Modelling sulphur chemistry in diffusion flames

• Main Author: Cerru, Fabien
• Published: Imperial College London 2008
• Subjects: 547.06
• Online Access: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.500009

In the present thesis, the effects of sulphur seeding on sulphur oxidation and flame structure are investigated. First, a detailed sulphur mechanism consisting of 12 sulphur containing species and 70 reversible reactions is derived and added to a larger detailed H-C-N-O mechanism that models hydrocarbon combustion. The functionality of the mechanism includes the pyrolysis and oxidation of hydrogen sulphide and sulphur dioxide oxidation. The mechanism is validated by comparison with existing laminar flame measurements. The sensitivity of reaction paths to alternative rate determinations has been analysed. The detailed mechanism produces good agreement with experimental data. In a second part, the detailed reaction mechanism is reduced to six independent scalars using steady state assumptions. The reduced mechanism provides good agreement over the range of conditions studied. Further simplifications are made for S02 oxidation and a two-step mechanism is derived. It retains good agreement with a more limited set of experiments. In a third part, the sulphur oxidation mechanism has been applied to model the effects of sulphur addition under turbulent flame conditions. The first order conditional moment closure (CMC) method for slender layers is implemented to model the turbulence-chemistry interactions and validated by comparison with the Sandia flames D to F. Excellent agreement in terms of conditional means and averaged values is observed for a low degree of local extinction. The effect of sulphur on turbulent flame structures and pollutant emissions is then evaluated for a natural gas diffusion flame with a moderate Reynolds number. Sulphur dioxide can be seeded into the fuel or the oxidiser. When sulphur dioxide is introduced in the oxidiser, the NOx emission level is decreased by 20 %. This feature is observed both experimentally and computationally.