An Application of Modal Decomposition to Supersonic Flow Over a Wall-Mounted Turret

Data driven modal decomposition techniques provide a means for extracting physically important features from high-dimensional, time-resolved dynamic systems. While the theoretical bases of these methods are well-developed, the interpretation of the obtained modes to understand complex unsteady flow physics can be difficult. In this research, the Proper Orthogonal Decomposition (POD) and Dynamic Mode Decomposition (DMD) techniques were employed to study the unsteady behavior of supersonic flow over a wall-mounted cylindrical turret. This flow problem contains a variety of time-dependent features including multiple unsteady shock waves and separated shear layers. The multi-scale and broadband nature of these features lend themselves well to be examination via the above modal decomposition techniques which extract modes based on characteristics such as energy content and frequency. The POD method was used to objectively rank the dominant features in the system while the DMD method highlighted structures based on their dynamics and their evolution with time. The analysis indicates that unsteady shock waves dominate many aspects of the unsteadiness in this flow, since the top POD mode contained over 40% total energy. Lower ranked POD modes yielded detail about structures in the shear layer and separation regions; however, these were mixed and broadband in character. The DMD modes revealed that the unsteady shock waves were oscillating at a peak nondimensional frequency of StL ∼ 0.3, about an order of magnitude lower than the next extracted mode. Since DMD modes contain only a single frequency, other modes extracted coherent oscillation patterns in the shear layers, separation regions, and turbulent boundary layers. Overall, it was found that DMD provided a better characterization of all unsteady features in the flow. Although the application was complex, modal decomposition was shown to be a useful tool for characterization of the features dominating the unsteady flows, which is vital for flow control techniques to be developed. Future work should focus on the development of reduced order models for use in flow control analyses.Ohio Space Grant Consortium STEM ScholarshipNo embargoAcademic Major: Aeronautical and Astronautical Engineerin