西南石油大学学报(自然科学版) ›› 2018, Vol. 40 ›› Issue (1): 173-180.DOI: 10.11885/j.issn.1674-5086.2016.11.25.02

• 石油机械与油田化学 • 上一篇    

基于改进K-means的潜油电泵质量评估方法研究

刘广孚1, 朱赫1, 周凯迪1, 简纪新2, 常云飞1   

  1. 1. 中国石油大学(华东)信息与控制工程学院, 山东 青岛 266580;
    2. 中国石化胜利油田分公司胜利采油厂, 山东 东营 370500
  • 收稿日期:2016-11-25 出版日期:2018-02-01 发布日期:2018-02-01
  • 通讯作者: 刘广孚,E-mail:liugf@upc.edu.cn
  • 作者简介:刘广孚,1966年生,男,汉族,河北辛集人,副教授,博士,主要从事潜油电泵井的参数监测与故障诊断研究。E-mail:liugf@upc.edu.cn;朱赫,1989年生,男,汉族,安徽宿州人,硕士研究生,主要从事采油设备的信号检测和处理方向研究。E-mail:zhuhe112473@163.com;周凯迪,1992年生,女,汉族,山东德州人,硕士研究生,主要从事采油设备的信号检测与处理方向研究。E-mail:zhou_kd1031@163.com;简纪新,1966年生,男,汉族,重庆忠县人,工程师,主要从事潜油电泵现场施工和日常维护及管理工作。E-mail:jianjixin@163.com;常云飞,1992年生,女,汉族,山东滨州人,硕士研究生,主要从事采油设备的信号检测与处理方向研究。E-mail:changyunfei1104@163.com
  • 基金资助:
    国家自然科学基金(61273160)

Study on the Quality Evaluation of Electric Submersible Pumps Based on Improved K-means

LIU Guangfu1, ZHU He1, ZHOU Kaidi1, JIAN Jixin2, CHANG Yunfei1   

  1. 1. College of Information and Control Engineering, China University of Petroleum, Qingdao, Shandong 266580, China;
    2. Shengli Oil Production Plant, Shengli Oilfield Subcompany, SINOPEC, Dongying, Shandong 370500, China
  • Received:2016-11-25 Online:2018-02-01 Published:2018-02-01

摘要: 针对潜油电泵使用寿命的问题,开展了基于潜油电泵井下振动测量的单节潜油电泵装配质量等级分类的研究,建立了单节潜油电泵的三维振动模型,提出了拐点法优化K-means聚类算法初始点的选取,采用三维合成方法求取电泵三个维度的合成振动加速度均方根值,利用改进K-means算法对239节潜油电泵的合成振动加速度均方根进行了聚类分析,建立了潜油电泵装配质量等级评价标准。对所测量的电泵进行质量跟踪,其结果验证了电泵装配质量等级分类的合理性。结果表明,潜油电泵合成振动加速度均方根小于1.30 m/s2的潜油电泵为优质电泵,使用寿命最长,宜用于高产井;处于1.30~1.72 m/s2的为质量良好电泵,使用寿命较长,可正常使用;处于1.72~2.35 m/s2的为质量合格电泵,使用寿命一般,应谨慎使用;大于2.35 m/s2为质量较差电泵,使用寿命较短,应维修后再使用。

关键词: 潜油电泵, 实验井, 振动, 加速度, K-means算法

Abstract: This study examines issues concerning the lifespan of electric submersible pumps and establishes assembly quality grade classifications based on underground vibration measurements. A 3D vibration model of a single electric submersible pump is established. The inflection point method is proposed to optimize the selection of the initial point for a K-means clustering algorithm. 3D synthetic method is employed to calculate the root-mean-square of the synthetic vibration acceleration in three dimensions. The improved K-means algorithm is used to perform a cluster analysis on the root-mean-square values of the synthetic vibration acceleration for 239 electric submersible pumps. Evaluation criteria for the assembly quality of electric submersible pumps are established. Quality tracking is carried out for electric pump measurements and the research results verify the rationality of the established classifications for electric pump assembly quality. According to the results, electric submersible pumps with root-mean-square synthetic vibration accelerations smaller than 1.30 m/s2 are excellent pumps with the longest life spans. These pumps are applicable for use in high-yield wells. Pumps with values between 1.30 and 1.72 m/s2 are considered as having good quality. Pumps with values between 1.72 and 2.35 m/s2 have a typical lifespan and should be used carefully. Those with values higher than 2.35 m/s2 are poor quality pumps with a short lifespan and should be repaired before use.

Key words: electric submersible pump(ESP), experimental well, vibration, accelerated velocity, K-means clustering algorithm

中图分类号: