Turbulent time-frequency spectral structures and the extraction of the desired turbulence component

Abstract

Turbulence is induced in a constant-volume bomb using the motion of a perforated plate. Statistical ensemble-averaged and the cycle-resolved analyses are performed for these raw velocity data. It is shown that the cycle-resolved analysis is a suitable method of characterizing non-stationary turbulent flow. Based on the cycle-resolved analysis, the velocity data are then processed using wavelet packet and moving time-window techniques to obtain turbulent time-frequency spectral structures. From the comparison of the turbulent time-frequency spectral properties, a method to induce turbulence with different characteristic parameters is developed. To extract the desired turbulence frequency content, a multiple cut-off frequencytechnique is proposed and good agreement is found between the results obtained from wavelet analysis and the multiple cut-off frequency method. The integral time scale of the turbulence component has been studied as well. It is shown that the integral time scale is the key characteristic time for the turbulence and this is a function of the turbulence frequency.