Weak-extinction limits of large-scale, bluff-body flameholders

Michael Robert Baxter, Purdue University

Abstract

The results of an experimental and analytical research program aimed at quantifying the weak-extinction characteristics of large-scale, bluff-body flameholders are presented. An existing experimental facility was modified to allow the supply of both homogeneous and heterogeneous mixtures of fuel and air to flameholders of various sizes and shapes similar to those found in turbojet afterburners and ramjets. The experimental results indicate that the weak-extinction limits of flameholders supplied with homogeneous fuel-air mixtures are governed mainly by inlet mixture temperature and to a lesser extent by mixture velocity and upstream vitiation level. Also, for a constant duct cross-sectional area, weak-extinction limits are extended by an increase in flameholder width up to a critical width after which blockage effects begin to significantly oppose the beneficial effect of increased flameholder width. The experimental configuration used to obtain the heterogeneous data closely resembles a typical afterburner system. The weak-extinction limits of the heterogeneous system differ significantly from the homogeneous results. Various nonhomogeneous phenomena provide fuel-concentrating effects which allow stable combustion at extremely low duct fuel-air ratios. Also, the heterogeneous results exhibit a dependence on flameholder geometry at marked variance with the homogeneous results. Different from the homogeneous system, the performance of the heterogeneous system deteriorates with increasing flameholder width. A series of analytical techniques were developed in support of the experimental program. A method for predicting the weak-extinction limits of homogeneous, bluff-body systems based on approximating the wake-region as a well-stirred chemical reactor was devised. The heterogeneous model extends the above with submodels to account for key physical transport processes such as droplet evaporation, turbulent diffusion and flameholder droplet capture. These submodels allow the calculation of the wake-region's effective fuel vapor-air ratio which is then inserted into the stirred-reactor model. Comparisons between the experimental results and analytically generated predictions lend credence to the approximations and assumptions underlying the analyses. The degree of correlation present in the comparisons indicates that the analytic techniques developed and described herein may be used to perform preliminary design evaluations or as a basis for more extended and refined models.

Degree

Ph.D.

Advisors

Lefebvre, Purdue University.

Subject Area

Mechanical engineering

COinS