A new shear-layer growth-rate compressibility-scaling parameter is proposed as an alternative to the total convective Mach number, M c . This parameter derives from considerations of compressibility as a means of kinetic-to-thermal-energy conversion and can be significantly different from M c for flows with far-from-unity free-streamdensity and speed-of-sound ratios. Experimentally observed growth rates are wellrepresented by the new scaling.

We report on the results from a set of incompressible, shear-layer flow experiments, at high Reynolds number (Re δ ≡ ρ ∆U δ T (x)/µ 2 × 10 5 ), in which the inflow conditions of shear-layer formation were varied (δ T is the temperature-rise thickness for chemically-reacting shear layers). Both inert and chemically-reacting flows were investigated, the latter employing the (H 2 +NO)/F 2 chemical system in the kineticallyfast regime to measure molecular mixing. Inflow conditions were varied by perturbing each, or both, boundary layers on the splitter plate separating the two freestream flows, upstream of shear-layer formation. The results of the chemically-reacting 'flip experiments' reveal that seemingly small changes in inflow conditions can have a significant influence not only on the large-scale structure and shear-layer growth rate, as had been documented previously, but also on molecular mixing and chemicalproduct formation, far downstream of the inflow region.

The mixing and flowfield of a complex geometry, similar to a rearwardfacing step flow but with injection, is studied. A subsonic top-stream is expanded over a perforated ramp at an angle of 30 • , through which a secondary stream is injected. The mass flux of the second stream is chosen to be insufficient to provide the entrainment requirements of the shear layer, which, as a consequence, attaches to the lower guidewall. Part of the flow is directed upstream forming a re-entrant jet within the recirculation zone that enhances mixing and flameholding. A control-volume model of the flow is found to be in good agreement with the variation of the overall pressure coefficient of the device with variable mass injection. The flowfield response to changing levels of heat release is also quantified. While increased heat release acts somewhat analogously to increased mass injection, fundamental differences in the flow behaviour are observed. The hypergolic hydrogen-fluorine chemical reaction employed allows the level of molecular mixing in the flow to be inferred. The amount of mixing is found to be higher in the expansion-ramp geometry than in classical freeshear layers. As in free-shear layers, the level of mixing is found to decrease with increasing top-stream velocity. Results for a similar configuration with supersonic flow in the top stream are reported in Part II of this two-part series.