Flow pattern classification in liquid-gas flows using flow-induced vibration

•A novel algorithm to identify the flow pattern using flow-induced vibration is presented.•The churn and slug flow presented a distinct frequency band activity.•The vibration main excitation cause in slug and churn flow is transient.•The RMS of the vibration signal envelope was used to identify the churn flow cases.•The Pearson correlation coefficient was used to identify the dispersed bubble cases. The multiphase flow is not only the most common flow in nature but also occurs in various major industrial fields. Furthermore, in many industrial plants, the single and multiphase flows generates vibration and noise. In the context of two-phase flows, a specific case of multiphase flow, the flow pattern determination is crucial to their analysis, and despite the recent progress and developments in flow-induced vibration for two-phase flows, it is still considered an open topic. This paper develops a novel algorithm for flow pattern classification using the vibration signal from a vertical pipe conveying a liquid-gas two-phase flow to determine the flow pattern. An experimental apparatus and procedure were developed to perform this investigation. The analysis in the frequency domain showed a distinct frequency band activity for slug and churn flows. The analysis in time domain showed a significant amplitude variation for these flow patterns. Finally, by using the RMS and Pearson correlation coefficient, it was possible to classify the studied cases accurately. The results present a non-intrusive technique to identify the flow pattern in two-phase liquid-gas vertical flows.