UC Irvine

Data centers are the largest sectoral consumer of electricity in the United States and in order to reduce their carbon emissions the introduction of advanced power and cooling technologies is a necessity. Solid-oxide fuel cells (SOFCs) are a promising technology that can provide high-efficiency electricity to data centers while utilizing existing natural gas infrastructure for fuel. Additionally, SOFC cogeneration with absorption technology can provide free chilling while meeting the requirements of server computers. This dissertation explores how the successful integration between solid-oxide fuel cells and absorption refrigeration can meet the principal demands of data centers, reduce their carbon footprint, all while running on existing fuel infrastructure. The feasibility of these systems’ integration is explored numerically and experimentally. A server row-level model is constructed based on the experimental data collected from the lab-scale SOFC and absorption chiller setup. The operation of the SOFC and absorption chiller is then optimized to minimize carbon emission intensity, levelized cost of utility, and maximize primary energy savings (PES) ratio. The integrated system is able to operate with a carbon emission intensity of 0.38 kg/kWh compared to the grid’s 0.44 kg/kWh, at a PES ratio of 25.8%, costing 75.7 $/MWh of utility production compared to the grid’s 58.5 $/MWh. Additionally, the integrated system can be further optimized to match the transient chilling requirement of the server computers. Adjusting the stack temperature and fuel utilization of the SOFC can control the chilling capacity of the absorption chiller allowing the SOFC to operate more efficiently when chilling is not required but still meet peak chilling demand. Successful deployment of this active control can reduce the carbon emissions of a server row by 27.2%. Additionally, the real-world feasibility is also explored in the case study of a 1.4 MW high-temperature fuel cell cogeneration plant on the UCI Medical Center. Operational and economic investigation of the system reveal that the primary driver of such an integrated system’s success is the reliability of the power produced and consideration of stack degradation throughout its lifetime. Finally, SOFC systems can incorporate blends of hydrogen or biogas to further reduce their carbon footprint and move towards a hydrogen economy. The production of green hydrogen can also come from solid-oxide technology at high efficiencies. This operation is explored experimentally with challenges such as delamination and electrode degradation analyzed. Ultimately, the thermal integration solid-oxide technology presents a promising step towards decarbonizing the digital infrastructure industry.