Preview

PDF, English - main document Download (968kB) | Terms of use

Abstract

The droplet size distribution and interaction of the liquid phase and the gas flow are key features in the modeling of evaporating spray flows, which are important because of their vast range of industrial and engineering applications. Two-phase effects and poly-dispersity of droplet size distributions dominate the structure of any spray and related applications such as spray flames, end products of spray drying processes, or medical applications. The spray dynamics depends on various physical processes such as droplet inertia, evaporation, and gas phase characteristics. Thus, it is important to have reliable models and numerical techniques in order to be able to describe the physics of two-phase flows, where the dispersed phase consists of droplets of various sizes that may evaporate, coalesce, breakup as well as have their own inertia and size-conditioned dynamics.

In the present thesis, an evaporating water/air spray is modeled using direct quadrature method of moments (DQMOM) and discrete droplet model (DDM) in an axisymmetric geometrical configuration. In DDM, the two-phase effects are captured by resolving the gas phase conservation equations considering the droplets as point sources. The system of conservation equations is closed using an extended k-epsilon model. The system of equations is solved using a hybrid finite volume - Lagrangian particle tracking method. DQMOM is not yet coupled to gas phase fully, rather the inlet gas flow properties are used to compute the drag force exerted on droplet velocity. For both DDM and DQMOM, appropriate initial and boundary conditions as well as the starting values for simulations are generated from experimental data, which have been carried out by the group of Prof. G. Brenn at TU Graz, Austria. The simulation results are compared with experiment and found in good agreement.

Furthermore, a turbulent methanol air jet spray flame is investigated. A detailed methanol/air combustion mechanism consisting of 23 species and 168 elementary reactions is implemented through a spray flamelet model. The process of molecular mixing is treated following probability density function (PDF) modeling, where two approaches are used i.e., presumed PDF and transported PDF. The standard beta distribution is used as the base case to describe the process of molecular mixing. Its shape parameters and distribution characteristics are known and well established.

A bivariate joint PDF of the mixture fraction and enthalpy is applied for turbulent spray flames. The PDF transport equation is deduced. The mixture fraction and enthalpy are described using an extended Interaction-by-Exchange-with-the-Mean (IEM) model and modified Curl's model. The PDF transport equation is closed through coupling with an extended k-epsilon model, and it is solved using a hybrid finite volume/Lagrangian Monte-Carlo particle method. The numerical results of the gas velocity, the gas temperature, and the Sauter mean radius are compared with experimental data from the literature and good agreement with the experiment is observed. Furthermore, the shapes of the PDF of the mixture fraction and enthalpy at different positions, which are computed by the transported PDF method, are presented and analyzed. For the sake of comparison, the presumed PDF method is also applied, where statistical behavior of mixture fraction is described using standard beta function. A comparison of the results of the transported PDF method using modified Curl's and IEM models with the standard beta function shows that the standard beta function fails to describe the statistical behavior of mixture fraction accurately. Effect of a four parameter modified beta distribution instead of a standard beta distribution are also discussed. A trivariate joint PDF of enthalpy, gas velocity and mixture fraction is proposed for future simulations, and its transport equation is derived, where the gas velocity is modeled using an extended simplified Langevin model.