Optical measurement of nitric oxide and hydroxyl radicals distributions in combusting diesel sprays
The development and combusting behaviour of a diesel spray were investigated to provide a deeper understanding of the formation of nitric oxide (NO) in diesel engines. To characterise the spray, the nozzle flow was measured by the rate tube technique. The sensitivity of the flow to injection pressure was shown to follow the theoretical behaviour. Penetrations of the liquid spray were measured by means of high speed video imaging. The innovative measurements of the liquid penetration during the combustion allowed combustion phases and liquid jet lengths to be associated. Hydroxyl (OH) radicals were acquired by planar laser-induced fluorescence (PLIF). Combined with high speed videos of the flame natural luminosity, they were used to identify precisely the evolution of combustion in time and space. The measured OH distributions compared favourably with results from simulations using the KIVA code. The OH radicals were shown to be present mainly in the mixing controlled phase, distributed in a thin layer around the vapour fuel in the jet, within the diffusion flame location. OH radicals could be seen as early as 0.4 ms before the pre-mixed heat-release spike and until the end of apparent heat release. In the conditions studied, the diffusion flame, therefore, spanned most of the combustion process, starting very soon after autoignition. Finally distributions of NO were acquired by LIF and compared with the evolution of combustion. NO was found to appear 0.5 to 1 ms after the development of the diffusion flame, on the lean side of the flame front, outside the region with a high density of OH radicals but also later on, downstream the spray, on the outskirts of the zone with high soot density. The formation rate of NO was found almost constant during the mixing controlled combustion, with a small increase at the end of injection, when the flame collapsed on the fuel spray. The observed increase was linked to a rapid cooling of the flame plume and the associated freezing of the thermal-NO mechanism. Varying injection pressures did not significantly affect the overall formation rate although peak NO densities were seen to gradually move downstream the flame plume with increased injection pressure. NO formation increased with the in-cylinder pressure in accordance with a higher density of air and higher local temperatures.