| description abstract | Atomization plays an important role in the gasification or combustion of biooils, where the atomizer parameters need to be properly controlled to efficiently atomize a highly viscous liquid at elevated pressures with imparting the least amount of kinetic energy to the discharged droplets because of evaporation and chemical reaction constraints. With a focus on biooil deployments in microgas turbines (MGTs), an aqueous surrogate of a preheated biooil, injected from an original equipment manufacturer (OEM) twinfluid atomizer, is used in the present study for spray size and velocity measurements at elevated pressures. The experiments were conducted in High Pressure Spray Facility of the National Research Council of Canada (NRC) using various optical diagnostics including laser sheet imaging (LSI), phase Doppler anemometry (PDA), and laser diffraction (LD). A scaling strategy was adopted to conserve the ranges of gastoliquid momentum flux ratio, M, at different working pressures, P. Over the range of conditions studied, it is found out that the cone angle of sprays is insensitive to P, but they decrease with increasing M. For a constant value of M, droplet mean diameters increase and their corresponding velocities decrease with increasing P, attributed to the effect of gastoliquid density ratio on the primary breakup of a liquid jet in a coaxial gas stream. Therefore, to estimate the Sauter mean diameter of spray droplets, D32, a correlation previously reported in the literature is modified by including the effect of system air density at elevated pressures, and a novel correlation is proposed based on four dimensionless groups, namely, gas Weber number and gastoliquid momentum flux ratio, density ratio, and viscosity ratio. The detailed results obtained in the present study could be used to define the optimal parameters required for twinfluid atomization of high viscosity liquids with various atomization gases under realistic operating conditions and to enhance the capabilities of their numerical simulations. | |
