Joint Learning of Commonality and Specificity Dictionaries for Identifying Flow Regime in Oil–Gas–Water Three-Phase Flow

Three-phase flow of oil, gas, and water is a complex flow process with various flow regimes, which is prevalent in the energy and petrochemical engineering. The diverse flow regimes of oil–gas–water three-phase flow result from the combined influence of several factors, including time evolution, spatial structure, and flow conditions. To explore the shared and distinct characteristics of different flow regimes in terms of fluid structure and flow parameters, commonality and specificity dictionaries are jointly studied to realize flow regime identification and analysis. Utilizing multisensor measurement data from the flow process, discriminant dictionaries capturing specificity and the shared dictionary capturing commonality of different flow regimes are established. The distinction between flow regimes is further emphasized through the application of the local constraint, while the low-rank constraint ensures an ideal shared subspace. The efficacy and merits of the approach in flow regime analysis are demonstrated by dynamic experiment results conducted in the horizontal loop of the experimental device.