Abstract
In this study, a two-phase flow experiment has been carried out to investigate the feasibility of using probability density function for distinguishing dispersed bubbly flow from the flow regime map. The conductivity sensors were installed and used to acquire the transient conductivity signals, which have been converted to void fraction values during the tests. The Matlab software was used to draw the probability density function of void fraction, and it appears clear differences among the probability density function of void fraction in different flow patterns. Based on this, the probability density function is used to identify the boundary of dispersed bubble flow. By obtaining the important features of the probability density function profiles such as peak location, maximum/minimum values and area ratios of covered regions, the boundary of dispersed bubble flow can be identified.