Acoustic Modes in an Open Box Cavity With Variable Depth Using Two Distinct Wind Tunnels

Abstract

Cavity resonances in ducts are a classical problem that has involved researchers from very different fields over time. More recently, the aerospace community became engaged again due to resonances found within the flow path of aero-engines, for example, in bleed cavities in the low-pressure compressor section. These resonances can lead to problems with the structural integrity of upstream components and thus warrant investigation. This study expands the previous dataset of an open box cavity with variable depth (D) and a rectangular opening with length (L), effectively testing L/D=[4,2,1/2]ratios. This parameter was indicated by Rossiter to be a strong predictive parameter in frequency generation. This article compares the results of the same cavity geometry within two wind tunnels of different dimensions and operational setups. The intention of the study is to isolate the Rossiter modes from any other geometric modes that are due to the wind tunnel’s test section geometry. This isolation allows the modification of the model of Rossiter and the cavity depth model of East to improve the predictive capability over a larger range of Mach numbers and for deeper cavities. One wind tunnel has an open loop continuous flow operation where the Mach number is set by changing the outlet static pressure. The second wind tunnel has a blowdown operation where both Mach and Reynolds numbers can be set independently by pulling a downstream vacuum and setting the upstream total pressure. The experiments were performed for an operating range from low subsonic to transonic Mach numbers. The analysis focuses on modifying the constants in both the Rossiter model and the cavity depth model proposed by East along with investigating the phenomenon of mode switching for Rossiter modes. The analyses show a good repeatability of the dataset between the two wind tunnels displaying strong resonances at similar operating points. The new proposed constants for the Rossiter model when using an L/D=2 are γ=0.15 and κ=0.58. For the deeper cavity with L/D=0.5, the constants proposed are γ=0.15 and κ=0.9. The modification of the cavity depth model by East to better predict the mode above Mach 0.3 is proposed by adjusting the constant B=2.5 and leaving A=0.65. An independence from Reynolds number of the acoustic frequency generation is demonstrated within the operating range. A mode-switching behavior is identified with the shallowest and deepest cavity showing multiple mode transitions within the operating range and a near constant Rossiter mode 1 for L/D=2.