Large eddy simulation of a turbulent jet diffusion flame using the filtered mass density function model
Abstract
In many practical combustion devices, rapid mixing of the fuel and oxidizer is often desired to maximize heat release rates and to reduce the overall combustion chamber size. Unfortunately, turbulent straining of high-speed flames can increase local heat loss rates beyond the combustion heat release rate, locally extinguishing the flame. Extinguished pockets of fluid can lead to reduced efficiency and increased pollutant emission if allowed to pass from the combustor before reigniting. Common time-averaged engineering models have difficulty simulating extinction and reignition because it is a complex, unsteady process that depends on the nonlinear interactions between turbulent eddies and the local thermochemical state in a flame. To aid in studying extinction and reignition for future model development, a high-order accurate large eddy simulation (LES) research code is developed that can capture these finite-rate chemistry effects in turbulent flames. The PDF transport-based filtered mass density function (FMDF) closure model is used for scalar transport and combustion, which is solved by a stochastic Monte Carlo method. The hydrodynamic solver uses a hybrid energy-preserving high-order finite difference scheme that provides a stable and non-oscillatory solution for scalar transport, and is coupled to the Monte Carlo solver in an internally-consistent fashion. To aid in simulating a turbulent jet flame, a stochastic technique is developed to generate coherent turbulent inlet boundary conditions. A version of the in situ adaptive tabulation (ISAT) algorithm is implemented to significantly speed chemical kinetics calculations with realistic combustion mechanisms and thermophysical properties. All numerical techniques are implemented in an efficient, scalable parallel formulation and the LES code capability for simulating realistic turbulent flames with local extinction and reignition is demonstrated.
Degree
Ph.D.
Advisors
Frankel, Purdue University.
Subject Area
Mechanical engineering
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.