Sensitivity Analysis of Auto-Ignition Simulation at Gas Turbine Operating Conditions

Abstract

The demand to reduce CO2 emissions favors the use of alternative hydrogen-rich fuels, which can stem from precombustion carbon capture or power-to-gas technologies. These fuels are characterized by a higher reactivity and reduced ignition delay time compared to natural gas. Therefore, current combustor designs need to be adapted to the new requirements. Numerical modeling greatly assists the further development of such systems. The present study aims to determine how far a sophisticated computational fluid dynamics (CFD) combustion method is able to predict auto-ignition at real engine conditions. Scale-resolving computations of auto-ignition were performed at elevated pressure (15 bar) and intermediate temperatures (>1000 K). The conditions are similar to those occurring in premixing ducts of reheat combustors. A nitrogen-diluted hydrogen jet is injected perpendicularly into a stream of hot vitiated air. The scale-adaptive simulation (SAS) method as proposed by Menter and coworkers has been applied. The chemistry is captured by direct inclusion of detailed kinetics. Subgrid fluctuations of temperature and species are considered by an assumed probability density function (PDF) approach. The results are compared with appropriate experimental reference data. The focus of the present work is set on the identification of the major sources of uncertainty in the simulation of auto-ignition. Despite the very challenging operating conditions, satisfactory agreements could be obtained within experimental uncertainties.