期刊名称：IOP Conference Series: Earth and Environmental Science
印刷版ISSN：1755-1307
电子版ISSN：1755-1315
出版年度：2019
卷号：272
期号：3
页码：1-7
DOI：10.1088/1755-1315/272/3/032030
出版社：IOP Publishing
摘要：New approach for classification of the state of technological pumping equipment is presented in the paper. The approach involves the use of pumping equipment parameters operative monitoring data for indirect fault identification. The proposed method is a part of developed integrated approach for decision support in the management of technological equipment of oil and gas fields. The method realizes a multi-stage classification scheme based on an ensemble approach to the intelligent data analysis. The scheme involves the creation of simple classifiers of the first level, which can be implemented on the basis of artificial neural networks or other effective classifying methods. The second level of the scheme is realized by a dynamically tunable aggregators of the first level solutions. The results of an experimental numerical study of the proposed approach and a number of data analysis techniques are presented. The obtained results allow to confirm that it is possible to detect different states of the pumping technological equipment more effectively by the usage of the proposed approach as a part of intelligent data driven diagnostics system.
