High-speed flow over an open cavity is associated with dynamically complex unsteady events, which warrant the need to better understand the flow in a reduced-order manner. The current work performs a modified version of Spectral Analysis Modal Method (Zhang et al. 2020) on supersonic cavity flows which enables significant computational speed-up and a more optimal way of encoding the large-scale structures. Synchronized measurements of non-time-resolved velocity field using particle image velocimetry and time-resolved single-point pressure measurements were conducted on an open cavity at a nominal free-stream Mach number of 1.4. The test section comprises a rectangular cavity of length-to-depth ratio of 6 and width-to-depth ratio of 6, which spans the full width of the wind tunnel. The datasets are analyzed using a combination of Proper Orthogonal Decomposition and stochastic estimation techniques to obtain the reduced-order modes. Reduced-order reconstructions at high-energy Rossiter frequencies show large-scale unsteady vortical structures indicative of Kelvin–Helmholtz shear layer instabilities convecting down the cavity. The coherence-based reconstruction enhances the extraction of Kelvin–Helmholtz instability dynamics by reducing contamination from incoherent turbulent motions. Furthermore, these interactions highlight significant shear layer–aft wall interactions along with flow ejection and in-sweep events, emphasizing the aft wall interactions of supersonic flow over an open cavity.
