Micro droplet evaporation and modeling

2017-02-17T02:11:32Z (GMT) by Nguyen, Dung Trung
The study of fuel droplets undergoing evaporation in convective environments has been made. Droplets studied have sizes less than 100 μm which are close to the range found in modern high pressure fuel spray systems. To enable this study, a co-flow reactor and a piezo-electric nozzle were used. The evaporation environments with varying degrees of convection were produced by use of an inverse flat-flame burner operating on a combination of CH4, H2, O2 and N2. Temperatures studied are up to 676 K which closely resembles that at the end of the compression stroke of a diesel engine. To measure the evaporation rate in these circumstances, digital inline holography was selected and developed to a magnified version which was combined with particle image velocimetry for measurement. This thesis presents the development of the measurement technique and a full analysis of potential sources that contribute to the measurement uncertainty. The development was based on use of a coherent collimated laser beam with a wavelength of 635 µm for illumination and a high speed camera with CCD array resolution of 12 µm/px for recording. To find the focus position of the droplet in the optical path at which the image is used to calculate the droplet size, a criterion of maximum area with a threshold value was chosen. Magnification factor used in this study was 3.996 which enables the measurement of droplets with sizes down to about 10 μm. Droplet sizes measured were found to have acceptable uncertainty with the largest error to be about ±4 µm resulting from optics imperfections. Although the cross-correlation of reconstructed images was found to be affected by the reconstruction quality, the uncertainty in velocity measurement was found negligible. The evaporation rate was estimated to have a typical error of 9%. Experiment was done for petroleum distillates n-nonane, n-decane and n-dodecane; and standard light diesel as a multicoponent fuel in two varying heating conditions. It was found that the evaporation process is not only dependant on the surrounding temperature but also largely influenced by the preheating effect. For current conditions, all fuels initially experience thermal expansion except for nonane. This expansion level as well as the preheating time was found to be affected by the droplet size which, in turn, was found to play no noticeable role on the evaporation rate for the low convections used. Amongst alkane fuels the evaporation rate was found to be higher for the lighter fuel but when the gas temperature is close to the peak used, their rates become virtually the same to be about 0.22 mm2/s. For all temperatures, diesel has the lowest evaporation rate which initially shows preferential vaporisation but then resembles the diffusion limit behaviour. For the peak temperature used, its rate is about 0.19 mm2/s. A simple infinite conductivity Lump model accounting for convection as well as the transient droplet preheating was formulated which was found to be efficient in describing the single component droplet behavior in this work in terms of the evaporation rate but slightly underpredicts the initial expansion and slightly overpredicts the droplet lifetime. If a proper set of properties is used, diesel may also be approximated by this model for certain conditions. This model indicated that the evaporation rate is controlled by the droplet temperature which tightly follows the gas temperature and that the droplet temperature is always well below the boiling point. Extension of the model to account for the Stefan flow effect and internal liquid temperature variation by using an effective conductivity concept revealed that for current conditions, the Stefan flow has very little effect in slowing down the process and that the initial strongly unsteady evaporation is due to the difference between the droplet surface and center temperatures. This difference is however not large for all cases even though the liquid thermal transport was found to be mostly by conduction. The Lump model was favored by the findings in this work that it could be used to approximate micro droplets undergoing evaporation in high temperature, strongly convective environments at least for single component fuels such as those used here.