ICA及其在气液两相流辨识中的应用  

ICA and Its Application in the Identification of Gas/Liquid Two-Phase Flow

作　　者：刘洪林[1] 李海山[1] 

机构地区：[1]大庆石油学院地球科学学院,黑龙江大庆163318

出　　处：《吉林大学学报（地球科学版）》2009年第1期31-36,共6页Journal of Jilin University：Earth Science Edition

基　　金：黑龙江省自然科学基金项目(E0317);黑龙江省科技攻关项目(GZ05A102)

摘　　要：应用独立分量分析(ICA)方法研究气液两相流的辨识问题。根据基于高阶统计量的改进J-HICA求解算法,利用给定仿真信号及不同流量油、气、水单相流体响应信号,模拟两相、三相混合信源进行分离实验,验证该方法的可靠性,并对生产井实测气液两相流体多路传感器响应信号进行分离。对比及分析分离结果表明,该方法成功地从混合信号中分离出表征气、液特征的源信号,达到了辨识目的。The independent component analysis （ICA） is applied to study the two-phase flow identification. According to the improved J - H algorithm of ICA which based on higher order statistics, the reliability of the method was conformed using given modulate signals and different fluid flow of monophase of oil, gas and water response signals to emulate tow phase and three phase mixed up signal sources. The field productive well logging gas and liquid two-phase signals which responded by mull-channel sensors were separated, compare and analysis indicate that signals which characterize the gas and liquid information respectively are successfully separated from mixed signals of well logging.

关 键 词：独立分量分析 高阶统计量 J—H算法 两相流 峭度 

分 类 号：P631.4[天文地球—地质矿产勘探]