Experimental studies of ammonia-hydrogen diffusion flames for kinetic model development

Abstract

The motivation for the work presented in this dissertation is the need to better understand the kinetics of ammonia combustion, and in particular that of ammonia-hydrogen fuel blends. Ammonia has been the subject of growing interest as a carbon-free fuel for storing renewable energy. The combustion of ammonia presents challenges in flame stability and emissions of unburned ammonia and nitrogen oxides. The limitations of current kinetic mechanisms for ammonia combustion are evident in the discrepancies between predictions and measured levels of nitrogen oxides across a wide range of flame conditions. A primary goal for this work was to provide in-flame and exhaust species measurements to extend database available for kinetic mechanism validation to non-premixed flames, for which little data exists. Two canonical non-premixed flames were studied in this work: (1) the counterflow diffusion flame, which can be modeled as a quasi one-dimensional system, and (2) the co-flowing diffusion flame, which can be modelled as an axi-symmetric two-dimensional system. Laser-induced fluorescence (LIF) was used to obtain non-invasive measurements of species in the flames studied. Saturated planar LIF was used for species profiles of nitric oxide (NO), imidogen (NH) and hydroxide radicals (OH). The saturated NO and OH profiles were calibrated by laser absorption, measured by fluorescence intensity in a bi-directional configuration. Saturated LIF NH profiles were calibrated via the OH absorption measurements. NH₂ radicals were also measured by planar LIF, and chemiluminescence used to obtain profiles of the exicited states OH* and NH₂*. For the co-flow flame, exhaust nitrogen oxide levels, including nitrogen dioxide (NO₂) and nitrous oxide (N₂O), were measured with a Fourier Transform Infrared (FTIR) spectrometer. In addition, extinction strain rates of the counterflow flame were measured for low hydrogen content fuel. Species and extinction measurements from this work were used to evaluate existing kinetic mechanisms for ammonia combustion, and to update a recently-published mechanism by Shrestha and coworkers. The updated kinetic model showed significant improvement in predicting the experimental results, and better performance than the selected existing mechanisms. The updated kinetic model also performed well when validated with published data from premixed ammonia-hydrogen flames. However, remaining discrepancies between predictions of NH and NO and the experimental results in this work indicate the need for additional work in understanding amine radical (NHᵢ) reactions and NO formation in ammonia flames.