摘要
抽油机的异常情况会使油田的产油效率降低,而不同的异常类型对应的抽油机示功图特征也各不相同,因此造成的损害程度也不同。针对以上问题,文中提出了一种抽油机井功图识别模型,该方法将支持向量机(SVM)用于抽油机井功图识别。首先利用改进的矢量曲线数据压缩方法(ICVDC)对抽油机井下示功图进行特征数据提取,在此基础上,采用"一对一"分类法建立基于支持向量机的井下示功图分类模型,进而对不同特征的示功图进行分类识别,并与其他识别分类模型进行了识别分类效果对比。实验结果表明,该方法分类准确度高,有效地解决了示功图的识别和分类问题,方便对油井设备等进行进一步的故障分析处理,从而大大提高抽油机的性能与效率,以此来达到油田提高采收率的目的。
The pumping unit anomalies will reduce oil production efficiency,while the different abnormal types corresponding to the characteristics of oil pumping machine indicator diagram also are not identical,so the damage degree is different. For above problem,put forward a well pumping unit work diagram recognition model,the method uses Support Vector Machine （ SVM） for pumping unit well figure identification. First use the Improved Curve of Vector Data Compression （ ICVDC） method to extract characteristics of the data of pumping unit downhole indicator diagram,on this basis,using the “one-against-one” classification to establish the downhole indicator diagram classification model based on SVM,with different features to identify the classification of the indicator diagram,and compare with other recognition classification model in classification effect. The experimental results show that,the method for the classification has high accuracy,effectively solving the problem of identification and diagnosis of the diagram,which is convenient for further analyzing and handling the fault of oil well equipment,thus greatly improving the performance and efficiency of the pumping unit,in order to achieve the purpose of oil field recovery improved.
出处
《计算机技术与发展》
2014年第8期215-218,222,共5页
Computer Technology and Development
基金
黑龙江省教育科学技术研究项目(12511010)
关键词
示功图
支持向量机
特征提取
分类
识别
indicator diagram
support vector machine
feature extraction
classification
identification
