基于支持向量机的潜油往复式油田抽油机工况诊断方法

Working Condition Diagnosis Method Based on SVM of Submersible Plunger Pump

潜油往复式抽油机是一种重要的新型油田抽油机,它的工况诊断水平直接影响其在油田生产中的应用效果。文章提出一种潜油往复式抽油机的工况诊断方法,在分析潜油往复式抽油机工作原理的基础上,结合泵体结构,得到潜油直线电机在不同工况下加载-卸载过程的不同特征。通过定义相关参量,提取各工况下潜油直线电机运行数据中的特征量。将其作为支持向量机(SVM)分类器的输入,建立基于SVM的潜油往复式抽油机工况诊断模型,并在模拟平台上对该方法的有效性和误判率等进行了验证试验和仿真分析。

The submersible plunger pump is a new device for lifting in onshore oil wells. Its level of working condition diagnosis has important influence on production of oilfield. This paper proposes a working condition diagnosis method of submersible plunger pump. Based on analyzing the potential oil reciprocating pumping works, combining with the pump structure, it occurs different characters in loading and offioading process of submersible linear motor in different working conditions. By defining correlation parameter, the characteristic quantity can be extracted from operation data of submersible linear motor. The support vector machines take these characteristic quantities as input, the working condition diagnosis model based on support vector machines（SVM） of submersible plunger pump is proposed. In this condition, effectiveness and misjudgment rate of working condition diagnosis method can be analysed and validated on simulation platform.