Evolution of coherent structure in turbulent wake

Abstract

In this research the generation and shedding processes of vortices in the near wake region and the coherent structure in a flow field are studied at two Reynolds numbers of 3900 and 9500 using proper orthogonal decomposition (POD) POD is a methodology used for the purpose of identifying large-scale eddies (such as Karman vortex) in lower-order modes and for recognizing small scale eddies in high-order modes that contribute to turbulence in the entire flow field The cylindrical wake possesses a large scale coherent structure which will be attenuated with distance from the cylinder POD and a spectrum analysis are used to identify the large scale coherent structure and its harmonics frequency In the high and low Reynold number experiments the coherent structure energy contributions of the three regions which are the upstream (0 5-5d) midstream (5-10d) and downstream (10-15d) regions of the flow field are used to identify the degradation of the coherent structure Regardless of whether the Reynolds number is high or low the energy contribution of the coherent structure is less than 5% in the downstream region Reynolds decomposition and period-time averaging process exhibit almost no differences in terms of downstream turbulence intensity and Reynolds stress Three POD parameters are analyzed in the study The first parameter was the number of samples where 6000 8000 and 10 916 samples were used to compare energy contribution of each mode However it reveals that it needs more samples to reach a statistically stationary result The second parameter was the number of modes used in the reconstruction of the flow field The Sirovich criterion and the Kaiser criterion were respectively used as the energy contribution bases in the flow field reconstruction The Kaiser criterion can't reveal the characteristics of small fluctuations The Sirovich criterion which accumulates more modes is a better choice The third parameter was the relationship between the Taylor microscale and the harmonic frequency It was found that when the harmonic frequency falls into the inertia subrange the peak of the harmonic frequency cannot be identified in the spectrum energy diagram

Date of Award	2020
Original language	English
Supervisor	Keh-Chin Chang (Supervisor)