Dynamics and Structure of Supersonic Multi-Phase Reacting Flows

Abstract

Proper understanding of high-speed multi-phase reacting flows in realistic fuels is critical for the design and operation of advanced propulsion and energy systems, manufacturing processes, munitions, and for prevention and mitigation of explosions in industrial settings. Over the past decade, significant progress has been made in understanding single-phase reacting flows such as high-speed gaseous turbulent flames and detonations, while high-speed subsonic and supersonic multi-phase systems though practically relevant, received little attention. The primary goal of this dissertation is to study a wide variety of subsonic and supersonic multi-phase reacting and non-reacting systems using high-fidelity numerical simulations to develop a deeper understanding of the structure and dynamics of coupled gas/particle systems. To accomplish this, we begin by developing a new versatile state-of-the-art multi-scale, multi-fidelity, multi-phase reacting flow solver capable of modelling both solid and liquid particles. An Eulerian-Lagrangian formulation is employed, in which the gas-phase is represented using an Eulerian grid while the liquid-phase or the solid-phase is modeled as discrete Lagrangian point particles. Novel Lagrangian particle tracking method based on symplectic integrators and Weighted Essentially Non-Oscillatory (WENO) based high-order interpolation schemes is presented to pro-vide higher-order spatial and temporal accuracy necessary to study a wide variety of multi-phase flows, in particular, high-speed flows characterized by flow discontinuities. State-of-the-art multi-phase models are implemented for droplet drag, droplet evaporation, droplet breakup as well as solid particle heat transfer. Extensive testing is carried out to characterize the numerical properties of the developed multi-phase solver in canonical flow problems and to benchmark the solver against data available in the literature. The dynamics of conductive solid particles and massless Lagrangian tracer particles is studied in the context of the Particle Image Velocimetry Technique (PIV) in two-dimensional (2D) reacting gas-phase detonations at atmospheric conditions and three-dimensional (3D) non-reacting homogeneous isotropic turbulence at engine relevant conditions. The aim of this study is to a) understand the stability, accuracy and limitations of the multi-phase solver in practically relevant high-speed complex flows and b) numerically assess the accuracy of the PIV technique for characterization of high-speed, compressible reacting flow regimes such as a 2D gas-phase detonation and for flow field characterization of systems with wide range of spatio-temporal scales, such as those encountered in a 3D high-speed subsonic turbulent flow, using synthetic PIV. Various aspects of the PIV technique such as its ability to probe the flow field non-intrusively, ability of the speed particles to sample the flow field uniformly, accuracy of the speed particle trajectories and finally the accuracy of the PIV reconstruction of the velocity field in comparison to the actual velocity field are critically assessed. The possible sources of error in the PIV technique for high-speed flows are illustrated. Several practical recommendations including optimal choice of the PIV parameters for high-speed PIV diagnostics are proposed based on this study. Finally, a new spray detonation model with realistically complex chemical kinetics, detailed multi-species transport and state-of-the-art multi-phase models is developed to study the dynamics and structure of a 2D freely propagating spray detonation in a jet fuel spray. The resultant detonation properties are contrasted with those of a purely gaseous detonation and key differences are highlighted. Quantitative insights into the detailed flow conditions encountered by the spray droplets are provided, in order to assess the accuracy, validity, and fidelity of existing droplet drag, evaporation, and atomization models. In particular, the thermodynamic conditions encountered by liquid droplets are critically analyzed along with droplet Reynolds, Mach, Weber, and Ohnesorge numbers. Several key limitations of the current spray models are highlighted and discussed in de-tail. Based on this study, future extensions of the physical models for the accurate and predictive modelling of spray detonations are proposed.