Probabilistic mapping of adiabatic horizontal two-phase flow by capacitance signal feature clustering

In order to better quantify two-phase flow regime transitions, a sensor is developed which measures the electric capacitance of two-phase flows. A large number of experiments are done with air–water flow in a 9 mm ID horizontal tube. Based on a multivariate analysis, the most suitable sensor signal parameters are selected for building a flow regime classifier. This classifier is based on a fuzzy c-means clustering algorithm together with a regression technique. The output of the algorithm is used to create a probabilistic flow regime map. A comparison between a visual classification based on high speed camera images and the outcome of the flow regime classifier shows a remarkable agreement. The flow regime transitions are further quantified and discussed based on the probabilistic information and the sensor signal characterization. Probabilistic mapping makes it possible to combine flow regime dependent correlations in the two-phase flow models for heat transfer and pressure drop with smooth and appropriately quantified transitions from one flow regime to another.