Recognition of gas-liquid flow regimes in helically coiled tube using wire-mesh sensor and KNN algorithm

Abstract

• Flow patterns in helically coiled tube are classified by visual observation. • Void fraction for entire cross-section is measured by wire-mesh sensor. • k -Nearest Neighbors algorithm is utilized to identify flow patterns objectively. • Influence of centrifugal force and secondary flow on phase distribution and flow transition is revealed. Helically coiled tube (HCT) is used extensively in compact types of heat exchangers. Complete knowledge of gas-liquid flow behavior in HCT is essential in designing the heat transfer equipment involving phase change. Although the gas-liquid flow patterns in straight tubes have been widely studied, investigations on flow pattern recognition in HCT are still scarce. In this paper, based on the backlight imaging tomography and self-designed wire-mesh sensor, air-water experiments are carried out to study the characteristics of gas-liquid flow in a transparent HCT. Bubble, plug, slug, wavy and annular flow are observed, and their void fraction at the cross-section of the tube are measured and analyzed using the statistic parameters of time-trace, PDF and CPDF. Owing to the simplicity of implementation, the void fraction data are subsequently utilized along with the k -Nearest Neighbors algorithm (KNN algorithm) to identify the flow patterns objectively. Finally, a flow regime map is proposed and compared with the previous straight flow map and HCT map in the literatures. Accordingly, the influence of curvature-induced centrifugal force and secondary flow on the phase distribution and flow-regime transition is revealed in the present study.