Date of Completion

4-13-2016

Embargo Period

4-7-2016

Advisors

Baki Cetegen, Xinyu Zhao

Field of Study

Mechanical Engineering

Degree

Master of Science

Open Access

Open Access

Abstract

Detailed chemical kinetics is important for high-fidelity reacting flow simulations. The major challenge of incorporation of detailed chemistry in large-scale simulations is primarily attributed to the high computational cost, induced by the large number of species and reactions, as well as the severe chemical stiffness. Methodologies are therefore needed to facilitate the use of detailed chemistry in large-scale combustion simulations. In the present study, a linearized error propagation (LEP) model is developed to eliminate unimportant species and reactions from detailed chemistry. In the LEP model, the reduction errors are analytically approximated and formulated for perfectly stirred reactors (PSR). The performances of LEP in development of local and global reduced models are compared with previous approaches including directed relation graph (DRG) and DRG with error propagation (DRGEP). It was shown that LEP can effectively control the reduction errors in selected target species and global flame properties, such as ignition delay time. The skeletal models obtained by LEP are validated in PSR, auto-ignition and 1-D premixed flames.

Chemistry calculations can be further accelerated through dynamic adaptive chemistry (DAC) and in-situ adaptive tabulation (ISAT). DAC can expedite the time integration of chemical kinetics by using local skeletal models that can be substantially smaller than global skeletal models. ISAT can reduce the number of time integration by tabulating and re-using the previous solutions. Their relative performances are investigated for homogeneous charge compression ignition (HCCI) combustion and partially-stirred reactors (PaSR). It was shown that, compared to ISAT, the performance of DAC is mostly independent of the nature of combustion simulations, e.g., steady or unsteady, premixed or non-premixed combustion, and its efficiency increases with the size of chemical kinetic models. DAC is particularly suitable for transient combustion simulations with large chemistry models, while ISAT can be more efficient for simulations where chemistry calculations can be frequently retrieved from the ISAT table. Moreover, a combined approach of ISAT and DAC, namely ISAT-DAC, is developed and demonstrated to accelerate the chemistry calculations. The incurred errors in temperature and species concentrations by ISAT-DAC are well controlled and the performance of ISAT is shown significantly enhanced by DAC.

Major Advisor

Tianfeng Lu

Share

COinS