Modeling of self-heating monolithic catalytic reformer for hydrogen production for fuel cells

Dissertation for Master's degree

In the field of chemical engineering, thermokinetic orientation

Abstract

With the increase in the use of fuel cells in the industry and in the form of on-site applications, there is a need to develop hydrogen production units. In this research, a monolithic catalytic reformer in which autothermal reforming of methane takes place is modeled in three dimensions. The catalyst used in this modeling is 5%. This modeling includes the simultaneous solution of survival equations in which the reactions that have taken place are also given an effect. One channel of this monolithic reactor has been used as a computing domain. The results of this modeling are in good agreement with the laboratory data available in the sources. This model has been used to estimate the performance of the reformer in other operating conditions. The studied parameters include the molar ratio of oxygen to methane input (O2/CH4), the molar ratio of water vapor to input methane (H2O/CH4) and the temperature of the input gas to the reformer. Finally, after examining the effect of the mentioned parameters, it was concluded that in order to achieve the maximum amount of hydrogen in the investigated range in terms of operational parameters, the molar ratios of O2/CH4 and H2O/CH4 input to the reactor should be selected as 0.445 and 3.8, respectively. Also, the temperature of the gas entering the reactor should be 600 °C.

Key words: hydrogen, autothermal reforming, methane, monolithic reactor

1-1- Introduction

Fuel cells directly convert the chemical energy of a fuel into electrical energy. Fuel cells, due to their high power density, harmless by-products for the environment and fast recharging, are considered as one of the new technologies for energy production in the future and a suitable alternative for energy production from conventional methods. The most important advantage of fuel cells, compared to reciprocating and Stirling engines, is the possibility of achieving higher efficiency in converting fuel to electricity, which is especially suitable in polluted areas. For fuel cells, hydrogen is the preferred fuel. The advantage of using hydrogen in a fuel cell is its high reactivity for the electrochemical reaction of the anode and its non-polluting nature. However, hydrogen does not exist as a gaseous product in nature. For this reason, water, fossil fuels and other materials with high hydrogen density must be used, which can be a difficult and expensive process. Also, storing hydrogen, especially for use in vehicles and household applications, is not yet easily possible. For this purpose, the use of fuel processing systems has been suggested to produce hydrogen needed for fuel cells on site. The use of these fuel processing systems makes it possible to combine the high energy density of fuels and the high power density of the fuel cell and creates a system with high efficiency. So far, a lot of research has been done to investigate fuel processing systems in the form of laboratory work and modeling.

There are three reforming methods for hydrogen production, which include steam reforming (SR) [1], partial oxidation (POX) [2] and autothermal reforming (ATR) [3]. Steam reforming is endothermic and partial oxidation is an exothermic process. Reactants for self-heating reforming include steam, oxygen and fuel. In fact, self-heating reforming is a combination of reforming with steam and partial oxidation. Self-heating reforming is the preferred method for use in a vehicle due to the lack of an external heat source and the formation of smaller amounts of soot. In this study, with the help of Computational Fluid Dynamics (CFD) [4] autothermal methane reformer has been modeled.

The most important goal of this research is the numerical study of autothermal methane reforming process with the help of 3D modeling. With the help of modeling results, changes in temperature and concentration of components can be studied at any point inside the reactor. The importance of this modeling goes back to providing information for the design of reforming systems, which can be used to avoid problems such as the formation of hot spots inside the reactor that lead to damage to the catalyst.. Therefore, CFD modeling helps to optimize reactor design and determine conditions that lead to improved fuel conversion efficiency. Also, the time and cost needed to implement new ideas and designs will be reduced.

Many researches for self-heating reforming of methane on conventional catalysts such as nickel, platinum, palladium, etc. has taken place In many of these studies, the catalyst used for partial oxidation and steam reforming is different. The modeling done in this research is mainly based on the relationship of reaction rates on conventional catalysts. In the research conducted by the author, so far the modeling of self-heating reforming of methane on 5% catalyst in a monolithic reactor has not been done. The aim of this research is to model autothermal reforming of methane on 5% catalyst with the help of computational fluid dynamics. The advantage of using 5% catalyst is that it can promote both partial oxidation and steam reforming reactions. In the modeling, modified speed equations for 5% catalyst have been used. The reactor selected in this research is a monolithic catalytic reactor. Monolithic reactors consist of a large number of parallel flow channels separated by solid walls. Monolithic reactors are suitable for mobile applications due to their high surface-to-volume ratio and low pressure drop.  However, monolithic reactor modeling is very costly and time-consuming. For this purpose, the behavior of one channel of the monolithic reactor is assumed to be almost the same as the behavior of the entire monolithic reactor, and the geometry of a channel is chosen as the computational domain. This modeling includes a three-dimensional model for the reactor, which includes the equations of conservation of mass, momentum, energy, and the survival of chemical species, as well as a model to consider the mechanism and speed relationship of reactions. These equations have been solved with the help of Fluent 6.3.26 software, which is based on finite volume calculations [5]. To consider the reaction speed relationship, programming in C++ environment has been used, which can be used for similar tasks. The results of this modeling have been compared with the laboratory work done for autothermal reforming of methane on 5% catalyst. In the following, the effect of changing operating parameters on the amount of hydrogen and carbon monoxide produced and the temperature profile inside the reactor have been investigated. The investigated operational parameters include the molar ratio of oxygen to methane (O2/CH4), the molar ratio of water vapor to methane (H2O/CH4) and the temperature of the gas entering the reactor. After presenting the introductions in the first chapter, the second chapter examines the reforming processes used for hydrogen production. Next, the kinetic models presented for methane reforming processes are presented, and finally, the modeling done for monolithic reactors is reviewed. The third chapter presents the characteristics of the monolithic reactor used for modeling. Also, assumptions and equations of conservation of mass, energy, momentum and survival of chemical species governing the modeling are presented. Finally, the kinetic equations used for autothermal reforming of methane on 5% catalyst are given. In the fourth chapter, the results of the modeling are discussed and the most optimal mode (within the investigated range) that leads to the highest amount of hydrogen production is presented. Finally, in the fifth chapter, the future solutions to improve the modeling results are presented. of developing onboard hydrogen production units is increasing. In this study, a three dimensional model is developed for a catalytic monolith reformer in which methane autothermal reforming occurs in it. Catalyst is selected to be 5% Ru-Al2O3. In this simulation, conservation equations are solved considering reactions occurring. A single channel of the monolith reformer is selected as the computational domain. Simulation results can satisfactorily predict the experimental data in literature.