High-Resolution Visualization Measurement of Vortex-Shedding at High Frequencies Using Sub-Nyquist-Rate Piv and Compressed Sensing

Abstract

Oscillating flows appear in various applications, such as vortex shedding in vehicles, wings, and structures, as well as pulsating flows in pipes. While time-resolved particle image velocimetry (PIV) is often employed for these measurements, high-speed cameras are required to capture high-frequency oscillations, which increases cost and reduces spatial resolution. This study employs a low-speed, high-resolution camera to achieve high-spatial-resolution visualization measurements. Compressed sensing was utilized to compensate for the lack of time resolution, thereby enabling the measurement of high-frequency velocity oscillations. Random sampling, essential for compressed sensing, was performed by randomizing the exposure times of two time-delayed cameras. Proper orthogonal decomposition was used to identify spatial modes, and correction of nonstationary amplitude modulation was performed using attractors of the time functions of the two paired modes. The proposed method successfully reconstructed the flow with a Reynolds number of 250 and a vortex shedding frequency of 73.5 Hz from 20 velocity field data points obtained at 30 Hz. The reconstruction error for the entire velocity field was 11%, effectively reproducing vortex shedding, advection, and diffusion behaviors. The newly developed correction method for amplitude modulation can be applied to other visualization and measurement techniques using compressed sensing.