|
|
|
Cherian A. Idicheria, Isaac G. Boxx and Noel T. Clemens Turbulent combustion is a challenging area of research because of the many physical processes that interact in complex (and interesting!) ways. For example, in any given flame, the structure of the flame may be influenced by: turbulence, buoyancy, chemistry, heat release, mass diffusion and heat transfer. In fact, it is the complex interaction of these physical processes that makes it so difficult to develop the fundamental understanding of how turbulent flames behave. Furthermore, it is this lack of understanding that makes it difficult to properly model - and thus predict - important quantities such as the levels of atmospheric pollutants produced by utility power generation systems and aircraft engines. The great advantage of using the microgravity environment is that it enables us to "turn-off" the effect of buoyancy. Buoyancy is often one of the most important (and complicated) parameter that affect the structure of a flame. Through comparison of flames under microgravity and normal gravity conditions, we expect that this work will lead to an improved understanding of flames in general, not just those in microgravity. In this study, our specific objective is to investigate the basic strain-reaction zone relationship and the flame structure of turbulent nonpremixed flames. We will use the 1.25 second drop-tower at UT-Austin and 2.2 second drop tower at NASA Glenn in Cleveland, Ohio to provide the low- the microgravity conditions, respectively. The UT drop-tower provides low- gravity levels of 20mg whereas the 2.2 second drop-tower provide 10-4g. The use of the drop towers required that we build a new drop rig that is essentially a completely self-contained experimental apparatus. Virtually the entire experimental apparatus fits within a box that has dimensions of 38" x 16" x 36". The experimental techniques include high speed imaging of soot emission, planar laser Mie scattering and particle image velocimetry.
The picture above shows the drop rig used in the normal-, low- and microgravity experiments. Till date, we have studied methane piloted ethylene, propane and methane flames in normal-, low- and microgravity conditions. Experiments were conducted at the two drop- towers and the data collected was used in analyzing jet flame startup, mean and RMS characteristics in varying gravity levels. Additional analysis of the image sequences using volume rendering helped in understanding flame tip dynamics and large-scale organization. The initial diagnostic employed was high speed imaging of soot emission. A Pulnix TM-6710 camera operated at either 350 frames per second (fps) or 235 fps depending on the flame studied, was used to obtain time resolved images.
A typical flame startup image sequence of ethylene obatined from a normal- and low-gravity experiment is shown above. The first set is the normal-gravity sequence and the second is in low-gravity at an exit Reynolds number of 7500. Comparison of the startup sequences showed that low-gravity flames take a longer time to reach steady state than normal-gravity flames. However, the difference was only about 100-200 ms. It is also seen that the flames in all three gravity environments reach a steady state in less than 400 ms regardless of the fuel used.
The figure above shows mean luminosity images obtained from steady state frames for the various fuels and Reynolds numbers studied. The corresponding "buoyancy parameter" (xL) that quantifies the relative importance of buoyancy is given at the top of each image. From the first image pair it can be seen that as the flame becomes more momentum dominated (xL~1) in normal-gravity, there is no significant difference to the one in low-gravity. However, there is a significant difference when there is large difference in values of xL. It can also be seen from the propane flame images that there is virtually no difference between flames in low-gravity and microgravity. This suggests that xL is a sufficient parameter to quantify the effects of buoyancy. Additional analysis also suggests that the structure of the large-scale turbulence reaches its momentum-driven asymptotic state for values of ξL less than about 3.
The above figure shows RMS intensity images (false
colored) obtained from steady state frames for the same set of conditions as
the mean flame images. These were
computed to determine if the trends that were observed in the mean images are
also seen in fluctuating quantities. RMS images for the ethylene flames at ReD=10,500
in normal- and low- gravity have noticeable similarities, but clear
differences are apparent such as the lower peak RMS values on the centerline
of the ξL=1.0 flame. However, more drastic differences arising due
to effects of buoyancy can be seen when comparing flames with a larger
difference in ξL. It is also interesting to notice the large
fluctuations present near the flame tip in the large ξL flames. Low
ξL flames all have qualitatively similar RMS contours: the
fluctuations peak near the periphery of the flame and remain low even at the
flame tip.
Differences in flame tip oscillation is studied by analyzing individual frames and also using volume rendering. The figure above shows volume renderings obtained from sequential images of ethylene at various Reynolds number. In this method sequential 2-D images (x-y) are stacked one behind the other (along z-axis) which corresponds to time t, and a 3-D volume (x-y-t) is regenerated using image processing. Consistent with the analysis of individual frames, the flame tip oscillates at a lower frequency in low-gravity than in normal-gravity for lower Reynolds numbers. However, these differences are less obvious in the higher Reynolds number cases. Also the slope of the wrinkles on these images gives an estimate of convection velocity of the structures propagating along the flame. The buoyant acceleration aiding convection at normal-gravity can be clearly seen by comparing the slopes of the wrinkles for the image pair at Reynolds number of 2500. The decreasing effect buoyancy has on convection velocity is apparent by studying the volume rendering images at higher Reynolds number. Flame movies Click on the links below to view sample movie files generated from the instantaneous images obtained from an experimental run. You will require an avi player and each file is about 400KB. The movies presented below are for turbulent propane flames at a Reynolds number of 8500.
normal-gravity low-gravity microgravity Ongoing work We are currently working on implementing laser based diagnostics like planar laser Mie scattering and particle image velocimetry. Publications:
back to research page |
|
|
|
|
