基于示功图的油井供液能力智能识别方法研究  

Intelligent identification method of oil well's liquid supply capacity based on dynamometer cards

作　　者：夏峰 陈夕松[1] 钱帅康 姜磊 XIA Feng;CHEN Xisong;QIAN Shuaikang;JIANG Lei(College of Automation,Southeast University,Nanjing Jiangsu 210096,China;Nanjing Richisland Information Technology Co.,Ltd.,Nanjing Jiangsu 210061,China)

机构地区：[1]东南大学自动化学院,江苏南京210096 [2]南京富岛信息工程有限公司,江苏南京210061

出　　处：《石油化工应用》2022年第2期93-98,共6页Petrochemical Industry Application

基　　金：江苏省重点研发计划项目“高性能原油在线调合平台研发”,项目编号:BE2019016;南京江北新区重点研发计划“高端原油调合调度一体化系统软件研发”,项目编号:ZDYF20200127。

摘　　要：油井供液能力作为游梁式抽油机间抽控制和冲次调节的重要依据,它的准确识别对降低油田生产能耗、提升抽油机系统效率和增加油田产油量意义重大。论文研究了基于数据驱动的深度学习回归模型,采用融合了特征金字塔、高斯热力图和注意力机制的深度学习网络提取示功图图形特征并识别功图的四个凡尔开闭点,根据识别出的凡尔开闭点量化油井供液能力,为后续抽油机的智能控制提供调节依据。实验结果表明,该模型能够准确识别功图的四个凡尔开闭点,实时估算油井供液能力。As an important basis for the pumping control and stroke adjustment of pumping units,the output capacity of oil wells is of great significance for reducing energy consumption,improving pumping unit system efficiency,and increasing oilfield production.This paper studies a data-driven deep learning regression model,and uses a deep learning network that combines feature pyramids,Gaussian heatmaps and attention mechanisms to extract graphical features of the dynamometer cards and identify the four valve opening and closing points of the power diagram and quantify the output capacity of the oil well based on these points,which is to provide an adjustment basis for the intelligent control of the pumping unit.The experimental results show that the model can accurately identify the four valve opening and closing points of the dynamometer cards,and estimate the output capacity of the oil well in real time.

关 键 词：抽油机 示功图 凡尔开闭点 供液能力 智能识别 

分 类 号：TE331[石油与天然气工程—油气田开发工程]