Categorization of liquid-liquid counter-current two-phase flow-structures in 11 mm ID tube at different conduit-orientations

Attempts are made to identify, characterize, and automatically categorize the liquid-liquid counter-current two-phase flow-structures in a circular-conduit of 11 mm diameter by varying the flow-orientation. Rigorous ex-perimentations are performed for a wide range of superficial kerosene and water velocities at seven conduit-orientations. Some unique liquid-liquid counter-current two-phase flow-structures, not reported in the litera-ture to date, are found in the present investigation. Interfacial distributions and the highly time-varying nonlinear attributes of the flow-structures corresponding to different combinations of the superficial water and kerosene velocities at each conduit-orientation are detected and recognized through high-resolution photography and high-speed videography. Objective descriptions of the flow-structures are accumulated in the form of time-series current and voltage signals using a differential-pressure-transmitter and in-house fabri-cated conductivity-probes, respectively. Statistical analyses are performed to parameterize the time-series re-petitive objective signatures. The non-parametric probability distribution plot and statistical parameters extracted from the signals are found to be very effective in detecting and distinguishing the individual flow-structures. Flow-structures are parameterized using those statistical parameters. A computational intelligence-based categorization methodology is developed to automatically categorize the obtained flow-structures considering these statistical parameters as the inputs. A good agreement is found for each orientation between the flow-regime groups based on visual observation and those obtained through the categorization methodology. It simply establishes that the flow-structures are perfectly characterized by the statistical parameters extracted from their objective-signatures.