Summary Produced polymer from a polymer flood alters the rheology of the produced fluids and leads to complex oil/water two-phase flow patterns within the production wellbore that challenges production profile monitoring. This study introduces a multiscale rescaled range fractional permutation entropy (MRSFPE) method to identify these flow patterns. By processing differential pressure signals from vertical pipe flow experiments, the MRSFPE algorithm extracts multiscale entropy features. A joint distribution plane constructed from the entropy mean and entropy rate accurately distinguishes five flow patterns, achieving an overall identification accuracy of 95.39%, outperforming traditional multiscale weighted permutation entropy (MWPE). Analysis using the Ohnesorge (Oh) and Weber (We) numbers reveals the flow pattern transition mechanism: Increasing Oh (dominant viscous forces) drives the evolution from bubbly to slug flow, with oil holdup (Yo) positively correlated to Oh. Furthermore, mapping entropy features onto the Oh-We space establishes a quantitative correlation between entropy characteristics (spatial complexity and temporal randomness), flow pattern structure, and the underlying force balance. This study not only provides a high-precision flow pattern identification tool for well production profile monitoring but also offers new theoretical basis and an analytical paradigm for optimizing parameters of polymer flooding technology and researching flow mechanisms.
Dong et al. (Sun,) studied this question.