2025年04月07日 星期一 首页   |    期刊介绍   |    编 委 会   |    投稿指南   |    期刊订阅   |    统合信息   |    联系我们
计量学报  2019, Vol. 40 Issue (5): 835-841    DOI: 10.3969/j.issn.1000-1158.2019.05.16
  流量计量 本期目录 | 过刊浏览 | 高级检索 |
基于随机森林算法的井下原油含水率软测量方法
王鹏1,2,3,龚盼1,2,冯定1,2,3,涂忆柳1,2,3
1. 长江大学 机械工程学院,湖北 荆州 434023
2. 湖北省油气钻完井工具工程技术研究中心,湖北 荆州 434023
3. 非常规油气湖北省协同创新中心,湖北 武汉 430100
Soft Sensing Method for Water Cut of Crude Oil Based on Random Forest Algorithm
WANG Peng1,2,3,GONG Pan1,2,FENG Ding1,2,3,TU Yi-liu1,2,3
1.College of Mechanical Engineering, Yangtze University, Jingzhou, Hubei 434023, China
2.Hubei Engineering Research Center for Oil & Gas Drilling and Completion Tools, Jingzhou, Hubei 434023, China
3.Hubei Cooperative Innovation Center of Unconventional Oil and Gas, Wuhan, Hubei 430100, China
全文: PDF (1124 KB)   HTML (1 KB) 
输出: BibTeX | EndNote (RIS)      
摘要 井筒产出液实时测量是获得井下产液信息的关键步骤,是实现智能井的重要前提,但由于井下环境的复杂性,油水混合物状态本身的不确定性,缺少较好的方法能在井下较好地完成产出液各成分含量的计量。针对井下油、水两相混合计量的问题,提出了一种结合机器学习算法的井下原油含水率软测量方法。结合采油工程需要,可将原油含水率以10%的间隔划分成11个类别,将不同原油含水率对应的物理属性作为测量对象,利用随机森林算法对已知样本进行学习分类,得到原油物理属性与含水率之间的关系,进而实现对原油含水率的测量。最后,通过设计的室内实验,对提出的软测量方法进行了验证,结果表明利用该方法预测原油含水率,得到的结果与实验值一致,可以初步满足工程对井下原油含水率测量的需求,为井下测量提供一了种新的思路。
服务
把本文推荐给朋友
加入我的书架
加入引用管理器
E-mail Alert
RSS
作者相关文章
王鹏
龚盼
冯定
涂忆柳
关键词 计量学原油含水率井下测量软测量随机森林算法    
Abstract:Real time measurement of wellbore output fluid is a key step in obtaining information from the downhole liquid production and an important prerequisite for the realization of intelligent well. However, there is no better way to measure the content of each component of the output liquid, because of the complexity of downhole environment and the uncertainty of oil-water mixture state. Aimed at the problem of mixed measurement of oil and water in the downhole, a soft sensing method for water cut of crude oil based on machine learning algorithm was presented. Combined with the needs of oil production engineering, water cut of crude oil was divided into 11 categories at 10% intervals. Taking physical properties corresponding to different water cut of crude oil as measurement object, the relationship between physical properties and water cut of crude oil was obtained and then the purpose of measuring the water cut of crude oil was realized, by learning and classifying known samples using random forest algorithm. Finally, the soft sensing method was verified by a designed indoor experiment. The results show that the predicted results by using the mentioned method agree with the experimental values , which can preliminarily meet the requirement of engineering for measuring water cut of underground crude oil and providing a new concept for downhole measurement.
Key wordsmetrology    water cut of crude oil    downhole measurement    soft sensing    random forest algorithm
收稿日期: 2018-07-10      发布日期: 2019-09-01
PACS:  TB937  
基金资助:长江大学青年基金(2016cqn30);中海石油(中国)有限公司湛江分公司横向项目(CCL2017ZJFN2272)
通讯作者: 冯定(1963-),男,安徽池州人,长江大学教授,主要从事油气装备及井下工具的设计、诊断及动态仿真研究。 Email: fengd0861@163.com     E-mail: fengd0861@163.com
作者简介: 王鹏(1983-),男, 河南新乡人,长江大学博士研究生,主要研究方向为石油机械及井下工具的设计、 诊断及动态仿真理论与技术应用。 Email: 94211936@qq.com
引用本文:   
王鹏,龚盼,冯定,涂忆柳. 基于随机森林算法的井下原油含水率软测量方法[J]. 计量学报, 2019, 40(5): 835-841.
WANG Peng,GONG Pan,FENG Ding,TU Yi-liu. Soft Sensing Method for Water Cut of Crude Oil Based on Random Forest Algorithm. Acta Metrologica Sinica, 2019, 40(5): 835-841.
链接本文:  
http://jlxb.china-csm.org:81/Jwk_jlxb/CN/10.3969/j.issn.1000-1158.2019.05.16     或     http://jlxb.china-csm.org:81/Jwk_jlxb/CN/Y2019/V40/I5/835
京ICP备:14006989号-1
版权所有 © 《计量学报》编辑部
地址:北三环东路18号(北京1413信箱)  邮编:100029 电话:(010)64271480
本系统由北京玛格泰克科技发展有限公司设计开发  技术支持:support@magtech.com.cn