Direct numerical simulation of aerodynamic fragmentation of liquid drops

点击翻译按钮获取中文摘要

The destruction of gas-liquid interfaces of a liquid phase dispersed as droplets in gas by fast inertial effects is called aerodynamic fragmentation or aero-breakup. Technical applications of such processes are widely spread, involve e.g. secondary fragmentation of sprays of fuels, melts, paints, coatings, or biomedical emulsions. Process optimization requires understanding of physical mechanisms and, eventually, a prediction capability of aerodynamic fragmentation (AF), which both currently are severely limited. Experimental research resulted in important main findings. Quantitative experimental data, however, are very difficult to obtain due to extremely small spatial and temporal scales. Individual drops can be investigated by direct numerical simulations (DNS, resolving all relevant spatial and temporal scales of the flow), provided advanced high-resolution discretization schemes and discretely conservative interface models are employed. Efficient computing strategies require tailored data-operation splitting algorithms and multi-resolution representations. When a liquid drop in a gas is impacted by a shock, the shock diffracts and reflects, giving rise to complex wave patterns inside and outside of the drop. Pressure-gradient, baroclinic and surface-tension effects deform the interface. Shear forces along the interface initiate contact-sheet instabilities, whereas capillary instabilities during AF are negligible (surface tension itself is not), as are Richtmyer-Meshkov instabilities due to the high acoustic impedance jump at the interface. Commonly, based on experimental observation, three typical breakup scenarios are distinguished: (i) bag breakup, (ii) sheet stripping, (iii) cata-strophic breakup, which strongly depend on data and flow state uncertainties. A classification of prevalent scenarios in different parameter ranges is highly relevant for applications. Each of the breakup scenarios is characterized by the interplay of different mechanisms, most importantly: (i) Rayleigh-Taylor piercing, (ii) shear-induced entrainment, and (iii) shear-layer instabilities. The objective of the current project is threefold. First, a deterministic DNS method, based on an available high-resolution model for compressible interfacial flows, including effects of surface tension and interfacial friction, needs to be adapted and validated for the applications in this project. Second, a phenomenological analysis of different breakup scenarios with respect to prevalent breakup mechanisms will be performed. A set of numerical simulations are planned whose results allow for clearly assigning breakup scenarios to pa-rameter ranges and for qualification of experimental findings. Third, in a second funding period, we will extend the Mach-number range and assess additional mechanisms that become relevant. Furthermore, uncertainty-propagation tools will be employed for quantifying the confidence.