Date of Award

12-2016

Degree Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Mechanical Engineering

First Advisor

John Abraham

Committee Chair

John Abraham

Committee Member 1

Peter H. Meckl

Committee Member 2

Steven Son

Committee Member 3

Haifeng Wang

Abstract

Lean-burn natural gas engines offer enhanced thermal efficiencies and reduced soot and NOx emissions. However, cycle-to-cycle variability in combustion that can result from unreliable ignition, variability in equivalence ratio and quenching is a challenge. Reliability of ignition can be improved by employing a dual-fuel ignition strategy in which a small quantity of diesel fuel is injected to initiate ignition. Computational studies of n-heptane/methane-air mixing layers are performed to provide insight into the fundamental physics of dual-fuel ignition. The results show that the characteristic time required for steady premixed flame propagation has three components: time for autoignition to occur, time for peak temperature to be achieved following autoignition, and time for steady flame propagation in the premixed fuel/air mixture to be achieved. The autoignition time correlates well with pressure and temperature of the unburned premixed charge. The time to achieve peak temperature is relatively short, but correlates with mixing layer thickness and premixed equivalence ratio. The time to achieve steady propagation correlates with mixing layer thickness and laminar flame speed and thickness.

Subsequent work focuses on turbulent flame propagation in lean homogeneous mixtures by employing direct numerical simulations (DNS) under conditions that are relevant to lean-burn engines. Attention is specifically focused on the turbulent flame speed (ST) as a parameter of interest because of its importance in modeling combustion in engines. The studies are carried out in the thin reaction zone (TRZ) regime of turbulent premixed combustion. Normalized turbulence intensity (urms/SL) varies from 2 to 25 and the ratio of integral length scale to flame thickness (L o/δL) varies from 3.2 to 12.8. Initial studies show that the normalized turbulent flame speed (ST/SL) depends on more parameters than urms/SL suggested by some models.

Although it is known that the turbulent flame speed varies with equivalence ratio, it is shown that the normalized turbulent flame speed does not change with equivalence ratio provided the Karlovitz (Ka) and Damköhler (Da) numbers are fixed. This suggests that Kaand/or Da are important parameters in characterizing the turbulent flame speed. Furthermore, ST/SL can be related to the flame area enhancement AT/AL and an efficiency factor Io which is close to unity. AT/AL is raised by increasing turbulent Reynolds number ReT and by reducing Ka. Increasing ReT leads to a broader spectrum of turbulent eddies that generate flame surface area. Increasing Karesults in fine wrinkling at the expense of larger scale wrinkling. This results in a net reduction in the effective surface area enhancement. Based on these insights, a correlation for ST that shows a dependence on Re T and Ka is proposed. Modeling of the Flame Surface Density (FSD) evolution is also considered. FSD is influenced by tangential strain rate and flame displacement speed. Surface averaged tangential strain rate is found to scale linearly with Ka. The effects of Ka on flame displacement speed are modeled using a Probability Density Function (PDF) based approach.

The effects of premixed combustion on turbulence are investigated. For flames in the TRZ regime, the turbulence kinetic energy (TKE) decays monotonically across the flame brush. Scaling analyses of the terms in the transport equation of TKE reveal that viscous dissipation is the dominant contribution in the TKE equation. The relative importance of the other terms in the TKE equation decreases with increasing Ka.

Share

COinS