High Fidelity Simulation of the Impact of Density Ratio on Liquid Jet in Crossflow Atomi- zation

Atomization of liquid fuel jets by cross-flowing air is critical to the performance of many aerospace combustors. Recent advances in numerical methods and increases in computational power have enabled the first principle, high fidelity simulation of this phenomenon. In the recent past we demonstrated for the first time such simulations that were comprehensively validated against experimental data obtained at ambient conditions. At combustor operating conditions, however, both temperature and pressure are significantly elevated. In this work we perform a computational study of the impact of reduced liquid-gas density ratio due to increased air density associated with operating pressure elevation on the atomization physics. A previously validated ambient condition case is used as the baseline for comparison with three cases with decreasing density ratios. The density ratio is independently varied by adjusting the gas density and velocity together so that the momentum flux ratio and Weber number are maintained constant. Results indicate a significant modification of the atomization process at lower density ratios. Although the global-scale jet penetration and trajectory are not significantly modified by the conditions, both the process of liquid breakup and the degree of atomization are altered. The trends in the degree of atomization represented by the liquid volume to area ratio extracted from in the simulation results agree with the observations from a recent experiment at elevated pressure conditions. Further effort is still required to understand the detailed physical mechanisms for atomization at different density ratios. * Corresponding author: [email protected] ILASS Americas 28th Annual Conference on Liquid Atomization and Spray Systems, Dearborn, MI, May 2016 Introduction Liquid Jet atomization In a Crossflow (LJIC) using aerodynamic forcing is a critical process occurring in the liquid fuel injection step during the operation of aircraft engine combustors. The increasingly strong requirements and regulations on improving the aeroengine combustor efficiency and reducing pollutant emissions have driven the increases in the combustor inlet pressure and temperature for an enhanced liquid fuel evaporation rate and fuel-air mixing. Since the fuel/air properties are highly sensitive to the operating pressure and temperature, e.g. the air density being strongly dependent on the pressure, the sensitivity of the atomization process to the operating conditions is strong. Thus understanding and optimizing LJIC in such elevated conditions has become an important subject in the liquid atomization research. In the current study we focus on the dependence of atomization process on the liquid-gas density ratio altered by pressure conditions. Traditional liquid atomization research has relied on experimental approaches that were mostly constrained to ambient pressure conditions due to the complexity and high cost of experiments at elevated conditions. Global features such as liquid jet trajectory and penetration and far-field spray distribution were measured and reported using a variety of empirical correlations [1-6]. A number of detailed experiments focusing on near-field atomization details [7-11] shed some light towards understanding the fundamental multiphase breakup mechanisms, despite the fact that whether we could extrapolate these understandings to assess LJIC at elevated pressure condition is still questionable. Results from only a few high pressure experiments of LJIC were reported in the literature [12-15]. Becker and Hassa [12] studied the breakup of kerosene jet in crossflow with pressure up to 15 bar. They explored the impact of pressure on liquid atomization regime, jet penetration and lateral dispersion, and droplet size distribution. However, the effects of elevated pressure were lumped into the effects of increased air momentum flux or Weber number as a result of an increase in air density. In fact, the density ratio and the Weber number are two controlling parameters that can be independently varied, with the effects of the former being rarely studied and the impact of the latter being relatively well understood from the ambient condition work. The observed changes in high pressure atomization in terms of a reduction in jet penetration and reduced sensitivity of droplet sizes to Weber number variations [12] can be explained by the shift into a higher Weber numbers shear breakup regime. And such physical link has been established during the ambient condition investigations. In Bellofiore et al. [14], a large number of flow conditions at 10 bar and 20 bar were tested, allowing the extraction of the density ratio effects independent of the Weber number. The impact of pressure on spray trajectory, plume width and coverage area was reported, yet the large degree of data scattering causes difficulty in extracting the detailed impact of density ratio, as pointed out by Herrmann et al. [16]. In a recent LJIC experiment by Song et al. [15], the air pressure was elevated from 2.07 to 9.65 bar, and the impact of density ratio was independently investigated by comparing data at fixed momentum flux ratio and Weber number. Jet breakup regime and mean droplet size downstream were shown to have a strong dependency on the density ratio while the dependence of penetration/trajectory on density ratio seemed to be weak. At very high pressure, the reduced liquid-gas density ratio may become comparable to the density ratio existing in many Gaseous Jets In Crossflow (GJIC) applications. The physics of GJIC has been extensively studied [17-20] in terms of a complex set of interacting vortex system. While the knowledge developed from GJIC studies may be borrowed for understanding the large scale vortical flow structures in LJIC at high pressure, the multiphase breakup phenomena unique to LJIC have to be understood by accounting for the phase separation caused by the presence of surface tension. Due to the complex multiphase multiscale physics involved in the liquid atomization process, traditional modelling approaches have encountered severe difficulties in capturing the impact of operating conditions. The applicability of the phenomenological models [21, 22] calibrated at ambient condition is questionable when the operating conditions are elevated. High fidelity simulation of liquid atomization [23-28] has emerged to provide a very promising path for detailed investigations of the impact of operating conditions without reliance on the experimental calibration due to its firstprinciple nature. Yet because of the challenges of overcoming the numerical instabilities that typically occur when the liquid/gas density contrast is high, a number of high-fidelity simulations of LJIC have been conducted at reduced density ratios only [25, 29], which happened to reflect the scenarios at elevated pressure conditions. Numerical study of the impact of density ratio on LJIC atomization was initiated by Herrmann and coworkers [16] considering two density ratios rρ=10 and rρ=100, both lower than the typical liquid-gas density ratios at ambient condition. Reducing density ratio was found to cause the decrease of liquid core penetration together with an increased bending and transverse spreading. It also increases the column wavelength and the mean droplet size [30]. The decrease in density ratio also leads to an increase in the normalized crossflow droplet velocity and a decrease in the normalized transverse droplet velocity, due to the increase in the relative Stokes number controlled by size. However, in addition to the limit on the density ratio set by the numerical instability challenges, in Herrmann’s work [16, 30] , the variation of density ratio was configured by altering the liquid density, not exactly the same as an air density change, which is the dominant response of pressure change. To which degree such configurations represent the high pressure condition needs further verification. Recently, simulations of high density-ratio LJIC at ambient condition have been performed by our team and successfully validated against near-field experiment [8] in terms of detailed column features and droplet formation [28, 31]. The solver with enhanced interface tracking capabilities allows stable simulations at density ratios covering a broader range of pressure conditions, thus enables a comprehensive study of the impact of density ratio changes associated with operating pressure variations. In this work, we present the results from three LJIC simulation cases with increasing air density reflecting the trends with pressure increases. For the current investigation, all other impacts of pressure are ignored. The cases were set up based upon a previously validated ambient condition case. The density ratio was independently varied while the momentum flux ratio and the Weber number were fixed to be constants. Previous ambient condition simulation case was used as the baseline for investigating the impact of decreasing density ratio. In the following, previously adopted formulation and numerical methods are briefly highlighted. The computational configurations of LJIC with varying density ratios are described. The impact of density ratio on the qualitative feature and quantitative degree of atomization is presented. Finally, summary and conclusions