Classification of gas–liquid flow patterns by the norm entropy of wavelet decomposed pressure fluctuations across a bluff body

Abstract

Identification of gas–liquid flow patterns remains one of the paramount needs in multiphase flow metering. It is hardly possible to realize accurate measurement and control of parameters in a gas–liquid flow system without a clear understanding of its flow pattern. Here we explore the characterization of gas–liquid flow patterns using the norm entropy extracted from the wavelet decomposed pressure fluctuations across a bluff body. Experiments on air–water two-phase flow at ambient temperature and atmospheric pressure are carried out in the bubble, plug, slug and annular flow patterns. On the basis of the experimental results, two original flow-pattern maps are constructed: one is coordinated with the average norm entropy versus the total mass flow rate, and the other is the average norm entropy versus the volumetric void fraction. Verification tests demonstrate that the overall identification rates of the flow-pattern maps developed exceed 95%. This approach provides an effective and simple solution to the classification of gas–liquid flow patterns.