Modelling, Simulation, and Analysis of Molten Carbonate Fuel Cell Systems

Abstract

Fuel cells convert hydrogen into electrical energy with high efficiency and low pollutants. Among many types of fuel cells, the molten carbonate fuel cell (MCFC) is considered in this work. MCFC works at a temperature of around 650oC that enables efficient recovery of residual pressure. In addition, no precious metals are required as fuel catalysts. Fuel cells require primary resources, such as fossil fuels in order to produce hydrogen which is the main material within the fuel cell stack for generation of electricity. In this study, diesel is used as a primary resource. Diesel is a potential fuel, produced by the fractional distillation of crude oil. It contains sulfur components, such as dibenzothiophene (DBT), benzothiophene (BT), and 4,6-methyldibenzothiophene. Sulfur components can poison the catalyst in a steam reformer as well as the fuel cell itself. That is why a pre-treatment (desulfurization) is required in order to reduce the sulfur content. Adsorptive desulfurization was used in this work. It typically uses an adsorption process with a particular adsorbent. Examples of commonly used adsorbents for desulfurization are activated carbon, zeolite silica gel, etc. Upon desulfurization, the sulfur concentration in the diesel feed was reduced from 10 to 0.1 ppm. This secured input into the steam reformer without interfering with the performance of the fuel cell. Desulfurized diesel will pass through several other processes, such as: steam reforming, water gas shift, as well as a purifying unit. Subsequently, the produced hydrogen is used as an input to the MCFC unit, and 2 MW of electrical energy is generated. All processes ranging from diesel desulfurization and conversion to electricity generation in the MCFC unit were modeled and simulated using Aspen HYSYS, Techno economic analysis was also performed in order to calculate the cost of power generation. Desulfurization is an important part of the fuel cell system simulation. However, in Aspen HYSYS (AspenTech) this process cannot be properly simulated. Because of that, the desulfurization process was simulated using COMSOL Multiphysics (Altsoft) with the computational fluid dynamics (CFD) concept that can be used for distribution of diesel concentration in the reactor when the desulfurization process occurs. In the fuel cell system simulation, a molar flow 1300 kg/h in the diesel feed is required to generate 2 MW of electricity in the MCFC module. The percentage of hydrogen fuel reacted in the fuel cell was 89 %, while the efficiency of the fuel cell stack was 52.33 %. In the desulfurization part, initial sulfur concentration in diesel feed was 10 ppm. Upon entering the column, the adsorption will begin to occur. It can be seen from the CFD results that the color will start to change which indicates the direction of sulfur adsorption by the adsorbent. Differences in surface area on the inlet (small) to the larger surface area and there are adsorbents in the column, causing the drag force. Friction between diesel and the adsorbent also, resulting in reduction of pressure and velocity of feed. Since the velocity is reduced, the contact time between diesel and the adsorbent is longer and this causes sulfur to be adsorbed to the desired extent. Final concentration of sulfur at the outlet was 0.1 ppm. Within techno economic analysis, working capital and fixed capital investment (FCI) were estimated at $7,000,000 and $28,700,000, respectively. Utility cost, total installed cost and manufactured cost were $951,610, $19,600,000 and $57,388,359. Desulfurization segment of the work is very important to the whole process, and since an adsorbent was used for that purpose, the adsorbent price will have a highest effect on the overall cost. In the sensitivity analysis, the parameters were adsorbent price, annual interest rate and project life. The price of the adsorbent was obtained from Zigma Aldrich (300 $/kg) and compared the Net Present Value (NPV) and payback period with the adsorbent price of 225 $/kg and 375 $/kg, variate annual interest rate 10 %, 20 %, and 30 %. For project life, we compared 10 years and 20 years. With the adsorbent price at 300 $/kg, NPV was 2.79 $ million dollars and the payback period was 6.6 years. If the adsorbent price decreases to 225 $/kg NPV would increase significantly: to 23.4 million $ with a payback period of 3.3 years. Meanwhile, if the price gets to 375 $/kg, then the NPV is 7.81 million dollars with an undefined payback period. Project life of 10 years gives NPV of 2.79 million $ with a 6.6 years payback period. However, if the project life increases to 20 years, NPV becomes 10.57 million $ with a 6.3 payback period. Annual interest rate of 10 % yields an NPV of 2.79 million $ with a 6.6 years payback period. An increase of annual interest rate has little effect to NPV. Annual interest rate of 20 % yields an NPV of 3.91 million $ with a 9.3 years period, whereas 30 % annual interest rate yields an NPV of 8.1 million $, but with an undefined payback period. From the comparison, it was concluded that the adsorbent price has the highest impact on NPV and the payback period.