| Oil/water two-phase flow widely exists in pipeline in the petroleum industry andchemical engineering. The identification of flow pattern is the key to the study ofoil/water two-phase flow, but the difficulty where, and it has great significance both intheory and application.Complexity measure, as an important method of nonlinear time series analysis,attracts much attention from different fields recently. In this paper, we introduced theorigination, methods and meanings of complexity theory concisely. The sequencesfrom logistic map and some general signals with different stochastic degrees wereused to analyze the arithmetics of the four complexity measures including LZComplexity, Fluctuation Complexity, Spectral Entropy and Approximate Entropy. Wecompared the difference and the recognition capability of the four complexitymeasures to different signals, such as period signal, stochastic signal, mixed stochasticsignal and chaotic signal. It is shown that the three complexity measures including LZComplexity, Spectral Entropy and Approximate Entropy taken the stochastic signal asmost complex and the period signal as most simple, which described the structurefeatures of sequence, whereas the Fluctuation Complexity taken the stochasticsequence of uniform distribution as not complexity and reflected the evolution statusof sequence in some extent.Finally, we applied the four complexity measures to characterize the flow patternsof oil/water two-phase flow in vertical upward pipes using the conductance timeseries fluctuating signals. It is shown that all the four complexity measures had littlechanges with variations for water cut ranging from 61% to 91%, which belongs to thesame oil-in-water flow pattern, but for the transitional flow pattern with 51% watercut, the four complexity measures showed irregular sudden changes with the totalflow rate variations. We conclude that the four complexity measures behavior as asensitive 'indicators' of transitional flow pattern.
|
PDF Full Text Request