Objective characterization of interfacial structures in two-phase flow

In view of establishing a detailed and reliable measurement technique for characterizing the interfacial structures and identifying flow regimes in two-phase flow, two objective approaches are presented. First, the state-of-the-art four-sensor conductivity probe technique is presented to obtain the detailed local information. The newly designed four-sensor conductivity probe accommodates the double-sensor probe capability. Hence, it can be applied in a wide range of two-phase flow regimes spanning from bubbly to churn-turbulent flows with a measurement error of approximately ± 10%. The signal processing scheme is developed such that it accounts for the missing bubbles and defective signals. Furthermore, it categorizes the acquired parameters into two groups based on bubble cord length information. Local information on the void fraction, interfacial area concentration, Sauter mean diameter, interface velocity for each group of bubbles was obtained successfully. Second, a global measurement technique using the non-intrusive impedance voidmeter and neural networks is presented. In this method, an advanced non-intrusive impedance voidmeter provides global interfacial structure information to neural networks which are used to identify the flow regimes. Both supervised and self-organizing neural network learning paradigms performed flow regime identification successfully. In the application of this global method, two approaches are presented, namely: One based on the Probability Density Function (PDF input method), and another based on the ordered set of void fraction measurements which were acquired in a very short time period (instantaneous direct signal input method). The direct signal input method minimizes the time required for identifying the flow regime