On discriminating sizes of CFD generated bubbles with signal processing analysis

The aim of this research work is to propose a signal processing based technique to classify and estimate bubble sizes. Bubbles are generated by computational fluid dynamics simulations. Local measuring points for velocity and pressure are set, and the captured data are analyzed using signal processing. The signal analysis includes the generation of templates for signal measurements of different bubble sizes using the short-time Fourier transform. Euclidean distances between templates of the different bubble size classes are subsequently computed. An inter/intra-class distance based matrix methodology is proposed to assess the discriminability of the Fourier-based template representation. The results indicate that the proposed technique based on signal processing can lead to the discrimination of bubble sizes with the information of bubbles passing through a single sensor point. Moreover, the model presented in this paper suggests that the analysis window size may play a highly important role in the discriminability according to the range of target bubble sizes.