Data Mining Methods Application to Solve Oil Wellproduction Flow Regimes Problem Classification

The oil well production flow regimes classification problem is consid-ered using neural networks and data mining methods. Functional separability of classes definition is suggested in order to simplify of neural networks architecture and to lower the calculation complexity. A multiphase measuring system is installed in oil wells and registers the set of primary data, used to show, that the property the classes functional separability of the subject area under consideration makes it possible to use simpler neural networks solving the classification problem. The determination importance multiphase flow regime is explained to maintain oil well operating efficiency.