Modeling studies of baffle-type combustors

Abstract

Geometrically similar ramjet-type combustion chambers of 1.61, 4.026, and 6.065-inch diameter and lengths of 5.25 D and 4.25 D were operated at Reynolds numbers ranging from 24,000 to 230,000 and ambient pressure levels from 0.2 to 1.5 atmospheres, with propane fuel. Flameholders were annular V-gutters having a leading-edge included angle of 30° and producing 44% blockage of the combustor cross section. Excepting the wall thickness of some burners, all dimensions of burners, flameholders, upstream piping and mixing section were scaled in proportion to burner inside diameters. Analysis indicated the importance of keeping the following parameters constant if the model experiment was to reproduce prototype performance: o a. Reynolds number, DU ϱ/μ . b. Chemical loading parameter, kϱ n−1 D/U . c. Mach number, M U/C . d. External heat-loss group, λ G / D [( L w /λ w )+(1/ h c+r,0 )]. e. Helmholtz resonator group (D 3 /V) 1/2 /M . Analysis indicated the probable importance of group d above—the external heat-loss group—and and a poorer performance to be expected of larger burners modeled by use of groups a and b alone, because of abnormally high external loss from large burners. An experimental study of the effect of wall losses, using emissivity control or radiation shielding, confirmed the analysis and explained some of the deviations in burner efficiency results of modeling based on groups a and b alone. This indicated the importance of including group d if the prototype walls run hot. Longitudinal, tangential, or radial oscillations that may occur in modeled combustors are scaled only if the Mach number is kept constant. As indicated above, this is automatic with modeling based on Reynolds number and the chemical loading parameter if, but only if, the pressure exponent n is 2.