基于SVM的油水两相流相态识别方法  

Recognition Method of Oil-water Two-phase Flow Based on SVM

作　　者：李昱萱 刘兴斌[1,2] 韩连福[1] 付长凤[1] LI Yuxuan;LIU Xingbin;HAN Lianfu;FU Changfeng(Northeast Petroleum University,Daqing,Heilongjiang 163000,China;Daqing Logging and Testing Services Company,Daqing,Heilongjiang 163153,China)

机构地区：[1]东北石油大学,黑龙江大庆163000 [2]大庆油田有限责任公司测试技术服务分公司,黑龙江大庆163153

出　　处：《石油管材与仪器》2021年第2期91-95,共5页Petroleum Tubular Goods & Instruments

基　　金：黑龙江省自然科学基金“化学驱水平小井筒内携砂油气水三相流流动参数测量方法研究”(编号:LH2020E012)。

摘　　要：目前国内各陆上油田大多处于高含水开发期,多数油井均为油水两相的混合流动状态,准确识别油水两相流分界面非常必要。为了提升识别率,采用基于小波包分解提取特征参数,运用Python编程语言搭建支持向量机(SVM)训练模型的方法实现油水分界面辨识,经过对训练及测试结果的分析,油水两相流分相识别率达到了100%,证实SVM在油水两相辨识方面有很好的适用性。At present,most of the major oilfields on the mainland are in the high water-cut development period,and most of the oil wells are in the mixed flow state of oil-water two-phase flow.It is necessary to identify the phase state of oil-water two-phase flow accurately.In order to improve the recognition rate,the method of extracting characteristic parameters based on wavelet packet analysis and using Python programming language to build a support vector machine(SVM)training model is used to implement the recognition of oil-water interface.After the analysis of the training and test results,the phase recognition rate reaches 100%,which proves that SVM has good applicability in oil-water phase recognition.

关 键 词：油水两相流 地面计量 小波包 特征提取 支持向量机 

分 类 号：TE35[石油与天然气工程—油气田开发工程]