摘要
在石油开采中,能够对抽油井井下的故障进行预测和诊断,并计算油井的产油量,从而及时了解和掌握采油系统的工况,实现采油系统的自动化监控和科学管理,是当前迫切需要解决的一个问题。本文针对如何实现油井故障自动诊断进行研究,通过实测示功图、提取功图特征值,并结合神经网络,以及能耗计算辅助判断故障类型的方法,从而建立油井故障诊断专家系统,通过油田实采数据验证,使诊断结果正确率基本达到95%,结论表明该方法是一种较为成功的尝试,有较高的实用推广性。
In oil exploitation, the capability of pumping well downhole fault prediction and diagnosis, and calculating the oil production, so as to timely know and master the working condition of the oil exploitation system, realize the automation monitoring and scientific management of oil exploitation system, is an urgent problem to be solved in the oil industry. Therefore, the research of pumping well fault diagnosis technology has a high practical value, and it has also been an important task to do- mestic and oversea oil exploitation engineering technicist. In this paper, the research is according to the measured indicator diagram, extracting the characteristic values of indicator diagrams and combining with the neural network technology to establish the oilweU fault diagnosis expert system, and realize the oilwell fault automatic diagnosis. The accuracy of fault diagnosis is basicaUy up to 95 percent.
出处
《电子设计工程》
2015年第18期119-122,共4页
Electronic Design Engineering
关键词
示功图
有杆抽油系统
专家系统
神经网络
故障诊断
能耗计算
dynamometer card
sucker rod pumping system
expert system
neural network
fault diagnosis
energy consumption calculation
