期刊名称：Journal of Theoretical and Applied Information Technology
印刷版ISSN：1992-8645
电子版ISSN：1817-3195
出版年度：2013
卷号：48
期号：2
页码：933-938
出版社：Journal of Theoretical and Applied
摘要：In this paper, a hybrid particle swarm optimization based on the natural selection (NPSO) was presented and used to optimize the parameters of Least Square Support Vector Machine (LSSVM). The NPSO algorithm overcomes the shortcomings of premature convergence and poor local search capability of traditional Particle Swarm Optimization (PSO). Then a classification model of oil-gas-water three-phase flow patterns was established based on NPSO-LSSVM to identify three typical water-based flow patterns of oil-gas-water three-phase flow including bubbly flow, slug flow and bubbly-slug flow. By combining the statistics analysis, Hilbert-Huang transformation, complexity measure analysis, chaotic recurrence quantification analysis and chaotic fractal analysis, the conductance fluctuation signal of oil-gas-water three-phase flow in the vertical pipe was analyzed. The nine feature parameters reflecting the changes of oil-gas-water three-phase flow were extracted and used as the input vectors of the NPSO-LSSVM classification model. Simulation results showed that the correct identification rate of the oil-gas-water three-phase flow patterns was 94%, and it indicated that the classification model proposed in this paper was reasonable and had a practical value.
