Dynamic Estimation of Large-Scale Flow Events in Open Cavity Flows

Flow over an open cavity has a dynamically complex flowfield, generating pressure fluctuations that can amplify shear stress and cause structural damage to the stores within it and the aircraft as a whole. These pressure loads are linked to large-scale resonant shear layer phenomena. Consequently, it is crucial to establish this link by characterizing the large-scale flow events conditional upon pressure measurements to improve the prediction and control of the detrimental pressure fluctuations. In this study, a link between the pressure measurements (to control) and large-scale shear layer events (to actuate) will be established using a combination of two data-driven techniques, Linear Stochastic Estimation (LSE) and Dynamic Mode Decomposition (DMD), in an innovative approach called sensor-based conditional DMD (conDMD). The sensor-based conDMD approach applies DMD directly to pressure measurements, contrary to traditional DMD applications that are performed on velocity. This method is computationally more efficient and inherently low-ranked. The resultant modes are then combined with LSE to obtain conDMD modes of velocity. The conDMD modes represent dynamic, large-scale flow structures convecting at specific frequencies and growth rates that are correlated with pressure fluctuations. The DMD eigenvectors represent the flow structures, and the eigenvalues provide the growth rate and frequency of each mode. The estimation of large-scale flow events from extracted pressure modes establishes a direct link between coherent flow structures and discrete surface pressure, offering an innovative method for real-time flow analysis and control. In conclusion, sensor-based conDMD effectively estimates the dynamics of large-scale flow structures associated with pressure fluctuations in open cavity flows, potentially allowing for better flow prediction and actuation mechanisms.