Study of exhaust gas recirculation (EGR) on the operation of liquid-fueled gas turbines

Abstract

Gas turbines, one of the main producers of electricity, and propulsion for transportation, are significant contributors to the US greenhouse gas (GHG) inventory. The goal of this thesis is to study the process of removing GHGs (i.e., CO2) from industrial/marine propulsion scale gas turbines by leveraging carbon capture while maintaining low NOx and CO emissions. Recirculating the exhaust gas back into the intake of the gas turbine, in a process termed Exhaust Gas Recirculation (EGR) poses a pathway to reduce the scale of industrial and shipboard carbon capture and storage, where weight, space, power, and cost are significant constraints, by enriching the CO2 concentration in the exhaust. However, the amount of EGR and the associated exhaust CO2 concentrations are limited by the onset of combustion instabilities. This occurs because EGR displaces intake air with oxygen-depleted exhaust, reducing the oxygen available for stable and efficient combustion. Earlier works have studied the incorporation of EGR in natural gas-fired turbines. The industrial and maritime sectors have demonstrated growing interest in integrating carbon capture systems with gas turbines operating on liquid hydrocarbon fuels, including diesel, and emerging sustainable fuel alternatives. This study is the initial attempt to understand EGR limits on gas turbine combustion with liquid fuels through the development of a chemical reactor model (CRN) comprised of a series of joined perfectly stirred reactors (PSR), representing different flame and flow zones within the combustor. The CRN is a numerical investigation tool that provides plenary species composition and temperature data at a fraction of the time and cost of conventional reacting flow computational fluid dynamics (CFD) simulations. Chemical reactor networks (CRNs) typically developed for gaseous fuel applications have shown potential to predict gas turbine emissions and operational limits and have been incorporated into gas turbine technology Research and Development (R&D). Nonetheless, challenges exist with implementing CRN approaches with liquid fuels due to the inherent multi-dimensional nature of liquid fuel evaporation and combustion. To incorporate liquid fuels in a CRN model, a stratification method was employed, where a PSR was created to model a portion of the fuel flow as a lean mixture representing the premixed/vaporized fraction of the fuel in parallel with a PSR operating at stoichiometric mixture fractions to simulate the portion of the non-premixed/vaporized fuel flow and its combustion. The C3 chemical mechanism was implemented in this modeling approach to simulate the diesel fuel chemistry (using a simple surrogate) while incorporating chemical reactions to predict NOx and CO emissions. This approach provided an avenue of control within the reactor network to predict emission levels, which was compared and validated against Solar Turbines Centaur 40 (SoLoNOx combustor) engine test emissions data. Following model validation, Exhaust Gas Recirculation (EGR) was introduced, revealing a decreasing trend in NOx emissions with increasing EGR rates, accompanied by a corresponding rise in CO emissions. At elevated EGR levels, CO concentrations reached thresholds associated with significant combustion instability and eventual flameout. The findings indicate that the SoLoNOx (Dry Low Emissions) liquid-fueled combustor exhibits greater EGR tolerance compared to the conventional liquid-fueled combustor.