Atomization and Sprays, Vol.20, No.5, 435-451, 2010
LARGE EDDY SIMULATION OF DROPLET STOKES NUMBER EFFECTS ON MIXTURE QUALITY IN FUEL SPRAYS
This paper provides a continuing study to the recent study by the authors and the results presented herein are based on the same numerical experiments enriching the previously obtained picture (Vuorinen, V, Hillamo, H., Kaario, O., Larmi, M., and Fuchs, L., Atomization Sprays, vol. 20, pp. 93-114, 2010). Large eddy simulations (LES) and Lagrangian particle tracking (LPT) techniques are applied to simulate turbulent fuel spray evolution using high spatial and temporal resolution (Delta x < 10(-4) m and Delta t < 10(-7) s), hence providing detailed information on the anisotropy of the spray problem. The aim of this paper is to apply LES/LPT to identify possible causes to observed fuel spray heterogeneities in nonevaporating fuel sprays (Cao, Z., Nishino, K., Mizuno, S., and Toni, K., Exp. Fluids, vol. 29, pp. S211-5229, 2000; Hillamo, H., Kaario, O., and Larmi M., SAE Paper No. 2008-08PFL-552, 2008). In this context, the numerical setup is new since it avoids the dense spray region by emulating the dilute spray part as a droplet laden jet. The apparent role of droplet diameter (d) to spray dispersion, spray structure, and the mixture quality can then be investigated in detail. This paper is motivated by previous investigations in which smaller nozzle hole diameters and higher injection pressures together are known to yield smaller droplets, improved mixing, and higher levels of entrainment. The main emphasis is on monodisperse sprays, but for comparison, also some polydisperse sprays and a single-phase jet (which is slightly loaded with tracer particles) are considered. The droplets are assumed to be spherical; they are of constant size and do not interact with each other (i.e., assuming only two-way momentum coupling). The droplets have a Stokes number within the range of 0.07 <= St(p) <= 2.56, corresponding to diameters between 2 <= d <= 12 mu m. The results include (i) visualization of spray structures, combined with (ii) quantitative and statistical analysis of the results. The spray cloud shape and its internal structure are shown to depend on Sty, and the apparent differences in spray dispersion are quantified using two mixing indicators that both depend strongly on St,, and imply that the smaller droplets yield better mixing than the larger droplets: (i) a diffusion coefficient, and (ii) a mixing index.