NUMERICAL STUDY OF AMMONIA-HYDROGEN COMBUSTION CHARACTERISTICS IN SWIRL COMBUSTOR USING COMPUTATIONAL FLUID DYNAMICS

• Main Author: Zuhair Altair, Firas
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/79423
• Institution: Institut Teknologi Bandung

The current energy usage has increased by 62% per person since 1965 for various purposes, such as electricity, transportation, and heating. The primary source of energy is fossil fuels, leading to air pollution and CO2 emissions contributing to global warming. It is estimated that temperatures will rise from 1.9°C to 3.5°C by 2100. A solution to reduce emissions is transitioning to renewable energy sources like hydropower, solar, and wind, despite their limited availability. The technology of ammonia-fueled gas turbines presents an attractive alternative. Ammonia is widely used in various sectors, and its production continues to increase, reaching 150 metric tons in 2022. Therefore, ammonia has the potential to serve as an environmentally friendly fuel for energy and transportation. This approach could be an efficient solution to address global energy challenges and climate change. Therefore, in this research, the investigation was carried out under swirl flow conditions for a mixture of ammonia and hydrogen at atmospheric pressure. The analysis utilized a combustion chamber designed by the Clean Combustion Research Center (CRCC). This combustion chamber employs four tangential inlets and two axial inlets to mix fuel and air before entering the combustion zone. The objective of this analysis is to determine the CHEMKIN mechanism that provides accurate predictions for NO species and the influence of Reynolds and Swirl numbers on flame shape and NO emissions. The method employed is Computational Fluid Dynamics (CFD) using ANSYS Fluent software, with the turbulence model being the Reynold Stress Model (RSM) and Partially Premixed as the combustion model. Boundary conditions at the inlet include mass flow rates, progress variable, and fuel fractions, while at the outlet, pressure, progress variable and fuel fractions are specified. The walls are assumed to be adiabatic and no-slip conditions are applied. In this combustion simulation, various CHEMKIN mechanisms were used, including GRI-Mech 3.0, Reduced GRI-Mech 3.0, Okafor Mech, and Otomo Mech for ammonia fractions ranging from 0.7 to 1.0 and equivalence ratios ranging from 0.4 to 1.1. Additionally, the analysis included variations in Reynolds numbers ranging from 5,000 to 30,000 and Swirl numbers ranging from 0.25 to 1.5. The numerical results obtained indicate that Otomo Mechanisms with the FGM-Diffusion flame and reactor network combustion model is the most accurate compared to other CHEMKIN mechanisms and combustion models, with an average relative difference of 150% from experiment. This model demonstrates the highest accuracy at an ammonia fraction of 0.8, with an average relative difference of 13%, while at ammonia fraction of 0.9, the average relative difference of 13% from experiments. On the other and, combustion variables such as flow pattern and temperature are not influenced by the choice of combustion model and CHEMKIN mechanism. Therefore, an ammonia fraction of 0.8 and an equivalence ratio of 1 were selected as good combustion stability and minimal unburned fuel residue From the variations in Reynolds number and Swirl number for the geometry, it is observed that the total pressure loss increases with higher Reynolds numbers and Swirl numbers, with the largest total pressure loss recorded at 2,980 Pa. Additionally, the flow patterns and temperature distribution exhibit relatively consistent shapes and values in the injector plane region. Regarding NO emissions, it is noted that increasing the Reynolds number results in higher NO emission values, with the smallest NO emission value recorded at 2,067 ppm.