Phase-Averaging Methods for a Naturally Oscillating Flow Field

Abstract

The presented study examines various methods for phase-averaging the naturally oscillating flow field of an enlarged fluidic oscillator acquired by a high-speed PIV system. Because of the absence of an external trigger, phase-averaging the acquired data is challenging. Mathematical and physical methods are categorized and described. The results of these methods are evaluated for their accuracy in capturing the natural flow field. It is found that the mathematical methods, especially the method of proper orthogonal decomposition, produce reasonable qualitative results. However, compared to the physical methods, shortcomings in quantitative accuracy are revealed. The physical methods require a time-resolved reference signal. Two possibilities to identify the oscillation periods in the reference signal are described and compared. It is found that applying an autocorrelation on the reference signal improves the period identification due to consideration of a locally fluctuating mean value. This period identification method and according phase-averaging yields the best results regarding the minimum fluctuation of the oscillation period lengths. The according procedure is described in detail and applied to the internal and external flow field of the fluidic oscillator.