Objective phase-space identification of coherent turbulent structures in 1D time series data

Abstract

Listen

Translate article

Turbulent flow has been known to contain coherent flow structures. The size, shape, timescale and dynamics of these coherent structures are important in understanding turbulent flow processes such as pollutant mixing and particle transport. A range of methods exist for the detection of coherent structures in 2D or 3D data, but there are limited objective methods available for 1D data. Existing 1D techniques are subjective, require calibration with manual visualizations, or are based only on detecting velocity extremes and thus only detect structures according to a limited definition. A technique is presented here for objectively detecting structures using a phase-space method modified from a previously described despiking method. This method requires no calibration or subjective input, and identifies structures based on extremes in velocity, acceleration and jerk. This method gives turbulence statistics comparable with previous methods while recognizing a broader and more realistic definition of the physical properties of coherent structures by considering also the first and second derivative of velocity.