Statistical characterization of the interfacial behavior captured by a novel image processing algorithm during the gas/liquid counter-current two-phase flow in a 1/3 scaled down of PWR hot leg

This study addresses a statistical characterization of the interfacial behaviors during gas/liquid counter-current flow in a large scale of pressurized water reactor’s hot leg typical geometry on the basis of both time and frequency domain analyses. Here, the visualizations were captured by high-speed cameras. Furthermore, the image processing algorithm on the basis of pixel identification was developed to identify the time-series interfacial dynamics of the liquid holdup fluctuations, comprising the wave growth and its movement. The obtained data were further analyzed by using various advanced statistical tools comprising the probability density function, power spectral density function, and also discrete wavelet transform. Additionally, chaotic levels of the flow were obtained through the Kolmogorov-entropy analysis.
Particular results reveal that a typical mechanism of flooding begins with the appearance of a wavy interface which further develops into either a roll or large wave, and later blocks the entire cross-sectional area of the conduit around the bend region. This later stage indicates the inception of flooding which is characterized by the increase of the water level close to the bend region. Next, the higher the pressure of system, the higher the frequency occurrence of slugging, while the injected gas flow rate obtained a different trend. However, the Kolmogorov entropy enables to correlate the stabilizing effect due to the fluid resistance which corresponds to either the pressure of the system or the physical properties of the fluid. In addition, there were no significant differences between the flooding mechanisms for both the air/water and steam/water flows.