支持向量机的无杆抽油机示功图自动识别技术  

Study on automatic recognition of indicator diagram of rod- less pumping unit based on support vector machine

作　　者：朱春梅[1] 陕梅辰[1] 智玉杰 王朝霞 刘娇[1] 

机构地区：[1]北京信息科技大学机电工程学院,北京100192 [2]中国石油集团工程设计有限责任公司北京分公司,北京100085

出　　处：《北京信息科技大学学报（自然科学版）》2014年第1期16-20,共5页Journal of Beijing Information Science and Technology University

基　　金：国家自然科学基金资助项目(61273173);北京市自然科学基金资助项目(4122029);北京市教委科技计划重点项目(KZ201211232039);北京市属高等学校人才强教计划资助项目(PHR201108251)

摘　　要：针对抽油机示功图人工目测误差大、效率低的现状,研究了基于支持向量机理论的示功图自动识别技术。提出了基于支持向量机理论对示功图进行自动识别的可能性,提供了基于支持向量机理论的示功图自动识别步骤,并对无杆抽油机正常、供液不足、游动凡尔泄漏以及油管漏失4种工况下的示功图进行了自动识别研究。识别结果表明:基于支持向量机理论的示功图自动识别方法能对以上4种工况的示功图进行准确有效的识别。To solve the problem of big error and low efficiency in the visual observation of indicator diagram for rod-less oil pumping machine, the technology of automatic recognitionis studied based on support vector machine. The possibility of automatic recognition of indicator diagram based on support vector machine theory is supported. Procedures of automatic identification based on support vector machine theory are provided. The research on indicator diagram under different conditions such as, normal, fluid insufficient, dropout of traveling valve and tubing dropout is carried out to rod-less pumping unit. The recognition results show：based on the indicator diagram theory of support vector machines, automatic identification method can identify accurately and effectively, which is significant for improving the automation and intelligent analysis of running state monitoring of rod-less pumping unit.

关 键 词：无杆抽油机 支持向量机 示功图自动识别 

分 类 号：TH7[机械工程—仪器科学与技术]