Date of Award

Spring 2003

Document Type


Degree Name

Doctor of Philosophy (PhD)


Mechanical Engineering

Committee Director

Surendra N. Tiwari

Committee Member

Gerald L. Pellett

Committee Member

Sushil K. Chaturvedi

Committee Member

Taj O. Mohieldin

Committee Member

Arthur C. Taylor, III


In this study, a detailed 2-D numerical technique is utilized to investigate the structure, and flame extinction and restoration characteristics of laminar hydrogen-air opposed jet diffusion flames, using both plug (uniform) and parabolic inflow boundary conditions for 3 millimeter jet tubes, spaced 6 millimeters apart and imbedded in low velocity coflows.

First, by using the most recent of two chemical kinetic models, excellent agreement was obtained between calculated distributions of temperature and major species, and published UV laser Raman scattering measurements, for 50% and 100% hydrogen-air flames over a range of input strain rates. Agreement with measured OH profiles was reasonably good at high strain rates, but generally less so at low strain rates. Also, the numerically simulated extinction limit of the 100% hydrogen-air diffusion flame was predicted within −5.1% (Jachimowski kinetic model) and +4.7% (Yetter et. al model) of a published grand-average measurement of global applied stress on the airside (average input velocity/tube-diameter), obtained using parabolic inflow profiles with a 2.7 millimeter tube opposed jet burner system.

Second, the study showed counterflow flame extinction limits for 100% hydrogen-air were most consistent when compared using flame core maximum radial strain rates on the center-line. By this measure, flame extinction occurred at similar strain rates (within 9.6%) for the two very different inflow boundary types. Also, the respective radial strain rates were linearly proportional to both centerline maximum axial strain rate and global applied stress rate, up to and just before the extinction state. Thus, both of these measurable reference rates provided a suitable relative basis for characterizing flame extinction limits.

However, the ratio of radial to axial strain rates varied significantly with input flow boundary types. The plug input flame resulted in a slightly smaller (0.92×) radial flame core strain rate on the centerline than the maximum axial strain rate, and this measure can be compared to the parabolic inflow flame which had a relatively larger (1.47×) radial strain rate. Furthermore, the respective radial/axial ratios (0.92 and 1.47) were not even close to the “classic” 0.50 ratio for the near-extinction state derived from the simplest 1-D stream-function approximation (Heimentz or potential flow) method.

Finally, a previously observed ring-shaped post-extinction 100% hydrogen-air flame was numerically simulated when the stretching limit of the parabolic input velocity (for a 3 millimeter opposed jet) was exceeded beyond the critical extinction point. The post-extinction flame was tri-brachial, with fuel-rich and fuel-lean diffusion and premixed branches, displaying a quite different flame structure than the typical counterflow diffusion flame.