Repository logo
Log In(current)
  1. Home
  2. Colleges & Schools
  3. Graduate School
  4. Doctoral Dissertations
  5. Direct Numerical Simulations of Flame Propagation in Stratified Mixtures at Auto-ignitive Conditions Using Accelerated Computing
Details

Direct Numerical Simulations of Flame Propagation in Stratified Mixtures at Auto-ignitive Conditions Using Accelerated Computing

Date Issued
December 15, 2019
Author(s)
Desai, Swapnil
Advisor(s)
Ramanan Sankaran
Additional Advisor(s)
Derek Splitter, Charles Finney, Zhili Zhang
Permanent URI
https://trace.tennessee.edu/handle/20.500.14382/26950
Abstract

Direct numerical simulation (DNS) of auto-ignition under turbulent conditions has played a very important role in improving the fundamental understanding and advancement of combustion technologies for practical applications. However, very little is known of the nature of combustion in a reactive fuel/air mixture that is conducive to both spontaneous ignition and premixed deflagration. As such, characterizing the precise nature of combustion and the relevant propagation speed remains a challenge. This study attempts to address these questions by performing fully resolved numerical simulations of preheated fuel/air mixtures at elevated pressures using a newly developed DNS code called KAUST Adaptive Reactive Flows Solver (KARFS). Unlike a periodic box setup that has been used in most of the previous DNS studies, an inflow-outflow configuration representing a statistically stationary reaction front has been employed to understand the unsteady flame dynamics at auto-ignitive conditions.The first part of the dissertation is devoted to a discussion on parametric mapping of propagation speeds of auto-ignitive dimethyl-ether/air as well as dimethyl-ether/methane/air mixtures at elevated pressures under the influence of monochromatic thermal and composition/reactivity stratification using a one-dimensional, statistically stationary configuration. Thereafter, the implementation and effectiveness of Weighted Essentially Non-oscillatory (WENO) schemes in performing DNS of turbulent reacting flows is demonstrated with various non-trivial model problems. In addition, the scalability and performance portability of KARFS is presented on heterogeneous (CPU + GPU) system architectures. Finally, as a more extensive parametric quantification of the effects of thermal and composition stratification on turbulent flame propagation, results from DNS of a turbulent premixed flame in an auto-ignitive dimethyl-ether/air mixture conducted at elevated pressure are presented and discussed. The outcomes of this dissertation are expected to provide a fundamental understanding of the combustion mode transition and relevant propagation speeds in modern engines utilizing mixed-mode combustion.

Subjects

Flame Speed

Multi-stage Ignition

Turbulence

WENO

High Performance Comp...

Degree
Doctor of Philosophy
Major
Energy Science and Engineering
File(s)
Thumbnail Image
Name

utk.ir.td_12822.pdf

Size

22.72 MB

Format

Adobe PDF

Checksum (MD5)

c9fff310fd9ce66c88a32ccec71f8207

Built with DSpace-CRIS software - Extension maintained and optimized by 4Science

  • Privacy policy
  • End User Agreement
  • Send Feedback
  • Contact
  • Libraries at University of Tennessee, Knoxville
Repository logo COAR Notify