to the oxide concentration was proportional to the CO2 concentration.

MFC

Temperature Sensor

Gas to Furnace (CO2/N2)

Furnace

MFC

Inlet 2

Inlet 1

PC

Specimen Dilution Gas (N2)

Outlet MFC

Gas Sensor

Gas out

Fig. 1 Schematic drawing of exit-gas measurement system. Remark: Only the abstract is available, because the authors chose to publish elsewhere. Please see Presentations on www.EFCF.com/LIB or contact the authors directly.

Development of systems & balance of plant components Chapter 04 - Session A13 - 57/58

th

12 European SOFC & SOE Forum

www.EFCF.com/Lib ISBN 978-3-905592-21-4

5 - 8 July 2016, Lucerne/Switzerland

A1320 (Abstract only, published elsewhere) Experimental study on the fuel ejector for solid oxide fuel cell system Kanghun Lee (1), Sanggyu Kang (1, 2), Youngduk Lee (1), Kook-Young Ahn (1,2) (1) Korea Institute of Machinery and Materials (KIMM); Gajeongbukro 156; Daejeon/Republic of Korea (2) University of Science and Technology (UST), Gajeongro 217; Daejeon/Republic of Korea Tel.: +82-42-868-7267 Fax: +82-42-868-7284 [email protected]

Abstract The anode-off gas from the solid oxide fuel cell (SOFC) stack can be reutilized to improve the system efficiency. That is, because the anode-tail gas from the SOFC is including the unreacted fuel as well as high amount of steam, which can be reused as a fuel for SOFC stack and steam for methane steam reforming reaction (MSR), respectively. Recirculation of the SOFC anode-off gas can obtain the benefit from its high operating temperature. Ejector is one promising way for the recirculation of the SOFC anode-off gas due to robustness at high temperature and low cost. However, the amount of suction flow is not controlled easily compared to the regenerative blower. In this study, one SOFC system using the fuel ejector has been developed. And the fuel ejector has been designed, manufactured, and evaluated its performance. The effects of the operating pressure and temperature on the ejector performance are identified. Furthermore, the effect of geometric parameter of nozzle exit position on the ejector characteristics has been investigated. This study is useful to optimize the design of the ejector and establish the optimal operating scheme of ejector.

Remark: Only the abstract is available, because the authors chose to publish elsewhere. Please see Presentations on www.EFCF.com/LIB or contact the authors directly.

Development of systems & balance of plant components Chapter 04 - Session A13 - 58/58