A real time acquisition and processing program was developed to record acoustic emission (AE) signals

to extract AE wave packets

and to calculate AE parameters of continuous bubbles in their rising stages after released from a 3-mm-diametered nozzle. AE signal wave packets of 1.42 m/s gas velocity at orifice was classified by kmeans clustering analysis method. For each category

statistical characteristics of AE signal parameters was analyzed and time/frequency domains of AE signals were analyzed by fast Fourier transform and wavelet transform method. Identification of AE signals was effectively achieved with assistance of moving bubble images. Signal classification capability were further studied under different gas velocities. The results show that AE signals of all categories have somewhat similarity after processed by time frequency analysis and k-means clustering method

AE signals of moving bubbles in their rising stages can be efficiently studied in combination with high-speed image-capturing technique

and flow patterns of moving bubbles can be identified

including bubble oscillation

breakup

coalescence and so on.