Application of Frequency Dependent Proper Orthogonal Decomposition to Time-Resolved Schlieren Data Extracted From a Supersonic Multi-Stream Rectangular Jet*

Frequency dependent proper orthogonal decomposition has been used in recent work to examine the structure of turbulence in various flow regimes and involves decomposing the flow into energy-ranked coherent structures along with extracting energy spectra. The resulting modes were used to investigate wavepackets of the Kelvin-Helmholtz type and the Orr type mechanism, which are differentiated by their modal and non-modal behaviors respectively. In a similar manner, this technique of low dimensional modeling is applied to time-resolved schlieren data extracted from an experiment revolving around the investigation of a multi-stream rectangular nozzle flow. An instability arising from the merging of the core stream and additional bypass stream (third stream) inside the nozzle has been proposed as the physical mechanism driving a high frequency 34 kHz signal seen throughout the entire flow-field including the far field acoustics. Modal and spectral analysis through application of this technique to the schlieren data set can reveal information regarding the behavior of this instability, how it is propagated further downstream, and the propagation of information from other flow structures in the region of the nozzle exit.