Προσομοίωση και πειραματική διερεύνηση της λειτουργίας κελιών καυσίμου στερεού οξειδίου χρησιμοποιώντας ως καύσιμα: υδρογόνο, μεθάνιο και προϊόν αεριοποίησης βιομάζας

Abstract

This completed and submitted PhD thesis provides a simulation and experimental investigation of the new alternative energy technology Solid Oxide Fuel Cell (SOFC), analyzing its performance on different kinds of fuels with special focus on Biomass Gasification derived producer gas. The development and application of several theoretical and experimental methodologies for these investigations are presented. The main target, to demonstrate the technical feasibility of the integration of the SOFC with another alternative energy technology, i.e. Biomass Gasification, in a real experimentally combined system, was successfully achieved. The approach to the thesis is divided into three steps. The first step (Chapters 2 and 3), SOFC modeling and simulation studies, provides the theoretical background and understanding of SOFC operation characteristics and the inherent losses, which determine the performance of an operating SOFC. The operation of a SOFC is a complex interaction of physico-electro-chemical processes and its performance depends highly on the selected operational parameters such as type of fuel and temperature as well as on the cell’s geometry and its microstructure. All of these interdependent processes can be modelled to a certain degree of detail which is crucial for the understanding of the underlying mechanisms responsible for the performance of an SOFC. These detailed models not only support experimental findings, but more importantly are a cost-effective tool for SOFC developers in designing new cell properties by conducting theoretical experiments, i.e. simulations with varying model parameters. As a second step (Chapter 4), an experimental investigation with hydrogen and hydrocarbon based fuels was conducted which allows a performance comparison employing the theoretical findings of the modeling and simulations chapters. The use of synthetic gases from bottles and controlled operational parameters allows ruling out disturbing experimental factors, which are encountered in real tests, such as contaminants in the fuel gases. The final step is the experimental demonstration of SOFCs fuelled with real producer gas from four different laboratory and pilot-scale biomass gasifiers (Chapters 5 to 7) and the associated analysis of carbon deposition (Chapters 8 and 9), considered as a major source of anode contamination in those tests. A literature review revealed that until now, power production via SOFC fuelled by product gas from biomass gasification has been investigated mainly in theory or experimentally with synthetic gas mixtures (from gas bottles) only. In order to investigate experimentally the feasibility of the combined systems - SOFC and Biomass Gasifier -, real tests were planned and carried out within the cooperation of the EU funded BioCellUS project as part of this dissertation. All of the gasifiers visited during the test campaign were of different design and represented most of the available technologies ranging from air-blown fixed bed downdraft gasifiers over allothermal steam blown fluidized bed gasifier to steam and oxygen blown circulating fluidized bed gasifier. Experimental conditions were ranging from high-quality and clean to low quality and highly tar contaminated producer gas. The experimental setup of the combined systems is presented in a very transparent way and can serve as a guideline for similar tests in the future as well as for endurance and scale-up tests. Furthermore, a proof of concept for potential alternative energy production system is achieved. Chapter 2 presents the first modelling efforts made in this work which are extended in the following Chapter 3. Within the framework of SOFC models, up to recently modelling efforts were either based on empiric global kinetic approaches for the methane steam reforming and water gas shift reactions or on thermodynamic equilibrium assumptions. These approaches are rather easy to handle but lack general validity as well as a closer insight into the underlying heterogeneous chemistry. For the scope of this dissertation -the investigation of SOFC operation on different fuels apart from hydrogen such as methane, syngas and biomass derived producer gas -a lately published detailed elementary heterogeneous catalytic reaction mechanism (HCR), implicitly accounting for the kinetics of methane reforming, water gas shift and Boudouard reactions, was used. Through this work, the detailed HCR was for the first time employed in a quasi-2-D model being capable of simulating the thermal impact and the depending losses for a cross-flow configured planar electrolyte supported SOFC. A comparison with a commonly used global reforming model allows SOFC modelers to make their choice between a simple global or detailed methane reforming kinetic model. The HCR predicts incomplete reforming which is also observed in experiments, while the global reforming approach overpredicts complete reforming, incomplete reforming on the one hand results in a lower cell performance, but on the other hand causes a smoother solid temperature profile over the cell reducing thermally induced stresses. A decrease in methane inlet content results in smaller deviations between the model predictions of the detailed HCR and the simple global reforming approach. A reliable reforming model is crucial to optimize thermal management of an SOFC in order to avoid thermal stress induced failures. Chapter 3 presents a detailed flexible model of planar SOFCs, which allows the simulation of steady-state performance characteristics, i.e. voltage-current density (V-j) curves, and dynamic operation behavior. Its special capability of simulating electro-chemical impedance spectroscopy (EIS) makes the model an important tool for analyzing SOFC fundamentals as well as for design, materials and operational parameters optimization. Such a tool allows SOFC developers to quickly refine different design options, optimize design parameters, and thus reduce the amount of costly experiments. The dynamic model is based on physico-chemical governing equations and allows spatial discretization for 1-D (button cell approximation) up to quasi 3-D (full size anode supported cell in cross-flow configuration) geometries. Different fuels based on hydrogen, methane and syngas with inert diluents can be investigated. The model is implemented on the commercially available modeling & simulations platform gPROMS. All the necessary equations, parameters, boundary and initial conditions are presented so that it can easily be reproduced by the readers. The model is applied in a detailed parametric analysis of the SOFC inherent losses (overpotentials) in an attempt to deconvolute the impedance spectrum of an SOFC. Each of the considered main transport processes can be attributed to an impedance arc. The mass transport above the electrodes produces a low frequency impedance arc called Nemst impedance. Mass transport through the porous electrodes (due to diffusion) causes the concentration impedance arc in middle frequency range of the EIS. The high frequency arc is related to the double-layer charge transport and activation overpotentials. Chapter 4 presents the experimental investigation of SOFCs fuelled with synthetic hydrogen and methane/syngas containing gases. The purpose of these tests is the evaluation and comparison of the electrochemical performance of an anode supported SOFC (ASC) operating on the different fuels. The focus of the analysis is the current limiting behavior occurring at high currents well below complete fuel utilization. Gas chromatography (GC) measurements of anode exhaust gas allow a closer analysis of tests with syngas/methane containing fuels. Results from the theoretical modelling and simulations approach are confirmed. The high frequency arc is experimentally confirmed to be related to activation of the electrochemical reactions, where the cathodic contribution prevailed. At high current operation (fuel utilization > 70%), current limitation sets in which manifests itself as a sudden drop of the V-j-curve. It is related to hydrogen depletion at the triple phase boundary (TPB) (concentration overpotential) due to the anodic porous electrode gas diffusion process. These observations are confirmed by the EIS measurements at high currents, where the middle to low frequency arc, related to the concentration overpotentials, is much larger than for the undiluted hydrogen case. Methane or syngas fuelled SOFCs have a lower performance at high currents than under equivalent hydrogen operation. This is not only due to incomplete reforming, which is responsible for a decreased Nemst potential, but also due to increased diffusion losses (concentration overpotentials) arising from the larger molecules involved. Chapter 5 presents the first real tests of combined SOFC and Biomass Gasification operation in the open literature. The purpose of this test is the operation for a prolonged duration (150 hours) on a rather clean product gas in order to establish a transparent and detailed proof of concept for the combined SOFC-Biomass Gasifier system. The test was carried out at the two-stage fixed bed downdraft biomass gasifier “Viking” from the Danish University of Technology (DTU), which fulfilled these previously stated requirements of producing a “clean” gas without tars, whose influence on SOFC performance was not known so far. The SOFC voltage performance is sustained throughout the duration of the experiment. The gasifier operation inherent gas composition fluctuations do not have any effect on the successful SOFC operation. The feasibility of the integrated SOFC and Biomass Gasification system is established with this test. Chapter 6 presents the next step forward, i.e. the testing of an SOFC fuelled with real product gas from two different Biomass Gasifiers under more severe conditions. The tests were carried out in connection with an air-blown fixed bed downdraft gasifier and a steam-blown allothermal fluidized bed gasifier (“BioHPR”). The influence of several possible harmful but realistic product gas characteristics on SOFC performance and durability is assessed. This includes gas impurities such as tars at low to high levels whose influence on SOFC anode material during operation is not well known yet but reported to be improper for use in gas-turbines or gas engines. Continuous testing with an approximate duration of up to one day on tar reformed product gas as well as on medium tar level gas, and finally a shorter test (7 h) with a high tar load of 3000 mg Nm⁻³ show, that tars do not cause noticeable problems to Ni-GDC anode operation. For the first time in open literature, the feasibility of SOFCs running on tar-laden biomass derived product gas could be assessed. Furthermore, the effects of producer gas particulates reaching the SOFC anode resulting in carbon deposition as well as SOFC degradation under very harsh conditions, i.e. at high fuel utilization together with high steam content and fluctuating gas composition and flow are presented. Chapter 7 presents the SOFC tests carried out under most severe real conditions, i.e. with a very high tar laden product gas from a circulating fluidized bed biomass gasifier. The ideal and most efficient solution for the investigated combined system would be as less as possible gas pre-treatment prior the fuel gas entering the SOFC to make direct use of the anodes internal reforming capabilities. Thus it needs to be shown that tars do not necessarily harm SOFC operation as it was until now assumed, but instead could be allowed to enter the cell without the necessity of an upstream tar reforming equipment. Tests have been carried out with and without tars entering the cell and a detailed tar analysis is given. The very high tar content (> 10 mg Nm⁻³) and short periods of H2S breakthrough (< 1.5 ppmv) did not influence SOFC performance negatively and no carbon or other product gas trace constituents contamination of the anodes was found when the SOFC membranes were examined with SEM/EDS after the tests. The cell operated at a slightly lower voltage when fuelled with non pre-reformed producer gas, compared to reference tests with pre-reformed gas. This difference is attributed to the lower heating value of the non-reformed gas entering the SOFC resulting in a lower Nemst potential. Chapter 8 presents the detailed post-experimental optical and chemical analysis of the SOFC membranes tested on real and partially contaminated biomass gasification producer gas (from tests of Chapters 5 to 7) in order to assess the impact of real producer gas on fuel cell material. Own procedures had to be developed for the Scanning Electron Microscopy (SEM) and Energy Dispersive X-Ray Spectroscopy (EDS) investigation. Carbon deposits, which were originally anticipated to arise from the producer gas tars, could not be detected. Anode contamination due to traces of sulphur from H₂S content in the producer gas, or evaporated zinc from the zinc oxide desulphurization beds, can also be excluded. The anodes’ functional layers were found to be intact concerning its microstructural parameters like porosity, particle size and composition distribution. Chapter 9 presents an additional experimental investigation with deliberately induced carbon deposition on the Ni-GDC anode, which could not be observed in the tests conducted on real tar laden biomass producer gas. For this purpose, Ni-GDC anodes were exposed to a methane rich dry atmosphere at carbon deposition favourable conditions. Due to the absence of water vapour, catalytic dissociation of methane produces solid carbon within the anode. Those tests are assumed to result in more serious carbon deposition than during electrochemical operation, where the flow of oxygen ions to the anode provides a chemical potential for oxidation of carbon deposits. The amount and location of the carbon deposits could qualitatively be determined by SEM/EDS analysis of the anode microstructure. Some of the samples were so heavily contaminated with carbon that the 3 different anode layers delaminated at the layer interfaces, which have the weakest bonding strength. Carbon deposited in large amounts within the fine microstructure of the functional anode layer, which consists of fine nickel particles, catalyzing heterogeneous hydrocarbon reforming and electrochemical reactions. An in-depth literature review on carbon deposition, which can occur on catalytic surfaces such as SOFC anodes, reveals different types of carbon deposits and their effects. Prior to this thesis, many challenges and unanswered questions existed, which were tackled and partially resolved by this work. SOFC modeling efforts in this level of detail as presented here were scarce, especially the combination of multi-dimensional spatial discretization, a detailed heterogeneous reaction mechanism for methane and syngas fuels, multi-component diffusion and the simulation of electrochemical impedance spectroscopy combined into one work. Up to now this has only been undertaken by a few research groups, mainly on non-commercial codes, which are not open to the public. Information on the interpretation of electrochemical impedance spectra was ambiguous and sometimes misleading and experimental investigations to back the theory up, especially concerning the use of hydrocarbon fuels, were rare. A detailed parametric analysis employing the developed model brought the deconvolution of the impedance spectra a step forward. The theoretical results could be experimentally confirmed to a certain extent. The successful experimental combination of SOFC and Biomass Gasification is new. Most available works published up to now on combinations of SOFC and Biomass Gasification were based on theoretical investigations or on experiments with simulated syngas from bottles. An important question whether the SOFC could deal with producer gas tars without the deposition of detrimental carbon could positively be answered by the experimental demonstration campaign. Since nowhere in literature a detailed standardized method for post-experimental SOFC SEM/EDS analysis was found, especially regarding the detection of carbon deposits, an own analysis procedure needed to be established.