The control of the injected spray is important when optimizing performance and reducing emissions from diesel engines. The research community has conducted extensive research especially on smaller four stroke engines, but so far only little has been done on sprays in large two stroke engines. The latter is the subject of this dissertation. The theory and experimental findings on diesel sprays are investigated, including e.g. spray parameters and droplet break up. It is found that no complete theory is yet present and large challenges lie ahead. Generally, there is fairly good consensus on which physical quantities have an effect on the spray characteristics, but there are some discrepancies. It is found that small differences in setup have large effects on the spray characteristics, especially differences in nozzle layout have shown a large effect. The nozzles of large two stroke engines both have different scales and other designs than those used in the literature, so extending results from the literature will require experiments on this particular type of setup. Numerical investigations of diesel sprays are performed using the Eulerian/Lagrangian engine CFD code Kiva. In agreement with other authors it is found that cell sizes applicable in Kiva are inadequate to capture the scales of the spray. Different approaches on compensating for the too large grid are tested, and it is concluded that the problem of artificial diffusion of momentum is the most critical to be solved. A criterion is derived for an absolute maximum grid size for achieving reasonable momentum transfer from liquid to the gas phase. Even smaller cell sizes are required for e.g. resolving gradients in the flow field. An experimental setup is constructed, investigating sprays at atmospheric conditions, using nozzles and an injection system identical to those of large two stroke diesel engines. Specially designed single hole nozzles enables in nozzle pressure measurements. A number of experimental methods are successfully tested, including a high speed imaging system using reflected light, a low cost shadowgraph system, integral mass measurements and transient total momentum measurements. A device for qualitative measurements of spray momentum distribution is also tested but needs further improvements. The measurements are used to quantify the spray by e.g. finding spray angle, penetration and discharge coefficients. A cross correlation analysis is successfully conducted on data series of spray width, giving the average velocities of the spray borders at different positions. The experimental methods are used on four configurations of single hole injectors. It is found that hole type and upstream conditions has an effect on both the amount of injected mass and momentum and on spray angle. Some configurations had the nozzle slide removed to get very stable upstream conditions, which results in a spray with asymmetrical spray distribution. There is a dense main spray carrying the main part of the momentum, and besides that there is a disperse region traveling with a lower velocity. The source for the asymmetrical spray is the upstream conditions, including differences in nozzle phenomena. For nozzles with the slide, typical turbulent sprays are observed but border velocity calculations indicate a disperse region close to the nozzle. The asymmetrical spray distributions means that care should be taken when using spray parameters like spray angle, for e.g. inlet conditionsfor CFD computations.
Main Research Area:
Meyer, Knud Erik, Sorenson, Spencer C, Mayer, Stefan