Modeling and simulation of solid oxide fuel cells (SOFCs): emphasizing effects of material properties on cell performance

Continuous reliance on fossil fuels and the current power production technologies have led to serious insecurities in terms of producing petroleum, climatic changes and urban air pollution-related hazards. Even if the ever-increasing improvement of energy efficiency and the reduction of the spread of power production system-related pollutions continue, it is expected that insecurity hazards relevant to producing petroleum which include the emission of greenhouse gases and air pollutants would increase under prevalent conditions. Fuel cell is an emerging technology that is to a great extent promising for the achievement of higher energy efficiency and lower environmental load. Therefore, a responsible use of energy is proportional to the high energy efficiency and its minimum impacts on the environment. In order to meet the requirements of Kyoto protocol, fuel cell was considered as an emerging technology. Fuel cells come in many varieties, but two of them having received much attention among governments, industries and scientific communities are: Proton-exchange Membrane and Solid Oxide Fuel Cells. Proton-exchange cells are mostly used in the transportation sector as a substitute for internal combustion engines. Solid Oxide Fuel Cells, on the other hand, is one of the high-temperature fuel cells which convert chemical energy to electrical power through electrochemical reactions at high temperatures. These cells have high efficiency, fuel flexibility, long-term stability, and lower air pollution potentials compared to the other power sources. In addition, they can be installed in combined cycle plants for heightening the efficiency level and lowering environment pollution. Due to functioning at high temperatures, Solid Oxide Fuel Cells are often used in power plants and in heavy machineries. For these FCs the electrolyte is a solid ionic conductor such as Yttria-stabilized Zirconia (YSZ) operating at 900-1000°C. Many efforts have been spent to find materials able to operate at lower temperatures (500-800°C), in the hope that overall FC production and manufacturing costs could fall down. Due to the lower operating temperature, these SOFCs are called intermediate temperature solid oxide fuel cells (IT-SOFCs). A rapid overview of materials state-of-the-art is sketched in the following: 1) Electrolytes – CeO2 based oxides have been extensively investigated as IT-SOFC electrolytes and appear promising; nevertheless, they show a low chemical stability under reducing atmosphere at T ≥ 600°C. LaGaO3 doped with Sr and Mg ions (LSGM) attracted the attention of scientific and technical communities for its high ion-conducting properties at 600-800°C and its wide electrolytic domain. 2) Electrodes - Both electrodes must have high electro-catalytic activity and electronic conductivity to minimize the actual electrical resistance. The choice of anode and cathode depends on the specific electrolyte. SOFCs could have different geometries. Laboratory scale experiments usually adopt anode-supported button cell. “Anode-supported” because the anode is the thickest and strongest layer placed on the bottom of the cell, providing the mechanical support. The present thesis regards simulation and modelling of the performance of a SOFC. In the first 0D simulation, a polarization modeling of Solid Oxide Fuel Cells (SOFCs) has been developed considering activation overpotential (modified Butler-Volmer equations), ohmic resistance and concentration overpotential. For this study the assembly of Ni-YSZ / YSZ / LSM-YSZ was considered. Then, a multi-parametric analysis has been carried out to highlight which parameters are most relevant in the overall cell performance as measured by polarization curve. In particular, the following parameters were accounted for the optimization process: exchange current density (J0), temperature (T), charge transfer coefficient (a), and leakage current density (JLeak). Optimization of these parameters was investigated with the aid of Matlab software which was used to compute optimized quantities by varying current density (J ) and limiting current density (JL). The model has been tested against another simulation approach and experimental data of the literature. By optimizing the above mentioned parameters it was found that the obtained results compare well with the simulated results of the literature. The second validation was less satisfactory, but still acceptable. Then, a 2D isothermal axisymmetric model of an anode-supported Solid Oxide Fuel Cell has been developed. The model which is based on finite element approach comprises electronic and ionic charge balance, Butler-Volmer charge transfer kinetic, flow distribution and gas phase mass balance in both channels and porous electrodes. The model has been validated using available experimental data coming from a single anode-supported cell comprising Ni-YSZ / YSZ / LSM-YSZ as anode, electrolyte and cathode, respectively. Hydrogen and steam were used as fuel inlet and air as an oxidant. The validation has been carried out at 1 atm , 1500 ml min-1 air flow rate and three different operating conditions of temperature and fuel flow rate: 1073 K & 400 ml min-1, 1073 K & 500 ml min-1 and 1003 K & 400 ml min-1. The polarization and power density versus current density curves show a good agreement with the experimental data. Some of the most important results in post-processing part include: cell potential and current density distribution, anode and cathode activation loss, anode and cathode charge transfer current density, electrolyte loss, mass flow fractions, velocity and pressure gradient profiles. A parametric analysis has been carried out to highlight which parameters have main effect on the overall cell performance as measured by polarization curve, focusing on the influence of some geometrical and physical characteristics, and especially effective material properties. Finally a 1D model only considering the cell area “without channels” with the same conditions of 2D axisymmetric model has been developed; this 1D model needs some improvement towards high current density range. In future developments, in addition of using new kind of materials for the cell and different fuel composition, thermal analysis and more complex chemical reactions will be accounted for to expand the model reliability.

Questa tesi ha come argomento la simulazione e modellazione del funzionamento di una cella a combustibile ad ossidi solidi, (Solid Oxide Fuel Cell, SOFC). Inizialmente è stato sviluppato un modello 0D della curva di polarizzazione di una SOFC, considerando le sovratensioni per attivazione (equazione di Butler-Volmer modificata) e concentrazione e le resistenze ohmiche. In questa parte della tesi è stata considerata una cella tradizionale Ni-YSZ / YSZ / LSM-YSZ. Successivamente, è stata effettuata un’analisi multi-parametrica per evidenziare quali parametri siano maggiormente rilevanti sulle prestazioni di cella, valutate in base alla curva di polarizzazione. In particolare, i parametri considerati sono stati: densità di corrente di scambio (exchange current density, J0), temperatura (T), coefficiente di trasferimento di carica (charge transfer coefficient, a) e correnti di dispersione (leakage current density, JLeak). L’ottimizzazione di questi parametri è stata effettuata utilizzando il software Matlab, calcolando le quantità di interesse al variare della densità di corrente (current density, J) e della densita di corrente limite (limiting current density, JL). Questo modello è stato validato per confronto con un altro modello e con dati sperimentali della letteratura. I risultati ottenuti sono in ottimo accordo con quelli del modello alternativo della letteratura , mentre nel confronto con i dati sperimentali i risultati sono stai meno soddisfacenti ma comunque accettabili. Successivamente, è stato sviluppato un modello 2D assialsimmetrico, isotermo, di una cella SOFC anodo-supportata. Il modello, basato sulla analisi agli elementi finiti, è stato sviluppo tenendo in considerazione i bilanci di carica, le cinetiche di trasferimenti di carica (Butler-Volmer), le fluidodinamica, i bilanci di massa in fase gassosa sia nei canali di alimentazione che negli elettrodi porosi. Il modello è stato validato utilizzando i dati sperimentali ottenuti su una cella singola anodo-supportata, costituita da anodo in Ni-YSZ, elettrolita in YSZ e catodo in LSM-YSZ. La cella è stata alimentata da idrogeno umidificato come riducente (fuel) e aria come ossidante. Fissata la pressione operativa ad 1 atm ed il flusso in ingresso di aria a 1500 ml min-1, il modello è stato validato in tre diverse condizioni operative: 1073 K e 400 ml min-1 di idrogeno, 1073 K e 500 ml min-1 di idrogeno e 1003 K e 400 ml min-1 di idrogeno. Le curve polarizzazione-densità di corrente e densità di potenza-densità di corrente ottenute sono in ottimo accordo con le curve sperimentali. Alcuni dei dati più importanti ottenuti sono: distribuzione del potenziale di cella e della densità di corrente, perdite di attivazione anodica e catodica, densità di corrente di trasferimento di carica all’anodo e al catodo, perdita ohmica nell’elettrolita, gradienti di velocità e pressione dei gas, e gradienti delle frazioni molari dei singoli componenti nei gas. Per determinare quali fattori fossero più rilevanti sulle prestazioni totali della cella, è stata condotta un’analisi parametrica, concentrandosi sull’influenza di alcuni parametri geometrici, caratteristiche fisiche, e proprietà dei materiali sulla curva di polarizzazione. Infine, è stato testato anche un modello 1D considerando solo l’area della cella, escludendo i canali di alimentazione; questo modello, testato nelle stesse condizioni del modello 2D, necessita di qualche miglioramento nella simulazione della parte di curva di polarizzazione ad alta densità di corrente. In vista di un miglioramento del modello, saranno da tenere in considerazione l’analisi dei gradienti termici, l’implementazione di miscele diverse di combustibili e la simulazione di nuovi materiali.