No abstract

Predictive models of heating and evaporation of fuel droplets in the dense region of sprays are essential to optimize the design of combustion chambers in internal combustion engines. This topic is addressed here, based on an experimental study using lines of equally-spaced droplets evaporating in a high temperature chamber (540°C). The experimental setup allows controlling several key parameters including the droplet size, velocity and the inter-droplet distance. The volume-averaged temperature of the droplets is measured using two-color laser-induced fluorescence, while the droplet size and velocity are deduced from a double-pulse shadowgraphy method. The combination of these measurement techniques allows evaluating the heating and evaporation rates of single-component droplets made of ethanol, n-dodecane, n-decane and isohexane. Nusselt and Sherwood numbers are estimated from the experiments and compared with the existing correlation concerning the isolated droplet, which allows quantifying the effects of droplet interactions on the heat and mass transfers. The droplet spacing appears to have a strong influence on the size and temperature evolutions. However, it also seems necessary to consider the development of a boundary layer around the chain of droplets. As the thickness of the boundary layer increases with the distance from the injector, forced convection has a more and more limited influence on the heat and mass transport. Inside the boundary layer, the transfers are mainly governed by diffusion and convection by the Stefan flow. In a first approach, these effects are partially incorporated in a reduced parameter using the concept of volume of influence. Then, a more detailed study based on numerical simulation is carried out. The Navier-Stokes equations and the vapor transport equation are solved for various periodical arrangements of droplets. A parametric study of the influence of the main characteristic numbers involved allows to infer a correlation for Nusselt and Sherwood numbers which is finally validated against the measurements.